Proc. Natl. Acad. Sci. USAVol. 81, pp. 347-351, January 1984Biochemistry

Single catalytic site model for the oxidation of ferrocytochrome cby mitochondrial cytochrome c oxidase

(multiphasic kinetics/steady-state kinetics/electrostatics)

SAMUEL H. SPECK*, DANIEL DYE, AND E. MARGOLIASHtDepartment of Biochemistry, Molecular Biology, and Cell Biology, Northwestern University, Evanston, IL 60201

Contributed by Emanuel Margoliash, September 21, 1983

ABSTRACT A single catalytic site model is proposed toaccount for the multiphasic kinetics of oxidation of ferrocy-tochrome c by cytochrome c oxidase (ferrocytochrome c:oxy-gen oxidoreductase, EC 1.9.3.1). This model involves nonpro-ductive binding of substrate .to sites near the catalytic site oncytochrome c oxidase for cytochrome c, decreasing the bindingconstant for cytochrome c at the catalytic site. This substrateinhibition results in an increase.in the first-order rate constantfor the dissociation of the ferricytochrome c-cytochrome c oxi-dase complex, the rate-limiting step in the steady-state turn-over of electrons between cytochrome c and cytochrome c oxi-dase in the spectrophotometric assay, yielding increases in theinitial rate as well as the Michaelis constant-namely, multiplekinetic phases.

The kinetics of oxidation offerrocytochrome c by mitochon-drial cytochrome c oxidase (ferrocytochrome c:oxygen oxi-doreductase, EC 1.9.3.1; also called cytochrome oxidase)have been under investigation since 1930. Two major prob-lems appear to have remained at the origin of continued dis-agreements on the interpretation of the well-established ki-netic behavior. The first is that, even though the dependenceof the initial velocity on the concentration of ferrocyto-chrome c under a variety of conditions follows the Michae-lis-Menten relation, under certain conditions the time coursefor the reaction is strictly first order. Thus, Keilin (1) ob-served that the reaction with cytochrome oxidase exhibited ahyperbolic dependence on substrate concentration. Subse-quently, Stotz et al. (2) and Borei (3) proposed that the reac-tion involved the formation of an enzyme-substrate com-plex. In 1949, Slater (4) confirmed and extended the observa-tions of these authors, demonstrating that at any singleconcentration of enzyme [provided in the form of a Keilin-Hartree heart muscle particle preparation (5)], the kineticsindeed fit the Michaelis-Menten relation. However, the oxi-dation of ferrocytochrome c monitored spectrophotometri-cally (namely, in the absence of any added reducing agent)was generally observed to follow a first-order time course(6-13). Because the observed first-order rate constants de-crease with increasing cytochrome c concentration ratherthan remaining unchanged, as one would expect for a simplebimolecular collisional mechanism with no precursor-pairformation, Smith and Conrad (13) suggested that the hyper-bolic dependence of the initial velocity on cytochrome c con-centration was due to inhibition of the reaction by cyto-chrome c itself.That this was not the only possible explanation of these

kinetics followed from Minnaert's (14) elegant analysis, inwhich it was shown that a first-order time course could arisefrom equal binding of ferro- and ferricytochromes c to cyto-chrome oxidase, with ferrocytochrome c forming a produc-tive complex while ferricytochrome c acted as a competitive

inhibitor. This hypothesis was supported by Yonetani andRay (15), who demonstrated that (i) under conditions inwhich first-order kinetics are observed, the K1 for ferricy-tochrome c is the same as the apparent Michaelis constantfor ferrocytochrome c and (ii) at more alkaline pH values,where the kinetics deviate from a first-order time course, theK1 for ferricytochrome c is no longer equal to the apparentKm for the reaction.

Indeed, inspection of the Michaelis-Menten equation,which takes into account binding of product to the enzyme,leads to the expression:

dp _ Vmax 15= kobS,dt LKm + (s + P)when Km is a good approximation of Ks, the equilibrium sub-strate binding constant, and the binding of substrate andproduct are equivalent, namely K, = Kp (16). Thus, kobs is afirst-order rate constant that is distinguishable from an ordi-nary first-order rate constant because it is a function of s +p, the total concentration of substrate and product. This ex-plains the observed decrease in kbs with increasing cyto-chrome c concentration and its dependence on the total con-centration of cytochrome c (13, 17) rather than on that of theferric or ferrous protein. Clearly, the conclusions that thisbehavior is not that of a Michaelis-Menten system and that,for such a system, one would necessarily observe a deviationfrom a first-order time course (17) are unwarranted.Although this analysis is able to explain the observed

