Resistance of Nitric Oxide Dioxygenase and Cytochrome c Oxidase to Inhibition by Nitric Oxide and Other Indications of the Spintronic Control of Electron Transfer

Abstract

Heme enzymes that bind and reduce O₂ are susceptible to poisoning by NO. The high reactivity and affinity of NO for ferrous heme produces stable ferrous-NO complexes, which in theory should preclude O₂ binding and turnover. However, NO inhibition is often competitive with respect to O₂ and rapidly reversible, thus providing cellular and organismal survival advantages. This kinetic paradox has prompted a search for mechanisms for reversal and hence resistance. Here, I critically review proposed resistance mechanisms for NO dioxygenase (NOD) and cytochrome c oxidase (CcO), which substantiate reduction or oxidation of the tightly bound NO but nevertheless fail to provide kinetically viable solutions. A ferrous heme intermediate is clearly not available during rapid steady-state turnover. Reversible inhibition can be attributed to NO competing with O₂ in transient low-affinity interactions with either the ferric heme in NOD or the ferric heme-cupric center in CcO. Toward resolution, I review the underlying principles and evidence for kinetic control of ferric heme reduction via an O₂-triggered ferric heme spin crossover and an electronically-forced motion of the heme and structurally-linked protein side chains that elicit electron transfer and activate O₂ in the flavohemoglobin-type NOD. For CcO, kinetics, structures, and density functional theory point to the existence of an analogous O₂ and reduced oxygen intermediate-controlled electron-transfer gate with a linked proton pump function. A catalytic cycle and mechanism for CcO is finally at hand that links each of the four O₂-reducing electrons to each of the four pumped protons in time and space. A novel proton-conducting tunnel and channel, electron path, and pump mechanism, most notably first hypothesized by Mårten Wikström in 1977 and pursued since, are laid out for further scrutiny. In both models, low-energy spin-orbit couplings or ‘spintronic’ interactions with O₂ and NO or copper trigger the electronic motions within heme that activate electron transfer to O₂, and the exergonic reactions of transient reactive oxygen intermediates ultimately drive all enzyme, electron, and proton motions.

1. Background and Introduction

1.1. NO Toxicity and Inhibition of Heme Enzymes

NO has been known to be toxic to living organisms since the synthesis and discovery of the gas by Sir Humphrey Davy well over 200 years ago [1]. NO, and the orangish-brownish NO₂ gas formed in the reaction of NO with atmospheric O₂, was quickly recognized as harmful and not euphoric, anesthetic or analgesic, and seemingly innocuous [2], like N₂O. Nevertheless, NO serves as an important intermediate in microbial denitrification [3] and, at high nanomolar steady-state concentrations, requires detoxification by NO reductases (NORs) and NO dioxygenases (NODs) [4,5]. Mammals also synthesize NO, which is ultimately oxidized to form nitrate and nitrite under normal physiological conditions. NO is released by leukocytes and is elevated in various pathological conditions involving inflammation [6,7,8,9,10]. NO is produced primarily by inducible NO synthases (NOSs) [11], secondarily via enzymatic and non-enzymatic reduction of NO₂⁻ [12], or pharmacologically through the metabolism of vasorelaxants such as nitroglycerin and isoamyl nitrite [13]. Indeed, an understanding of the physiology of nitroglycerin, first described by the early discovery of Ascanio Sobrero and subsequent research of Constantin Hering and coworkers in the mid-19th century [14], has finally been achieved. The elucidation of the role and mechanism of NO or the ‘endothelium-derived relaxing factor’ in regulating blood pressure was recognized with the Nobel Prize in Medicine or Physiology in 1998 [15]. Similar to microbes, mammals, and apparently all life forms, metabolize and detoxify NO [5,16]. Hemoglobin (Hb) and myoglobin (Mb) are important and common scavengers of NO that catalytically dioxygenate NO to form nitrate [4,17]. When tightly-coupled to reductases, globins enzymatically dioxygenate NO [4,5,17,18]. It is noteworthy that the non-enzymatic reaction of NO with oxyHb was first reported by Sir Humphrey Davy over 200 years ago, with the visible browning of red blood upon exposure [1], and was further characterized by Doyle and Hoekstra in 1981 [19]. OxyHb and oxyMb react with NO with measured bimolecular rate constants in the range of 2 to 4 × 10⁷ M⁻¹ s⁻¹ [19,20,21]. Short-lived caged intermediates in an O-atom rearrangement mechanism are hypothesized [17,20,21,22,23].

NO inhibits or inactivates enzymes either directly or indirectly through NO-derived reactive species, including ONOO⁻, NO₂, ⁺NO, or NO⁻ [24,25]. At sufficient concentrations, NO can inactivate or activate metalloenzymes, thiol-dependent enzymes, di-iron enzymes, and tyrosine-dependent enzymes within living organisms. Iron-sulfur [9,26,27,28] and heme-containing enzymes, including cytochrome c oxidase (CcO) [29,30] and cytochrome P450s [31,32,33], were identified early on as sensitive targets for NO toxicity. Several NO-sensitive iron-sulfur and heme targets have been found to be utilized as NO sensors for diverse functional roles by various organisms. Indeed, blood pressure regulation ultimately occurs due to ferrous heme-NO formation that activates the soluble guanylate cyclase [11,34]. NO impacts mammalian iron metabolism via reactions with the cytosolic aconitase iron-sulfur center [35,36]. In some scenarios, reversible NO inhibition of CcO may serve a beneficial function by maximizing tissue O₂ distribution or sensing O₂ [37,38].

1.2. Cellular NO Detoxification Pathways, NODs, NORs, and Other Pathways

Similar to the evolution of NO-sensing functions, it should not be surprising that sensitive and critical targets of NO toxicity have forced the evolution of mechanisms for NO resistance, including damage repair. Furthermore, it should not be unexpected that an NO resistance mechanism would provide a chemical and genetic route for the evolution of enzymes specialized for NO metabolism and detoxification. According to the dictum eloquently stated by Max Perutz [39] and reiterated by Matti Saraste and others [40,41], evolution transforms one system to produce another through small changes based on a shared chemistry. The NO and O₂ metabolizing enzymes (i.e., NODs, NORs, and O₂ reductases), sensors, or O₂ transport-storage proteins are perfect examples where this is made abundantly clear. A proviso to this condition is that any progress in understanding observed reactions and other phenomena requires the discernment of genuine biological functions among shared chemistries via comprehensive studies of kinetics, mechanism, and genetics in the context of physiology. The effort demands a critical examination of many, if not all, of the empirical observations, possible interpretations, a recognition of the reproducible and consistent patterns, and the contradictions. Only after extensive searching and ‘wandering’ [42] can a logical essence or understanding emerge from ‘the gardens of the mind’ [43]. Understanding the function and evolution of NO and O₂ systems has been one focus and goal of my life’s work and of this review.

1.3. The Paradoxical Function(s) of flavoHbs in NO Detoxification

Evidence for an O₂-dependent nitrate-producing NO metabolic enzyme in soil microbes, including Escherichia coli (abbreviated as E. coli), was reported by Ralf Conrad and coworkers in a series of publications from 1989 to 1996 [44,45,46,47]. Unfamiliar with Conrad and coworkers’ publications, my group observed an NO-inducible, O₂-dependent, and cyanide-sensitive NO metabolic activity in E. coli while investigating the NO-sensitive citric acid cycle enzyme aconitase [48,49]. In cell-free extracts, the activity required NAD(P)H, O₂, and FAD and generated nitrate. The NOD activity was isolated, sequenced, and identified with flavoHb [49]. The requirement for FAD for NOD activity was unexpected and troublesome, but it was advantageous for FAD affinity chromatography and rapid isolation. Similar to the structurally related cytochrome P450 reductase [50], the FAD cofactor is apparently released from the reductase domain during ammonium sulfate fractionation but readily rebinds and reconstitutes the apo-enzyme [51].

A simple enzymatic cycle was readily proposed based on the well-known reaction of NO with oxy-globins to produce nitrate (Equations (1) and (2)) [19,20,21,52]. NAD(P)H reduced the FAD in the reductase domain (Equation (3)), which reduced the ferric heme in the globin domain (Equation (4)) with one NAD(P)H consumed per two NO molecules (Equations (5) and (6)). Hausladen et al. independently demonstrated the inducible and protective NO metabolic activity of E. coli flavoHb, reported the reaction stoichiometry (Equation (6)), and briefly decided [53] (see Section 3.3.2 below) in favor of the NOD mechanism (Equations (1)–(6)) [54].

O₂ + Fe²⁺ ⇌ Fe³⁺O₂⁻

(1)

Fe³⁺O₂⁻ + NO → Fe³⁺ + NO₃⁻

(2)

NAD(P)H + H⁺ + FAD → FADH₂ + NAD(P)⁺

(3)

FADH₂ + Fe³⁺ → Fe²⁺ + FADH + H⁺

(4)

FADH + Fe³⁺ → Fe²⁺ + FAD + H⁺

(5)

2 NO + 2 O₂ + NAD(P)H → 2 NO₃⁻ + NAD(P)⁺ + 2 H⁺

(6)

Several earlier investigations had pointed to a role for flavoHbs in NO metabolism or detoxification. Numerous investigations have subsequently supported that biological role [5,55,56,57]. However, the mechanism, and thus molecular function, continues to be debated [56]. Poole and coworkers had shown that E. coli flavoHb is induced by nitrite or by bolus NO addition and postulated a role in nitrogen oxide metabolism and a NOR function [58]. Nakano also reported increased flavoHb expression in Bacillus subtilis with growth on nitrite and postulated a role in nitrite reduction [59]. Bärbel Friedrich’s lab had determined earlier that the Alcaligenes eutrophus (aka Ralstonia eutropha) flavoHb bore no detectable NOR activity but, nevertheless, was essential for NO reduction to N₂O during denitrification [60]. The high NO affinity of the ferrous heme and a rapid autooxidation of the oxy complex of the yeast and A. eutrophus flavoHbs purportedly discouraged Austen Riggs, Hao Zhu, John Olson, and Yi Dou, and likely others, from an earlier foray into investigations of flavoHb NO metabolic activity [61,62] involving an ‘NO-induced oxidation’ similar to oxyHb or oxyMb [19,20]. In our eventual collaboration with Riggs, Zhu, Olson, Li, and Dou, the two flavoHbs, along with the E. coli flavoHb, were analyzed side-by-side and were found to have remarkably similar transient ligand binding kinetics and steady-state NOD kinetics [62]. A ping-pong mechanism with O₂ first binding the ferrous heme and NO reacting with the complex was not entirely supported by the data. The problem of high-affinity NO binding to ferrous heme in place of O₂ in the proposed NOD mechanism (Equation (1)) was now patently evident [51,62], but the contradiction or paradox was never resolved in our joint efforts. With the prohibitively high-avidity of NO for the ferrous heme now measured and published, Hausladen et al. readily argued against the NOD mechanism (Equations (1)–(6)). A more favored alternative ‘denitrosylase’ mechanism with a reverse binding order for NO and O₂ was argued, and a new enzyme nomenclature was suggested [53,63] (Equations (7) and (8)).

NO + Fe²⁺ ⇌ Fe³⁺NO⁻

(7)

Fe³⁺NO⁻ + O₂ → Fe³⁺ + NO³⁻

(8)

Further confounding the mechanism and function of flavoHb, an ancillary anoxic NOR activity that was demonstrated by Kim et al. [64,65] but was contradicted by the observations of Cramm et al. [60] was put forth to explain several anaerobic induction and protection phenomena. Indeed, the phenomena continue to point to NO-related functions for flavoHb that are both unrelated to NO dioxygenation and occur in the absence of any demonstrable NOR activity of flavoHb within cells [66]. Anoxic NO-dependent physiological effects invariably continue to lead to the supposition of NO reduction [60,67]. However, the absence of trace and infinitesimal O₂ is rarely, if ever, demonstrated. In this regard, an NOD activity consumes both NO and O₂, and either O₂ or NO may be important under hypoxic conditions in the absence of other more dominant O₂ or NO consumption activities. NO dioxygenation by globins may remove O₂ and protect O₂-sensitive targets in legumes [68], in Ascaris [69], or increase NO reduction by adventitious NO/O₂ reductases, including some that are readily inactivated by O₂ [70]. Under microaerophilic conditions, trace O₂ may inhibit the dedicated NOR while NO may inhibit the dedicated O₂ reductase. An overexpressed NOD activity would partly relieve both inhibitions by contributing to NO and O₂ depletion. A sulfur trioxide anion radical scavenging function has also been demonstrated for flavoHbs [71] and may explain phenomena related to oxidative stress, such as the protection of aerobic E. coli against paraquat toxicity [72] or yeast against copper toxicity [73]. Notwithstanding the numerous intriguing, yet unexplained, physiological phenomena, progress in deciphering the biological NO detoxification mechanism(s) and molecular function(s) of flavoHbs and related single domain Hbs continues to be achieved through kinetic, structural, and other biophysical investigations [18,74].

The NOD vs. denitrosylase mechanism has been debated for well over two decades [5,53,56,63], with the concept of an NO reduction-dependent denitrosylase mechanism and dual aerobic–anaerobic function sporadically relegating the concept of a limited aerobic NOD mechanism and function [56,67]. Nevertheless, progress on many fronts has illuminated an NOD mechanism with subtle, albeit unorthodox, interactions of the ferric heme with O₂ and NO [18,74,75] that may settle the lingering controversy and confusion. The revised NOD mechanism depends upon spin-orbit coupling (SOC) phenomena, which I dub ‘spintronics’ in close analogy to the quantum principles recently deployed in spintronic devices, including those being designed for efficient electrochemical O₂ reduction in fuel cells [76,77,78]. The proposed mechanism also employs motions similar to those demonstrated long ago in cooperative O₂ binding to red blood cell Hb and allostery. In Section 3.3.4, I focus on the essential role of spintronics and motions for the NOD mechanism. It is hoped that greater knowledge of the NOD mechanism may ultimately be applied to the design of targeted antibiotics [79] and other therapies or, perhaps in the distant future, to bioengineering spintronic devices.

