HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 277, Issue 17, 14894-14901, April 26, 2002

This Article Abstract Full Text (PDF) All Versions of this Article: 277/17/14894    most recent M111922200v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Mills, D. A. Articles by Ferguson-Miller, S. Search for Related Content PubMed PubMed Citation Articles by Mills, D. A. Articles by Ferguson-Miller, S. Social Bookmarking                      What's this?

Membrane Potential-controlled Inhibition of Cytochrome c Oxidase by Zinc^*

Denise A. Mills, Bryan Schmidt, Carrie Hiser, Erica Westley^§, and Shelagh Ferguson-Miller^¶

From the Department of Biochemistry, Michigan State University, East Lansing, Michigan 48824-1319

Received for publication, December 14, 2001, and in revised form, January 31, 2002

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Like many voltage-sensitive ion pumps, cytochrome c oxidase is inhibited by zinc. Binding of zinc to the outside surface of Rhodobacter sphaeroides cytochrome c oxidase inhibits the enzyme with a K_I of  5 µM when the enzyme is reconstituted into phospholipid vesicles in the presence of a membrane potential. In the absence of a membrane potential and a pH gradient, millimolar concentrations of zinc are required to inhibit. This differential inhibition causes a dramatic increase in the respiratory control ratio from 6 to 40 for wild-type oxidase. The external zinc inhibition is removed by EDTA and is not competitive with cytochrome c binding but is competitive with protons. Only Cd²⁺ of the many metals tested (Mg²⁺, Mn²⁺, Ca²⁺, Ba²⁺, Li²⁺, Cs²⁺, Hg²⁺, Ni²⁺, Co²⁺, Cu²⁺ Tb³⁺, Tm³⁺) showed inhibitory effects similar to Zn²⁺. Proton pumping is slower and less efficient with zinc. The results suggest that zinc inhibits proton movement through a proton exit path, which can allow proton back-leak at high membrane potentials. The physiological and mechanistic significance of proton movement in the exit pathway and its blockage by zinc is discussed in terms of regulation of the efficiency of energy transduction.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Zinc is observed to have an unusually strong and specific inhibitory effect on a number of proton and ion channels, but the physiological significance of this inhibition is the subject of much debate (1). The controversy is heightened by the difficulty in accurately assessing free zinc levels in tissues and subcellular compartments (1, 2). In mitochondria, zinc is known to inhibit the bc₁ complex at submicromolar levels (3), and similar strong inhibition is observed in bacterial photosynthetic reaction centers (3, 4). In each case, the zinc inhibition is reversible and blocks a proton pathway. In the reaction center, where addition of Zn²⁺ limits a proton uptake step, the site for zinc binding was determined by x-ray crystallography (4) allowing the identification of the predominant route for proton uptake and two aspartates that participate in that proton transfer pathway.

Cytochrome c oxidase (CcO)¹ requires protons as a substrate, using 4 protons for the reduction of O₂ to form 2H₂O and translocating 4 protons across the membrane for each O₂ reduced. A total of 8 protons are therefore taken up from the interior per 4 electrons donated externally by cytochrome c, the electron donor. Crystal structures of cytochrome c oxidase have helped identify two proposed channels for proton uptake in subunit I (5-7), referred to as the D and K channels, because an aspartate and a lysine, respectively, are important for their activity (Fig. 1). The roles of these channels are still debated in terms of the number and destination of the protons they conduct and how the individual proton uptake events are coupled to electron transfer. Even less well understood is the pathway for proton release, although several proposals have been made (8, 9). A hydrogen-bonded network, the H channel, has been suggested as a potential proton pathway that spans the membrane in the bovine oxidase (7), but its presence in bacterial oxidases has not been confirmed (10).

View larger version (38K): [in this window] [in a new window]   Fig. 1.   The proposed proton paths and external surface amino acids in R. sphaeroides cytochrome c oxidase. A, a side view showing the metal centers (CuA, CuB, and Mg) and hemes are depicted with possible proton paths: K path (blue) with residues (bottom to top) Ser-299, Lys-362, Thr-359, and Tyr-288; Thr-352; D path (red) with residues Asp-132, Asn-121 and Asn-139, Ser-197, and Glu-286; H path (dashed brown arrow) with residues (yellow) Glu-450, His-456, Ser-425 and Ser-504, Gln-471, Arg-52, Ser-497, and Tyr-414. A tentative connection (dotted black arrow) from the D path to possible exit paths (black solid arrows) is shown. The outside amino acids that are also depicted in view B are colored magenta. Arg-482 (which interacts with the propionates of heme a, close to Arg-481) and His-334 (CuB ligand) are shown in gray but not labeled. B, top view showing the outside amino acids discussed in this report with side-chain color matched to the color of the amino acid number. The pathways, D, K, and H, are labeled in Subunit I (gray helices) corresponding to the apparent "pores" identified by Iwata et al. (5). Also shown are the 2 hemes (black wire frame), Cu and Mg (red- and yellow-filled circles, respectively). Subunit III is yellow and subunit IV is orange with associated phospholipids (green wire frame). Subunit II has been omitted for clarity, with the exception of M263II and CuA. The figure was made using Rasmol. (The coordinates for RsCcO were generously provided by M. Svensson-Ek, L. Rodgers, J. Abramson, S. Tonroth, P. Brzezinski, and S. Iwata, personal communication.)