pseudo-first-order kinetics, it is unable to account for severalexperimental results that constitute the second major obsta-cle to a simple interpretation of the kinetic behavior of cyto-chrome oxidase. This comprises the continued increase inrate at high concentrations of cytochrome c (18) and the vari-ability of the Michaelis constant, which was found to be de-pendent on the range of substrate concentrations used (18,19). To explain these anomalies, Nicholls suggested (18) thesimultaneous reaction of a second molecule of cytochrome cwith cytochrome oxidase. That this indeed appeared to bethe case followed from the observation of clearly biphasickinetics as monitored by a polarographic assay usingN,N,N',N'-tetramethylphenylenediamine (TMPD) andascorbate as reductants (20). On this basis Ferguson-Milleret al. (20) proposed the existence of two catalytic sites oncytochrome oxidase for cytochrome c. The Michaelis con-stants for the high- and low-affinity phases of the reaction,determined under low ionic strength conditions (25 mM Trisacetate, pH 7.8), are 2 x 10-8 M and approximately 10-6 M,respectively (20). Because the Km for the high-affinity phase

Abbreviations: TMPD, N,N,N',N'tetramethylphenylenediamine;Sm, ferrocytochrome c bound to a noncatalytic site on cytochromeoxidase; Pm, ferricytochrome c bound to a noncatalytic site on cyto-chrome oxidase.*Present address: Dana-Farber Cancer Institute, Harvard MedicalSchool, 44 Binney Street, Boston, MA 02115.tTo whom reprint requests should be addressed.

347

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Proc. NatL. Acad. Sci. USA 81 (1984)

is small, it can only be observed at low cytochrome c con-centrations, which are within the experimental range for thepolarographic assay but difficult to attain spectrophotometri-cally. Thus, it was only recently, with the advent of sensitivespectrophotometers, that the presence of a high-affinityphase of the reaction could be confirmed spectrally (17, 21,22).That a second molecule of cytochrome c can react with the

high-affinity cytochrome c-cytochrome oxidase complex hasbeen further substantiated by analysis of the presteady-statekinetics (23). The time course of the rapid kinetic oxidationof ferrocytochrome c by purified cytochrome oxidase doesnot fit a single exponential when cytochrome c is in slightexcess. This may result from a change in the process that israte limiting or, alternatively, from the reaction of more thanone molecule of substrate with the enzyme. Several studieshave concluded that the rate constant for the dissociation ofthe high-affinity cytochrome c-cytochrome oxidase complexis small under low-ionic-strength conditions (<25 mM), withvalues ranging from 2 to 10 s-1 (24, 25). The reaction of asecond molecule of cytochrome c with the high-affinity cyto-chrome c-cytochrome oxidase complex is much faster undermost conditions than the rate of dissociation of this complex,consistent with the proposed existence of a second catalyticsite for cytochrome c on the enzyme (23).

In addition, equilibrium binding measurements with a par-tially phospholipid-depleted cytochrome oxidase prepara-tion demonstrated that two molecules of cytochrome c bindto cytochrome oxidase with different affinities (20). The ap-parent dissociation constants determined from a Scatchardanalysis have values similar to the apparent Km values deter-mined polarographically for the two phases of the reaction(20, 26-28). This close correlation between the Michaelisconstants for the reaction and the binding constants was ob-served also for the reaction of a number of chemically modi-fied cytochromes c displaying a wide range of affinities andreactivities (27).The domain on the surface of cytochrome c that is in-

volved in both the high- and low-affinity reactions has beenmapped by steady-state and presteady-state kinetic analyseswith a series of singly substituted lysine derivatives of horsecytochrome c (26, 27, 29). Although an early analysis ofsteady-state data had indicated that the high- and low-affinitydomains were different (26), a careful, rapid kinetic studydemonstrated that both the high- and low-affinity reactionsutilize the same interaction domain on the surface of cy-tochrome c (29). Thus, they only can be distinguished onthe basis of affinity and not mechanistically, at least at thelevel of cytochrome c. The nature of the interaction do-mains/catalytic site(s) on cytochrome oxidase has only beenpartially resolved. Several studies of covalent cytochromec-cytochrome oxidase complexes indicate that the "high-af-finity" binding site for cytochrome c on cytochrome oxidaseis the only catalytic site (30-34). Although some ambiguousresults have been obtained (see Discussion), it is generallyaccepted that crosslinking of a single molecule of cyto-chrome c to either subunit II or III of purified oxidase com-pletely blocks electron transfer from exogenously added fer-rocytochrome c. Whereas the binding of the first molecule ofcytochrome c appears to be to subunit II (30-34) (and possi-bly III) of cytochrome oxidase, the binding of a second mole-cule may be to phospholipid associated with the complex(32). Vik et al. (35) showed that extraction of the tightlybound cardiolipin from purified cytochrome oxidase canabolish the low-affinity phase of the reaction while maintain-ing some high-affinity activity. Furthermore, the steady-state oxidation of cytochrome c by mitochondrial mem-brane-bound cytochrome oxidase, monitored spectrophoto-metrically, has been shown to exhibit kinetics with morethan one low-affinity phase (20, 36). This observation has

