Nitrite and nitric oxide (NO), two active and critical nitrogen oxides linking nitrate to dinitrogen gas in the broad nitrogen biogeochemical cycle, are capable of interacting with redox-sensitive proteins. The interactions of both with heme-copper oxidases (HCOs) serve as the foundation not only for the enzymatic interconversion of nitrogen oxides but also for the inhibitory activity.
1. Introduction
Proton motive force (pmf) is essential for bacteria to grow and survive under non-replicating conditions by providing energy for a wide range of crucial processes
[1][2][3]. The pmf (electrochemical potential) consists of two gradients: the chemical proton or pH gradient (∆pH) and the membrane potential generated by the transport of electrical charge (∆ψ). Bacteria are capable of generating the pmf by a variety of mechanisms; among them, the most efficient one is through oxygen reduction
[3]. The oxygen-reducing enzymes (terminal oxidases) that contribute to the pmf generation are classified into two main groups: heme-copper oxidase (HCO) (also called heme-copper oxygen reductase (HCOR)) superfamily and
bd-type quinol oxidase (
bd QO) family
[4][5].
The HCO superfamily is composed of three subfamilies, A, B, and C, as well as the structurally-related nitric oxide (NO) reductases (NOR)
[6][7]. A-family HCOs include cytochrome
c oxidases (C
cOs), such as
aa3 from eukaryotic mitochondria, and some prokaryotes (often as
caa3, where
c represents a cytochrome
c subunit), and QOs, such as
bo3 of
Escherichia coli [8][9][10]. HCOs of B and C families are present only in prokaryotes, with
ba3 and
cbb3 as respective representatives
[9][11]. HCOs are highly efficient and specialized in pmf generation during the exothermic four-electron reduction of O
2 because of the proton-pumping mechanism
[12][13]. In contrast,
bd QOs, found exclusively in prokaryotes to date, do not pump protons and are thus less efficient in energy conservation, but play an important role in mediating viability under various stress conditions
[14][15].
Although HCOs of A, B, and C families are diverse in terms of subunit composition, electron donor, and heme type, they house a similar signature active site, the so-called binuclear center (BNC), where the reduction chemistry occurs
[7]. Located in a subunit with 12 membrane-spanning helices, this BNC consists of two magnetically coupled redox-active metal centers, a high-spin heme (
a3,
o3, or
b3), and a copper ion (Cu
B)
[7][8][11][16]. In all HCOs, these two metal centers are in proximity, with the two metals (Fe and Cu) only ~5 Å apart
[7][17]. During the oxygen reduction, the BNC experiences an oxidative-to-reductive phase transition involving several intermediate states
[18].
Nitrogen is essential to all life and is a constituent element of amino acids, proteins, and nucleic acids. After fixation, nitrogen as nitrogen gas (N
2), the most abundant element in the atmosphere, can be converted to ammonium and a variety of nitrogen oxides, among which nitrite (NO
2−) and nitric oxide (NO) are the most common and bioactive species
[19][20] (
Figure 1). Given that bacteria are able to catalyze all steps of the nitrogen cycle, they are crucial for the inter-conversion of different nitrogen oxides
[21]. Both nitrite and NO are involved in diverse physiological processes in bacteria functioning as important cellular signaling molecules, substrates of metabolic enzymes, and inhibitory agents modulating protein activity
[19][22]. Although a significant portion of phenotypic changes caused by nitrite are NO-independent, it is widely accepted that NO is the molecule largely underpinning the physiological influences; nitrite impacts the physiology of living organisms in part by serving as a biochemical circulating reservoir for NO
[19]. Additionally, NO can be engaged in cellular physiological and pathological processes through a complex cross-talking with two other gasotransmitters, carbon monoxide (CO) and hydrogen sulfide (H
2S)
[23][24][25][26][27]. Meanwhile, nitrite can be converted back to NO through the one-electron-oxidation of NO.
Figure 1. Redox cycle for nitrogen driven by prokaryotes. Shown are the major biological nitrogen transformation pathways, each of which are represented by lines in the same color, and the relative oxidation state at which they occur.