A reversible metal inhibitor that selectively binds to one of the proton paths would be an excellent tool for elucidating the function of the different proton transfer channels in oxidase. Although studies in 1988 (11) showed zinc inhibition of bovine heart oxidase reconstituted into proteoliposomes, the estimated inhibition constant of ~20-40 µM was quite high. However, in the reconstituted Escherichia coli bo₃ oxidase, 10 µM ZnSO₄ gave significant inhibition (12). Recent studies on purified RsCcO show two different inhibitory effects on steps in the reaction cycle (13) with high and low affinities for Zn²⁺. Similarly, Paracoccus denitrificans oxidase shows evidence of Zn²⁺ sites that affect proton uptake steps in the reaction cycle similar to those identified in the purified Rs oxidase (14).

The studies reported here demonstrate that there is a zinc binding site on the external side (P side) of the oxidase that is strongly inhibitory in the presence of a membrane potential (, negative inside), but whose inhibitory effect is much diminished when the electrical potential is removed. The zinc inhibition of RsCcO in lipid vesicles is readily reversed by a water-soluble chelator, demonstrating that the zinc site is external and solvent accessible. We suggest that the potential-sensitive inhibitory effect involves blockage of a proton exit path.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Preparation of Cytochrome c Oxidase-- Cytochrome c oxidase was purified from Rhodobacter sphaeroides as previously reported (15) using an Ni²⁺-NTA-agarose (Qiagen) affinity step followed by a further fast protein liquid chromatography (Amersham Biosciences, Inc., AKTA-519) purification procedure with tandem DEAE-5PW columns (Toso-Haas) (16). Bovine heart oxidase was purified by the Yoshikawa method (17). Site-directed mutants were made previously (M263L^II (18); D132A^I (19); D407A^I and R481K^I (20); H277A^I (21); R234H/C^II, D412A^I, and H411Q^II (22, 23); and H93C/N^I 2 and subunit III-less oxidase (24)) and purified in the same way as wild-type.

Reconstitution of cytochrome c oxidase vesicles (COVs) was performed using a dialysis method, giving a final concentration of 20 mg/ml asolectin (recrystallized from Associated Concentrates using the procedure of Sone et al. (25)) and 2 µM oxidase in 50 µM HEPES-KOH, pH 7.4, + 44 mM KCl + 38 mM sucrose. Alternatively, purified COVs were made using an oxidase construct with a His-tag on subunit II (htIICcO) that is able to bind to an Ni²⁺-NTA affinity chromatography column for isolation and concentration of COVs with correctly oriented oxidase, as described in Hiser et al. (16). Basically, detergent was removed by a Sephadex G-25 column equilibrated with 75 mM HEPES-KOH, pH 7.4, +14 mM KCl. The COVs were collected and diluted into 20 mM HEPES-KOH, pH 8.0, 27 mM KCl and 38 mM sucrose, and the pH was adjusted to 8.0 with KOH before addition to Ni²⁺-NTA-agarose. Vesicles containing correctly oriented oxidase bound to the column and were eluted with 20 mM HEPES-KOH, pH 7.4, 100 mM histidine, 38 mM sucrose. Histidine and buffer were removed by concentration of the COVs on a Centriplus 100 centrifugal spin filter (Amicon) and dialyzed against 2 × 1000 volumes for 6 h each of 50 µM HEPES-KOH, pH 7.4, 45 mM KCl, and 44 mM sucrose. The concentration of oxidase in the purified vesicles was calculated from the dithionite-reduced spectrum at 605 nm using an extinction coefficient of 33.35 mM¹ cm¹ after background baseline subtraction (26).

ZnSO₄ was from Columbus Chemical Industries, Inc., and all other metals used were of ACS grade. Na-EDTA (Invitrogen) at pH 7.4 was used as a chelator.

Steady-state Measurements-- Oxygen consumption was measured with a Gilson oxygraph at 25 °C using ascorbate, TMPD (Kodak), and horse heart cytochrome c (Sigma), which further was purified using carboxymethylcellulose ion-exchange chromatography (27).

Zinc inhibition constants (K_I) were calculated from the fits of the data using non-linear least-squares fitting in Microcal Origin to 1 or 2 inhibitory binding sites using the equations,

(Eq. 1)

for uncontrolled COV measurements, and,

(Eq. 2)

for controlled COV and purified enzyme.

The data from the steady-state kinetics of cytochrome c reaction with controlled COVs was fitted in Microcal Origin to Michaelis-Menten plots using the equation for two cytochrome c binding sites (28),

(Eq. 3)

Stopped-flow Measurements-- Measurements of cytochrome c oxidation were made in an Olis-rsm stopped-flow apparatus with COVs. Scans were collected (1000/s), and from these the rates of cytochrome c oxidation (550 nm) were fit either using Global analysis or by single-exponential fitting to the kinetic traces using Microcal Origin. The electron donor was pre-reduced horse heart cytochrome c made by sodium dithionite reduction followed by desalting and depolymerizing through a Sephadex G-75 (Amersham Biosciences, Inc.) gel filtration column into 0.5 mM HEPES-KOH, pH 7.4, + 45 mM KCl + 1 mM Na-EDTA.