been partially reproduced by reconstitution of purified cyto-chrome oxidase into phospholipid vesicles (36). In the lightof these results, the existence of a second catalytic site oncytochrome oxidase for cytochrome c is questionable andwould not account for the additional low-affinity kineticphases with membrane-bound cytochrome oxidase.We propose here a single-catalytic-site kinetic model that

can readily explain the observed multiphasic steady-stateoxidation of cytochrome c by cytochrome oxidase. A prelim-inary account of this work has appeared (37).

EXPERIMENTAL PROCEDURESCytochrome c. Horse cytochrome c-was prepared by the

procedure of Margoliash and Walasek (38) as modified byBrautigan et al. (39). Prior to enzymic assay, the cytochromec was fully reduced with minimal dithionite and chromato-graphed on Sephadex G-50 superfine (Pharmacia) in 100 mMTris-acetate (pH 7.5) to separate any polymeric material (26).When necessary to ensure that a fully reduced preparation ofdepolymerized cytochrome c was obtained, the cytochromec was reduced again with dithionite, passed over a smallSephadex G-25 column that had been pretreated with a buff-er containing 1 mM ascorbic acid, and thoroughly washed toremove the reducing agent. This material was appropriatelydiluted and stored at -80'C.

Purified Cytochrome c Oxidase. Beef cytochrome c oxi-dase was prepared by the method of Hartzell and Beinert(40). The enzyme was stored at a concentration of about 0.6mM in liquid nitrogen. Prior to assay, fresh enzyme was di-luted into 100 mM Tris-acetate (pH 7.25) containing 0.2% do-decyl D-maltopyranoside (Calbiochem), 0.2% Tween 20 (Sig-ma), 1 mM EDTA, and 50% (vol/vol) glycerol. The enzymecould be stored at a concentration of 1 ,uM at -20°C for sev-eral weeks without any detectable loss of activity. Storage ofmore dilute oxidase was avoided.

KINETIC MODELSpectrophotometric Assay Model. As discussed above, the

existence of high- and low-affinity kinetic phases for thesteady-state oxidation of ferrocytochrome c by cytochromeoxidase has been established both by the ascorbate/TMPDpolarographic assay and by a spectrophotometric assay. Inthe former assay, cytochrome c is kept reduced by ascorbateand TMPD, and the steady-state rate of consumption of mo-lecular oxygen as a function of cytochrome c concentrationis followed; in the latter assay, the rate of conversion of fer-ro- to ferricytochrome c is monitored directly. The spectro-photometric kinetics are more conducive to a steady-staterate analysis because of the simplicity of the system as com-pared with the polarographic kinetics, which involve the ad-ditional substrates ascorbate and TMPD.A single-catalytic-site scheme, based on a form of sub-

strate inhibition, is outlined in Fig. 1. As previously men-tioned, the rate-limiting step, under low-ionic-strength con-ditions, in the steady-state oxidation of ferrocytochrome cby cytochrome oxidase is the rate of dissociation of the ferri-cytochrome c-cytochrome oxidase complex (24, 25). Themost definitive evidence in this regard is that the presteady-state reaction of purified ferrocytochrome cl with the ferri-cytochrome c-cytochrome oxidase complex is first order,yielding a rate constant of 2.5 s-1 in 5 mM phosphate (pH7.0) (25), which is at least 2 orders of magnitude slower thanthe pseudo-first-order rate constant for the reaction of puri-fied ferrocytochrome cl with ferricytochrome c under thesame conditions (23). The slow rate constant for the disso-ciation of the ferricytochrome c-cytochrome oxidase com-

plex determined in this fashion correlates well with the slowsteady-state turnover of the high-affinity phase of the oxi-dase kinetics, as monitored spectrophotometrically under