1.4. The Paradoxical NO Resistance of CcO and Other O₂ Reductases

Observations of O₂-dependent, cyanide-sensitive, and protective NO metabolism by E. coli and various mammalian cell lines and tissues originally prompted my lab to investigate the possible role of the terminal oxidases (i. e. CcO, cytochrome bo, and other O₂ reductases) in NO metabolism within E. coli [48] and mammalian cells [16]. A role for the mitochondrial CcO (aka cytochrome aa₃) in cellular NO metabolism had been hypothesized [80,81]. And, at the time, the role of globin-type NOD activities in NO metabolism discussed above in Section 1.3 had not yet been discovered. The exquisite NO sensitivity of cytochrome aa₃ was confirmed, but a role for cytochrome aa₃ in NO metabolism was not substantiated [16]. Nor did the E. coli cytochrome bo or cytochrome bd catalyze measurable NO metabolism within E. coli [48]. Cyanide was without effect on the respiration of E. coli at levels that clearly inhibited NO detoxification [48,82]. Moreover, E. coli strains deficient in either of the two terminal respiratory oxidases showed normal levels of the constitutive and inducible aerobic NO consumption activities, whereas flavoHb-deficient E. coli express essentially no constitutive or inducible aerobic NO consumption activity [49]. NO metabolism by some mammalian cells is remarkably robust [16], suggesting independent enzymes. The ratio of NO and O₂ consumption rates of various cultured mammalian cell lines under normoxia ranges from 0.1 to 1.1. Similar to O₂ respiration, NO consumption is cyanide-sensitive [16,80]. However, inhibition of O₂ respiration and the function of the cyanide-sensitive cytochrome aa_3, with rotenone, antimycin A, myxothiazole, and oligomycin did not proportionally inhibit the NO consumption activity, thus indicating little, if any, role for cytochrome aa₃ in NO metabolism, contrary to what Moncada and coworkers argued [83]. Furthermore, cells showed an O₂-independent NO consumption activity that equaled 10–20% of the aerobic activity, suggesting the expression of an unidentified NOR, perhaps similar to the eukaryotic P450-type NOR [84]. Purportedly, cytochrome aa₃ does not function within cells as an NOR [85] as early measurements with the purified enzyme suggested [86]. More recently, Giles et al. demonstrated a lagging and CcO-independent NO metabolic activity of cardiac mitochondria with an estimated turnover of 1.3 s⁻¹ relative to CcO [87].

NO scavenging by globin-type NODs, various NORs, or reactive oxygen species now appears as an important mechanism for the NO resistance of terminal oxidases in microbes [70,82] and mammals [16,18,87,88]. However, this limited view ignores the important issue of what happens when terminal oxidases, or other metabolic pathways (e.g., cytochrome P450s), are indeed inhibited or essentially inactivated by excess NO accompanying inflammation (e.g., septic shock).

Several O₂ reductases, including the mitochondrial cytochrome aa_3, bind NO with high affinity [89,90] and are very sensitive to competitive inhibition by NO [16,29,30,82,91,92,93,94,95,96]. But, as already discussed, these O₂ reductases do not robustly metabolize NO within cells [16,48,87]. Other O₂ reductases, such as the di-heme cytochrome bd, are more resistant to NO inhibition [96,97,98,99]. Indeed, the cytochrome cbb₃ does reduce and metabolize NO in addition to O₂ [100]. In the case of heme-copper oxidases, including cytochrome aa₃ (CcO) and cytochrome bo₃, NO inhibition is rapidly reversible [29,37,94]. Yet, for these O₂ reductases, measured association and dissociation rate constants suggest that NO should bind tightly to the ferrous-heme intermediate in place of O₂ and inhibit the enzymes irreversibly. This underrecognized contradiction and paradox suggests either the absence of a ferrous-heme intermediate for NO and O₂ binding during normal cellular catalysis or a very rapid release of NO. A review of the investigations of mechanisms of NO inhibition and reversal provides potential explanations involving NO reduction or oxidation by the enzyme (Section 3.3). However, after analysis and pondering, none of these explanations are kinetically adequate or consistent, thus leading us back to the first explanation, namely, the paucity of the ferrous heme intermediate R in the catalytic cycle and the proposal of an unorthodox ferric-O₂ intermediate (Section 4.3.4) similar to that postulated in the NOD mechanism (Section 3.3.4) [18,74,75]. Supporting this supposition, the R intermediate is only significantly observed within intact mitochondria under severe hypoxia [101] and even more so with uncouplers and impaired membrane potentials [102]. A comparative analysis of the enzyme kinetics and structures of the two O₂-metabolizing heme enzymes, namely NOD and CcO, further indicates the necessity and likely operation of similar electron-transfer control motifs. For the O₂ reductases, a heme Fe³⁺O₂Cu²⁺ spintronic electron-transfer control also provides, in my judgement, an aesthetically satisfying mechanism for chemically, energetically, and structurally linking the countercurrent movements of electrons and protons.

2. Purpose and Scope

The NO sensitivity-resistance paradox provides a clear window for viewing the structure–function of NOD and CcO and opens up a unique perspective not provided by isolated kinetic measurements, spectroscopic intermediate studies, and static structures. A thorough and critical review of the investigations of NO binding and of the strategies that these two enzymes are thought to employ to resist a far more detrimental irreversible NO inhibition, as predicted from NO avidities, leads to the necessity of considering alternative mechanisms and introducing adequate concepts. Data and theory supporting an unusual, but apparently common, O₂-ferric heme interaction and enzyme intermediate that controls heme spin state, protonation, and electron transfer and that averts ferrous heme-NO formation are discussed and related to structural models, allostery, and motions. The concept and derived model provide a unique perspective for understanding proton pumping by CcO as well.

“No paradox, no progress”—Niels Bohr (1885–1962).

“For an understanding of the phenomena, the first condition is the introduction of adequate concepts; only with the help of the correct concepts can we really know what has been observed.”—Werner Heisenberg (1901–1976).

3. NOD Resistance to NO Poisoning

The NOD function, distribution, reaction, and mechanism of various members of the globin superfamily and their associated reductases have been widely reviewed, and the reader is referred to specific sources for additional information [5,18,55,56]. In my analysis, I consider the well-studied flavoHb-type NOD as an archetype for all NODs. While the function of globins as NODs has received considerable attention, the topic of NO resistance to irreversible inhibition has surprisingly received considerably little attention despite the glaring paradox it presents. As already stated, globins with associated reductases should not function as enzymatic NODs as initially envisioned [4,49] and more reservedly conceived [53,54]. A viable mechanism to explain the widely observed NOD activities and reversible NO inhibition has only recently emerged.

FlavoHbs catalyze NO dioxygenation with exceptional efficiency. At the core of the reaction, an enzyme-bound superoxide radical couples with the NO radical to form nitrate (Equation (1)). The bimolecular rate constant of the reaction estimated from k_cat/K_m (NO) is 2.4 × 10⁹ M⁻¹ s⁻¹, which approaches the diffusion limit and is 50–100 fold larger than the rate constant measured for HbO₂ and MbO₂ [19,20]. Globin-NODs serve the additional role of enzymatically catalyzing O-atom rearrangement in nitrate formation [17]. While tetrameric red blood cell HbFe²⁺O₂, or rather HbFe³⁺O₂⁻, catalytically dioxygenates NO, it acts stoichiometrically and not enzymatically. Herein lies the root of the paradox. An obvious difference is that the flavoHb-type NOD enzyme has a structurally-linked reductase domain, whereas the erythrocyte metHb reductase is not coupled to the oxidation event, at least not in the tetramer. Furthermore, in order to function in O₂ transport-storage, the red blood cell Hb and muscle Mb are reduced to the ferrous state regardless of the presence of O₂. The interactions of various globins with specific reductase partners are likely choreographed for the structural and functional motions required for NOD activity [18].

3.1. O₂ vs. NO Binding to Globins

As a rule, NO shows manyfold greater affinity for ferrous heme in globins than O₂ [4,89,103,104,105]. The E. coli flavoHb-type NOD is typical in this regard [4,51,62,75,106]. The bimolecular association rate constant measured for O₂, k’, is 3.3 × 10⁷ M⁻¹ s⁻¹, and the first-order dissociation constant, k, is 0.2 s⁻¹ (Equation (9)). The calculated half-time for O₂ dissociation is 3.5 s, whereas catalytic turnover occurs in roughly 7 ms at 20 °C and 1.5 ms at 37 °C. The calculated equilibrium dissociation constant, K_D, is 6.1 nM. Thus, by this criterion, half-saturation of the enzyme is expected at nanomolar O₂. But the K_m(O₂) value is 20 µM at 20 °C and 120 µM at 37 °C. NO associates with the ferrous heme with a roughly similar bimolecular rate constant of 3.1 × 10⁷ M⁻¹ s⁻¹ (Equation (10)). However, NO binds ~1000-fold more tightly with a dissociation rate constant of 0.0002 s⁻¹. The calculated K_D for NO is 6.5 pM, and the half-time for NO release is 58 min at 20 °C, or roughly 500,000 times longer than turnover at 20 °C. Our initially proposed enzyme mechanism [49] ignored the well-known high-affinity binding of NO to ferrous heme. Transient ligand binding kinetic measurements obtained in collaboration with John Olson’s lab [51,62] demonstrated high-affinity NO binding and produced additional glaring inconsistencies in the proposed mechanism that have long demanded an explanation. More recently, a resolution to the contradictions, inconsistencies, and discordant data and interpretations (Section 3.2 and Section 3.3) has been achieved with evidence of an O₂-sensing electron transfer (ET) switch and stepwise motions in the NOD mechanism (see Section 3.3.3) [18,74,75].

Fe²⁺ + O₂ ⇌ Fe²⁺O₂

(9)

k’ = 3.3 × 10⁷ M⁻¹ s⁻¹, k = 0.2 s⁻¹, t_1/2 = 3.5 s, K_D = 6.1 nM

Fe²⁺ + NO ⇌ Fe²⁺NO

(10)

k’ = 3.1 × 10⁷ M⁻¹ s⁻¹, k = 0.0002 s⁻¹, t_1/2 = 58 min, K_D = 6.5 pM

3.2. NO Inhibition of NODs

Steady-state kinetic investigations of the wild-type flavoHb and engineered mutants have been most revealing of the enzyme mechanism, in particular, where contradictions between the steady-state kinetics and elementary rate constant measurements have arisen. For example, NO reacts with the flavoHb ferrous heme with a rate constant of 2.6 × 10⁷ M⁻¹ s⁻¹ limited only by the rate of NO migration through the protein matrix to the iron, but the reaction of NO during steady-state catalysis is nearly 100-fold faster with a rate constant of 2.4 × 10⁹ M⁻¹ s⁻¹, as calculated from the k_cat/K_m(NO), thus suggesting an alternative pathway for NO migration. Indeed, substitution of the conserved G8 position valine in the putative gated tunnel for NO with a more bulky leucine residue decreased the k_cat/K_m(NO) value and presumably NO migration three-fold, while a smaller alanine increased the value ~two-fold [75]. Clearly, a consideration of motion in the protein structure during catalytic turnover is critical for a reliable interpretation. In this case, as in others, it was imperative to assign greater weight to the steady-state kinetic results than to the various potentially artificial, suboptimal, or preconceived states used for pre-steady-state transient kinetic measurements. For example, the measured K_m value of 20 µM for O₂ at 20 °C is 3300-fold greater than the K_D for O₂, and interaction or binding of O₂ paradoxically appears rate-limiting. At 37 °C, the K_m (O₂) value is even larger at 120 µM [51,62,75]. As stated above, the k_cat/K_m(NO) value is 100-fold larger than the value measured for NO association with the ferrous heme, supporting an NO migration route that opens for NO association during turnover [75]. And, most importantly, NO inhibits competitively with respect to O₂ and not irreversibly as suggested by the ferrous heme affinity measurements. NO inhibits the enzyme by 50% at an NO:O₂ ratio of approximately 1:100 at 37 °C and 1:20 at 20 °C. The inhibition pattern observed in 1/v vs. 1/O₂ plots [51,62,75] indicates competitive substrate inhibition in a ping-pong mechanism [107]. The observed cooperativity of O₂ and NO in steady-state kinetics and the effects of mutations demanded a mechanism that was better informed by structural dynamics [75].

3.3. Reversal of NO Inhibition of NODs

Most, if not all, oxy-globins catalyze NO dioxygenation, and globins tightly coupled to specific reductase partners often show enzymatic NOD activity [51,57,62,75,108,109,110]. Globins that are not tightly coupled to a reductase, such as the tetrameric red blood cell Hb, are reduced and irreversibly inhibited for subsequent NO dioxygenation with NO binding to the ferrous heme and thus prevent further O₂ binding [16,62]. On the other hand, globins that are tightly coupled to a reductase, such as the monomeric cytoglobin [110], are reversibly inhibited by NO in competition with O₂ [16,51,75,108,111,112]. NO inhibition is not progressive as expected for the small dissociation constants for ferrous heme-NO. The NO:O₂ ratio for 50% inhibition of the flavoHb-NODs and the flavoHb-like mammalian NOD activity, presumably a Hb monomer tightly coupled to the cytochrome b₅ or P450 reductase [18], is remarkably similar at approximately 1:100 at 37 °C [51,62,75].