Proton Pumping Measurements-- Proton pumping measurements were made as described previously (16). Scans were collected, and kinetic traces for cytochrome c oxidation at 550 nm or phenol red changes at 557 nm (isosbestic point of cytochrome c) were extracted after averaging at least three data sets and creating difference spectra (reduced minus oxidized). A small mixing artifact at 0-200 ms was subtracted from the phenol red changes (16). Rates of proton uptake or release were measured by fitting the kinetic traces for phenol red at 557 nm to one exponential with Microcal Origin. The cytochrome c oxidation rates were obtained from Global fitting analysis by the Olis software to a single-component exponential and were similar to those obtained from the kinetic traces at 550 nm, but the latter were more perturbed by the influence of the phenol red absorbance.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Membrane Potential Effects on Zinc Inhibition-- The inhibition with zinc was examined with cytochrome c oxidase reconstituted in lipid vesicles to physically separate the outside of the enzyme (P side) from the inside (N side). Measurements of cytochrome c oxidation with reconstituted cytochrome-c oxidase (COVs) were made: (a) in the controlled condition (no ionophores added; i.e. with pH and ), (b) with valinomycin (no ), and (c) in the uncontrolled state with uncoupler (no pH or ) (Fig. 2). It was immediately obvious that zinc was more effective in inhibiting the controlled state, in stopped-flow measurements with few or many turnovers. This was also true in steady-state measurements of oxygen consumption, where increasing ZnSO₄ concentrations revealed a K_I of ~5 µM, reaching ~75% inhibition in the controlled state at 300 µM ZnSO₄ (Fig. 3A).

View larger version (18K): [in this window] [in a new window]   Fig. 2.   Inhibition of reconstituted R. sphaeroides cytochrome c oxidase activity by ZnSO4. COVs (0.05 µM aa3 final concentration) ± 240 µM ZnSO4 were mixed in the Olis-rsm stopped-flow apparatus with 10 µM cytochrome c2+ and scans collected (1000 scans/s) in 50 mM HEPES-KOH + 24 mM KCl, pH 7.4. Kinetic traces at 550 nm are shown from an average of three data sets. Controlled state (A), with 2 µM valinomycin (B), or uncontrolled state with valinomycin + 10 µM CCCP (C).

View larger version (19K): [in this window] [in a new window]   Fig. 3.   Differential inhibition by nickel and zinc of purified enzyme compared with COVs. Steady-state measurements of ZnSO4 (solid lines) or NiSO4 (dashed lines) inhibition were made as described under "Experimental Procedures" with increasing metal added immediately prior to the addition of TMPD, ascorbate, enzyme, and 30 µM horse heart cytochrome c. A, controlled (no ionophores); B, uncontrolled states (+1 µM valinomycin + 6 µM FCCP), COVs (14 nM aa3), 5.7 mM ascorbate, and 0.28 mM TMPD were added. C, purified enzyme (2 nM aa3), 0.05% lauryl maltoside, 1.4 mM ascorbate, and 1.1 mM TMPD were added. The inset to B shows the full range of metal concentrations tested with the uncontrolled COVs. The data from at least two separate measurements is fitted to 1 or 2 binding sites as under "Experimental Procedures."

Zinc also inhibited after the addition of nigericin, which removes the pH leaving a , both in steady-state and stopped-flow measurements where activity was decreased ~60% with 100 µM zinc. Zinc was much less inhibitory when valinomycin or uncoupler was added (K_I ~ 1 mM) (Fig. 3B). In the purified, unreconstituted enzyme, the activity profile with increasing zinc (Fig. 3C) fit to two inhibition constants (K_I ~ 9 µM and ~ 400 µM). The effect of zinc on bovine heart-controlled COVs (data not shown) was less potent, giving inhibition constants of ~ 25 µM and in the millimolar range, similar to the earlier report (11).

Inhibition by Zn²⁺ was reversed by addition of Na-EDTA under all conditions and within the dead-time of mixing in the stopped-flow apparatus, indicating that zinc binds on the outside of the enzyme surface and is solvent-accessible.

In the controlled state, slight additional inhibition at very high concentrations of zinc (K_I ~ 1-2 mM) was discerned from biphasic fitting of the data of Fig. 3A, but this was likely due to vesicle aggregation, which can be monitored by light-scattering changes. Examination for these scattering effects showed substantial vesicle aggregation with 5 µM poly-L-lysine but no measurable effect up to 100 µM Zn²⁺, indicating that vesicle aggregation is not a factor in the zinc inhibition at the lower concentrations.

