In Escherichia coli Ammonia Inhibits Cytochrome bo3 But Activates Cytochrome bd-I

Interaction of two redox enzymes of Escherichia coli, cytochrome bo3 and cytochrome bd-I, with ammonium sulfate/ammonia at pH 7.0 and 8.3 was studied using high-resolution respirometry and absorption spectroscopy. At pH 7.0, the oxygen reductase activity of none of the enzymes is affected by the ligand. At pH 8.3, cytochrome bo3 is inhibited by the ligand, with 40% maximum inhibition at 100 mM (NH4)2SO4. In contrast, the activity of cytochrome bd-I at pH 8.3 increases with increasing the ligand concentration, the largest increase (140%) is observed at 100 mM (NH4)2SO4. In both cases, the effector molecule is apparently not NH4+ but NH3. The ligand induces changes in absorption spectra of both oxidized cytochromes at pH 8.3. The magnitude of these changes increases as ammonia concentration is increased, yielding apparent dissociation constants Kdapp of 24.3 ± 2.7 mM (NH4)2SO4 (4.9 ± 0.5 mM NH3) for the Soret region in cytochrome bo3, and 35.9 ± 7.1 and 24.6 ± 12.4 mM (NH4)2SO4 (7.2 ± 1.4 and 4.9 ± 2.5 mM NH3) for the Soret and visible regions, respectively, in cytochrome bd-I. Consistently, addition of (NH4)2SO4 to cells of the E. coli mutant containing cytochrome bd-I as the only terminal oxidase at pH 8.3 accelerates the O2 consumption rate, the highest one (140%) being at 27 mM (NH4)2SO4. We discuss possible molecular mechanisms and physiological significance of modulation of the enzymatic activities by ammonia present at high concentration in the intestines, a niche occupied by E. coli.


Introduction
Cytochrome bo 3 and cytochrome bd-I are terminal oxidases in the aerobic respiratory chain of Escherichia coli [1]. Both enzymes catalyze the same redox reaction, the electron transfer from ubiquinol-8 to molecular oxygen giving rise to ubiquinone-8 and water [2,3]. In both cases this reaction is coupled to the formation of an electrochemical proton gradient across the bacterial cytoplasmic membrane [4][5][6]. Nevertheless, the bo 3 oxidase shows a higher energy transduction efficiency than the bd-I oxidase because the former utilizes the proton pumping mechanism [4,7].
The 3D structures of the proteins were determined [8][9][10]. Each one is composed of four different subunits, however the cytochromes are structurally and evolutionarily unrelated. Cytochrome bo 3 is a member of type A-1 of the heme-copper oxidase superfamily [11][12][13][14][15][16]. It carries the ubiquinol binding site, two hemes, b and o 3 , and a copper ion [17]. The latter, denoted Cu B , forms together with heme o 3 a binuclear site in which the oxygen chemistry takes place. The bd-type oxidases form their own family, distinct from the heme-copper superfamily, and the E. coli cytochrome bd-I belongs to the L subfamily of that family [3,18]. The bd-I enzyme has no copper but contains a binding site for ubiquinol and three hemes, b 558 , b 595 , and d. Heme d serves as the site for the O 2 reduction reaction.
Cytochrome bo 3 and cytochrome bd-I are expressed in E. coli under normal and low aeration conditions, respectively [19]. This is consistent with the fact that bo 3 is a lowoxygen-affinity oxidase [20], whereas bd-I is a high-oxygen-affinity oxidase [21]. The

Reagents and Purification of Cytochromes bd-I and bo 3 from E. coli
Tris Base was purchased from Fisher BioReagents. Other chemicals were purchased from Sigma-Aldrich. Cytochromes bd-I and bo 3 were isolated from the E. coli strains GO105/pTK1 and GO105/pJRhisA, respectively, as described by [35][36][37]. In the case of cytochrome bd-I, the fractions with an absorbance ratio of A 412 /A 280 ≥ 0.7 eluted from a DEAE Sepharose Fast Flow anion exchange column were pooled and concentrated [36]. Cytochrome bo 3 preparations were a kind gift of Marina Verkhovskaya (University of Helsinki). Being a His-tagged fusion protein, cytochrome bo 3 was purified by immobilized metal chelate affinity chromatography on Ni-NTA Agarose. The fractions eluted from the column containing pure four-subunit cytochrome bo 3 were pooled and concentrated [37]. The sample quality was evaluated by measuring enzyme activity and absorption spectra. All prepared samples showed high oxygen reductase activity (V max of about 150 mol O 2 /mol cytochrome bd-I/s and 60 mol O 2 /mol cytochrome bo 3 /s in the presence of the electron donor/mediator couple 10 mM dithiothreitol (DTT) and 0.25 mM 2,3-dimethoxy-5-methyl-6-(3-methyl-2-butenyl)-1,4-benzoquinone (Q 1 ) at pH 7.0) and characteristic absorption spectra both "as prepared" and dithionite-reduced. The concentration of cytochrome bd-I was determined from the dithionite reduced-minus-"as prepared" difference absorption spectra using ∆ε 628-607 of 10.8 mM −1 cm −1 [38]. Cytochrome bo 3 concentration was estimated from the Soret absorption band of the oxidized enzyme using ε 407 of 182 mM −1 cm −1 [39].