Given the particular importance of the inter-conversion of nitrite and NO, bacteria have evolved a variety of enzymes to catalyze the transformation of nitrite and NO, including NOR
[19][28]. In addition, although the HCOs of all A, B, and C families differ from NOR in their metal ions within the BNCs, they are also profoundly implicated in the biology of nitrogen oxides. The direct reactions of HCOs with nitrite and NO, which have been known for a long time, provide a mechanistic understanding of the interplay between the enzymes and the two nitrogen oxides
[29][30]. On one hand, it is well known that eukaryotic C
cOs mediate the reduction of nitrite to NO under hypoxic conditions
[31][32]. On the other hand, bacterial HCOs (
ba3,
caa3,
bo3) of the A and B families are capable of catalyzing the reduction of NO to N
2O, whereas C-family HCOs (
cbb3) convert nitrite to N
2O
[19][33][34][35]. On the other hand, both nitrite and NO are bacteriostatic agents due to their ability to inhibit proteins, especially hemoproteins
[36][37]. A great body of evidence has been accumulated showing that nitrite and NO react with purified HCOs and
bd QOs in vitro and inhibit cell respiration in vivo
[22][30][31][38][39][40][41][42]. While there are common reaction mechanisms involved in the inhibition by nitrite and NO, considerable discrepancies have been observed between their cellular targets identified to date
[22]. Thus, the extent of the effective mechanisms elucidated by in vitro analyses is in living cells is still a matter of study.
2. Bacterial Terminal Oxidases for pmf Generation
Proton translocation across the membrane is of crucial importance for sustaining the cellular activity in all organisms. However, due to the polarity characteristics, protons are unable to pass through the phospholipid-bilayer membrane freely by diffusion like small non-polar molecules. Proton pumps are special and efficient hydrogen ion transporters that move protons across the membrane from the low-concentration side to the high-concentration side to form pmf, which is subsequently utilized to drive the production of adenosine triphosphate (ATP), the cell’s chemical energy currency, by ATP synthase
[1][2][3].
In aerobic bacteria, transmembrane proton pumping is closely related to the oxidative phosphorylation process, especially with the terminal oxidases in the respiratory chain. The bacterial terminal oxidases, including HCOs of the A, B, and C families and
bd QOs, catalyze the four-electron reduction of oxygen to water using quinol or cytochrome
c as the electron donors. The main role of most HCOs in microbial metabolism is to conserve energy
[43][44][45], and
bd QOs are thought to contribute to nitrosative stress tolerance, hydrogen peroxide detoxification, or prevention of H
2S toxicity, especially in pathogenic bacteria
[42][46][47][48][49]. To date, a number of high-resolution structures of terminal oxidases from each group have been reported, which have greatly enhanced our understanding of the exact working mechanism of these enzymes.
2.1. HCOs
HCOs are the most extensively studied terminal oxidases. Despite the variety in the composition of the electron donor, polypeptide, and heme group type, all HCOs possess a conserved redox center composed of a low-spin heme and a heteronuclear heme-copper center (binuclear center, BNC) consisting of a high-spin heme and a copper (Cu
B)
[44][50]. Based on the amino acid sequences and the proton-pumping pathways, members of the HCO superfamily are divided into three families as A, B, and C
[51] (
Figure 2A).
Figure 2. Prosthetic group arrangements and proton pathways of typical bacterial HCOs (
A) and bd QOs (
B). (
A) Representative structures of HCOs of A (divided into A1 and A2), B, and C subfamilies. Protein peptides, heme cofactors, and ions are shown as cartoons, sticks, and spheres, respectively. SUs I of families A and B, SU III of family A and CcoN of family C are colored in green; SUs II of families A and B, and CcoO of family C are colored in cyan; SU IV of family A, SU Iia of family B, and CcoO of family C are colored in magenta; the 30-mer peptide in family C HCO is colored in yellow. The blue dashed arrows indicate the proton pathways inside each HCO, with the amino acid residues at the entry point of each pathway marked with dashed cycles. (
B) Structures of Cyt bd-I and bd-II QOs. Protein peptides and heme cofactors are shown as cartoons and sticks with subunits CydA and CydB colored in green and cyan, and CydX in bd-I QOs and CydS in bd-II QOs colored in magenta, respectively. The purple dashed arrow indicates the ‘water-molecule chain’ observed between residues Asp119 (subunit A) and Asp58 (subunit B) in bd-I QOs. The blue dashed arrows indicate two proposed proton pathways in bd-II QOs. Figures are prepared with PyMOL (Molecular Graphics System, LLC)
https://www.pymol.org (accessed on 20 December 2021).