Discerning reversible NO inhibition and sleuthing mechanisms of inhibition reversal has been hampered by non-uniform assay methods and conditional kinetic values. For example, the extensively characterized wild-type E. coli flavoHb-NOD [75] shows respective turnover numbers of 580 s⁻¹ and 160 s⁻¹ at temperatures of 37 °C and 20 °C. The respective K_m(O₂) values are 120 µM and 20 µM. In contrast, Mills et al. [108] independently reported respective V_max and K_m(O₂) values of ~10 s⁻¹ (calculated) and 47 µM for the E. coli flavoHb-NOD at 37 °C. Using kinetic equations derived for the E. coli flavoHb-NOD [75], we can estimate that the NOD activity would be inhibited >95% by the excessive and non-physiological 36 µM NO supplied at the O₂ saturation of 200 µM used in the assays. Rather low NOD activities of 4.5 s⁻¹ and 9.4 s⁻¹ have also been reported for Aspergillus oryzae flavoHbs at room temperature in a 50 mM sodium citrate buffer at pH 5.5 [113]. A Staphylococcus aureus flavoHb-NOD turnover number of 66 s⁻¹ was reported [114]. Frey et al. [57] reported NOD turnover numbers for various overexpressed bacterial flavoHbs in the range of 78–128 s^−1, including a value of 91 s⁻¹ for the E. coli flavoHb at room temperature with a uniformly inhibitory condition of 10 µM NO at normal O₂ saturation. The method of NO delivery may confound NOD assay results, given that NO is capable of inhibiting the NOD activity reversibly. Furthermore, under potentially uncoupled reducing conditions, NO may also inhibit NODs less reversibly or irreversibly. In future investigations, standardized methods and conditions for the NOD assay [115] should be used to ensure reliable kinetic comparisons and mechanistic interpretations.

3.3.1. NO Reduction by flavoHb-NOD

Numerous phenomena have suggested a NOR activity and function for flavoHbs. Such an activity (Equations (11)–(13)) would explain the NO resistance of the flavoHb-NOD. For full reversibility, the NOR turnover number would simply need to equal the NOD turnover number, and the NOD would never appear inhibited with respect to NO consumption. Moreover, in determinations of reaction stoichiometry, the NO-inhibited NOD would yield one N₂O per two NO instead of two NO₃⁻. There would be no O₂ dependence of the NOD activity. For a sub-NOD turnover rate, NO inhibition would appear progressive at all NOD velocities greater than the NOR turnover rate. Measured NOR turnover numbers of various flavoHbs range from 0.02 to ~1 NO s⁻¹ [51,62,64,107,108] or from 0.2 to 1% of the maximal NOD turnover. Clearly, flavoHb-NOR activity provides little, if any, role in the NO resistance of the NOD activity or reversal of NO inhibition.

Fe²⁺ + NO → Fe²⁺NO

(11)

Fe²⁺NO + e⁻ → Fe²⁺ + NO⁻ k = < 1 s⁻¹

(12)

2 NO⁻ + H⁺ → N₂O + OH⁻

(13)

In addition, the overexpressed flavoHb shows no measurable NOR activity and no protection of the NO-sensitive aconitase within anoxic E. coli [66]. Yet, observations by multiple investigators using different organisms have suggested an O₂-independent function for the inducible flavoHbs. The interested reader should consult the numerous relevant publications [49,60,64,66,108,114,116,117,118,119,120,121,122,123]. There are many possible explanations, yet no common activity or function has emerged. I have already introduced a couple of explanatory scenarios above (Section 1.3). The possibility of a two e⁻ N₂-yielding N₂O reductase activity (Equations (14) and (15)) requires testing. Many organisms express dedicated NORs that generate N₂O [3,70,84]. The abundant flavoHb, like copper and cobalamin enzymes, may bind and reduce N₂O to N₂. E. coli grown anaerobically with nitrate or nitrite reduces N₂O to N₂ [124], but the identity of the catalyst remains unknown. Univalent reduction of N₂O generates a toxic hydroxyl radical (Equation (16)) [2] and a gas-binding flavoHb with a two-electron reducing capacity is, as previously pointed out by Poole and coworkers [58], an attractive candidate.

Fe³⁺ + N₂O → Fe³⁺N₂O

(14)

Fe³⁺N₂O + 2 e⁻ + 2 H⁺ → Fe³⁺ + N₂ + H₂O

(15)

N₂O + e⁻ + H⁺ → N₂ + OH

(16)

3.3.2. Denitrosylase Mechanism of flavoHb

Denitrosylation or O₂ nitrosylation has been suggested as an alternative catalytic reaction mechanism of Ascaris Hb and flavoHb for O₂ and NO metabolism and detoxification [63,69]. The basic premise of the proposed mechanism is that heme-Fe²⁺NO or a nitrosylthiol of cysteine is the species that reacts with freely diffusing O₂ to form nitrate (Equations (17) and (18)). In the denitrosylase mechanism, formation of the heme-NO (Equation (11)) is not viewed as an irreversible inhibitory step. Rather, the ferrous-heme reduces and activates NO for reaction with O₂ (Equations (17) and (18)). The deficiencies of the proposed mechanism have previously been enumerated [5,18]. Chiefly, the rate constant for the reaction of heme-Fe²⁺NO or heme-Fe³⁺NO⁻ is grossly insufficient for both the proposed catalytic mechanism and for the reversal of NO inhibition of the NOD mechanism. NgbFe²⁺NO, MbFe²⁺NO, and HbFe²⁺NO show rate constants of <1 M⁻¹ s⁻¹ [125,126,127]. Nevertheless, interest and enthusiasm for the potential relevance of the reaction mechanism persists [104] with the demonstration of a larger rate constant of 6.3 × 10³ M⁻¹ s⁻¹ for the reaction of structurally unrelated ferrous nitrosylated nitrobindin with O₂ [128]. In terms of flavoHb-NOD activity, denitrosylation (Equation (18)) with the smaller rate constant of 1 M⁻¹ s⁻¹ falls short by over a million-fold for reversal of NO inhibition in the steady-state under the condition of 50% NOD inhibition, while the larger rate constant is insufficient by a factor of 4000-fold [18]. The rate constant required for the full reversal of inhibition (Equation (17)) is simply that measured for NO dioxygenation (Equation (2)) or 2.4 × 10⁹ M⁻¹ s⁻¹.

Fe²⁺NO ⇌ Fe³⁺NO⁻

(17)

Fe³⁺NO⁻ + O₂ → Fe³⁺ + NO₃⁻ k_ox = ~1 M⁻¹ s⁻¹

(18)

3.3.3. NO and O₂ Ligand Exchange Reaction or a Kinetic Aberration

In an attempt to resolve the NO inhibition paradox presented by the initially proposed NOD mechanism (Equations (1)–(6), (9), and (10)), Das and Meuwly suggested a possible mechanism for NO and O₂ exchange (Equation (19)) at the ferrous heme of Mycobacterium tuberculosis HbN-NOD [129]. However, neither the measured rate constants nor the structures support a simple ligand exchange or tunnel selectivity for NO and O₂ entry [130,131,132].

Fe²⁺NO + O₂ → Fe²⁺O₂ + NO

(19)

Following the first data-driven confrontation with NO resistance paradox [51], we considered the possibility that the large apparent K_m(O₂) value determined for the NOD activity relative to the unusual 4000-fold smaller K_D(O₂) value measured for the reduced flavoHb represented a competition of O₂ with NO that prevented NO binding and presumably irreversible NO inhibition. In this case, the true k’_NOD or k_cat/K_m(NO) value would have to greatly exceed the measured value of 2.4 × 10⁹ M⁻¹ s⁻¹ [51,75], which is near the diffusion limit for the reaction. Following the early reaction scheme design and kinetic modelling with some favorable rate constant estimates (e.g., k’_NOD of 0.6 × 10⁹ M⁻¹ s⁻¹ for 20 °C), a smaller but nevertheless troubling >10-fold discrepancy between the measured apparent K_m(O₂) value and the predicted K_m(O₂) value remained [51]. Those early kinetic discrepancies were finally eliminated with a revised reaction scheme (Section 3.3.4) that incorporates an O₂-sensing ferric heme trigger or switch for the ET step and O₂ activation [75]. The new scheme also explains reversible NO inhibition.

3.3.4. Heme-Fe³⁺O₂ Quantum State Switch in the flavoHb-NOD Catalytic Cycle

In the ensuing years of the unresolved paradox, significant advances in structure determinations, ET pathway modeling, and gas migration pathways had been achieved by investigators studying homologous globin-NODs, and those contributions were invaluable in formulating the structural model and resolving the discrepancies in the early model. Following our exhaustive kinetic and structural investigations of the E. coli flavoHb-NOD mechanism, a revised reaction scheme (Figure 1) that incorporates an O₂-sensing ferric heme trigger or switch for the ET step and O₂ activation was proposed [75], which, by all accounts, and the reticence, appears to have resolved the otherwise intractable paradox. In retrospect, intimate knowledge of the O₂⁻, O₂, and NO-mediated inactivation and dynamic reactivation (i.e., inhibition) of [4Fe-4S]²⁺ (de)hydratases [133], where ferric iron is clearly involved, should have led to the model (Figure 1) from the start.

In the NOD catalytic cycle model, ET transfer to the ferric heme is activated by weak O₂ and NO SOC [18,74]. Kinetic and structural evidence support controlled ET from the FADH₂ to the ferric heme (Figure 1, step 6). The existence of a spin crossover (SC)-driven switch (Figure 1, step 5) is supported primarily by theory but also by the combined effects of O₂ and key side-chain substitutions on the ferric heme spin state, steady-state kinetics, and ligand binding kinetics [74,75]. The stepwise ordered motions and kinetics of the ligands and flavoHb-NOD structure during the catalytic cycle have been previously described in detail elsewhere [18,75] and are briefly summarized in the legend to Figure 1

Another aspect of the model requiring emphasis here is the role of stored energy in the proposed ET switch mechanism. The ET switch requires protein motions and an estimated +2 kcal/mol of energy [74]. The NO dioxygenation reaction (step 7) produces NO₃⁻ and −30 kcal/mol of free energy that is utilized in the ET switch mechanism. The NO₃⁻ anion is produced in the hydrophobic distal pocket and migrates to the CD-loop and is captured in an anion hole of the unfurled spring-like CD loop (step 8). Energy is presumably stored through steric repulsion forces in the loop. To trigger the ET switch, O₂ must displace the distal leucine (E11) residue away from its van der Waals contact with the ferric iron. For this motion, O₂ is assisted by the structurally stored spring energy of the CD loop with loop re-furling and NO₃⁻ expulsion (step 2). The protein motions required for O₂ to set the trigger for the ET switch constitute the slow step in the catalytic cycle. In Figure 2, NOD cycle clocks are shown for 20 °C and 37 °C, with the approximate timing of each step of the cycle denoted for the respective 6.7 ms and 1.5 ms turnover times. In the model, transient O₂ binding to the ferric heme slowly occurs following the duration of step 2, and it is during this step that the primary gate and short tunnel for rapid NO entry also opens (step 3). Reversible NO inhibition occurs when NO competes with O₂ for transient binding to the ferric heme (Equations (20) and (21)).

heme-Fe³⁺ + NO ⇌ heme-Fe³⁺NO

(20)

heme-Fe³⁺ + O₂ ⇌ heme-Fe³⁺O₂

(21)

The respective on and off rates measured for NO and the resting ferric flavoHb heme are 4.4 × 10⁷ M⁻¹ s⁻¹ and ~4000 s⁻¹ at 20 °C, and the estimated K_D(NO) value is relatively large at ~100 µM [51]. The calculated half-time for NO dissociation is 170 µs, or a 20 million-fold shorter time than measured for the ferrous heme. However, from steady-state kinetic and structural studies [75], we have determined that NO accesses the heme via a gated short tunnel during turnover (step 3) with a much larger rate constant of 1.6 × 10⁹ M⁻¹ s⁻¹ at 20 °C, thus revealing a smaller and more relevant K_D(NO) value of 2.5 µM. From the kinetic competition of NO with O₂ and a 50% inhibition at an NO:O₂ ratio of ~1:20 at 20 °C, we approximate a K_D(O₂) value of 50 µM for catalysis, which is somewhat larger than the 20–27 µM K_m(O₂) value determined for 20 °C [51,75]. The estimated K_D(NO) and half-time for NO dissociation can explain both the potency and rapid reversibility of NO inhibition. Interestingly, the measured rate constant for O₂ entering the 20 Å long hydrophobic tunnel of the resting reduced flavoHb and binding the ferrous heme is 3.3 × 10⁷ M⁻¹ s⁻¹ [51] or ~49-fold smaller than for NO entry during turnover. If we assume a similar rate constant for O₂ entry and interaction (Figure 1, step 1) and a K_D(O₂) value of 50 µM, the estimated off rate for O₂ is 1500 s⁻¹ at 20 °C, and the half-time for O₂ dissociation is roughly 0.5 ms or 7000-fold shorter than for the ferrous heme. Through this analysis, it becomes clear why the SOC of NO with the heme-Fe³⁺O₂ complex (step 5) is essential for the ET switch mechanism. Dissociation of O₂ from the heme following reduction would allow NO to bind and inhibit irreversibly. Observed cooperativity of NO and O₂ in the steady-state has also suggested an interaction between NO and O₂ that triggers the ET [75].

The proposed O₂ binding to ferric heme is unusual but is supported by theory and experiment [134,135,136,137]. Bonding of O₂ to the bare ferric heme is thought to occur through the overlap of the filled or half-filled O₂ 2π*_‖ orbital with the empty iron 3d²_z orbital and through back-bonding of the half-filled 3d_yz orbital with the half-filled O₂ 2π*_⊥ orbital [135,137]. In flavoHb and other globins, the same orbital overlap, electrostatics of ferric iron, and hydrogen bonding from distal residues strengthen the weak O₂ interaction [74]. Calculated K_D(O₂) values are consistent with the calculated bond energies and measured K_m(O₂) values [74].

In addition to weak electrostatic bonding, ferric heme and O₂ interact magnetically through SOC. Similar to spintronic O₂ reduction devices, SOC likely contributes to bonding through electron density currents, and ferromagnetic coupling reportedly favors currents more than antiferromagnetic coupling [76,77,78]. We can also find analogies of the weak electrostatic bonding in the SOC of O₂ and ferric iron in advanced quantum dot theory. In electrostatic coupling between quantum dots, only the charge degrees of freedom of the electrons are involved. Within each quantum dot, the spin degree of freedom is then coupled to the charge degree of freedom via SOC. SOC is expected to be first order in the electrostatic interaction and second order in the spin [138]. Hence, electrostatics and SOC will interact in the bonding of O₂ to the ferric iron, with the charge degrees of freedom, electron densities, and electrostatic forces dominating, with spins playing a secondary role by influencing electron currents and densities.