Measurement Techniques and Assay Conditions.
O 2 consumption of cytochromes bd-I and bo 3 was measured using an Oxygraph-2k high-resolution respirometer (Oroboros Instrument, Innsbruck, Austria) equipped with two 1.5-mL chambers. UV-visible absorption spectra were recorded in an Agilent Cary 60 UV-Vis spectrophotometer. Assays were performed at 25 • C in 100 mM Tris-phosphate (pH 8.3) or 100 mM potassium phosphate (pH 7.0) buffer containing 0.1 mM EDTA, 2.5 µg/mL catalase, 10 mM DTT, 0.25 mM Q 1 , and either 0.05% N-lauroyl-sarcosine (cytochrome bd-I) or 0.02% dodecyl-β-D-maltoside (cytochrome bo 3 ). The concentrations of cytochrome bo 3 and cytochrome bd-I used in the oxygraphic measurements were 20 nM and 7.8 nM at pH 8.3, and 12 nM and 3.9 nM at pH 7.0 respectively. In the spectroscopic measurements, the concentrations of cytochrome bo 3 and cytochrome bd-I were 4.8 µM and 3.2 µM respectively. The pH of the stock solutions of (NH 4 ) 2 SO 4 and K 2 SO 4 was adjusted to the desired values (8.3 or 7.0). To generate the reduced state of cytochrome bo 3 or cytochrome bd-I, a few grains of solid sodium dithionite were added. The oxidized state of cytochrome bd-I was produced by incubating the "as isolated" enzyme with 33 µM tetrachloro-1,4-benzoquinone Antioxidants 2021, 10, 13 3 of 15 for 10 min [40]. To remove excess oxidant, the sample was centrifuged at 4 • C and the yellow pellet discarded.

Data Analysis
Data analysis was carried out using Origin (OriginLab Corporation). To compare the (NH 4 ) 2 SO 4 titration data obtained in oxygraphic and spectroscopic experiments, they were fitted to the standard hyperbolic equation y = A max ·x/(K dapp + x) using a built-in approximation function ("hyperbola function") in "advanced fitting tool" in the Origin program. A max and K dapp parameters were allowed to vary. K dapp is an apparent dissociation constant. In oxygraphic experiments, A max is either a theoretical maximum percent inhibition (cytochrome bo 3 ) or a theoretical maximum percent activity (cytochrome bd-I). In spectroscopic experiments, A max is a theoretical maximum absorption change. R-square (R 2 ) and standard deviation reported by the Origin program are shown in the figure legends. Since the high ionic strength may affect the activity of an enzyme [41], in order to take into account this possible effect, K 2 SO 4 at the same concentration was added to the enzyme for each condition as a control.

Effect of (NH 4 ) 2 SO 4 on O 2 Reductase Activity of E. coli Cytochrome bo 3
The effect of (NH 4 ) 2 SO 4 on the O 2 -reductase activity of the isolated cytochrome bo 3 from E. coli was examined by measuring the O 2 consumption rates before and after addition of the effector at pH 8.3 or 7.0. Figure 1A shows that at pH 8.3 the addition of 50 mM (NH 4 ) 2 SO 4 rapidly inhibits the O 2 reductase activity of cytochrome bo 3 by 51%. To take into account the effect of increasing ionic strength on enzyme activity, K 2 SO 4 was added to the oxidase under the same conditions as a control. As shown in Figure 1A, 50 mM K 2 SO 4 at pH 8.3 inhibits cytochrome bo 3 to a much lesser extent (by 18%). At pH 7.0 (NH 4 ) 2 SO 4 does not inhibit the O 2 reductase activity of the bo 3 enzyme ( Figure 1B). Figure 1C shows that with the increase in (NH 4 ) 2 SO 4 concentration, the inhibitory effect at pH 8.3 is progressively increased. The maximum inhibition observed (after subtraction of the corresponding control value with K 2 SO 4 ) was 40% at 100 mM (NH 4 ) 2 SO 4 .
We also studied the effect of (NH 4 ) 2 SO 4 on O 2 consumption by E. coli mutant cells expressing cytochrome bo 3 as a single terminal oxidase. Supplementary Figures S1B and S2 show that within the limits of the experimental error the addition of (NH 4 ) 2 SO 4 up to 27 mM at pH 8.3 caused no significant change in the respiration of the intact cells.