Bacterial A-family HCOs include
aa3-type C
cOs (
aa3-HCO, in some cases
caa3-HCO), which exhibit high structural relations to their mitochondrial counterparts, which contain only
a-type hemes, and the
bo3-type QOs (
bo3-HCO) from
E. coli (
Figure 2A). Most often, A-family HCOs contain three subunits, named SU-I, II, and III. SU-I is highly conserved among all HCOs and typically composed of 12 transmembrane helices (TMHs), which hold the BNC
[43]. SU-II contains a membrane-anchored cupredoxin domain functioning for harboring the mixed-valence di-nuclear copper (Cu
A) acting as the primary electron acceptor
[43]. The divalent cations (Mg
2+ or Mn
2+) located at the interface between SU-I and SU-II and close to the high spin heme in HCOs of A and C families are not essential for proton pumping; however, their exact functions remain unclear
[52][53]. SU-III is present in most bacterial A-family HCOs and possibly influences the oxygen reduction as well as the internal proton flow
[43]. An additional subunit (SU-IV) is also identified in A-family HCOs, with its function a mystery yet
[10][11][16].
B-family HCOs comprise similar subunit compositions but with low sequence homology to their A-family counterparts (
Figure 2A). In contrast to the canonical composition of 12 TMHs in SU-I of A-family HCOs, SU-Is of B-family HCOs, as seen in C
cOs from
Thermus thermophilus and
Aquifex aeolicus, possess 13 and 14 TMHs, respectively
[54][55]. The SU-II of B-family HCOs resembles its A-family counterpart in that both contain a membrane-anchored cupredoxin domain; however, an additional subunit (SU-IIa) consisting of a single helix is identified in the former
[55]. The His-Tyr cross-link in A-family HCOs, which functions to fix Cu
B in a certain configuration and distance from heme
a3 at the BNC
[56], is also conserved in B-family HCOs.
C-family HCOs are highly divergent from HCOs of the former two families in protein sequence (
Figure 2A). To date, only
cbb3-type C
cOs are reported in this group
[11]. The SU-I of C-family HCOs contains a His-Tyr cross-link as well, but with the two residues residing at two separate helices different from the situations in HCOs of A and B families. In addition, C-family HCOs lack the di-nuclear copper site (Cu
A) but utilize two auxiliary heme-binding subunits (CcoO and CcoP) to receive electrons from reduced cytochromes
[11]. C-family HCO contains an extra subunit CcoQ, a small non-heme protein that is not required for catalytic activity but has a role in the assembly of the HCO complex
[57].
The proton pumping in A-family HCOs is performed via two pathways, D-pathway and K-pathway, named accordingly by the conserved and functionally critical residues (Aspartate and Lysine, respectively) near the entry site of each pathway (
Figure 2A). In order to pump protons, A-family HCOs utilize the internal ‘proton wires’ to transfer the electronic charges in a way similar to the Grotthuss mechanism
[3][58]. Within the longer D-pathway, a series of acid residues and water molecules jointly form a consecutive chain through hydrogen bonds and connect the entrance aspartate to the gating residue close to the BNC
[59]. Due to the crucial role of the water molecules, a water gating proton pumping mechanism was thereby proposed in D-pathway
[60]. A-family HCOs could be further divided into two types as A1 and A2, based on the residue composition at the hydrophobic end of the D-pathway. Type A1 is featured with a conserved glutamate within the motif XGHPEV on helix VI. However, in type A2, this residue is replaced by consecutive tyrosine and serine in a YSHPXV motif. HCOs of both types A1 and A2 have a covalent bond between one of the Cu
B-coordinated histidines and a tyrosine on the same helix
[61]. The D-pathway is responsible for transporting six protons, four of which are pumped to the positive side (P-side) of the membrane; the remaining two are donated to the active site for use in oxygen reduction
[43]. By contrast, the shorter K-pathway typically consists of a few highly conserved polar residues and only qualifies to supply two protons to the catalytic site during the initial reduction of the BNC
[59]. A conserved binding domain (carboxyl group) for amphipathic compounds adjacent to the entrance of the K-pathway tends to play a role in organizing the water chain, which supports proton uptake
[9][62][63].
The canonical K- and D- pathways are absent in either B-family or C-family HCOs; instead, an alternative K-pathway analogous to that in A-family HCOs is exploited
[11][54] (
Figure 2A). This K-pathway in B-family HCOs consist of a series of conserved polar residues that form a proton channel. Most of these residues reside within SU-I, with an additional glutamate at the entry site on SU-II. C-family HCOs possess an alternative K-pathway structurally similar to that in B-family HCOs. Within the
cbb3-HCOs from
Pseudomonas stutzeri and
Rhodobacter sphaeroides, the proton pathway propagates through a few polar residues with the terminal residue tyrosine cross-linked to one of the histidine ligated to Cu
B [11][64].