Energy differences in the ferric heme spin states (s = 1/2, 3/2, and 5/2) are relatively small but show preferences in ligand binding presumably due to electrostatics and SOC-induced charge currents. O₂, NO, and water bind to low-spin states (s = 1/2 and 3/2) and anions with greater charge density preferably bind with ferric iron in the high-spin state (s = 5/2) [139]. Free heme stably binds O₂ in the s = 3/2 spin state [134,137]. In the E. coli flavoHb, the energy differences between the ferric heme spin states are also small, with an estimated +2 kcal/mol difference between the s = 1/2 and s = 5/2 states, with smaller energy differences between the s = 1/2 and s = 3/2 states [74]. In the proposed O₂ sensing ET switch activation mechanism (Figure 1, step 5 and Figure 3), O_2, with its two half-filled parallel spin 2π* orbitals, binds and spin couples with the half-filled iron 3d orbital antiferromagnetically. SOC induces SC of the iron to the s = 3/2 state, achieving more favorable antiferromagnetic spin–spin and electrostatic interactions. SOC between metal atoms and O₂ increases electron density currents to O₂ as evidenced by low electrode overpotentials in O₂ reduction reactions [76,77,78] and would thus increase O₂ bonding. In the model, due to the collision of NO (s = 1/2) with the O₂ molecule, NO loses angular momentum, gains spin (s = 3/2) [140], ferromagnetically couples with O₂, induces SC of the iron to the high spin s = 5/2 state to achieve spin parity in the antiferromagnetic SOC of iron with O₂ and NO, and the heme propionates and via the strong heme dipole, which transfers electron density to O₂. Any loss of electron density from the propionates would decrease the carboxylate group pKa values and favor ET. The effect is somewhat analogous to the Bohr effect, where increased O₂ binding correlates with the loss of Hb protons and protonation of Hb decreases O₂ binding [141]. The change in iron spin also causes or correlates with the ruffling and saddling of heme in general [142] and Hb heme [143,144].

In the model, the energy from NO dioxygenation is stored in the flavoHb CD loop structure (Figure 1, steps 7 and 8), which energizes the protein and the ET switch mechanism (step 5) in the ready position for the high-spin (s = 5/2) ferric heme (Figure 3) triggered motion. NO may not stay coupled with O₂ during the duration of the ET switch mechanism and ET. Regardless, a reaction of internalized NO with the Fe³⁺O₂⁻ complex (step 7) is evidenced by a steady-state k’’_NOD value of 0.9 to 1.9 × 10¹⁰ M⁻¹ s⁻¹ that exceeds the calculated diffusion-limit for free NO and flavoHb-Fe³⁺O₂⁻ [75]. Chung et al. have discussed the important role of SOC in the NOD-catalyzed reaction (step 7) and other O₂ activation reactions [145]. A similar ultrafast reaction rate constant of NO with free O₂⁻ of 1.6 × 10¹⁰ M⁻¹ s⁻¹ has been measured in what appears to be a diffusion-restricted caged reaction, although Koppenol has argued that the constant represents a free diffusion reaction [146].

A survey of the flavoHb-NOD and other globin-NOD structures has revealed a common motif in the mechanism of ET switching [18]. An ET pathway via the heme propionate d is influenced and ‘switched’ by the charge status and electrostatic field at the carboxylate group. In the model, SC from low-spin to high-spin ferric heme (step 5) with saddling and vice versa with flattening is thought to instigate the motions in the protein that affect the charge status and electrostatic field that turn the ET switch ON and OFF [74].

Comparisons of the E. coli and Saccharomyces cerevisiae flavoHb X-ray crystallographic structures [147,148] have provided insight into the ET path [149,150] and ET switch mechanism. In the proposed ferric flavoHb-NOD switching mechanism [18,74], the S. cerevisiae flavoHb structure models the ON state and the E. coli flavoHb structure models the OFF state (Figure 4). In the OFF state conformation, the water molecule O-atom p orbitals create an electrostatically repulsive Coulomb blockade for ET between the 5.6 Å distant FAD C8 methyl group C atom and the heme d propionate carboxylate O-atom. In the ON state conformation, the blockade water molecule moves, and the positively charged lysine (F7) ammonium group binds and neutralizes the negatively charged heme d propionate carboxylate group for ET. In addition, the flavoHb nine amino acid carboxy-terminus forms a short 3₁₀ helix with the positive end of its electrostatic dipole oriented toward the neutralized carboxylate group and favoring the ET [18]. Motion of the switch to the ON state (Figure 1, step 5) is triggered by saddling of the high-spin heme and motion of the iron atom toward the proximal histidine (F8) that conveys rotational and translational forces on the F-helix. Motion is mostly driven by the energy and forces stored in the O₂-activated conformation (step 2) and triggered by the NO-induced SC (step 5). The switch returns to the OFF state following ET (step 6) and NO dioxygenation (steps 7 and 8). Kinetic and mutagenesis studies support the structural model [74,75].

Figure 4. Proposed ET switching mechanism. Key structural elements of the switch mechanism are shown for the E. coli flavoHb (PDB ID: 1GVH) (Left) and S. cerevisiae flavoHb (PDB ID: 4G1V) (Right). Reacting O₂ and NO molecules drawn into the structural model (Right) are shown at 40% of their van der Waals radii. Water molecule shown in aqua and all other colors follow the CPK scheme. Iron and water O-atoms shown at 50% and ~100% of their van der Waals radii, respectively. Pathway for ET shown by the green arrow. Carboxy-terminal 3₁₀-helix dipole shown as a magenta arrow pointed from the negative to the positive pole of the dipole (Right).

Figure 4. Proposed ET switching mechanism. Key structural elements of the switch mechanism are shown for the E. coli flavoHb (PDB ID: 1GVH) (Left) and S. cerevisiae flavoHb (PDB ID: 4G1V) (Right). Reacting O₂ and NO molecules drawn into the structural model (Right) are shown at 40% of their van der Waals radii. Water molecule shown in aqua and all other colors follow the CPK scheme. Iron and water O-atoms shown at 50% and ~100% of their van der Waals radii, respectively. Pathway for ET shown by the green arrow. Carboxy-terminal 3₁₀-helix dipole shown as a magenta arrow pointed from the negative to the positive pole of the dipole (Right).

Analogies to the proposed switching mechanism can be found in the low-to-high spin state change in cytochrome P450 BM3 with the binding of the fatty acid substrate laurate [151,152]. An effect on heme redox potential has been suggested to be important (for ET) with a 100 mV more oxidizing high-spin state heme [153]. Also, with laurate binding and SC, water moves to a nearby position for a potential proton donor role in O₂ bond scission [154].

A direct analogy and scientific precedent to the globin-NOD structural changes with O₂ ligand binding and heme spin state changes is also found in the classic studies of the red blood cell Hb cooperativity and allostery [42,143,144,155,156]. However, in the well-studied Hb, the more familiar ferrous heme undergoes low-spin to high-spin crossovers that affect O₂ binding and the ‘relaxed’ and ‘tense’ globin structures. Whether O₂ affects structure or structure affects O₂ binding has been debated, but it is clear that both are thermodynamically related. In the case of the globin-NODs, energy stored in the CD loop spring-like structure appears critical for the O₂- and NO-triggered structural changes [18,74,75]. On close inspection, the core of NOD enzyme structural motions appears precursory to the evolution of the O₂-transporting red blood cell Hb cooperativity and allostery [75,155,156]. Indeed, the long-storied history and dedicated efforts in understanding the O₂ cooperativity, motions, and allostery of Hb [42] have facilitated investigations and understanding of molecular, atomic, and electron motions in the globin-NOD mechanism [18].

4. CcO Resistance to NO Poisoning

Heme-copper cytochromes aa₃, bo, and ba₃ are all sensitive to reversible NO inhibition, while the cytochrome bd-type oxidases show far less sensitivity [96,97,98,99,157]. Sensitivity to NO inhibition has been attributed to the slower release of NO from the binuclear heme-copper type oxidases compared to the cytochrome bd-type oxidases [157]. The cytochrome bd-type oxidases utilize heme iron for O₂ or NO binding, lack a distal metal, but bear glutamate residues both distal and proximal to the heme iron that apparently increase the dissociation of NO. The rapidly reversible NO inhibition of CcO and other heme-copper oxidases prompted my foray into the literature and proposal of a mechanism for resistance similar to that described for the flavoHb-NOD (Section 3). To my knowledge, the evidence for a role of the oxidized CcO in binding O₂, NO, and CO during catalysis, like the flavoHb-NOD, has been only cursorily considered and discussed before [74].

4.1. O₂ vs. NO Binding to CcO

Long after Keilin and Hartree’s seminal studies of reduced cytochrome oxidase and O₂ [158], Greenwood and Gibson put forth a mechanism in 1967 for O₂ binding BNC_red [159] and later amended the mechanism to allow potential O₂ binding to Cu⁺_B in competition with CO [90]. O₂ binds to BNC_red rapidly with a second-order rate constant of 0.4 to 3.5 × 10⁸ M⁻¹ s⁻¹ (Equation (22)) [90,159,160,161]. NO also binds to BNC_red with a roughly equivalent rate constant to that for O₂ at 0.4 to 1 × 10⁸ M⁻¹ s⁻¹ (Equation (23)) [90,161]. An unusually large dissociation rate constant for O₂ was approximated at 5 × 10⁴ s⁻¹ when the rapid reduction of O₂ by BNC_red was accounted for [90]. The NO dissociation constant is much smaller, with reported values of 0.004 s⁻¹ at 20 °C and 0.012 s⁻¹ at 37 °C [93]. The calculated equilibrium dissociation constant (K_D) values for O₂ and NO at 20 °C are roughly 500 µM and 40 pM, respectively. These early investigations by Gibson, Greenwood, and others revealed that following O₂ binding to BNC_red, O₂ is reduced within roughly a millisecond [90,159,160,161]. In contrast, following NO binding to BNC_red, NO dissociates very slowly with a half-time of ~1.5 min. The early data and proposed CcO enzyme mechanism [159] provided a clear, albeit potentially misleading (see Section 4.3), mechanism for envisioning irreversible poisoning of CcO and respiration by NO.

In retrospect, the large K_D estimated for O₂, as well as the early observation of Antonini et al. of ‘pulsed’ and ‘resting’ enzyme state effects on electron-transfer rates to cytochrome a₃ and O₂ [162], also revealed important deficiencies in the simple model for O₂ binding and reduction. From these early investigations, it was apparent that an activation step for ET was required and that the O₂ affinity, as measured and calculated, was much lower than expected for ferrous heme binding. Standing alone, these two early observations demanded an explication (see Section 4.3.5) that is not provided by current models of the CcO mechanism [163,164]. Granted the enormity and complexity of CcO investigations, some observations will remain unexplained. However, these two early observations, activated ET and low O₂ affinity, appear fundamental and still require clarification.

Fe²⁺Cu⁺ + O₂ ⇌ Fe²⁺O₂Cu⁺ ⇢ ⇢ O₂ reduced + Fe³⁺ Cu²⁺

(22)

k’ = 1 × 10⁸ M⁻¹ s⁻¹, k = 5 × 10⁴ s⁻¹

Fe²⁺Cu⁺ + NO ⇌ Fe²⁺NOCu⁺

k’ = 0.4–1.0 × 10⁸ M⁻¹ s⁻¹

(23)

k = 0.0025 s⁻¹ @ 20 °C and 0.012 s⁻¹ @ 37 °C; t_1/2 = 58 s and 277 s

4.2. NO Inhibition of CcO

In 1994, Cleeter et al. [30] demonstrated potent NO inhibition of CcO in quick freeze-thawed mitochondria using the NO-generator S-nitrosoglutathione and suggested reversible binding to the BNC copper to account for the previously reported negligible effect of NO-releasing macrophages on CcO activity following cell harvests [10,165]. Brown and Cooper demonstrated potent inhibition of respiration and CcO by authentic NO within cells and, more surprisingly, showed that NO inhibition is indeed competitive with respect to O₂ and is rapidly reversible [29]. NO was observed to inhibit CcO by 50% at NO:O₂ ratios as low as 1:500 in neuronal cells [29], but ratios as large as 1:40 have been reported with more oxidizing redox states of mitochondria and CcO [95]. With observations of higher NO:O₂ ratios and weaker inhibition, an important additional consideration is NO removal by autooxidation and non-enzymatic or enzymatic NO detoxifying pathways [16,88]. Remarkably, in order for NO and O₂ to appear competitive, NO must be released from binding with a rate constant equal to or exceeding CcO turnover or on the order of 100 s⁻¹. Otherwise, NO would rapidly and irreversibly inhibit the CcO as indicated by the on and off rates for NO binding to reduced CcO (Section 4.1). In some reports, NO inhibition is more slowly reversed (t_1/2 = ~15 s), suggesting multiple mechanisms for inhibition and/or multiple mechanism(s) for release from NO inhibition (Section 4.3) [88,92]. Inhibition reversal mechanisms involving either oxidative or reductive NO metabolism have the potential to also provide a mechanism for significant cellular or mitochondrial NO consumption via the relatively abundant CcO [80,166]. Bona fide NO metabolizing enzymes, such as the NODs, also require careful attention in the assessment of CcO-catalyzed NO metabolism and NO inhibition reversal kinetics within cells or mitochondria [16,87].