Effect of (NH 4 ) 2 SO 4 on Absorption Spectra of E. coli Cytochrome bo 3
The finding that (NH 4 ) 2 SO 4 can inhibit cytochrome bo 3 pushed us to explore the effect of ammonia on absorption spectra of the isolated cytochrome bo 3 . Figure 2A shows the spectral changes induced by the addition of (NH 4 ) 2 SO 4 at millimolar concentrations to the oxidized cytochrome bo 3 at pH 8.3. The ammonia-induced spectrum showed a red shift of the enzyme Soret band with a maximum at 416 nm and a minimum at 400 nm. In the visible spectrum, some weak intensity broad bands were displayed, the most pronounced of which was a band with a minimum around 630 nm. The spectral changes caused by the ligand are possibly due to its binding to the heme o 3 -Cu B binuclear site. The observed changes could be also due to a small reduction of the enzyme. However, the reduced-minus-oxidized spectrum in the Soret region of cytochrome bo 3 showed a maximum at 428-430 nm [39,42] rather than 416 nm. The magnitude of the Soret spectral changes increased as (NH 4 ) 2 SO 4 concentration was increased ( Figure 2B). Analysis of the titration curve yields K dapp of 24.3 ± 2.7 mM (NH 4 ) 2 SO 4 and the value for maximum absorption change at 416-400 nm A max of 22.1 ± 0.6 mM −1 cm −1 . It has to be noted that the titration profile ( Figure 2B) is similar to that for the plot of percent inhibition versus (NH 4 ) 2 SO 4 ( Figure 1C). The addition of (NH 4 ) 2 SO 4 to the dithionite-reduced cytochrome bo 3 under identical conditions brought about no spectral change.   reductase activity of cytochrome bo 3 measured at pH 8.3 at increasing concentration of (NH 4 ) 2 SO 4 . The effect of increasing ionic strength on enzyme activity is taken into account for each data point by subtracting the percent inhibition value in the presence of K 2 SO 4 (control) from that in the presence of (NH 4 ) 2 SO 4 at the same concentration. Experimental data (filled circles) are shown together with their best fit (solid line) to the hyperbolic equation (see Materials and Methods), giving a maximum percent inhibition value A max of 58.4 ± 10.1%, and K dapp of 44.2 ± 16.9 mM (NH 4 ) 2 SO 4 (8.9 ± 3.4 mM NH 3 ) (mean ± standard deviation, n = 3, R 2 = 0.83663). O 2 reductase activity of cytochrome bo 3 is sustained by 10 mM DTT and 0.25 mM Q 1 . Enzyme, 20 nM (A,C) or 12 nM (B). In the absence of (NH 4 ) 2 SO 4 , V max values are 28 ± 2 and 60 ± 5 mol O 2 /mol enzyme/s at pH 8.3 and 7.0, respectively. changes increased as (NH4)2SO4 concentration was increased ( Figure 2B). Analysis of the titration curve yields Kdapp of 24.3 ± 2.7 mM (NH4)2SO4 and the value for maximum absorption change at 416-400 nm Amax of 22.1 ± 0.6 mM −1 cm −1 . It has to be noted that the titration profile ( Figure 2B) is similar to that for the plot of percent inhibition versus (NH4)2SO4 ( Figure 1C). The addition of (NH4)2SO4 to the dithionite-reduced cytochrome bo3 under identical conditions brought about no spectral change.

Effect of (NH4)2SO4 on O2 Reductase Activity of E. coli Cytochrome bd-I
Next, we studied the influence (NH4)2SO4 on the O2-reductase activity of the isolated cytochrome bd-I from E. coli under the same conditions as used for the bo3 oxidase. We found that in contrast to cytochrome bo3, cytochrome bd-I is not inhibited by millimolar concentrations of (NH4)2SO4 at pH 8.3. Furthermore, under these conditions, activation of the catalytic activity of the enzyme was observed.
As shown in Figure 3A, at pH 8.3 the addition of 25 mM (NH4)2SO4 increased the rate of O2 consumption of cytochrome bd-I by 39%. In the control with 25 mM K2SO4, the