All HCOs are electrogenic proton pumps, and the internal and intramolecular electron transfer pathways for each HCO family have been extensively studied
[43][65][66][67][68]. The catalytic cycle of A-family HCOs includes two phases, an oxidative phase and a reductive phase, involving several intermediate states of the active site, fully oxidized (
O, Fe
3+ Cu
B2+), single-electron reduced (
E, Fe
3+ Cu
B+ or Fe
2+ Cu
B2+), and two-electron reduced (
R, Fe
2+ Cu
B+) (
Figure 3)
[18]. Upon O
2 binding to
R, a short-lived new complex
A is formed, which delivers electrons rapidly to bound O
2 for the cleavage of the dioxygen bond, forming intermediates
P and
F in sequence
[69]. Both
P and
F are a ferryl derivative of (Fe
4+ = O Cu
B2+), but the former carries Y244 in the radical form, whereas the latter has Y244 reduced and protonated
[31][70][71][72]. Eventually, the fully oxidized state
O is regenerated from
F after receiving an additional electron from CuA/heme
a. During the oxygen reduction, the first proton pumping event occurs during the
P to
F transition, and the reaction cycle completes when
R is regenerated from
E with two product water molecules released, and two protons pumped
[73]. The source of the electron may be a tyrosine or tryptophan radical close to the active site; however, the exact identity of the amino acid, which provides this electron, is still under debate
[43].
Figure 3. The catalytic cycle of HCO and the interplay with NO through the two pathways. The catalytic cycle of HCO is schematically reported with the indication of the redox and the oxygen ligation state of the BNC (heme a3-CuB active site). In the reductive phase, the oxidized species O is fully reduced to R by two single-electron donations via formation of the half-reduced intermediate E. In the oxidative phase, upon reaction with O2, R converts to P and F, and O is regenerated eventually through further electron transfer. The nitrite-bound derivative (Fe3+ CuB2+ NO2−) and the nitrosylated adduct (Fe2+ CuB+ NO) are generated by the reactions of these intermediates with NO. Tyr, CuB-interacting residue Y244, with an asterisk representing the radical form.
The proton-pumping loading sites in all HCOs are still controversial
[74]. The histidines ligating the heme iron and Cu
B, as well as the A- and D- propionates of heme
a3 at the catalytic center have been proposed as candidate proton loading sites
[75]. A hydrophilic cavity above the hemes, housing divalent cations or water molecules at the interspace of SU-I and SU-II, is possibly the beginning of the water exit pathway
[10][11][16].
2.2. bd QO
bd QO is a quinone-type terminal respiratory enzyme that distributes widely in bacteria and archaea. Unlike HCOs that have two hemes and a copper in the active sites,
bd QOs accept electrons from quinones (ubiquinol or menaquinol) to reduce oxygen to water using three hemes (
Figure 2B)
[14][76]. A few isoforms of the
bd QO family contain only
b-type hemes, which are less sensitive to inhibition by cyanide
[77]. Initially,
bd QO was considered to consist of two subunits only, named CydA and CydB, encoded by a single operon
[78]. However, a small single-transmembrane subunit (CydX or CydS) encoded by the third gene in the
cyd operon is found to be not only functionally essential but also involved in the assembly of the enzyme complex in some bacteria in recent years
[79][80][81]. Based on the structural differences of the quinol binding sites (Q-loop),
bd QOs are subdivided into L-subfamily (long Q-loop) and S-subfamily (short Q-loop)
[14][82].
bd QO lacks a counterpart proton-pumping mechanism as in HCOs; instead, they generate a pmf through the transmembrane charge separation and the coupled Q-cycle
[83]. There are two potential proton pathways, CydA and CydB pathways, through which protons could pass from the cytoplasm to the high-spin heme site (
b595) in
bd QO from
Geobacillus thermodenitrificans [15]. The proton transfer from heme
b595 to the oxygen reduction site via heme
d is possibly facilitated by water molecules or the heme propionates of heme
b595 [15]. To date, no oxygen channels have been found in
bd QOs, implying that oxygen molecules likely approach heme
d laterally from the alkyl chain interface with the membrane
[15].
3. Roles of HCOs in the Transformation of Nitrogen Oxides
In addition to carrying out O
2 reduction, HCOs have been shown to be deeply implicated in the biotransformation of multiple nitrogen oxides. It is well recognized that mitochondrial
aa3-HCO is capable of reducing NO
2− to NO under hypoxic conditions
[20][35]. This reduction, which involves only one electron, differs from the four-electron reduction of O
2 to H
2O and is not involved with proton pumping
[35]. Unlike their eukaryotic counterparts, bacterial HCOs, including the
caa3-HCO and
ba3-HCO from
T. thermophilus,
cbb3-HCO from
P. stutzeri, and
bo3-HCO from
E. coli, are able to catalyze the reduction of NO to N
2O
[36][84][85]. In addition,
cbb3-HCO of
P. stutzeri is also able to reduce NO
2− to N
2O directly
[33].