4.3. Reversal of NO Inhibition of CcO

In 1980, Gary Brudvig and Sunney Chan first reported reductive and oxidative reactions of CcO with NO, generating either N₂O or nitrite and involving the BNC [86]. Following Brown and Cooper’s seminal report on the NO sensitivity of CcO, the competition of NO and O₂, and rapid reversibility of NO inhibition [29], several independent groups focused their investigations on the reductive (Section 4.3.1) and oxidative (Section 4.3.2) mechanisms for NO metabolism by CcO that may account for the unexpected reversal or resistance to NO inhibition and that may also contribute to significant NO metabolism by mitochondria, cells, and tissues [80,81,85,91,92,93]. In their culminating investigation, Cooper and coworkers proposed a mixed inhibition model to explain reversibility of inhibition in which NO can either bind reduced heme and dissociate (competitive) or bind cupric ion to be oxidized (non-competitive) [94]. Indeed, Torres et al. reported NO-induced CcO Soret band increases at 441 nm during turnover and assigned the increase to a₃-Fe²⁺NO complex formation with a Soret maximum of 444 nm [91]. Similar spectra and interpretations were reported by Sarti et al. [93]. In retrospect, it is noteworthy that the Soret maxima for the flavoHb heme-Fe²⁺NO and Fe³⁺NO complexes are found at nearly indistinguishable wavelengths of 420 nm and 419 nm [51,62], thus suggesting an ambiguity. The possibility of NO competing with O₂ for binding a CcO cycle intermediate with an oxidized Fe³⁺Cu²⁺ center (Section 4.3.3) does not appear to have been considered [93,94,167,168,169,170]. Since the early investigations of reactions of NO with BNC_ox [86,171], subsequent investigations appear to have focused solely on the oxidative Cu²⁺ reaction with NO [81,166] and ignored the likelihood and significance of NO binding or transiently interacting with a₃-Fe³⁺ (Section 4.3.2).

4.3.1. NO Reduction by CcO

In 1996, Borutaité et al. suggested a role for CcO in mitochondrial NO reduction [80] that was consistent with the in vitro observations of Brudvig and Chan [86]. Brudvig and Chan had reported N₂O formation following prolonged incubation of CcO with 1 mM NO, 10 mM ascorbate, and the ET mediator p-phenylenediamine and suggested an enzymatic NOR mechanism (Equation (24)). Furthermore, the recognition of the evolution of CcO from the cytochrome bc-type NOR [3,40,41] had suggested an ancestral and residual NOR function of CcO. However, Sarti’s group subsequently reported that purified CcO has no detectable NOR activity [85], and a similar conclusion was recently reported by Balaban’s group for cardiac muscle mitochondria [87]. Clearly, the low or negligible NO reduction activity of CcO would be inadequate to rapidly reverse and prevent progressive NO inhibition of the O₂ reduction activity in the steady-state. Other heme-copper oxidases including the cytochrome ba₃ and cytochrome caa₃ of Thermus thermophilus do catalyze the reduction of NO [172]. The Pseudomonas stutzeri cytochrome bb₃ oxidase with significant homology to the cytochrome bc-type NOR also shows a notable NOR activity of approximately 2 s⁻¹ [100].

Fe²⁺NOCu⁺NO + 2 e⁻ + 2 H⁺ → Fe²⁺Cu⁺ + N₂O + H₂O

(24)

The possible NOR function of CcO and other O₂ reductases has also been important for delineating pathways for NO metabolism in cells and mitochondria especially under the condition of hypoxia or anoxia [16,80,83,87], where NOD activities or MbO₂-catalyzed NO dioxygenation fail [16,87,88,173] and knowledge of NORs is lacking. The reported failure of CcO to reduce NO [85,87] contradicts the findings and interpretations of Borutaité et al. [80] and suggests the existence of other mechanisms.

4.3.2. NO Oxidation by CcO

The results of numerous investigations by two independent groups and schools of thought over a period of more than two decades have supported a CcO-catalyzed oxidation of NO to nitrite as an important mechanism for NO inhibition reversal [81,88,166,167,169,174]. In the mechanism, NO is proposed to react with cupric ion in the oxidized intermediate ‘O’ with a rate constant of 10⁴ to 2 × 10⁵ M⁻¹ s⁻¹ and to rapidly form a cuprous-nitrite complex (Equation (25)) [88,175]. Reduction of BNC_ox rapidly ejects the bound nitrite (Equation (26)) [81,166,174] and produces the reduced CcO for O₂ binding in the catalytic cycle. The oxidation reaction is relatively slow at physiological NO concentrations, temporarily removes a fraction of the CcO from the O₂ reduction cycle, and produces non-competitive inhibition in steady-state kinetic analysis primarily at higher [NO] [94]. In the proposed model, it is the fully reduced CcO intermediate R that is thought to be susceptible to competitive inhibition by NO (Equations (22) and (23)) [88,94].

Recall, however, that the calculated half-time for NO dissociation from the reduced heme is 58–173 s. The inhibition by competitive binding of NO to BNC_red would be progressive and essentially irreversible in the time frame of the CcO kinetic assays barring the progressive non-competitive inhibition expected with NO oxidation [94]. There stands a glaring contradiction in the current models suggesting reversible competitive NO inhibition of reduced intermediate R with relatively small rate constants for NO dissociation [88,94,95,169]. It is noteworthy that Antunes et al. applied a 10-fold larger NO dissociation rate constant value of 0.13 s⁻¹ and a 2.5-fold smaller NO association rate constant for an overall 25-fold smaller k_NOon/k_NOoff ratio in their kinetic modeling of simple reversible competitive inhibition of the reduced CcO [95]. The smaller ratio undoubtedly produced a more modest concave curvature and less progressive inhibition in the plots and simulations of -d[O₂]/dt.

Fe³⁺H₂OCu²⁺ + NO ⇌ Fe³⁺Cu⁺NO⁺ + H₂O → Fe³⁺Cu⁺NO₂⁻ + H⁺

(25)

k’ = 10⁴–2 × 10⁵ M⁻¹ s⁻¹

Fe³⁺Cu⁺NO₂⁻ + e⁻ → Fe²⁺Cu⁺ + NO₂⁻

(26)

While it has been argued that kinetic experiments showing large >100 s delays in the full reversal of inhibition of the isolated CcO following rapid NO depletion with HbO₂ support the formation of Fe²⁺NOCu⁺ [93], NO inhibition of respiration and CcO within mammalian cells is immediate, is not progressive, correlates contemporaneously with NO concentrations, and is reversed fairly rapidly [16,29]. In this case, CcO may be forced into a reduced state by unfavorable catalytic conditions (e.g., low O₂ or dysregulated ET) to produce a poorly reversible NO inhibition that appears as a rapid and progressive inhibition under sub-optimal conditions [92], produces bona fide Fe²⁺NOCu⁺ spectra [92], and shows small NO dissociation rate constants [93]. Similar effects of dysregulated ET are observed in assays of globin-NODs with small NO dissociation constants with heterologous reductases [111] or a non-functional dysregulated ET switch with glutamine substituted for the F7 lysine residue (Figure 4) [75]. A greater understanding of the controls on ET in CcO and the factors influencing switching are required (see Section 4.3.3). Again, recall that the R intermediate was reportedly not detected in actively respiring mitochondria but only seen in extremely hypoxic non-respiring mitochondria [101]. The known bimolecular association rate constants may be insufficient for a reaction with a low-abundance intermediate and low ligand concentrations within the time frame of turnover. For k’(NO) = 1 × 10⁸ M⁻¹ s⁻¹ and at 10 × 10^–3 s per cycle, the [R] would need to be ≥ 2 µM for 50% inhibition.

It should be noted that Pearce et al. [176] and Collman et al. [177] have argued for an entirely different NO inhibition reversal mechanism involving O₂ binding and reduction at the reduced Cu⁺, followed by the reaction of the O₂⁻ formed with Fe²⁺NO to form peroxynitrite. Moreover, it was further suggested that CcO catalyzes the reduction of peroxynitrite to nitrite [176,178]. Furthermore, the authors suggested that CcO significantly metabolizes NO via peroxynitrite formation and reduction within mitochondria and cells [176,179]. The extensive investigations of the reactions of NO with CcO to form nitrite have established a mechanism (Equations (25) and (26)) fully supported by enzyme kinetics [94] that does not support any additional roles for peroxynitrite or O₂⁻.

Finally, it is important to appreciate that only trivial amounts of NO are removed from CcO, mitochondria, and cells by the nitrite-generating NO inhibition reversal mechanism. CcO shows little, if any, role in the metabolism of NO by mammalian cells [16] or E. coli [48]. Recall that CcO was initially considered potentially important for NO metabolism by cells and mitochondria [16,80,83,166]. In fact, it was the supposition that initiated and supported my initial work on NO metabolism [16]. It is now well-appreciated that dedicated NODs and NORs metabolize NO efficiently and, at least partially, protect the NO-sensitive CcO [16,87] and other O₂ reductases [82] from both a rapidly and slowly reversible NO inhibition.

4.3.3. Heme-Fe(III)O₂ Quantum State Switch and the CcO Catalytic Cycle

The similarities in the NO resistance of the NOD and CcO functions have suggested a common enzyme strategy involving O₂, NO, or CO binding or interaction with ferric heme. With want of extensive knowledge of the complex structures and mechanisms of CcO and other terminal oxidases, I hesitantly, but with further study, ardently set my sights on developing an internally consistent model for the CcO catalytic cycle. The resulting model introduces an ET control mechanism and builds upon the principles encountered or developed in investigations of the globin-type NOD structure–function, electron transfer, kinetics, and O₂ chemistry (Section 3). Concurrently, in order to better explain the proton pumping mechanism first demonstrated by Mårten Wikstrom in 1977 [180], I describe a potential proton tunnel from the matrix, and I also introduce an O₂ and reduced oxygen intermediate-driven countercurrent proton-coupled electron transfer (PCET) pathway. A de novo model-building approach that I, like others [163,164,168,181,182,183,184,185], have taken necessitates the subsequent pursuit of the following critical questions.

(i)

Can the model explain all empirical observations?

(ii)

Does the model agree with existing theory, or overlap with other models?

(iii)

Does the model suggest definitive experiments?

(iv)

Does the model need to be revised or rejected?

The rapid reversibility of the NO inhibition of CcO and competition with O₂ suggested a CcO catalytic cycle and mechanism in which O₂ binds weakly and interacts transiently with BNC_ox to initiate turnover (see Figure 5 and discussion below). Indeed, in 2011, Wikström and coworkers used DFT calculations to conclude that O₂ binding to BNC_ox is “nearly isoergonic with dioxygen binding to a reduced BNC” [186]. Apparently, the authors did not grasp the implication of univalent O₂ reduction (Fig. 1A in Kaila et al.) for CcO catalysis (see Figure 5) or rejected the possibility due to (i) thermodynamic considerations, (ii) a failure to observe UV–visible spectral changes (Fig. 6 in Kaila et al.), (iii) an unawareness of the role of SOC in redox reactions and the burgeoning field of spintronics, and (iv) unforeseen mechanisms for the membrane potential to affect proton tunneling, O_W (or O_H) dehydration, and ET (see below). In this model, NO competes with O₂ for the oxidized center, but only O₂ with two parallel spin (s = 1/2) 2π* electrons is able to couple with ferric heme iron (s = 1/2) and cupric ion (s = 1/2) in the center (O_D) to trigger SC in the low spin a₃ heme (see Figure 6 and discussion below) and to facilitate electron transfer to the ferric heme via arginine 438 and the d propionate of the ferric a₃ heme (see Figure 7 and discussion below). In the full model, the ET switch mechanism constitutes the energetic proton pump in an orchestrated system with an electrostatically-driven proton influx tunnel and an exiting PCET pathway of the concerted EPT type [187], with bifurcating unidirectional proton and electron current controls (see Section 4.3.4). The beginnings of the model can be found in an earlier pondering over the evidence for O₂ binding to BNC_ox [74].

CcO Catalytic Cycle

The proposed CcO catalytic cycle (Figure 5) is simply envisioned as four univalent O₂ reduction steps, each coupled with a pumped proton and with each of those steps followed by a protonation of the reduced oxygen intermediate produced in the prior step. The reduction of O₂, or a reduced oxygen intermediate, releases the free energy that is required to pump the proton against the pH gradient. The proposed cycle links each O₂ or oxygen intermediate reducing electron with each proton in time and space, potentially resolving the apparent timing conundrum [188,189].

Figure 5. Proposed CcO catalytic cycle. Competitive and non-competitive inhibition pathway for NO originating at BNC_ox (top) with O₂ binding BNC_ox followed by cycle intermediates. The cycle model is formulated with alternating successive exergonic reductive-proton pumping steps (red brackets) and proton-consuming steps (black). Standard free energy changes shown for each reduction step are from electrolytic O₂ reduction measurements [190] or are calculated (*, see text). Iron spin states shown are, in some cases, unknown or hypothetical. In the model, iron and copper ions are shown in their oxidized states with the assumption of SOCs and electron density distributions between the metal ions and the reduced oxygen intermediates. The nomenclature of intermediates is regularized for a simplified model description where O is oxidized, R is reduced, S is superoxide, P is peroxide, and F is ferryl. The subscript A designates the anionic form, and the subscript H designates the protonated state. D stands for dioxygen, and T represents tyrosyl. Corresponding common designations for cycle intermediates [163] are given in parentheses: O_W (O hydrated), S_A (A), F_A (P_R), F_H (F), O_T (P_M), O_A (O_H), and R, (R).

Figure 5. Proposed CcO catalytic cycle. Competitive and non-competitive inhibition pathway for NO originating at BNC_ox (top) with O₂ binding BNC_ox followed by cycle intermediates. The cycle model is formulated with alternating successive exergonic reductive-proton pumping steps (red brackets) and proton-consuming steps (black). Standard free energy changes shown for each reduction step are from electrolytic O₂ reduction measurements [190] or are calculated (*, see text). Iron spin states shown are, in some cases, unknown or hypothetical. In the model, iron and copper ions are shown in their oxidized states with the assumption of SOCs and electron density distributions between the metal ions and the reduced oxygen intermediates. The nomenclature of intermediates is regularized for a simplified model description where O is oxidized, R is reduced, S is superoxide, P is peroxide, and F is ferryl. The subscript A designates the anionic form, and the subscript H designates the protonated state. D stands for dioxygen, and T represents tyrosyl. Corresponding common designations for cycle intermediates [163] are given in parentheses: O_W (O hydrated), S_A (A), F_A (P_R), F_H (F), O_T (P_M), O_A (O_H), and R, (R).