Effect of (NH 4 ) 2 SO 4 on O 2 Reductase Activity of E. coli Cytochrome bd-I
Next, we studied the influence (NH 4 ) 2 SO 4 on the O 2 -reductase activity of the isolated cytochrome bd-I from E. coli under the same conditions as used for the bo 3 oxidase. We found that in contrast to cytochrome bo 3 , cytochrome bd-I is not inhibited by millimolar concentrations of (NH 4 ) 2 SO 4 at pH 8.3. Furthermore, under these conditions, activation of the catalytic activity of the enzyme was observed.
As shown in Figure 3A, at pH 8.3 the addition of 25 mM (NH 4 ) 2 SO 4 increased the rate of O 2 consumption of cytochrome bd-I by 39%. In the control with 25 mM K 2 SO 4 , the increase in the rate was significantly lower (by 11%). At pH 7.0, there was no effect of (NH 4 ) 2 SO 4 on the O 2 reductase activity of cytochrome bd-I ( Figure 3B). At pH 8.3, the O 2 reductase activity of cytochrome bd-I increased with an increase in ammonia concentration ( Figure 3C). Maximum activation in enzyme activity, 140%, was observed following the addition of 100 mM (NH 4 ) 2 SO 4 (after subtraction of the control with K 2 SO 4 ). increase in the rate was significantly lower (by 11%). At pH 7.0, there was no effect of (NH4)2SO4 on the O2 reductase activity of cytochrome bd-I ( Figure 3B). At pH 8.3, the O2 reductase activity of cytochrome bd-I increased with an increase in ammonia concentration ( Figure 3C). Maximum activation in enzyme activity, 140%, was observed following the addition of 100 mM (NH4)2SO4 (after subtraction of the control with K2SO4). Values are expressed with reference to the activity measured before sulfate addition taken as 100%. The effect of increasing ionic strength on enzyme activity is taken into account for each data point by subtracting the activity value in the presence of K2SO4 (control) from that in the presence of (NH4)2SO4 at the same concentration. Experimental data (filled circles) are shown together with their best fit (solid line) to the hyperbolic equation (see Materials and Methods), giving a maximum activity value Amax of 152.9 ± 6.3%, and Kdapp of 28.6 ± 8.7 mM (NH4)2SO4 (5.7 ± 1.7 mM NH3) (mean ± standard deviation, n = 3, R^2 = 0.92789). O2 reductase activity of cytochrome bd-I is sustained by 10 mM DTT and 0.25 mM Q1. Enzyme, 7.8 nM (A,C) or 3.9 nM (B). In the absence of (NH4)2SO4, Vmax values are 30 ± 6 and 152 ± 15 mol O2/mol enzyme/s at pH 8.3 and 7.0, respectively.
We also studied the effect of (NH4)2SO4 on O2 consumption by E. coli mutant cells containing cytochrome bd-I as the only terminal oxidase. Supplementary Figures S1A and S2 show that at pH 8.3 the addition of (NH4)2SO4 up to 27 mM increased respiration of the Values are expressed with reference to the activity measured before sulfate addition taken as 100%. The effect of increasing ionic strength on enzyme activity is taken into account for each data point by subtracting the activity value in the presence of K 2 SO 4 (control) from that in the presence of (NH 4 ) 2 SO 4 at the same concentration. Experimental data (filled circles) are shown together with their best fit (solid line) to the hyperbolic equation (see Materials and Methods), giving a maximum activity value A max of 152.9 ± 6.3%, and K dapp of 28.6 ± 8.7 mM (NH 4 ) 2 SO 4 (5.7 ± 1.7 mM NH 3 ) (mean ± standard deviation, n = 3, R 2 = 0.92789). O 2 reductase activity of cytochrome bd-I is sustained by 10 mM DTT and 0.25 mM Q 1 . Enzyme, 7.8 nM (A,C) or 3.9 nM (B). In the absence of (NH 4 ) 2 SO 4 , V max values are 30 ± 6 and 152 ± 15 mol O 2 /mol enzyme/s at pH 8.3 and 7.0, respectively.
We also studied the effect of (NH 4 ) 2 SO 4 on O 2 consumption by E. coli mutant cells containing cytochrome bd-I as the only terminal oxidase. Supplementary Figures S1A and  S2 show that at pH 8.3 the addition of (NH 4 ) 2 SO 4 up to 27 mM increased respiration of the intact cells. Maximum acceleration of the O 2 consumption rate (140%) was observed at 27 mM (NH 4 ) 2 SO 4 . In the control with K 2 SO 4 added at the same concentrations, there was no significant change in O 2 consumption by the cells.

Effect of (NH 4 ) 2 SO 4 on Absorption Spectra of E. coli Cytochrome bd-I
Finally, we found that ammonia affects the absorption spectrum of the isolated cytochrome bd-I in the fully oxidized state. Figure 4A displays the spectral changes caused by the addition of (NH 4 ) 2 SO 4 at millimolar concentrations to the oxidized bd-I enzyme at pH 8.3.
Antioxidants 2021, 10, x FOR PEER REVIEW 7 of 15 intact cells. Maximum acceleration of the O2 consumption rate (140%) was observed at 27 mM (NH4)2SO4. In the control with K2SO4 added at the same concentrations, there was no significant change in O2 consumption by the cells.