Despite having been known for a long time, the interactions of nitrite with mitochondrial
aa3-HCO have not yet been addressed. Instead, the current understanding of the subject derives mainly from the reactions of nitrite with synthetic heme/copper assemblies
[32][86][87] (
Figure 4A). During catalytic turnover, the ferrous heme of the
a3-Cu
B BNC functions as the electron donor, while the Cu
B center serves as a Lewis acid for the cleavage of the N-O bond of nitrite
[32]. The overall reaction is a one-electron reduction of nitrite, during which an oxygen atom derived from the nitrite is transferred to the BNC, resulting in an oxo-bridge Fe
3+-O-Cu
2+ intermediate. Interestingly, this oxo-bridge Fe
3+-O-Cu
2+ could also oxidize NO back to nitrite. It is speculated that the inter-conversion of nitrite and NO has important implications in the modulation of cellular O
2 balance
[32][86][87]. When O
2 is limited, nitrite interacts with the BNC for reduction, generating NO. This NO molecule can reversibly inhibit the oxygen reduction at the same site, leading to O
2 accumulation. The reverse reaction is composed of the same steps traversed backward, and as a result, NO is converted back to nitrite for future use, and the enzyme is freed to catalyze the four-electron reduction of O
2 to water. The overall reaction is regarded as an adaptive mechanism alleviating NO-mediated respiratory inhibition
[88].
Figure 4. Scheme of a heme-copper assembly mediated oxidation of NO to nitrite and structures of ligand-absent and NO-binding BNCs of CcOs. (
A) A μ-oxo heme-FeIII-O-CuII complex facilitates NO oxidation to nitrite, forming reduced heme and CuII-nitrito complexes. This scheme is modified from the figure in reference
[26] with ChemDraw. (
B) Spatial structure of the BNC of Cyt caa3 oxidase from
T. thermophilus HB8. The copper atom (CuB) is coordinated by three histidine residues. The distance between CuB and heme-iron is less than 5 Å. (
C) X-ray structure of the NO-bound CcO from bovine CcO. The distances between CuB and oxygen atom from NO, heme-Fe, and nitrogen atom from NO are 2.5 Å and 1.8 Å, respectively. Figures of B and C are prepared with PyMOL (Molecular Graphics System, LLC)
https://www.pymol.org (accessed on 20 December 2021).
Multiple lines of evidence suggest that eukaryotic and bacterial HCOs have different reactivities to nitrogen species
[89]. In contrast to eukaryotic mitochondrial HCOs, the bacterial counterparts react with nitrite and NO, producing N
2O as the end product
[33][34][89] (
Figure 4B,C). Upon the addition of NO to oxidized
ba3-HCO (
O) of
T. thermophilus, a six-coordinate heme Fe
2+-NO species has been detected, suggesting that a hyponitrite (HONNO
−) ion bound to the BNC in the
E state (Fe
3+ Cu
B+) is transiently formed
[90]. Further investigations have demonstrated that the binding of two NO molecules to the BNC is accompanied by protonation of the heme
a3-NO species and the electron transfer from Cu
B+, leading to the concomitant formation of the N-N bond. Eventually, N
2O and H
2O are released after additional H
+ is added and the N-O bond is cleaved. This mechanism has been observed in reactions of NO with
caa3-HCO,
bcc-
aa3 supercomplex (formed by cytochrome
bcc and
aa3-HCO in
Mycobacterium tuberculosis), and bacterial NORs, suggesting a possibility that it is conserved in all enzymes of the HCO superfamily
[91][92][93][94]. Despite this, the NO reductase activities of all HCOs tested so far are substantially lower than those observed with bacterial NORs, implying that their contribution to NO reduction is likely to be limited unless NORs are absent
[85][95]. In addition, the
ba3-HCO of
T. thermophilus also interacts with nitrite to form a ferrous heme
a3-nitro complex in the BNC, but NO or N
2O is not produced above the detection limit, suggesting that this enzyme is likely to be susceptible to nitrite inhibition, as discussed below
[89].
On the contrary, under reducing conditions, nitrite reacts with
cbb3-HCO to form a six-coordinate ferrous heme
b3-nitrosyl complex
[33]. The binding of NO
2− to heme
b3 triggers electron transfer from the heme to the substrate, leading to its double protonation reduction to NO and release of a H
2O molecule.