In the proposed catalytic cycle (Figure 5), I have renamed some of the well-known cycle intermediates with subscript A representing an anion and subscript H representing the protonated form for the sake of logic and simplicity in presentation and discussion. The series of univalent reductions of intermediates in catalytic sequence is named for the oxidized center with water (O_W), oxidized center with dioxygen (O_D), oxidized center with superoxide anion (S_A), oxidized center with peroxide anion (P_A), ‘oxidized’ center with ferryl (s = 1) [191], hydroxide anion bound to Cu²⁺ (F_A), and the oxidized center with hydroxide anion (O_A). Included as intermediates of protonation are the corresponding S_H, P_H, F_H, and O_W. The existence of the tyrosyl radical intermediate [192] was accounted for and included as an oxidation product or side-reaction of ferryl (F_H), named O_T, and is shown in an equilibrium with F_H. As an anion, O_T is proposed to serve a critical function in proton pumping (see below). For the simplified nomenclature, proposed intermediates roughly correspond to empirically characterized intermediates with traditional nomenclature [163,164,168,184,193,194] as follows: O_W corresponds to hydrated O, S_A corresponds to A, F_A corresponds to P_R, and O_A corresponds to O_H. There is no ferrous intermediate (R) or cuprous intermediate (E_H) in the normal cycle put forth. Importantly, the resting CcO with an oxidized center is EPR silent with antiferromagnetic SOC of the ferric iron (s = 1/2) and cupric ion (s = −1/2) [171] is proposed to transiently bind O₂ and be rapidly reduced. Spin states for the remaining intermediates are hypothesized based upon ferric heme spin state preferences with various ligands [139] and have not been examined in, or correlated with, the literature. Energies assigned to the reduction-coupled proton pump steps are taken from the standard free energy changes measured with electrodes [190] except for O₂, which is endergonic, and would be unfeasible for proton pumping at −180 mV or +4.2 kcal/mol. For intermediate O_D and O₂ reduction, a rough estimate of −6 kcal/mol is assigned to the step. In Blackmore and Gibson’s investigation of O₂ binding and rapid reduction, the rate constant of 35,000 s⁻¹ for reduced oxygen intermediate formation and the relationship ∆G = -RT ln k is used to estimate the actual free energy change during O₂ reduction. The favorable thermodynamics are unexpectedly due to SOC and spintronics in a₃-Fe³⁺O₂Cu²⁺ (see below). Without correction for temperature and concentration effects; each step in the proposed cycle (Figure 5) appears to generate sufficient free energy to pump a proton against an [H⁺] gradient of 2 pH units or +3 kcal/mol where ∆G = −2.3 RT ∆pH. The large excess of energy is likely expended in conformational changes and heat generation. For an in-depth discussion of free energy for the proposed catalytic cycle with O₂ binding to BNC_red, the reader should consult Wikström and Verkhovsky [195], Wikström, Rich, and Gennis [163], and Blomberg [196]. The proposed cycle does not hypothesize or require the additional two-electron reduction of BNC_ox before O₂ binding. Also, differences in proposed intermediates (e.g., P_A and P_H) and the sequence (e.g., O_T or P_M) will require resolution.

Competitive and non-competitive or slowly reversed NO competition are proposed to occur with NO initially competing for binding and interaction with the ferric iron of the O_W (or a dehydrated O) intermediate. The association of O₂ and NO and hence migration through the protein to the heme iron is equally rapid with similar rate constants of ~1 × 10⁸ M⁻¹ s⁻¹ as measured for the ferrous cytochrome a₃ (Equations (22) and (23)) [90,159]. A 50% inhibition and competition of NO with O₂ at a 1:500 ratio [29] suggests a much larger dissociation rate constant for O₂ than NO. If the turnover rate is 100 s⁻¹ and the slow step is the trigger for O_W reduction and requires the 10 ms, then for only 50% of the time or 5 ms would O_W be available for O₂ collision and interaction. The remaining 5 ms an NO molecule would be bound and interacting. If we assume a half-time for NO binding of 5 ms and an association rate constant of 1 × 10⁸ M⁻¹ s⁻¹, we can calculate an NO dissociation rate constant of 139 s⁻¹. The O₂ dissociation rate constant would be roughly 500 times larger or 70,000 s⁻¹ and half-time for binding of an O₂ molecule would be ~20 µs. The association lasts roughly 2 billion-fold longer than a simple collision time of 10 fs. Relative to human experience, a brief second of contact would proportionally result in a 63 year association. At 1 µM O₂, we can calculate that 100 molecules of O₂ will collide in a second and interact ~20 µs each time resulting in a turnover rate of 100 s⁻¹ or 10 ms per cycle. With a supraphysiological and toxic [NO] of 2 nM, 20 molecules of NO have the opportunity to compete with O₂ by binding more tightly and longer to O_W in that same 10 ms. From the analysis, we see that the large respective K_D values for O₂ and NO of 0.7 mM and 1.4 µM can be misleading. It is not necessary that the measured K_m(O₂) of ~1 µM equal the K_D(O₂) value, as is often supposed [38,197,198]. This relation is only true when substrate (i.e., O₂) binding is the rate-limiting step in catalysis. On the other hand, the K_i(NO) value is expected to approximate the calculated K_D(NO) value or k’(NO)/k(NO) of ~0.12 nM (Equations (25) and (26)), and it does [37,94] when ET in CcO is apparently dysregulated in vitro, as apparently also occurs in hypoxic uncoupled mitochondria [102]. The thought experiment above suggests an in vivo K_i(NO) value closer to 1.4 µM. Blackmore and Gibson calculated a remarkably similar large O₂ off rate of 50,000 s⁻¹ in their O₂ binding and reduction studies [90]; however, that large off rate was assigned to an A (Fe²⁺O₂Cu⁺) intermediate and not to the putative O_D intermediate (see Section 4.3.5 below). In the model (Figure 5), non-competitive NO inhibition can be explained, at least in part, by the reaction of NO with the cupric copper to form bound nitrite, which is then released with reduction (Section 4.3.2, Equations (25) and (26)) [88,94,174]. However, reduction now forms the R intermediate, and the R intermediate is also subject to irreversible or very slowly reversible inhibition (Section 4.3 and Section 4.3.2, Equation (23)).

Spintronics in BNC_ox Binding of O₂

If we ignore the contribution of rotational and translational motions of O₂ and thus entropy [199], in the more restricted confines of the active site, we can calculate a relatively weak binding energy of −4.5 kcal/mol, where ΔG⁰ = RT ln K, R = 1.99 cal/mol K, K is in molar units, and the K_D (O₂) value is 7 × 10^–4 M from the approximations above. Weak binding can be attributed to the limited electrostatic forces between Fe³⁺ and O₂ and Cu²⁺ and O₂ and the spin-orbit interactions. The estimated short half-time for bridging of Fe³⁺ and Cu²⁺ of roughly 20 µs would still suffice for the postulated ferric heme SC and doming motion. In the Mb, the heme iron SC and doming occur in 2.4 ps [200]. However, the available CcO structures, unlike the flavoHb-NOD structures (Section 3.3.4) [18,74], reveal little [201] or no evidence for doming or other movements near the cytochrome a₃ propionates, begging the question of how O₂-regulated ET switching control could even occur.

The magnetic SOC forces between O₂ (s = 1) and BNC_ox Fe³⁺ (s = ½) and Cu²⁺ (s = −½) are expected to be negligible compared to the electrostatic forces creating the weak binding energy. Nevertheless, SOC or Nature’s version of spintronics may be critical for both O₂ binding and the putative ET switch mechanism. The proposed role of SOC and hence spintronics in O₂ binding BNC_ox and the ET switching mechanism is explained by Figure 6. In the model, O₂ binds to BNC_ox with antiferromagnetic SOC to the ferric heme iron and ferromagnetic coupling to the cupric copper. Ferromagnetic SOC greatly facilitates electron density currents between the O₂ and cupric ion and increases the oxidizing potential of triplet O₂. The principle is exploited in the engineering of spintronic apparati where SOC lowers the overpotentials in the electrocatalytic reduction of chemiadsorbed O₂ [76,77,78]. Moreover, electron conductivity was found to be greater with ferromagnetic SOC than antiferromagnetic SOC [77]. SOC is expected to increase electron density currents between the ferric iron 3d orbital and the O₂ 2π* orbital and increase the electrostatic interaction and incipient O₂ binding. In BNC_ox SOC, O₂ gains oxidizing potential by sharing electron density with the cupric ion, increases spin, causes ferric heme SC from s = ½ to s = 3/2 to achieve spin and magnetic parity, and withdraws electron density from the porphyrin chiefly residing at the a₃ propionate d and a carboxylate groups. The effect would be similar to the Bohr effect in Hb where increased O₂ binding correlates with a deprotonation [141]. In the proposed mechanism, protonation of both propionate carboxylate groups would diminish the electron density in the a₃ carboxylate groups and decrease the large porphyrin dipole and O₂ binding.

Figure 6. Proposed spintronics in BNC_ox binding O₂. O₂ collision and rapid transient SOC with the ferric heme and cupric ion (grey arrows). Low spin to high spin crossover (SC) induced in the ferric heme (thick green arrow) with antiferromagnetic (AFM) SOC to O₂ and ferromagnetic (FM) SOC of O₂ with the cupric ion. Electron density current migration from O₂ to the cupric ion and from the ferric heme to the coupled O₂ and Cu²⁺ (magenta arrow). Alternative antiferromagnetic SOC of cupric ion with O₂ (yellow highlight) with a low energy spiral spin inversion (SI) of the cupric ion (double-headed green arrow).

Figure 6. Proposed spintronics in BNC_ox binding O₂. O₂ collision and rapid transient SOC with the ferric heme and cupric ion (grey arrows). Low spin to high spin crossover (SC) induced in the ferric heme (thick green arrow) with antiferromagnetic (AFM) SOC to O₂ and ferromagnetic (FM) SOC of O₂ with the cupric ion. Electron density current migration from O₂ to the cupric ion and from the ferric heme to the coupled O₂ and Cu²⁺ (magenta arrow). Alternative antiferromagnetic SOC of cupric ion with O₂ (yellow highlight) with a low energy spiral spin inversion (SI) of the cupric ion (double-headed green arrow).

In an alternative scenario (Figure 6, yellow highlight), the unpaired cupric d electron undergoes a spin inversion for antiferromagnetic SOC to one of the unpaired O₂ 2π* electrons concomitant with ferric heme SC. The cupric ion is unusual in that it can undergo a spiral spin inversion with a very minimal energy input of +2.9 meV or +0.069 kcal/mol in cupric oxide with the rotation by an electromagnetic force [202,203]. However, in the alternative SOC arrangement, there is less spin parity and less electron density current between O₂ and Cu²⁺. Importantly, there is a specificity for O₂ conferred in the proposed SOC. The otherwise similar NO (s = ½ or 3/2) would show unique SOCs and would not be sufficiently oxidizing to withdraw electron density from the ferric a₃ heme.

As illustrated (Figure 6), the bridged SOC at BNC_ox may also increase the O₂ electron affinity by altering the stability of the triplet state. An analogous reaction has been proposed to occur when two NO radicals simultaneously couple with the O₂ diradical to form iso-N₂O₄ bearing a single O-O bond, which rapidly homolyzes to form two NO₂ molecules [204]. In the transient interaction, two NO molecules simultaneously transfer electron density to O₂ through SOC. In contrast, the ferric heme and the cupric center have far less electron density to share or transfer. Thus, the bridged coupling is not expected to produce the long peroxide-like O-O bonds observed in oxidized CcO and other heme-copper oxidases with X-ray crystallography and discussed earlier [184,186,205] and more recently observed with cryo-EM [206]. In another example, O₂ forms dimers at low temperatures and can magnetically couple with a total dimer spin of S = 0, 1, or 2 [207]. The stable arrangement of the parallel spin 2π* electrons of triplet O₂ can also be ‘destabilized’ by the input of high-energy light to form the more oxidizing singlet O₂ molecule via SCs, but the parallel spin configuration is also apparently weakened through low-energy interactions of O₂ with specific molecules [208]. Simultaneous destabilizing or coupling interactions of the 2π* electrons apparently increase the electron affinity and reactivity of O₂.

In the proposed ET switching mechanism, O₂ SOC (Figure 6) withdraws electron density from the propionate d carboxylate group and removes the Coulomb blockade for ET from arginine 438 guanidinium nitrogen to propionate d carboxylate group of the ferric a₃ (Figure 7). Withdrawing electron density from the propionate a carboxylate group facilitates the migration of the positively charged proton from the propionate a carboxylate O-atom to the propionate d carboxylate O-atom via the bridging water. The positively charged proton and the negatively charged electron migrate simultaneously to the propionate d carboxylate O-atom for ET. The vacated propionate a carboxylate simultaneously receives or ‘pumps’ a proton from the hydrogen-bonded water 1 (or hydronium) (Figure 7) and the matrix. Hence, the barrier to ET is lowered to form a bound reduced oxygen intermediate simultaneously with the migration of the proton (from the prior pump cycle) (Figure 5). The free energy for pumping a single proton from the water channel is derived from the reduction potential (ΔG = −nFΔE) for each of the O_D, S_H, P_H and O_T cycle intermediates (Figure 5). The migrating electron and the electron density of the negatively charged reduced oxygen intermediates (S_A, P_A, F_A) and the hydroxide form (O_T) donates electron density and momentarily increases the pK_a of propionate a and d carboxylates. The negatively charged propionate d carboxylate would momentarily block further ET from arginine 438 and would be ready for proton transfer from propionate a via the bridging water molecule. Protonation of the S_A, P_A, and F_A or O_A cycle intermediates and formation of the respective O_D, S_H, P_H, and O_T cycle intermediates repeat the cycle. Note the proposed importance of O_T intermediate and the tyrosyl radical for donating electron density to the ‘motor’ for pumping protons. Each of the protonated intermediates, like O_D, is oxidizing and electron density withdrawing. Finally, PCET, shown operating between the cytochrome a and cytochrome a₃ d propionates, facilitates ET and also forms the energetic ‘motor’ in the more complete proton pumping mechanism proposed for CcO (see Section 4.3.4). Recall, arginine rapidly tautomerizes [209], and tautomerization may facilitate concerted EPT. Concerted EPT and other types of PCET eliminate large ET barriers associated with charge [187]. Here, the charge transfers are countercurrent, rapid, electrostatically choreographed, and temporally complex and will require advanced computational analysis and models to resolve [187]. Cytochrome a serves as a harbor for an electron, and water molecules and the propionates serve as harbors for protons in the scheme. The concerted motion of electrons and protons is extremely fast [187].