Effect of (NH4)2SO4 on Absorption Spectra of E. coli Cytochrome bd-I
Finally, we found that ammonia affects the absorption spectrum of the isolated cytochrome bd-I in the fully oxidized state. Figure 4A displays the spectral changes caused by the addition of (NH4)2SO4 at millimolar concentrations to the oxidized bd-I enzyme at pH 8.3. In the Soret region, the ammonia-induced spectrum showed a red shift with a maximum at about 423-428 nm and a minimum around 395-408 nm. In the near-IR region, there were two maxima at ~623 and ~673 nm, and a broad minimum at ~740 nm. The changes may report the interaction of ammonia with the ferric heme d. The magnitude of the changes increased with increasing concentrations of (NH4)2SO4. Analysis of the absorption titration curves measured at 426-408 and 623-595 nm ( Figure 4B,C) yields In the Soret region, the ammonia-induced spectrum showed a red shift with a maximum at about 423-428 nm and a minimum around 395-408 nm. In the near-IR region, there were two maxima at~623 and~673 nm, and a broad minimum at~740 nm. The changes may report the interaction of ammonia with the ferric heme d. The magnitude of the changes increased with increasing concentrations of (NH 4 ) 2 SO 4 . Analysis of the absorption titration curves measured at 426-408 and 623-595 nm ( Figure 4B,C) yields K dapp of 35.9 ± 7.1 and 24.6 ± 12.4 mM (NH 4 ) 2 SO 4 , respectively. The titration profile ( Figure 4B,C) is similar to that for the plot of percent enzyme activity versus (NH 4 ) 2 SO 4 ( Figure 3C).
No spectral change was observed, when under the same conditions, (NH 4 ) 2 SO 4 was added to cytochrome bd-I in the dithionite-reduced state.