The question of whether Zn²⁺ inhibition is mediated by binding to the phospholipid membrane rather than the oxidase was addressed by altering the phospholipid composition of the vesicles. When the asolectin vesicles were supplemented with 25% phosphatidylcholine (neutral charge) or 25% phosphatidylserine (negative charge), the resulting COVs gave an unchanged inhibition constant with Zn²⁺ (K_I  5 µM), arguing against a lipid-mediated effect. Further evidence that Zn²⁺ inhibition involves direct interaction with the RsCcO rather than the membrane comes from comparison of normal and htIICcO-purified vesicles, where the lipid-to-CcO ratio is decreased 10-fold. The decrease in lipids did not affect the K_I for Zn²⁺. Specific zinc inhibition of the controlled state was observed whether the histidine-tag, used for purification of the overexpressed enzyme, was on subunit I and therefore on the inside of the COVs or on subunit II at the C-terminal end (htIICcO) on the outside of the COVs (16), implying that the His-tag is not also involved in zinc inhibition.

Metal Specificity-- A comparison of nickel and zinc in the controlled state (Fig. 3A) showed that, although nickel appears to bind with high affinity to the outside of oxidase its inhibition is limited, only inhibiting to 20% even at high metal concentrations. Unlike the Ni²⁺ binding on the outside of COVs, Ni²⁺ strongly inhibits with medium affinity (~40 µM) the purified enzyme (Fig. 3C) but lacks the low affinity binding site seen with Zn²⁺.

Other divalent cations were found to be either non-inhibitory beyond ionic strength effects (Mg²⁺, Mn²⁺, Ca²⁺, Ba²⁺, Li²⁺, Cs²⁺, Tb³⁺, Tm³⁺), or slightly inhibitory (Hg²⁺>Ni²⁺>Co²⁺), with only Cd²⁺ being as effective as Zn²⁺ (data not shown with the exception of Co²⁺, Ca²⁺, and Zn²⁺ in Table I). These results indicate a specific interaction and not, for example, an effect of ionic strength. This was verified by the measurement of steady-state turnover of the oxidase with high concentrations of Ca²⁺, which did not remove the inhibition by zinc (data not shown). Neither was it an effect on the interaction of cytochrome c with the oxidase, as shown by examining the kinetics of interaction of cytochrome c in the absence or presence of zinc (Fig. 4). This result clearly shows that addition of zinc in the micromolar range to controlled COVs reduces the V_max from 900 to 600 e/s/aa₃. However, from the negative inverse of both of the slopes in the Eadie-Scatchard plot and from the fitting of the Michaelis-Menten plot (as under "Experimental Procedures," not shown), there was no significant effect on the K_m values for cytochrome c (K_a = 0.04 µM and K_b = 0.8 µM) with or without zinc. These apparent Michaelis constants are similar to those published for free enzyme (28).

View larger version (13K): [in this window] [in a new window]   Fig. 4.   Zinc is not competitive with the substrate cytochrome c. Measurements of oxygen consumption activity were made with oxidase reconstituted into vesicles with no ionophores added (controlled state) in the Gilson oxygraph, as described under "Experimental Procedures," in the absence (, upper line) or presence (, lower line) of 29 µM ZnSO4. The slope of the linear fits are 1/Km, and the Vmax is the x intercept.

The purified oxidized enzyme or COVs after turnover with reduced cytochrome c were examined by EPR (electron paramagnetic resonance): the Cu_A and heme a spectra revealed no changes with zinc addition up to millimolar concentrations (data not shown).

Zinc Inhibits Various Low Activity CcO Mutants-- To investigate the nature of zinc inhibition, mutants with already low activity due to blockage at varying sites were examined for additive Zn²⁺ effects. A mutant of the D channel path (D132A^I), has very low activity, normally 5% of wild-type activity when comparing the rates of the purified enzymes, and does not pump protons (19, 29, 30). In vesicles under controlled conditions, D132A (15 e/s/aa₃) has 15% of the wild-type rate. D132A is inhibited by zinc to an even lower rate (4 e/s/aa₃) with a similar K_I as wild-type (Table I). The D132A^I mutant is rate-limited by blockage of proton uptake from the inside via the D-path, but there is substantial evidence that protons can be supplied from the outside at a slow rate that is stimulated by the membrane potential (19, 30). The rate of proton uptake from the outside for D132A^I COVs, observed as an alkalinization of the vesicle exterior in the controlled condition, is the same as the electron transfer rate (Table II) whether or not Zn²⁺ is added. In the wild-type COVs, this apparent proton leak rate, under controlled conditions, is much slower than the electron transfer rate, because proton uptake from the inside is supporting activity and pumping. But the addition of Zn²⁺ slows the electron transfer rate to that of the alkalinization rate, similar to the D132A^I mutant. A mutation of the Cu_A ligand, M263L^II, has very low activity due to a 100-mV increase in the Cu_A redox potential (31, 32). Zn²⁺ also inhibited this mutant to a similar extent (to 5 e/s/aa₃ with 250 µM Zn²⁺) as D132A^I (Table I). Also, Zn²⁺ inhibits a mutant, R481K, which already has high respiratory control.