Figure 7. Proposed ET switch mechanism and proton pumping. (Left) OFF switch position in cycle with no ligand, water, or electron density donating reduced intermediate anions (yellow box) bound to BNC_ox. (Right) ON switch position with transient electron density withdrawing cycle intermediates (green box) showing the structure of the ET pathway (black arrows), proton pathway (red arrows), and the heme dipole (magenta arrow). The water molecules bridging the propionate a and d groups (W_B), the proton donor water (1). Water molecule O-atoms (aqua) are shown at ~100% of their van der Waals radii. CcO structure models 2ZXW and 3ASO were used for the respective OFF and ON ET switch positions.

Figure 7. Proposed ET switch mechanism and proton pumping. (Left) OFF switch position in cycle with no ligand, water, or electron density donating reduced intermediate anions (yellow box) bound to BNC_ox. (Right) ON switch position with transient electron density withdrawing cycle intermediates (green box) showing the structure of the ET pathway (black arrows), proton pathway (red arrows), and the heme dipole (magenta arrow). The water molecules bridging the propionate a and d groups (W_B), the proton donor water (1). Water molecule O-atoms (aqua) are shown at ~100% of their van der Waals radii. CcO structure models 2ZXW and 3ASO were used for the respective OFF and ON ET switch positions.

4.3.4. Tunnel and Countercurrent Charge Migration Path in Proton Pumping

Water channels and chains of water molecules have been identified for proton migration in CcO and other O₂ reductases and are required for both the reduction of O₂ to water and the pumping of protons from the mitochondrial matrix to the intermembrane space [210,211,212]. Moreover, catalytic cycle-linked models for the mechanism of electrogenic proton pumping, incorporating key principles [185,189,213,214,215,216], have been previously advanced. One current model suggests that electric fields drive proton pumping [201]. Another recent model proposes a specific role for Cu_B in proton pumping [217].

The question of O₂ sensing by BNC_ox and the resulting ET switch mechanism (Figure 7) led to further insight into proton pumping. The D, K, and H paths for proton delivery to the active-site for O₂ reduction have been previously defined [184,201,217,218,219] but a separate path for proton pumping via the proton-donating water #1 (Figure 7) (aka proton loading site [185,220]) remains poorly defined, and a route via either the D, K, or H path appears tortuous. A putative pathway for proton delivery to the proton donor water 1 at the a₃ propionate a carboxylate (Figure 7) is readily observed in the bovine (PDB ID: 2ZXW) (Figure 8 top), P. denitrificans (PDB ID: 7AU6), and E. coli (PDB ID:7CUB) heme-copper oxidase structures. The tunnel runs a nearly straight course perpendicular to the plane of the membrane, beginning at or near surface prolines (respective Pro14, Pro537/Pro538, or Pro546) and travels at an ~20° angle to the 28 amino acid alpha helix bearing the respective proximal His376 and His378 for the cytochrome a₃ and a iron atoms. In the static bovine CcO crystal structure, a stricture in the tunnel of ~4.7 Å is measured near the surface Thr488 and Asp486 residues. The bound internal water O-atom and the restricting amino acid carbon atoms show a negligible (~0.1 Å) overlap of their van der Waals radii. Hypothetically, protons would migrate via a water chain with the force of the strong electrostatic field created by interaction of the proton with the 28 amino acid helix dipole moment of (cos 20°) 98 D or 92 D [221] and the alkaline state of the pump cycle (Figure 7, left). Assuming an average travel distance of ~10 Å from the helix axis and ignoring potential heterogeneous shielding effects of local static dipoles and applying the field equation E = 1/4πε₀ (2p/r³), we can estimate that the positively charged proton would experience an appreciable axial electrostatic field of 2.8 × 10⁷ V/cm, which, over a ~30 Å axial travel distance, would equal a substantial free energy change of +0.084 V or −1.94 kcal/mol. If we ignore the comparable counteracting force and energy of the membrane potential [222], the proton would reach a velocity upwards of 2.9 km/s, sufficient for a CcO turnover of >5 × 10¹¹ s⁻¹. However, copious proton delivery may also dysregulate ET control by allowing ET in the absence of the O₂ trigger (Figure 6 and Figure 7). The putative water chain or proton wire is ~53 Å long and passes from Pro14 to Thr488 and along the proximal side of cytochrome a₃, the farnesyl group, and by His376 (Figure 8, bottom). Remarkably, a small number of bound waters or charged side chains are found in the putative proton tunnel in the bovine, P. denitrificans, and E. coli structures. This should not be surprising since a uniformity of the water molecules, proton-exchanging hydrogen-bonds, or unimpeded single water molecule rotations [223] in the water chain would be required for a low energy barrier for rapid proton hopping as in the classic Grotthus mechanism. Consider also the possible role of the strong membrane potential in normally loading the polar tunnel with polar water molecules. For a proton hopping distance of 3.3 Å [223] and water or hydronium rotational diameter of ~1.4 Å, we can expect a chain of 12 water molecules including water number 1 and the surface water (W_S) molecule (Figure 8). If confirmed, I suggest that the pathway be named the ‘Pro’ tunnel after the conserved surface Prolines and proton pump function. Together, the putative Pro tunnel and D, K, and H pathways may constitute the proton path(s) from the matrix for the four pumped and four consumed protons in the proposed cycle (Figure 5). The Pro tunnel may also serve in water removal and ‘dry out the surroundings of the BNC_ox‘ in O_W (aka O_H) (Figure 5) to form the dehydrated O intermediate [163], which would be expected to be more accessible to O₂. An electrostatically-driven water pumping or efflux mechanism that depends on the series of water dipole rotations with proton hopping [223], the alpha-helix dipole, the local membrane potential, and any cycle-dependent opening and closing motions of the stricture formed by Asp486-Thr488 can now be modelled. The Pro tunnel, or water pump, may also replenish the water required to sustain fatty acid oxidation and other matrix-localized water-consuming metabolic pathways.

A low-barrier (i.e., short gap and electrostatically favorable) pathway for electron migration from cytochrome c²⁺ and a feasible (i.e., short distance proton donor–acceptor) pathway for proton migration to the aqueous phase of the mitochondrial intermembrane space with countercurrent PCET of the concerted EPT-type [187] is apparent. The identified electron and proton pathways are illustrated by the extended model in Figure 9. From the cytochrome a propionate d, PCET occurs in only two short segments of the pathway to or from cytochrome c²⁺, namely, with tautomerization at histidine 204 and between cytochrome a propionate d and the arginine 439 guanidinium group. Between histidine 204 and the arginine 439 guanidinium group, the pathways bifurcate within the electric field created by the hydrated magnesium ion. The putative pathways are separated in space by a short 4 Å distance that may allow weak electrostatic interactions between the migrating proton and electron pairs. Two water molecules carry the proton from the arginine 439 guanidinium group to histidine 204. The protons move with the electrostatic field dipole created by Mg²⁺, thereby preventing the reverse migration of protons or a pump leak. In addition, the third water (Figure 9) is slightly beyond hydrogen bonding distances in at least two CcO structures (PDB IDs: 2Y69 and 5IY5), suggesting an additional control against reverse migration of protons and for preventing a possible alternative path for PCET. The third water (3) is not visible in the structure modelled in Figure 9 (PDB ID: 2ZXW) but is nevertheless present in other structures. A short 10 Å long water or hydronium ion channel is identified for proton migration from histidine 204 to a surface water molecule (number 4). The putative exit channel is bounded by the peptidyl carbonyl O-atoms of tryptophan 104, asparagine 203, and phenylalanine 206. In the model, the electrostatically paired charge migrates from cytochrome c²⁺ via the tryptophan 104 indole, the Cu_A center, histidine 204 imidazole, to the arginine 439 guanidinium group. The water or hydronium ion (number 4) at the surface [224] may facilitate ET from cytochrome c²⁺ to tryptophan 104. Electrostatic field dipoles (Figure 9, magenta arrows) and charge pairing neutralization synchronize and control the direction and timing of the movements consistent with the early proposals of charge neutrality by Rich et al. [225]. Ultimately, the ET switch mechanism (Figure 7) and O₂ binding and SOC (Figure 6) initiate the charge movements, and the reduction of O_D, S_H, P_H, and O_T intermediates provides the driving force (Figure 5). The simple structural model postulates a direct coupling [214] of proton and electron movements with switching and gating but no significant conformational changes.

4.3.5. Toward the Resolution of Unanswered Questions with Models

I had the good fortune and privilege of being one of about 100 scientists to observe the unveiling of long-awaited structures of the bovine CcO determined by Shinya Yoshikawa and coworkers [227,228] and the Paracoccus denitrificans CcO solved by Hartmut Michel and coworkers [229] at the Bioenergetics Gordon Research Conference in the Summer of 1995 at the Proctor Academy in New Hampshire. With the 13 transparencies representing the 13 subunits of the bovine CcO ceremoniously layered upon each other, thoughts of “What could be further accomplished? … the mechanism!” passed through my mind. Ten years later, I was asked to provide comments on a review article for a special issue of the Journal of Inorganic Biochemistry on heme entitled “Cytochrome c oxidase, ligands and electrons” by Professor Maurizio Brunori and coworkers. These and other events inspired a fascination with the complexity of CcO. Out of circumstance, my wandering with CcO began with measuring its NO sensitivity (resistance) within cells [16], followed by a 20 year interim dedicated to uncovering tightly-held secrets of the globin-NODs [18,74,75], to more recently studying the CcO structures and developing a model for the NO-resistant CcO mechanism with a beneficial working knowledge of the globin-NODs. I searched many other fields for adequate concepts and answers during this long period, including ET, porphyrins, structural dynamics, electronics, quantum chemistry, and eventually the spintronics field.

In their 2005 review, Brunori, Giuffrè, and Sarti [168] made a point by “pointing out where necessary unresolved facts or questionable interpretations” lie. I will begin by attempting to resolve, at least cursorily, these and other questions [195] as a post hoc appraisal of the usefulness of the mechanistic model (Figure 5, Figure 6, Figure 7, Figure 8 and Figure 9), a model built upon little known or unresolved facts, and a profound, but little recognized, paradox (discussed in Section 4.3). As succinctly and critically stated by Brunori et al. [168], “A pre-requisite for O₂ and CO binding, but not for NO, is the complete reduction of this binuclear site”. Since the work of Greenwood and Gibson [161] more than 60 years ago, all CcO mechanistic models, including those with the catalytic cycle linked to proton pumping, have been formulated with that problematic assumption. By 1975, Britton Chance and coworkers [193] very briefly evaluated O₂ binding to the oxidized binuclear site but apparently just as quickly abandoned the pursuit of the unorthodox idea as well as the limited but nonetheless intriguing data showing visible wavelength spectral changes at −97 °C. Roles for transient weak ligand (e.g., O₂, CO, NO, N₂, and H₂O) interactions with oxidized ferric heme in the gas phase [137] or in enzymes [18,74] are now appreciated.