Proposed Mechanism for Inhibition of Cytochrome bo 3 by Ammonia
The interaction of ammonia with a quinol oxidase has not been investigated before. Recently, von der Hocht et al. observed that in the presence of 20 mM (NH 4 ) 2 SO 4 at pH 9 the activity of the isolated aa 3 -type cytochrome c oxidase from Paracoccus denitrificans sustained by ascorbate, N,N,N',N'-tetramethyl-p-phenylenediamine and cytochrome c decreased by 22% [43]. They also reported that at pH 9 the addition of ammonia to the H 2 O 2 -generated F state led to its conversion into a novel P state called P N [43]. P and F are two transient ferryl intermediates formed sequentially during the catalytic cycle of both heme-copper and bd-type terminal oxidases [44][45][46][47][48][49][50]. In the case of cytochrome c oxidase, spectrally similar artificial P and F intermediates can also be produced by the addition of H 2 O 2 at different concentrations to the enzyme in the oxidized (O) state at alkaline pH [51,52]. In the new P N state [43], ammonia binds to Cu B , as shown by resonance Raman spectroscopy [53].
Here, we showed that the O 2 consumption of the isolated E. coli cytochrome bo 3 , sustained by DTT and Q 1 , was inhibited by (NH 4 ) 2 SO 4 at pH 8.3 ( Figure 1A). At the maximum concentration of (NH 4 ) 2 SO 4 used (100 mM), the enzyme activity decreased by 40% ( Figure 1C). The inhibition was not observed at pH 7.0. The pK a of ammonium in aqueous solution is known to be 9. 25 1.1 times). Thus, we can conclude that it is ammonia rather than the ammonium ion that inhibits cytochrome bo 3 .
At pH 8.3, ammonia caused a red shift in the Soret band of the oxidized cytochrome bo 3 (Figure 2A). The magnitude of the absorption changes increased with increasing the ligand concentration giving K dapp of 24.3 ± 2.7 mM (NH 4 ) 2 SO 4 that corresponds to 4.9 ± 0.5 mM NH 3 ( Figure 2B). The titration profile was similar to that for the inhibition of the enzyme activity by ammonia ( Figure 1C). Ingledew et al. earlier reported that the addition of cyanide induces a red shift in the Soret band of the oxidized cytochrome bo 3 [54]. The binding of a ligand to the high-spin heme brings about a red shift of the Soret band in enzyme absorption spectra [1,54,55]. On the contrary, no absorption change or a small blue shift in the Soret band was observed when a ligand binds to Cu B [1,56]. Thus, we suggest that ammonia binds to heme o 3 . This conclusion is supported by the fact that the ligand-induced red shift in the Soret band was accompanied by the loss of the broad charge transfer band around 630 nm (Figure 2A). The 630-nm band is characteristic of the fully oxidized binuclear site in which heme o 3 is in a high-spin state [54]. The decay of this band suggests the conversion of the high-spin heme o 3 into the low-spin ammonia complex. As a ferric heme usually binds an anion, we propose that NH 3 [55,57,58]. Compared to these two ligands, ammonia is a much smaller molecule. Rather, NH 3 can be compared to a water/hydroxide molecule that is a natural ligand of Cu B in several states of the catalytic cycle of heme-copper oxidases [59]. For this reason, NH 3 cannot be a bridging ligand at the binuclear site since a distance is around 4-5 Å. However, the binding of two NH 3 molecules at the oxidized binuclear site at a time, one to heme o 3 and the other to Cu B , may occur. Indeed, NH 3 is approximately the size of a water/hydroxide molecule. In some catalytic intermediates of a heme-copper oxidase, two molecules of H 2 O (or OH − ) can bind simultaneously at the binuclear site [59][60][61]. It could also be true for ammonia. We propose to designate the ammonia adduct of the oxidized cytochrome bo 3 as N.
Importantly, the binding of hydroxide (as opposed to water) with the oxidized heme at the binuclear site leads to the transition of the heme from the high-spin to the low-spin state [62]. This process is enhanced at alkaline pH. For instance, at pH 9, about 50% of the ferric heme a 3 is hydroxide-ligated whereas at pH 6.5, no hydroxide is bound to the heme [63]. The spectral shift caused by ammonia (Figure 2A) is similar to the effect of the formation of low-spin complexes of the initial high-spin heme with anionic ligands at the binuclear site. We suggest that when ammonia binds to the oxidized heme at alkaline pH, the complex with deprotonated ammonia (NH 2 − ) is formed, just as it happens with hydroxide. It has to be noted that a similar complex can be produced in cytochrome c nitrite reductase, before the release of the neutral ammonia, the final product of nitrite reduction, from the catalytic site [64]. Notably, the presence of a tyrosine residue near the heme is critically important. In cytochrome c nitrite reductase, the residue facilitates the transition of the heme from the high-spin to the low-spin state (via the stabilization of the radical form of the bound NH 2 − ) and serves as a proton donor/acceptor [64]. Surprisingly, all heme-copper oxidases also contain a conserved tyrosine residue that is part of the binuclear site. The tyrosine is bound to Cu B through a histidine ligand and is critical for the proton pumping function. At the same time, the structure of cytochrome bd that lacks the proton pump shows no tyrosine residue nearby heme d (or heme b 595 ) [9,10,65]. This could explain the higher sensitivity of the bo 3 oxidase to the inhibitory effect of ammonia as compared to the bd oxidase. Figure 5 shows a proposed molecular mechanism of inhibition of the catalytic activity of cytochrome bo 3 by ammonia. The ligand binds to the catalytic intermediates O and F thereby blocking the oxygen reduction reaction cycle of the enzyme.
Antioxidants 2021, 10, x FOR PEER REVIEW 9 of 15 may occur. Indeed, NH3 is approximately the size of a water/hydroxide molecule. In some catalytic intermediates of a heme-copper oxidase, two molecules of H2O (or OH − ) can bind simultaneously at the binuclear site [59][60][61]. It could also be true for ammonia. We propose to designate the ammonia adduct of the oxidized cytochrome bo3 as N. Importantly, the binding of hydroxide (as opposed to water) with the oxidized heme at the binuclear site leads to the transition of the heme from the high-spin to the low-spin state [62]. This process is enhanced at alkaline pH. For instance, at pH 9, about 50% of the ferric heme a3 is hydroxide-ligated whereas at pH 6.5, no hydroxide is bound to the heme [63]. The spectral shift caused by ammonia (Figure 2A) is similar to the effect of the formation of low-spin complexes of the initial high-spin heme with anionic ligands at the binuclear site. We suggest that when ammonia binds to the oxidized heme at alkaline pH, the complex with deprotonated ammonia (NH2 − ) is formed, just as it happens with hydroxide. It has to be noted that a similar complex can be produced in cytochrome c nitrite reductase, before the release of the neutral ammonia, the final product of nitrite reduction, from the catalytic site [64]. Notably, the presence of a tyrosine residue near the heme is critically important. In cytochrome c nitrite reductase, the residue facilitates the transition of the heme from the high-spin to the low-spin state (via the stabilization of the radical form of the bound NH2 − ) and serves as a proton donor/acceptor [64]. Surprisingly, all heme-copper oxidases also contain a conserved tyrosine residue that is part of the binuclear site. The tyrosine is bound to CuB through a histidine ligand and is critical for the proton pumping function. At the same time, the structure of cytochrome bd that lacks the proton pump shows no tyrosine residue nearby heme d (or heme b595) [9,10,65]. This could explain the higher sensitivity of the bo3 oxidase to the inhibitory effect of ammonia as compared to the bd oxidase. Figure 5 shows a proposed molecular mechanism of inhibition of the catalytic activity of cytochrome bo3 by ammonia. The ligand binds to the catalytic intermediates O and F thereby blocking the oxygen reduction reaction cycle of the enzyme.  [66]. NH3 presumably converts F into the PN state (o3 4+ =O 2− CuB 2+ -NH3) and O into the ammonia complex N (o3 3+ -NH2 − CuB 2+ and/or o3 3+ -NH2 − CuB 2+ -NH3), thereby leading to the inhibition of the oxidase activity.