View this table: [in this window] [in a new window]   Table II Measurements of cytochrome c oxidation and phenol red changes with controlled COVs in the stopped-flow Olis-rsm Under these conditions (no ionophores) there is no proton pumping but substrate protons are still consumed for the reduction of oxygen to H2O, and protons leak from the outside causing alkalinization. Rates are shown from the average of three data sets as either e/s/aa3 for cytochrome c oxidation (from Global analysis) or H+/s/aa3 for protons (calculated as under "Experimental Procedures") with errors from the standard deviation of the fits. Metals were added to a final concentration of 250 µM (CaCl2 or ZnSO4) with 0.1 µM oxidase with 5 µM cytochrome c2+. These measurements were made with Ni2+-NTA purified COVs as for Fig. 6.

Localization of the Zinc Binding Site-- In an effort to identify the zinc binding site, various mutants of outside sites were examined for their ability to be inhibited by zinc in the controlled state after reconstitution. The preferred ligands for zinc are cysteines, histidines, carboxyls, carbonyls, other charged groups, and H₂O (33). There are only a few conserved charged residues on the outside of oxidase, excluding the cytochrome c binding site on subunit II. The subunit I mutants H277L^I, H411Q^I, D412A^I, and D407A^I retained inhibition by zinc (Fig. 1). Mutations of a surface histidine (H93C/N^I) were made, above heme a close to the aspartate 51 in the bovine oxidase. This region has been identified as having an altered conformation in the reduced versus oxidized crystal structures of the bovine enzyme and was suggested to be the proton exit site (34). However, the reconstituted H93 mutants were still inhibited by zinc. Additionally, a mutant that lacked subunit III (Cox III()) (24) was inhibited by zinc, eliminating subunit III as a candidate for the external zinc binding site.

Respiratory Control and Zinc Inhibition-- In stopped-flow measurements of cytochrome c oxidation, zinc inhibition was decreased from ~85% at 240 µM Zn²⁺ to 5% after the addition of valinomycin, which equilibrates potassium across the membrane to dissipate the , to wild-type COVs (Table I). Upon further removal of both and pH with valinomycin and the uncoupler FCCP (uncontrolled condition), little inhibition was seen. This differential inhibition of the controlled state results in a very high respiratory control ratio (RCR = 44) for the wild-type COVs with Zn²⁺ added (Table I) whereas the normal RCR for wild-type is expected to be from 6 to 10. Certain oxidase mutants show unusually low RCRs, such as the subunit II mutant M263L^II, which normally has an RCR of 2-3. The RCR for M263L^II oxidase is increased with Zn²⁺ but only to 7, because the uncontrolled rate of electron transfer is extremely limited (31, 32). Other mutants, such as R481K^I, show unusually high RCR values (Table I) that are further increased by Zn²⁺. Reasons for the changes in RCR in various mutants are not entirely clear, but arginine 481 is part of a hydrogen-bonded network above the hemes and interacts with the heme propionates that have been implicated as participants in the exit path for protons (35). When the arginine is mutated to a lysine (R481K), it yields an enzyme that appears normal and is able to efficiently pump protons but is more strongly controlled by pH and ³ similar to a Zn²⁺-inhibited state. Zn²⁺ further inhibits the R481K^I mutant oxidase in the controlled condition (Table I), giving an exceedingly high RCR of 54. Interestingly, the ba₃ oxidase from Thermus thermophilus gives a low stoichiometry of proton pumping but has a naturally high RCR of ~40 (36).

The D channel oxidase mutant D132A^I, when reconstituted into vesicles, has unusual properties (19, 29, 30). Unlike the native enzyme, where activity is stimulated by removal of the /pH , this mutant oxidase is most active under the controlled condition. The D channel is severely compromised in D132A^I, but its remaining activity appears to be supported by protons from the outside (P side). When and pH are present, there is a driving force for inward proton movement from the outside, stimulating D132A^I activity (though it only achieves 15% of the normal activity). Because of this, the D132A^I mutant shows reverse respiratory control: i.e. RCR < 1. Interestingly, addition of Zn²⁺ to the D132A^I COVs led to restoration of normal respiratory control, due to strong inhibition of the controlled rate (Table I). The restoration of a normal RCR with D132A^I COVs was not observed with the addition of Ni²⁺ (K_I ~ 10 µM), which is much less inhibitory (maximum only 20% at mM Ni²⁺ concentration) and may not bind at the same site as Zn²⁺. It was shown in the bacterial reaction center that zinc and cadmium bound to the same sites, but that nickel and cobalt were bound differently and were less inhibitory (37).

Zinc Is Competitive with Protons-- A pH profile of the activity of COVs measured by stopped-flow in the controlled condition demonstrates a characteristic sigmoidal curve with an inflection point at approximately pH 6.8 (Fig. 5). However, after zinc addition this pK_a was no longer observed, with zinc inhibition increasing at higher pH and considerably less inhibition below pH 6.5. When these data are plotted as activity versus [H⁺] (Fig. 5, inset), Zn²⁺ inhibition shows a linear response with H⁺ concentration, suggesting direct competition of protons for the zinc site. This pH effect was observed with steady-state measurements (not shown). A 10-fold increase in zinc inhibition occurs in going from pH 6 to 7, as previously observed for zinc binding to the voltage-gated H⁺ channel from rat alveolar epithelial cells (38). Because of vesicle aggregation at pH less than 6, we were unable to appraise the effect of zinc at pH 5-6, which caused a 100-fold difference in the inhibition of zinc in the voltage-gated H⁺ channel. In earlier studies with bovine heart COVs (11), this pH dependence was not seen, perhaps reflecting a difference in the binding sites between the bacterial and mammalian enzymes.