348 Biochemistry: Speck et aL

Proc. NatL. Acad Sci USA 81 (1984) 349

k12S

SmEP Pm2E k12 etc.

kkSkkjgi kI1 k1 0

PmE 'k6 PmES PmEP k P,,E

k6S kg k7

8~~~ ~ ~ ~ ~|-7-

SrE SiES EP

k12S

StV2E=kk'7etc.

cytochrome oxidase as monitored spectrophotometrically. E, the

enzyme in its various oxidation states; S. the substrate ferrocyto-

chrome c; P, the product ferrcytochrome c; 5m and Pm, substrate

and product bound at noncatalytic sites.

similar conditions (21, 22). Thus, under conditions of low

ionic strength and low substrate concentrations (<0.1 ,uM)the rate of electron turnover in the steady-state reaction of

cytochrome c and cytochrome oxidase is slow (<10 elec-

trons s-1)* However, as the ferrocytochrome c concentra-

tion is increased, binding to negatively charged phospholipid

(and possibly protein) sites near the catalytic site may be-

come significant. The alteration of the electrostatic environ-

ment of the oxidase, resulting from binding of the strongly

positive charge of cytochrome c, leads to a decrease in the

site binding constant at the catalytic site with a concomitant

increase in the rate constant for the dissociation of the ferri-cytochrome c-cytochrome oxidase complex. This would be

manifested as an increase in the apparent Michaelis constant

and also an increase in the turnover rate, constituting a low-

er-affinity kinetic phase, unless the decrease in binding affin-

ity is so large as to lead to an overwhelming decrease in the

concentration of the enzyme-substrate complex.

The kinetic scheme in Fig. 1 shows two tiers correspond-

ing to two kinetic phases; as illustrated, it can be extended to

accommodate any number of kinetic phases. This mecha-

nism will provide multiple phases as long as the rate of disso-

ciation of product bound at the catalytic site is rate-limiting.

However, nonproductive binding of substrate or product

near the catalytic site will not only increase the rate of disso-

ciation of product from the catalytic site but also will de-

crease the bimolecular association rate constant for the bind-

ing of substrate. When the latter becomes rate-limiting, no

further increase in rate will occur on increasing the concen-

tration of substrate, and classical substrate inhibition should

be observed.

Also necessary for the scheme in Fig. 1 is the assumption

that product or substrate bound at the catalytic site is in rap-

id equilibrium with product or substrate bound near the cata-

lytic site, indicated by Sm or Pm in Fig. 1. This is required to

account for the fast bimolecular rate constant for the pre-

steady-state reaction of ferrocytochrome c with the ferrcy-

tochrome c-cytochrome oxidase complex [3.5 x 107 M-1

k-1 kS

sc at an ionic strength of 25 mM and pH 7.5 (28)], as com-

pared with the slow dissociation of the enzyme-product

complex discussed above (24, 25). This rapid equilibriummust be the result of two-dimensional diffusion of cyto-chrome c between the catalytic and membrane sites because,if the cytochrome c were to dissociate into free solution,

there should be no difference between the rate of reaction offerrocytochrome c with the enzyme-product complex andthe rate of dissociation of the complex (23).

Simplified Rate Equation and Computer Simulation. Itshould be noted that the kinetic scheme in Fig. 1 does nottake into consideration the various possible oxidation statesof cytochrome oxidase, each of which may differ in the rateat which it can accept electrons from cytochrome c. Howev-er, it appears unlikely that this is significant for the mecha-nism of the reaction of cytochrome c with cytochrome oxi-dase. Indeed, during steady-state, each molecule of the en-zyme turns over many times, so that the effective overallrate is a hybrid of the rates with the several forms of theoxidase. Furthermore, because the electron transfer ratewithin the enzyme-substrate complex is likely to be very fastas compared with the rate of either association or dissocia-tion of cytochrome c and the enzyme, for the purpose ofderiving an initial steady-state rate expression, one can sim-plify the kinetic scheme further, as follows:

E ES k3E+P

k4|k k8 k8 k-[5S

k6S -k7 , +PSmE

kSmES kS7 E + P

k-6

In this simplified scheme, the product terms have been re-moved as required for an initial steady-state analysis. Theinitial rate expression derived for the simplified kineticscheme is:

Ej{(k1k5k6k7 + k4k5k6k7)[S]3 + (kjk3k_5k6 + k3k4kL5k6+ k_1k4k6k7 + k3k4k6k7 + klk_4k5k7 + k1k6k7k8+ k4k5k7k-8)[S]2 + (klk3kL4k6 + klk3kL4k7+ klk3k-4k-5 + kik3k-6k-8+ klk3k7kL8 + k3k4k5k8)[S]}

VOk-lk4k-6 + k-lk4k7 + k3kL4kL6 + k3kL4k7+ k-lk_4k-5 + k3kL4k-5 + k-lk6kL8 + k3k-6kL8+ k3k7kL8 + k4L5k8 + (kLk_5k6 + k3kL5k6+ k_4k56 + kL4k5k7+ klk_4kL6 + klk_4k7+ klk_4kL5 + klk_6L8 + kjk7k-8 + k4kL5kL8+ kLk4kL6 + kLk4k7 + k3k4k-6 + k3k4k7+ k_1k4kL5 + k3k4kL5 + k1k_6k8 + k1k7k8+ k4kL5k8)[S] + (klk-5k-6 + k4k-5k6 + k-Lk4k6+ k3k4k6 + klk_4k5 + klk6k8 + klk5k-6 + k4k5k-8+ klk5k7 + k4k5k-6 + k4k5k7)[S]2 + (kik5k6+ k4k5k6)[S]3

This expression is the one derived for a biphasic reactionand should apply to the reaction of ferrocytochrome c withpurified cytochrome oxidase. From a consideration of thewealth of data that have been obtained from presteady-stateand steady-state kinetic analyses along with equilibriumbinding measurements, it was possible to simulate the ex-pected kinetics of reaction based on the rate expression giv-en above. We have assumed equivalent binding affinities forsubstrate and product to the enzyme. The values used for theindividual rate constants were taken to be as follows: k, = 1

Biochemistry: Speck et aL

Proc. NatL. Acad. Sci. USA 81 (1984)

1000

8080

x 60 tO00

T0

20

U)40 - \

0

0 10 20 30 40

TN(sWI)

FIG. 2. Computer-generated curve based on the initial rate equa-

tion derived for the simplified kinetic scheme. Computer simulationusing the rate constants given in the text; 0, experimental pointsobtained for the spectrophotometric assay with purified beef cyto-chrome oxidase.

X 109 M-1-s1 (29); kL1 = k3 = 10 s-1 [estimated from themaximal turnover number (TNma,,) of the high-affinity phase;also from ref. 24]; k5 = 3.5 x 107 M-1 s-1, determined fromthe presteady-state reaction of ferrocytochrome c with theferricytochrome c-cytochrome oxidase complex (28); kL5 =

k-6 = 45 s-1 (estimated from the maximal turnover numberof the low-affinity phase); k4 = 5 X 109 M-1's-1; kL4 = 15s-1; and k6 = 7 x 107 M-1lso1. The rate constants k4, k4,and k6 were optimized to fit the observed turnover values,assuming that k6 is similar in value to k5 and noting that theproduct of the equilibrium constants K4K6 must equal K1K5.The resulting computer-generated curve is shown in Fig. 2,along with experimental data obtained for the reaction of cy-tochrome c with purified cytochrome oxidase.

DISCUSSION

The question of the number of catalytic sites on cytochromeoxidase for ferrocytochrome c remains unresolved. Severallines of evidence are consistent with the presence of only asingle catalytic site (30, 34) and are difficult to reconcile witha two-site model (20, 26). Disulfide crosslinking of yeast iso-1 cytochrome c through cysteine-103 to subunit III of yeastcytochrome oxidase is able to block electron transfer fromexogenous ferrocytochrome c (30, 33). Disulfide crosslinkingof yeast iso-1 cytochrome c to beef cytochrome oxidase hasbeen reported to yield a 1:1 complex exhibiting no activitywith added ferrocytochrome c when monitored spectropho-tometrically, although in the polarographic ascorbate/TMPD assay, the rate of oxygen consumption increasedwhen free cytochrome c was added to the preparation (34).The significance of the latter observation is questionable be-cause, in the presence of the large excess of reducing agentused in the polarographic assay, the disulfide crosslink maynot be stable. Moreover, in our hands, attempts to generatethis material yielded mixtures containing various propor-tions of cytochrome oxidase carrying two molecules of disul-fide-linked iso-1 cytochrome c per cytochrome aa3 and un-modified cytochrome oxidase, which could be separated bygel filtration because of aggregation of the former. In addi-tion, crosslinking of horse cytochrome c to subunit II of beefoxidase through a photoaffinity-labeled lysine 13 completelyblocks electron transfer to the enzyme (31, 32). Further-more, this complex did not exhibit any activity in the ascor-bate/TMPD polarographic assay, in contrast to the results

reported for the yeast iso-1 cytochrome c disulfide-cross-linked beef cytochrome oxidase.

Binding of cytochrome c to phospholipid has been impli-cated in the low-affinity activity of cytochrome c with puri-fied cytochrome oxidase (32, 35). Extraction of the tightlybound cardiolipin depletes the low-affinity phase of the reac-tion, although the activity of the high-affinity phase also de-creases. The Vmax of preparations exhibiting no low-affinitykinetics was reported to be about 25% of that of the untreat-ed enzyme. There are two possible explanations for the de-crease in activity of the high-affinity phase: (i) the extractionof the tightly bound cardiolipin involves treatment with 5%Triton X-100 that may denature a fraction of the oxidase, or(ii) the maximal velocity of the high-affinity phase, which isdetermined in the ascorbate/TMPD polarographic assay bythe rate of TMPD reduction of the cytochrome c-cyto-chrome oxidase complex (28), may be sensitive to the phos-pholipid environment of the oxidase. The latter explanationis consistent with the disparity in the rate of turnover be-tween purified and Keilin-Hartree particle cytochrome oxi-dase observed in the ascorbate/TMPD assay (21). The mem-brane-bound cytochrome c-cytochrome oxidase complexappears to be more rapidly reduced by TMPD, suggestingthat the phospholipid environment may indeed be importantfor this reaction.

Additional evidence for phospholipid binding of cyto-chrome c being involved in the low-affinity reactions withcytochrome oxidase has been provided by reconstitution ex-periments (36). Incorporation of purified cytochrome oxi-dase into acidic phospholipid vesicles was able to recoverpartially the pronounced low-affinity kinetics of the innermitochondrial membrane-bound enzyme. The question re-mains, however, whether ferrocytochrome c bound to phos-pholipid site(s) is catalytically active. This seems unlikely,primarily because in almost all cases no activity is exhibitedby covalent 1:1 protein-to-protein complexes of cytochromec and cytochrome oxidase in the presence of added free cy-tochrome c. The alternative is that binding of cytochrome cto phospholipid or to protein sites near the catalytic site isnoncatalytic but does influence the binding of cytochrome cat the catalytic site, leading to the observed multiphasic ki-netics.Evidence for interaction between the high- and low-affini-

ty binding sites has been obtained from rapid kinetic studieson the reaction of horse cytochrome c with a 1:1 complex ofhuman or horse ferricytochrome c and purified beef cyto-chrome oxidase (28). It was shown that human cytochrome chas a higher affinity for beef cytochrome oxidase than hashorse cytochrome c. The reaction of a second molecule ofcytochrome c with the cytochrome c-cytochrome oxidasecomplex should be unaffected by the presence of the cyto-chrome c bound at the high-affinity site if the sites are nonin-teracting. However, when human cytochrome c was boundat the high-affinity site, the "on" constant for horse cyto-chrome c was decreased by a factor of 5 as compared to thereaction with the horse cytochrome c-beef cytochrome oxi-dase complex (28). This clearly demonstrates that the low-affinity reaction with purified cytochrome oxidase is influ-enced by the cytochrome c bound at the high-affinity site.A model for the reaction of cytochrome c with cytochrome

oxidase that is intermediate between single- and two-catalyt-ic-site schemes was proposed by Wilms et al. (25). This mod-el assumes that the postulated high- and low-affinity sites oncytochrome oxidase are not spatially distinct domains on theprotein complex but rather are represented by a single nega-tively charged catalytic site that can react with two mole-cules of cytochrome c. The high-affinity binding ofone mole-cule makes the binding of the second of lower affinity. Thisscheme makes it difficult to account for several low-affinitykinetic phases and, as discussed above, there is so far no

350 Biochemistry: Speck et aL

Proc. NatL Acad. SeL USA 81 (1984) 351

evidence to indicate that the first molecule of cytochrome chas anything but a rather specific binding site on the oxidase.

Errede and co-workers (17, 41) have described severaltwo-catalytic-site models for the cytochrome oxidase spec-trophotometric assay. As an additional alternative, they sug-gest that a second molecule of cytochrome c may react in aself-exchange reaction with the first molecule of cytochromec bound to the oxidase (41), rather than reacting at a secondcatalytic site. However, the mechanism used by this single-catalytic-site model does not appear to operate, as demon-strated by Veerman et al. (23). Indeed, with porphyrin cyto-chrome c, which is incapable of transferring an electron,bound in a high-affinity 1:1 complex with the enzyme, therate constant for the reaction of a molecule of ferrocyto-chrome c is indistinguishable from that determined for thesystem with native ferricytochrome c bound to the oxidase.This clearly precludes any inter-cytochrome c self-exchangereaction having a significant part in the electron-transferprocess.

In a recent analysis of the presteady-state reaction ofhorse cytochrome with anaerobic cyanide-inhibited purifiedbeef cytochrome oxidase, Antalis and Palmer (42) found thatthere is no distinction between the kinetics of reaction of thefirst and second molecules of ferrocytochrome c with the en-zyme under low-ionic-strength conditions (50 mM). This re-sult is in direct contradiction with those of van Gelder andco-workers (23, 25, 28, 43), who have observed that: (i) thereaction of ferrocytochrome c with the ferricytochromec-aerobic cytochrome oxidase complex under low-ionic-strength conditions (25 mM) yields a measurable bimolecularrate constant of 3.5 x 107 M-l s-1, whereas the reaction offerrocytochrome c with free aerobic cytochrome oxidase iscompleted within the mixing time of the stopped-flow appa-ratus (23, 28); and (ii) as one decreased the ionic strengthfrom a high value (250 mM), the bimolecular rate constantfor the reaction of ferrocytochrome c with cytochrome oxi-dase increased to the point at which it became unmeasurableat an ionic strength of about 60 mM (7 x 107 M-.s-1),whereas, at even lower ionic strengths, a much slower reac-tion was apparent (2.5 x 107 M-1-s-1 at an ionic strength of18 mM) that also increased with decreasing ionic strength.Both these experiments demonstrate that the reaction of thefirst molecule of cytochrome c with cytochrome oxidase isclearly different from the reaction of the second molecule.The discrepancy between these studies may stem from apossible fundamental difference between aerobic and anaer-obic cytochrome oxidase. Because the results of van Gelderand co-workers (23, 25, 28, 43) correlate well with thesteady-state parameters, the values for the various rate con-stants these authors obtained were used in the above simula-tion of the cytochrome c-cytochrome c oxidase reaction.

In summary, it seems clear from recent evidence thatbinding of cytochrome c to acidic phospholipid is involved inthe low-affinity kinetics of oxidation of cytochrome c by cy-tochrome oxidase. This binding or a similar type of bindingto protein does not need to be catalytic but rather can play aregulatory role in which it alters the binding affinity for thesubstrate at the catalytic site, making it possible for a single-catalytic-site model to account for the observed multiphasickinetics.

The authors congratulate Prof. David Shemin on the occasion ofhis 70th birthday. This work was supported by Grants GM29001 andGM19121 from the National Institutes of Health.

1. Keilin, D. (1930) Proc. R. Soc. London Ser. B 106, 418 -444.2. Stotz, E., Altschul, A. M. & Hogness, T. R. (1938) J. Biol.

Chem. 124, 745-754.3. Borei, H. (1945) Ark. Kem. Min. Geol. 20 (8), 1-215.4. Slater, E. C. (1949) Biochem. J. 44, 305-318.5. Keilin, D. & Hartree, E. F. (1947) Biochem. J. 41, 503-506.

6. Albaum, H. G., Tepperman, J. & Bodansky, 0. (1946) J. Biol.Chem. 163, 641-647.

7. Stannard, J. N. & Horecker, B. L. (1948) J. Biol. Chem. 172,599-608.

8. Cooperstein, S. J. & Lazarow, A. (1951) J. Biol. Chem. 189,665-670.

9. Brody, T. M., Wang, R. I. H. & Bain, J. A. (1952) J. Biol.Chem. 198, 821-826.

10. Hogeboom, G. H. & Schneider, W. C. (1952) J. Biol. Chem.194, 513-519.

11. Hess, H. H. & Pope, A. (1953) J. Biol. Chem. 204, 295-306.12. Frajola, W. J. & Vestling, C. S. (1954) J. Biol. Chem. 209,

677-686.13. Smith, L. & Conrad, H. (1956) Arch. Biochem. Biophys. 63,

403-413.14. Minnaert, K. (1961) Biochim. Biophys. Acta 50, 23-34.15. Yonetani, T. & Ray, G. S. (1965) J. Biol. Chem. 240, 3392-

3398.16. Cornish-Bowden, A. J. (1976) Principles of Enzyme Kinetics

(Butterworths, Boston), pp. 144-145.17. Errede, B., Haight, G. P. & Kamen, M. D. (1976) Proc. Natl.

Acad. Sci. USA 73, 113-117.18. Nicholls, P. (1965) in Oxidases and Related Redox Systems,

eds. King, T. E., Mason, H. S. & Morrison, M. (Wiley, NewYork), pp. 764-777.

19. Mochan, E. & Nicholls, P. (1972) Biochim. Biophys. Acta 267,309-319.

20. Ferguson-Miller, S., Brautigan, D. L. & Margoliash, E. (1976)J. Biol. Chem. 251, 1104-1115.

21. Smith, L., Davies, H. C. & Nava, M. E. (1979) Biochemistry18, 3140-3146.

22. Rosevear, P., van Aken, T., Baxter, J. & Ferguson-Miller, S.(1980) Biochemistry 19, 4108-4115.

23. Veerman, E. C. I., Wilms, J., Casteleijn, G. & van Gelder,B. F. (1980) Biochim. Biophys. Acta 590, 117-127.

24. Van Gelder, B. F., van Buuren, K. J. H., Wilms, J. & Ver-boom, C. N. (1975) in Electron Transfer Chains and OxidativePhosphorylation, eds. Quagliariello, E., Papa, S. & Palmieri,F. (North-Holland, Amsterdam), pp. 63-68.

25. Wilms, J., Veerman, E. C. I., Konig, B. W., Dekker, H. L. &van Gelder, B. F. (1981) Biochim. Biophys. Acta 635, 13-24.

26. Ferguson-Miller, S., Brautigan, D. L. & Margoliash, E. (1978)J. Biol. Chem. 253, 149-159.

27. Osheroff, N., Brautigan, D. L. & Margoliash, E. (1980) J. Biol.Chem. 255, 8245-8251.

28. Osheroff, N., Speck, S. H., Margoliash, E., Veerman, E.C. I., Wilms, J., Konig, B. W. & Muijsers, A. 0. (1983) J. Biol.Chem. 257, 5731-5738.

29. Veerman, E. C. I., Wilms, J., Dekker, H. L., Muijsers, A. O.,van Buuren, K. J. H., van Gelder, B. F., Osheroff, N., Speck,S. H. & Margoliash, E. (1983) J. Biol. Chem. 257, 5739-5745.

30. Birchmeler, W., Kohler, C. E. & Schatz, G. (1976) Proc. Natl.Acad. Sci. USA 73, 4334-4338.

31. Bisson, R., Azzi, A., Gutweniger, H., Colonna, R., Monte-cucco, C. & Zanotti, A. (1978) J. Biol. Chem. 253, 1874-1880.

32. Bisson, R., Jacobs, B. & Capaldi, R. A. (1980) Biochemistry19, 4173-4178.

33. Moreland, R. B. & Dockter, M. E. (1981) Biochem. Biophys.Res. Commun. 99, 339-346.

34. Fuller, S. D., Darley-Usmar, V. M. & Capaldi, R. A. (1981)Biochemistry 20, 7046-7053.

35. Vik, S. B., Georgevich, G. & Capaldi, R. A. (1981) Proc. Natl.Acad. Sci. USA 78, 1456-1460.

36. Speck, S. H., Neu, C. A., Swanson, M. S. & Margoliash, E.(1984) FEBS Lett., in press.

37. Speck, S. H., Neu, C., Swanson, M. S. & Margoliash, E.(1982) Fed. Proc. Fed. Am. Soc. Exp. Biol. 41, 3643 (abstr.).

38. Margoliash, E. & Walasek, 0. F. (1967) Methods Enzymol. 10,339-348.

39. Brautigan, D. L., Ferguson-Miller, S. & Margoliash, E. (1978)Methods Enzymol. 53, 128-164.

40. Hartzell, C. R. & Beinert, H. (1974) Biochim. Biophys. Acta368, 318-338.

41. Errede, B. & Kamen, M. (1978) Biochemistry 17, 1015-1027.42. Antalis, T. M. & Palmer, G. (1982) J. Biol. Chem. 257, 6194-

6209.43. Wilms, J., Dekker, H. L., Boelens, R. & van Gelder, B. F.

(1981) Biochim. Biophys. Acta 637, 168-176.

Biochemistry: Speck et aL