Proposed Mechanism for Ammonia-Induced Acceleration of the Cytochrome bd-I Activity
In contrast to cytochrome bo 3 , cytochrome bd-I at pH 8.3 was not inhibited by (NH 4 ) 2 SO 4 ( Figure 3A). Furthermore, the addition of the ligand led to an increase in enzyme activity. The highest enhancement of the rate of cytochrome bd-I-catalyzed reaction (140%) was achieved upon the addition of 100 mM (NH 4 ) 2 SO 4 ( Figure 3C). The fact that at pH 7.0 (NH 4 ) 2 SO 4 did not affect the enzyme activity ( Figure 3B) suggests that the activator was ammonia rather than the ammonium ion. Consistently, the addition of (NH 4 ) 2 SO 4 to intact cells of the E. coli mutant possessing cytochrome bd-I as the sole terminal oxidase at pH 8.3 increased cell respiration (Supplementary Figures S1A and S2). Maximum acceleration of the O 2 consumption rate (140%) was observed at 27 mM (NH 4 ) 2 SO 4 .
At pH 8.3, the addition of ammonia brought about spectral changes in the fully oxidized cytochrome bd-I, the amplitude of which increased with increasing the ligand concentration ( Figure 4). The titration curves ( Figure 4B,C) were similar to that of the ammonia-induced activity change ( Figure 3C). Surprisingly, the ammonia-induced difference absorption spectra ( Figure 4A) were similar to the difference absorption spectra recorded following addition of H 2 O 2 to the fully oxidized cytochrome bd-I [67,68]. In the reaction product, heme d was in the ferryl state [69]. As in the case of cytochrome c oxidase [51,52], after the addition of excess peroxide to cytochrome bd-I, the two ferryl species, P and F, were probably produced. P discovered by [47] is a heme d ferryl porphyrin π-cation radical intermediate [49]. Thus, the H 2 O 2 -induced difference spectra reported in [67,68] likely reflect a mixture of P and F.
It is known that in the air-oxidized cytochromes bd from E. coli and Azotobacter vinelandii heme d is mostly in the oxygenated form [7,40,70,71]. This state is often called compound A 1 (see Figure 6). Jünemann and Wrigglesworth reported that exposure of the A. vinelandii cytochrome bd in an air-oxidized state to alkaline pH leads to deoxygenation of heme d [71]. At alkaline pH, the heme d oxy-complex in the A 1 state is destabilized and may decay to the oxidized (O) state ( Figure 6). In contrast to cytochrome bo3, cytochrome bd-I at pH 8.3 was not inhibited by (NH4)2SO4 ( Figure 3A). Furthermore, the addition of the ligand led to an increase in enzyme activity. The highest enhancement of the rate of cytochrome bd-I-catalyzed reaction (140%) was achieved upon the addition of 100 mM (NH4)2SO4 ( Figure 3C). The fact that at pH 7.0 (NH4)2SO4 did not affect the enzyme activity ( Figure 3B) suggests that the activator was ammonia rather than the ammonium ion. Consistently, the addition of (NH4)2SO4 to intact cells of the E. coli mutant possessing cytochrome bd-I as the sole terminal oxidase at pH 8.3 increased cell respiration (Supplementary Figures S1A and S2). Maximum acceleration of the O2 consumption rate (140%) was observed at 27 mM (NH4)2SO4.
At pH 8.3, the addition of ammonia brought about spectral changes in the fully oxidized cytochrome bd-I, the amplitude of which increased with increasing the ligand concentration ( Figure 4). The titration curves ( Figure 4B,C) were similar to that of the ammonia-induced activity change ( Figure 3C). Surprisingly, the ammonia-induced difference absorption spectra ( Figure 4A) were similar to the difference absorption spectra recorded following addition of H2O2 to the fully oxidized cytochrome bd-I [67,68]. In the reaction product, heme d was in the ferryl state [69]. As in the case of cytochrome c oxidase [51,52], after the addition of excess peroxide to cytochrome bd-I, the two ferryl species, P and F, were probably produced. P discovered by [47] is a heme d ferryl porphyrin π-cation radical intermediate [49]. Thus, the H2O2-induced difference spectra reported in [67,68] likely reflect a mixture of P and F.
It is known that in the air-oxidized cytochromes bd from E. coli and Azotobacter vinelandii heme d is mostly in the oxygenated form [7,40,70,71]. This state is often called compound A 1 (see Figure 6). Jünemann and Wrigglesworth reported that exposure of the A. vinelandii cytochrome bd in an air-oxidized state to alkaline pH leads to deoxygenation of heme d [71]. At alkaline pH, the heme d oxy-complex in the A 1 state is destabilized and may decay to the oxidized (O) state ( Figure 6).   [72]. NH3 possibly promotes the formation of P from O, thereby leading to the acceleration of the oxidase activity. It is also possible that NH3 reacts with O 1 producing F. In the latter two reactions, NH3 serves as a two-electron donor being oxidized to NH2OH. The reaction of NH3 with O 1 is not shown for the sake of simplicity.  152 ± 15 mol O 2 /mol enzyme/s at pH 7.0). We hypothesize that the addition of ammonia to the oxidized cytochrome bd-I at pH 8.3 promotes the formation of P from O ( Figure 6), thereby increasing the enzyme activity. NH 3 may be oxidized to NH 2 OH serving as a two-electron donor in this reaction. NH 3 could also react with the one electron-reduced enzyme (O 1 ) with the production of F and NH 2 OH.