View larger version (21K): [in this window] [in a new window]   Fig. 5.   Protons compete with zinc binding. The cytochrome c oxidation rate of controlled COVs was measured (average of three data sets)  with no metal added or  with 240 µM ZnSO4 at different pH outside. COVs (0.1 µM aa3 final concentration) in 50 µM HEPES-KOH, pH 7.4, + 45 mM KCl were rapidly mixed, in the stopped-flow apparatus, with cytochrome c2+ in 50 mM of the appropriate buffer (MES, pH 6-7/HEPES, pH 7-8.5/CHES, pH 8.5-9) with constant ionic strength maintained by the addition of KCl, to give a final mixture of 25 mM buffer and 45 mM K+ in the range pH 6-9. The inset shows the same data plotted with substrate proton [H+] × 10 7 M concentration to show the linearity of the response when zinc is added.

Zinc Reduces the Rate and Extent of Proton Pumping-- For stopped-flow proton pumping measurements, COVs were used that had been purified through an Ni²⁺-NTA column so that vesicles with correctly oriented oxidase were separated from wrongly oriented and empty vesicles to minimize any artifacts from excess lipids (16). Proton pumping is normally measured in the presence of valinomycin where zinc is much less inhibitory (Table I). COVs show normal proton pumping with CaCl₂ (H⁺/e ~ 1) (Fig. 6B), but the rate and extent of proton release was diminished by 40-50% with Zn²⁺ (Fig. 6D). When uncoupler (FCCP) is added, normal alkalinization (increased phenol red absorbance) was observed with Ca²⁺, but again at a somewhat reduced rate with Zn²⁺ (Fig. 6, A and C, respectively). The D132A^I COVs still showed no proton pumping with the addition of Zn²⁺ (data not shown) despite the positive respiratory control ratio.

View larger version (18K): [in this window] [in a new window]   Fig. 6.   Proton pumping of COVs is not blocked by Zn2+ but is less efficient. Ni2+-NTA-purified htIICcO COVs were used in the Olis-rsm stopped-flow with 0.1 µM oxidase and 5 µM cytochrome c2+ and 100 µM phenol red dye in 50 µM HEPES-KOH, pH 7.4, + 45 mM KCl (average of three data sets). A and B, phenol red changes with 250 µM CaCl2; C and D, phenol red changes with 250 µM ZnSO4 as above; A and C, alkalinization after uncoupling by the addition of 2 µM valinomycin +5 µM FCCP (pumped protons are not seen, and the alkalinization is due to substrate protons consumed to make H2O); B and D, acidification due to proton pumping in the presence of 2 µM valinomycin to alleviate build-up of the . The arrow shows the decrease in absorbance with acidic outside.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

External Zinc Binding in R. sphaeroides CcO: pH Dependence, Metal Specificity, Relation to Other Zinc Sites-- Zinc inhibits cytochrome c oxidase reconstituted into vesicles by binding to an outside, solvent-exposed surface site, most strongly when there is a membrane potential present. Zinc binding does not compete with cytochrome c binding and does not appear to interact with the histidine-tag, because there is an equal response to zinc with or without the histidine-tag on the outside. However, zinc inhibition decreases with increasing proton concentrations (low pH) in a manner that suggests that zinc is competed off by protons (Fig. 5). This is reminiscent of the results obtained with the bc₁ complex (39), voltage-gated H⁺ channel (38), and in the N-methyl-D-aspartate receptor NR2A subunit, where, in the latter case, the pH-dependent inhibition by zinc was attributed to a single histidine (40). These results would suggest that, at the external site in oxidase, zinc is specifically bound to at least one amino acid ligand that is protonatable with a pK_a of 6-7. The pH dependence, plus the inability of other divalent cations to inhibit the reconstituted oxidase, argues against a non-protein binding site involving membrane phospholipids. A similar specificity for Zn²⁺ and Cd²⁺ is observed for various voltage-gated H⁺ channels and for the membrane proteins that contain proton paths that are inhibited by zinc (4, 38-41).

The external zinc binding site in oxidase is probably not the calcium site, not only because of the lack of competition with other metals, but because no inhibitory effect on activity has been shown with metals at the calcium site. In the Paracoccus and probably in the RsCcO, the calcium site remains inaccessible to solvent (42), unlike the zinc site. Additionally, zinc inhibition was observed with the reconstituted E. coli bo₃ terminal quinol oxidase (12), which does not have a calcium site (42). The rapid release of Zn²⁺ inhibition by EDTA demonstrates that the observed effects are not due to internalization of the Zn²⁺ under the influence of the membrane potential.