In their review, Brunori et al. raised four longstanding unanswered questions relevant to the model that I will address. Briefly, the key issues include (1) the low O₂ affinity of CcO, (2) the strong competition of NO with O₂ causing CcO inhibition despite the similar binding rate constants, (3) the intermediates and their formation, (4) the slow and fast rates of ET from the reduced cytochrome a to BNC_ox, and (5) whether ET or PT is rate-limiting. First, (1) the low affinities measured for O₂ may indeed be explained by the rapid reduction of O₂ and oxidation of BNC_red during the measurements, as Brunori et al. surmised. The model presented predicts a K_D(O₂) value of 0.7 mM for BNC_ox at 37 °C (Section 4.3.3), greater than the 0.32 mM estimated by Chance and coworkers [193]. Measured values for BNC_red, or in more likelihood BNC_ox [168], of 0.14–0.5 mM [90], 0.28 mM [230], and 0.285 mM [231] at 20 °C fall in that range. Furthermore, DFT calculations, albeit tentative, support weak O₂ binding to BNC_ox [186]. Here [168], the stated “Physiological meaningful affinities” reinforces a misunderstanding in enzyme kinetics since the O₂ capture is achieved through a transient ligand binding, SOC, ET switch activation (Figure 6 and Figure 7), and rapid O₂ reduction (Figure 5). (2) The data and model indicate that the competition between NO and O₂ occurs at BNC_ox, and the competition is explained by the kinetics of binding and turnover (Section 4.3.3). CO inhibition of CcO, like the flavoHb-NOD [18,62,74], can also be explained by the competition of CO with O₂ for BNC_ox. (3) The new model proposes O₂ binding to BNC_ox but otherwise shares intermediates and timing with the current models (Figure 5) [163,164,191,194,232]. Albeit, in the models examined, the order and presumed role of the tyrosyl radical intermediate, P_M or O_T (Figure 5), differs. Otherwise, the ET switch mechanism simply demands that the reduction of O₂ and oxygen intermediates precede their protonation since the anions form the motor or pump action (Figure 7). (4) The slow rate of reduction of BNC_ox observed in many studies may be attributed to the ET switch mechanism of the model (Figure 6 and Figure 7) (Section 4.3.3). The switch may also explain O₂ activation [181,233], ‘pulsed’ and ‘resting’ states [162], slow and fast phases of O₂ reduction [230,231,234], slow and fast electrogenic proton pumping phenomena in the absence of O₂ [188], and the unusual oxidative and reductive phases for proton pumping [235]. As proposed, ET requires ferromagnetic SOC of O₂ and Cu²⁺, antiferromagnetic SOC of a₃-Fe³⁺ with O₂Cu²⁺, SC in the ferric heme, electron withdrawal from the a₃ propionates, and proton migration from propionate a to propionate d. Also, the reduction requires PCET from the reduced cytochrome a to the oxidized a₃ propionate d. Anything that interferes with or dysregulates the ET switch mechanism in vitro will turn ET OFF (or ON). In addition, the proposed ET switch would also prevent the untimely reduction of the ferryl intermediate (F_H), formation of highly reactive hydroxyl radical, and potentially irreversible damage to CcO. Where a slow phase has been observed, and the electron supply does not appear limiting and may be dysregulated [234] or the O₂ SOC may be impaired by a hydrated (O_W) and protonated BNC_ox, the reversibility of function should be examined and the damage, protonation, and hydration assessed. Given the observed single turnover fast to slow phase transition [234], a failed dehydrating Ow → O cycle step that occurs with a non-existent membrane potential in the isolated CcO, with possible Pro tunnel/pump dysfunction, appears most attractive. Not recognizing, or accounting for, the existence of a sensitive and rapid ET gate or switch (Figure 7) and/or a proton uptake and water expulsion tunnel (Figure 8) and O_W (O_H) dehydration mechanism would cause incalculable confusion. For well over 50 years, the fast-slow phase a₃ reduction phenomenon [234] has been interpreted as evidence for the requirement of a 2 e- reduction of the BNC for O₂ binding [159,163]. The NO sensitivity-resistance of CcO discussed in Section 4.2 and Section 4.3 and the paucity of R in respiring mitochondria [101,102] strongly suggests otherwise. According to the model calculation estimates (Section 4.3.3), with regard to O₂ collision, SOC appears rate-limiting for turnover at 1 µM O₂, the approximate K_M(O₂) value [197,198]. O₂ appears efficient at triggering the ET switch mechanism during the steady-state given the estimated 1 O₂ transient binding interaction per 1 turnover for a 1/2 V_max = 100 s⁻¹. Indeed, given the underlying physics and chemistry, the trigger process would likely occur in picoseconds. At O₂ saturation with an ~3 ms turnover time, the mechanism for water dissociation from O_W may involve large motions and be rate-limiting for the catalytic cycle. As proposed (Figure 5), it appears unlikely that an O₂ collision alone would efficiently displace the water molecules. The process may require milliseconds and energy. For example, gating may be required. At least three different “conformations” of BNC_ox have been detected and studied by NO binding and EPR, and differences in O₂ accessibility have been discussed [236]. If we imagine a cycle clock for CcO similar to that drawn for the flavoHb-NOD (Figure 2), extrusion of water from O_W, or dehydrating O_W to form the fast ‘O’ state [163], would likely be apportioned the largest time. (5) The model provides a detailed map for charge migration (Figure 7, Figure 8 and Figure 9) and explains how the necessary unidirectionality of proton migration is achieved. It also explains how the electron migration path bifurcates from the proton migration path to allow strong electrostatic field forces from the magnesium ion to influence the charge movements (Figure 9). At present, the model is not sufficiently developed to describe the complex and rapid timing of synchronized countercurrent charge movements, but proton pairing with electrons through head-on or side-to-side meetings, or a countercurrent form of PCET, such as concerted EPT [187], is expected except for the storage and release of electrons from Cu_A or cytochrome a. An electron would migrate from one migrating positive charge to another migrating positive charge with little or no energy barrier and act as a conductor. A description of the synchronized motions may benefit from advanced computational modelling of PCET.

Experiments and theoretical work appear to support the general model (Figure 5, Figure 6, Figure 7, Figure 8 and Figure 9). I will only mention a few here. Arg438 and Arg439 are thought to be important for proton pumping and substitution mutations impair CcO function [237,238,239]. Cytochrome a and cytochrome a₃ propionates have long been considered important for ET and proton pumping [182,184,201,229,240,241]. The evidence for NO protection against cyanide, CO, or sulfide inhibition of CcO [242,243] may be explained by the competition with O₂ for BNC_ox and by the relative association and dissociation constants. More time and space will be required to thoroughly examine the literature and make all the proper attributions.

The change to the current catalytic cycle models [163,164,168,184,206,217,232] is small, but the improvements may help resolve some unanswered and puzzling questions. Revisions and improvements to consider include

(1)

O₂, NO, and CO binding to BNC_ox (Figure 5).

(2)

SOC and spintronics (Figure 6).

(3)

an ET switch mechanism (Figure 7)

(4)

the alpha helix dipole-driven and membrane potential-opposed Pro tunnel (Figure 8)

(5)

the ET-linked proton pump mechanism (Figure 7)

(6)

the existence of well-defined electrostatically-coordinated and concerted EPT pathways (Figure 9)

(7)

proton reverse flow and ET pathway control by the hydrated magnesium ion (Figure 9)

(8)

the proton or hydronium ion exit channel (Figure 9)

(9)

the free energy and mechanism for pumping first proton in the cycle comes from reduction of ferromagnetically coupled O₂Cu²⁺, not O₂ per se

(10)

other concepts mentioned but not listed here

The CcO model revisions and concepts that I have only tersely expounded upon can be readily tested, improved, expanded upon, and superseded with an in-depth analysis of new and existing structures and data, advanced molecular dynamics simulations, and state-of-the-art DFT calculations. As always, attention to both contradictions and agreements in the development of a comprehensive and coherent mechanistic model is essential. Physiological evidence and functional behavior should take precedence. For example, future investigations of the kinetic competition of NO, CO, cyanide, and sulfide with O₂ should test the model in vitro and within intact mitochondria with thorough kinetic analyses, including binding constants. Investigators of CO inhibition will note that, in 1951, Ball et al. [244] stated that CcO “exists in both an oxidized and reduced form, while CO and O₂ are competing only for the reduced form “. But recall again that the reduced form was not seen in respiring mitochondria [101]. Keilin and Hartree [158] had already recognized the contradiction between the light sensitivity of CO inhibition and the light insensitivity of the reduced CcO-CO complex as early as 1939. Keilin had stated earlier that ferric hematin bound CO but reduced hematin bound CO with higher affinity [245]. If Keilin had seen and confronted the light (in)sensitivity paradox then or if others [168,186,193] had pursued the relevance of BNC_ox transiently interacting with O₂, we can only imagine where CcO (and NOD) investigations would stand today. Indeed, CO reduces BNC_ox [246] and may also transiently and competitively interact with O_W (or O) to cause inhibition. SQUID-EPR [247], especially as applied to CcO [248], has the potential for measuring weak and transient O₂ interactions with BNC_ox at the low temperatures used by Chance et al. [193] as well as at higher temperatures. While spin parity and neutrality, as predicted for ferromagnetic SOC in the model (Figure 6), may preclude detection, the net loss of O₂ spin through SOC with BNC_ox may be apparent. In the classic experiments of Van Gelder et al. [249,250] and Brudvig et al. [236], fluoride was shown to weakly interact with BNC_ox by EPR or UV–visible spectroscopy. Competition between O₂ and another weak, yet detectable, ligand such as fluoride or CO may be used to indirectly measure the K_D(O₂) value for BNC_ox. We can also begin to ask whether SOC with electron-withdrawing O₂Cu²⁺ (Section 4.3.4) could produce the spectral changes recorded by Chance et al. [193]. Small differences in the Q_v or β (540 nm), α (590 nm), and CT (612 nm) bands of the ferric a₃ heme, similar to those recorded for the Mb ferric heme [74], may serve as sensitive indicators of a transient O₂ SOC. The failure of Kaila et al. [186] to observe spectral differences in CcO BNC_ox with 280 µM O₂ may be explained by the uptake of excess protons by the putative Pro tunnel (Section 4.3.4, Figure 8) and a hydrated BNC_ox (O_W). In the absence of an opposing membrane potential, protonation of the a₃ propionates would decrease the electron density of the heme required for O₂ binding. Indeed, the CcO Soret maxima (Kaila et al., Fig. 6) appears blue-shifted at ~421 nm vs. the ~424 nm recorded by Chance et al. [251], which may occur with propionate carboxylate protonation, hydration, or an increase in high-spin iron [74]. Water bound to the BNC_ox (O_W) may also cause a failure to observe the proposed O₂ SOC. Dehydrating BNC_ox to an O intermediate state [163] may be essential for significant O₂ interaction. A similar water and O₂ competition is apparent in ferric Mb [74]. Looking upstream of the ET switch, site-directed mutagenesis of the Pro tunnel similar to prior quantitative investigations of the D, K [219], and H pathways [218] can be readily utilized to test and analyze function. The model (Figure 8) and prior investigations predict a large influence of the membrane potential on proton uptake, delivery, and dysregulation of the ET switch mechanism (Figure 7). A collapsed membrane potential should lead to R intermediate formation and a greater NO sensitivity of CcO, and it apparently does (see above) [80,101,102]. When the membrane potential of heart mitochondria is collapsed with ADP in state 4 (note that authors call this state 3), the NO sensitivity of CcO increases manyfold [80]. What is the quantitative relationship of the membrane potential to R formation? How does the membrane potential compare to the helix dipole electrical potential energy (Figure 8)? Does the membrane or an external electrical potential affect a₃ propionate a and d protonation and R formation as supposed by the model? Could engineered mutations inhibit the Pro tunnel, protonation, and ET dysregulation in the isolated CcO? Could an altered Pro tunnel and protonation explain the >10-fold smaller a to a₃ ET rate observed with low phospholipid CcO [252]. Does C¹⁸O₂ formation via ¹⁸O₂ respiration, H₂¹⁸O formation, and fatty acid oxidation depend on CcO and the putative Pro tunnel or (de)hydrating water pump? Looking downstream, beryllium (Be²⁺), with a charge to radius ratio 1.5-fold larger than Mg²⁺ [226], may replace Mg²⁺ and alter proton pumping and the CcO function, possibly accounting, in part, for respiratory toxicity [253,254]. Further downstream, the exit channel for hydronium ions and protons provides a likely route also for HSO₃⁻ entry and for the initiation of the free radical chains of sulfite oxidation by copper. Given the relatively large rate constants estimated for CcO-catalyzed sulfite oxidation (k₂ = ~460 M⁻¹ s⁻¹) [71,255] and possible self-inflicted sulfur trioxide anion radical damage to CcO, the identified channel to the Cu_A site may be relevant to the susceptibility of individuals to the airway-constrictive, asthma-inducing, headache-causing, and rash-inducing effects of SO₂ in pollution, fumigated fruits, and treated wines. As always, revisions to the model and new concepts will be needed to accommodate old and new observations.

5. Pathology of Irreversible NO Poisoning

NO-resistant NODs and the NO resistance mechanism of CcO have clearly evolved to avoid more harmful NO poisoning of respiratory pathways, the citric acid cycle, etc., in pathological conditions including hypoxia, stroke, ischemia–reperfusion injury, septic shock, and inflammation. Conditions which lead to dysregulated ET or a slow leaky ET in either NODs or CcO to form their reduced forms, as observed in hypoxic cardiac muscle mitochondria [101], are expected to increase NO poisoning and damage to tissues. Under the reducing conditions of hypoxia or ischemia, stoichiometric NO scavengers, such as deoxyMb or oxyMb, would show a limited capacity for sustained NO scavenging and protection in most cells. Mb-rich cardiac myocytes may be the exception. with a greater capacity for sustained NO scavenging [87]. The O₂-independent NOR activities observed in cells or mitochondria [16,80] may be essential under these conditions. Numerous investigations of microbial NO toxicity and resistance have led to similar hypotheses and conclusions [70,170,256]. A greater knowledge of ET regulation and dysregulation in CcO and globin-NODs and the identification of the mammalian NORs may increase our understanding of NO pathology and resistance. Given the proposed role of impaired membrane potential in various conditions including aging and Alzheimer’s disease [222] or even extreme exercise and work, ET control dysregulation in CcO and an enhanced NO sensitivity of CcO [80] may be more generally important during fever, vaccinations, and other inflammatory conditions. For this reason, NO-resistant NODs, including Ngb [257,258], cytoglobin [259,260,261,262,263], and monomeric Hbs [264] and Mb [265] with their tightly-coupled reductases [18], may be even more critical in Alzheimer’s and other hypoxic NO stress conditions.

6. Prospective Summary

FlavoHbs and related globins with their coupled reductases should never function as NODs, but they do. The extreme NO affinity and competition with O₂ for ferrous heme binding should prohibit turnover, but it does not. Recent data and theory strongly suggest that NOD activity is resistant due to an allosteric ET gating mechanism in which the ferric heme is only reduced when O₂ is present. Together, NO and O₂ are proposed to elicit heme SC, doming, conformational changes, and electron transfer. By the same token, CcO should be irreversibly inhibited by NO, but it is not. The only explanation consistent with the reversibility of NO inhibition is a mechanism similar to that described for the NODs in which a heme-Fe³⁺O₂Cu²⁺ complex triggers the initial electron transfer to form Fe³⁺O₂⁻Cu²⁺. A subtle modification to the current mechanistic model for CcO and related terminal oxidases is suggested in which O₂ and reduced oxygen intermediates affect the heme spin state, heme propionate protonation, ET, and proton pumping. A greater understanding of the mechanism of the CcO is achieved through a thorough reappraisal of the NO resistance as well as the reinterpretation of assorted phenomena and structures. Given the ubiquity of NO, there are likely other examples of heme-based ET switching. The O₂-metabolizing P450s and NOSs are attractive candidates for further investigation.