Possible Physiological
Significance of the Difference between Cytochrome bo 3 and Cytochrome bd-I in Sensitivity toward Ammonia for E. coli Along with nitric oxide, carbon monoxide, and hydrogen sulfide, ammonia is considered as a "gasotransmitter" or endogenously generated gaseous signaling molecule [74]. A signaling role of NH 3 in cultured rat astrocytes has been reported [75]. The molecule can move across the plasma membranes both passively [76] and actively via the Amt/Rh family of ammonium/ammonia transporting membrane proteins, involving the electrogenic transport mechanism [77,78]. Ammonia is a degradation product of proteins, peptides, amino acids, and urea. It is mainly produced by the gut microbiota and the gut, liver, kidney, and muscle cells. An adult human gut generates 4-10 g of NH 3 daily [79]. E. coli is one of the most active ammonia-producers in the gut microbiota. Ammonia can also be recycled into amino acid synthesis [80]. NH 3 is a potent infochemical in bacteria-bacteria interactions. It is able to induce oxidative stress responses and increase resistance to antibiotics thus playing a role in defense mechanisms against antimicrobials [81]. At high enough concentrations, ammonia is toxic to cells, especially to neurons. For this reason, the plasma concentration of ammonia in healthy adults is maintained in the range of 10-35 µM [82]. In the intestines, the total ammonia concentration depends on intestinal segment and diet but in general it is about 1000-times higher than in blood [34]. The highest ammonia concentration in the body (27.2 ± 17.5 mM) is reported to be in the large intestinal lumen [34]. Intestinal pH shows high variability depending on the intestinal segment, diet, and regional distribution of microbiota, and there are conditions in which the pH values are in the alkaline region [83,84]. For example, in the distal ileum, the median pH is 8.1 [85]. In light of the above, we suggest that the difference between the two quinol oxidases in the sensitivity toward ammonia can have a physiological significance for E. coli and other enterobacteria. In contrast to the heme-copper oxidase bo 3 that is inhibited by NH 3 at alkaline pH, the bd-I oxidase under the same conditions is not inhibited but activated by the ligand. Thus cytochrome bd can sustain bacterial respiration in the presence of high concentrations of not only sulfide [26,27,86], but also ammonia.

Conclusions
In summary, we investigated the sensitivity of two physiologically important respiratory cytochromes of E. coli, bo 3 and bd-I, to ammonia at the level of both isolated enzymes and intact cells. It turned out that at pH 8.3 the isolated heme-copper bo 3 oxidase is partly inhibited by NH 3 . Surprisingly, under the same conditions, the isolated copper-lacking bd-I enzyme is not only resistant to but also activated by this gaseous signaling molecule. Consistently, respiration of intact cells of the E. coli mutant that relies on cytochrome bd-I as the only terminal oxidase is accelerated by NH 3 . With such a unique trait, the bd-type redox enzyme may provide E. coli and perhaps other bacteria with the ability to maintain the aerobic energy metabolism in the gut and other ammonia-rich environments.