Brzezinski's group (13) has found that zinc also inhibits the unreconstituted RsCcO with a K_I ~ 3 µM in the last and slowest step (F  O, >1 ms) of the cycle when CO is flashed off the fully reduced oxidase in the presence of O₂. This step has a high deuterium isotope effect, k_H/k_D of 8 (43), suggesting that it is rate-limited by a protonation event. Additionally, there is a medium affinity zinc binding site (K_I = 120 µM, P  F) that is presumed to be at a separate site. However, in the reconstituted oxidase there is no medium affinity site for zinc in the 10⁴ M range, although there is a similarly high affinity site for zinc. The high affinity site measured in the free enzyme is unlikely to be the same site as that observed with the reconstituted enzyme because of the release of inhibition by removal of the membrane potential and its much lower sensitivity to nickel (Fig. 2A).

Cu_A is situated in the soluble domain of subunit II close to the membrane/solvent interface and could be affected in its redox properties if zinc bound in the vicinity (12). However, in the purified, oxidized enzyme or in controlled COVs that were frozen after multiple turnovers with reduced cytochrome c and zinc, we see no perturbation of the EPR spectra of Cu_A and heme a. Additionally, the mutant M263L^II already has a very different Cu_A redox state (100 mV higher than normal) (31, 32), and yet it is further inhibited by zinc binding. If zinc is inhibiting by binding close to the Cu_A site, a further additive effect on the redox state of Cu_A would be required or competition with cytochrome c binding would be expected but is not observed.

The high affinity zinc site (0.5 µM) in the bacterial reaction center (3) is formed by 2 histidines, a glutamate, and one water ligand with tetrahedral coordination (4), similar in structure to a catalytic zinc site (44). Interestingly, nickel and cobalt are not bound to the same site as zinc in the bacterial reaction center, although they share one histidine ligand (4) and there is a difference in both the affinity and the mode of inhibition between zinc and nickel (45). This would suggest that the electronic configuration is important for coordination and metal selectivity. Nickel does inhibit the controlled state of COVs, but the inhibition is minor (10-20% maximum), suggesting some distinctive characteristics of its binding site compared with Zn²⁺. On the other hand, the inside binding has similar inhibition characteristics for both Ni²⁺ and Zn²⁺ (Fig. 2C) providing an important distinguishing feature between the exterior and interior interactions. The 5 µM affinity of zinc on the outside of RsCcO suggests similarity to the zinc site observed in the crystal structure of the chicken bc₁ complex, which has a K_I ~ 3 µM and appears to involve a histidine and an aspartate (46).

There are only a few histidines or charged amino acids on the outside surface that are conserved between bovine, RsCcO, and E. coli bo₃ oxidases and that are not critical for cytochrome c binding. Site-directed mutants of potential zinc binding residues exposed to the outside were examined for their ability to bind zinc in the reconstituted COVs in the controlled state: D407A^I, H277A^I, H411Q^I, D412A^I, and H93^I (within 5 Å of a glutamate 182^I) (Fig. 1), and the subunit III-less deletion strain (24). All of the mutant oxidases were inhibited by zinc. We are in the process of making mutations to test other potential ligands but recognize the handicap that there may be a different conformation of the region of interest in the presence of the membrane potential.

Mechanism of Zinc Inhibition-- It has been documented that the KA hippocampal neuronal channel only binds zinc at high affinity (~3 µM) when the membrane is hyperpolarized with a large negative (41). The suggestion is that a conformational change occurs in the mouth of the channel altering the accessibility of at least one of the key amino acids involved in zinc binding. Similarly, the reconstituted oxidase could undergo a conformational change in the presence of a that alters the proton exit path such that zinc binding is favored. Alternatively, zinc could remain bound in the absence of a but have little effect on activity in the latter state. It is difficult to directly measure zinc binding under these conditions, so we have not ruled out this possibility.

The normal assay for proton pumping by cytochrome c oxidase requires addition of valinomycin, thought to prevent the buildup of a membrane potential that would inhibit observable (net) proton extrusion. But valinomycin also reduces the inhibitory effect of zinc. Possibly, pumped protons coming out through the exit path effectively compete with zinc for binding. However, the D132A^I mutant, which does not appear to pump protons, is also less inhibited when there is no , arguing against this interpretation and in favor of the conformational effect of .

Conformational change in cytochrome c oxidase was previously proposed to involve at least two states differing in protonation of the enzyme (47, 48). This idea arose from experimental evidence suggesting that the oxidase is altered by the membrane potential or by the reduction of the redox-active metal centers, particularly Cu_A and heme a.

We propose that zinc inhibition in the controlled state is caused by prevention of the proton movement through the exit pathway which is essential for electron transfer. The requirement for charge neutralization of the electron transfer events involving protonatable groups in the oxidase has already been proposed (49, 50). Blockage of proton movement (from inside or outside) to and from key residues could inhibit the electron transfer rate or prevent the supply of protons to the active site or prevent a through-protein back-leak that could be a determinant of controlled respiration (51). Interesting recent results from Kannt et al. (14) raise this possibility as well. Inhibition by zinc on the inside of the Paracoccus COVs was reported to affect respiratory control. In fact, we were able to reproduce these findings with Rs COVs but also observed that addition of EDTA to the external medium removed the effect on RCR without altering the overall inhibitory effect. From this we conclude that the external binding of zinc is responsible for the effects on respiratory control.

In all of the systems that are inhibited by zinc, even where the zinc site is known, it has been difficult to define the exact mechanism of zinc inhibition. It has been proposed that zinc could physically block a channel (3, 39), alter the pK_a of groups important in H⁺ movement (45), prevent conformational mobility of a group (52), or control a gating event (53). At the fairly high concentrations required to get significant inhibition, metal binding to the oxidase is not likely to be physiologically important, but it may reveal the location of the proton exit path. We are continuing to investigate the mechanism of zinc inhibition.

Respiratory Control Is Modulated by Changes on the Outside-- Classically, the rate of electron transfer in mitochondria under controlled conditions (with a membrane potential) has been thought to be governed by the rate of passive leakage of protons across the lipid bilayer (54). The inhibition by zinc of this controlled rate, with metal specificity, micromolar affinity, and reversibility, argues strongly against this idea, as does the lack of effect of varying the nature and quality of the lipid membrane. The leakage of protons into or through the protein itself appears to be a more tenable hypothesis, also supported by evidence from native and mutant bacterial oxidases with widely varying RCR values. Most notable is ba₃ oxidase of T. thermophilus, which displays RCRs of 40-50 (36) when measured at 25 °C, well below the normal temperature for this thermophilic bacterium. It is reasonable to assume that the enzyme is in a rigid conformation at this low temperature, limiting motion that could be important for determining the level of controlled activity. Similarly, the interaction of zinc on the outside of the RsCcO oxidase could fix certain residues in an inflexible state.

The phenomenon of reverse respiratory control, exhibited by the D132A^I mutant of RsCcO, provides additional evidence of proton uptake from the outside that is stimulated by a membrane potential (29) and inhibited by zinc. This back-leak is able to support oxidase activity and thus feeds protons to some residue(s) that are normally supplied by proton uptake from the inside through the D pathway.

The control of proton back-leak through the oxidase has been suggested as a plausible mechanism of regulating the efficiency of cytochrome c oxidase (55). This mechanism invokes conformational control of the reversibility of the exit pathway, perhaps mediated in the mammalian enzyme by allosteric effectors such as ATP and phosphorylation (56). The data in this report raise the possibility that zinc could also be a physiological regulator of proton back-leak in CcO. A physiological role for zinc in inhibiting respiration of liver mitochondria has been postulated (2). Metallothionein has been shown to mediate zinc delivery to the mitochondrial intermembrane space, where it is capable of inhibiting respiration at micromolar concentrations.

The sum of available data indicates that zinc blocks proton movement in the exit path of CcO, in the presence of an electrical potential gradient and with a high degree of metal specificity. Identification of the zinc binding site on the exterior of CcO may give us an important clue regarding the location of the proton exit pathway and how it may contribute to oxidase regulation.

	ACKNOWLEDGEMENTS

We thank Dr. Honggao Yan for help with the equations for fitting the zinc and cytochrome c binding data. We also thank Sarah Cloutier for excellent technical assistance. Additionally, we acknowledge Dr. Peter Nicholls, Dr. Sasha Konstantinov, Dr. Peter Brzezinski, and Dr. Neil Law for helpful discussions on various aspects of zinc and oxidase.

	FOOTNOTES

^* This work was supported by National Institutes of Health Grant R37-GM26196 (to S. F. M.).The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

^§ Current address: St. Mary's Mercy Medical Center, Pathology Dept., Grand Rapids, MI 49503.

To whom correspondence may be addressed: Dept. of Biochemistry, Michigan State University, East Lansing, MI 48824-1319. Tel.: 517-353-3512; Fax: 517-353-9334; E-mail: millsden@pilot.msu.edu.

^¶ To whom correspondence may be addressed: Dept. of Biochemistry, Michigan State University, East Lansing, MI 48824-1319. Tel.: 517-355-0199; Fax: 517-353-9334; E-mail: fergus20@pilot.msu.edu.

Published, JBC Papers in Press, February 6, 2002, DOI 10.1074/jbc.M111922200

² C. Hiser and S. Ferguson-Miller, unpublished.

³ D. A. Mills, J. Qian, and S. Ferguson-Miller, unpublished data.

	ABBREVIATIONS

The abbreviations used are: CcO, cytochrome c oxidase; CCCP, carbonylcyanide-m-chlorophenylhydrazone; COVs, cytochrome oxidase-containing phospholipid vesicles; FCCP, carbonylcyanide-p-trifluoromethoxy-phenylhydrazone; htIICcO, cytochrome c oxidase with a His-tag on subunit II C-terminal; MES, 2-(N-morpholino)ethanesulfonic acid; Ni²⁺-NTA, nickel nitrilotriacetic acid; RCR, respiratory control ratio; Rs, R. sphaeroides; TMPD, N,N,N',N'-tetramethyl-p-phenylenediamine; RsCcO, Rhodobacter sphaeroides cytochrome c oxidase; CHES, 2-(cyclohexylamino)ethanesulfonic acid.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES