Effect of diethyl pyrocarbonate modification on spectral and steady-state kinetic properties of bovine heart cytochrome oxidase JOHND.

DORAN' AND

BRUCEC. HILL^

Biochem. Cell Biol. Downloaded from www.nrcresearchpress.com by University of Auckland on 02/08/15 For personal use only.

Division of Biochemistry, Department of Biological Sciences, University of Calgary, Calgary, Alta., Canada T2N 1N4 Received October 2, 1991 DORAN,J. D., and HILL,B. C. 1992. Effect of diethyl pyrocarbonate modification on spectral and steady-state kinetic properties of bovine heart cytochrome oxidase. Biochem. Cell Biol. 70: 565-572. The histidine-specific reagent diethyl pyrocarbonate has been used to chemically modify bovine heart cytochrome oxidase. Thirty-two of sixty-seven histidine residues of cytochrome oxidase are accessible to modification by diethyl pyrocarbonate. Effects on the Soret and a bands of the heme spectrum indicate disturbance in the environment of one or both of the heme groups. However, diethyl pyrocarbonate modification does not alter the 830-nm absorbance band, suggesting that the environment of Cu, is unchanged. Maximal modification of cytochrome oxidase by diethyl pyrocarbonate results in loss of 85-90% of the steay-state electron transfer activity, which can be reversed by hydroxylamine treatment. However, modification of the first 20 histidines does not alter either activity or the heme spectrum, but only when 32 residues have been modified are the activity and heme spectral changes complete. The steady-state kinetic profile of fully modified oxidase is monophasic; the phase corresponding to tight cytochrome c binding and low turnover is retained, whereas the high turnover phase is abolished. Proteoliposomes incorporated with modified oxidase have a 65% lower respiratory control ratio and 40% lower proton pumping stoichiometry than liposomes containing unmodified oxidase. These results are discussed in terms of a redox-linked proton pumping model for energy coupling via cytochrome oxidase. Key words: cytochrome oxidase, histidine modification, electron transfer, proton pumping, diethyl pyrocarbonate. DORAN,J. D., et HILL,B. C. 1992. Effect of diethyl pyrocarbonate modification on spectral and steady-state kinetic properties of bovine heart cytochrome oxidase. Biochem. Cell Biol. 70 : 565-572. Nous avons utilisC le ditthyl pyrocarbonate, rCactif spkcifique de l'histidine, pour modifier chirniquement la cytochrome oxydase du coeur de boeuf. Trente-deux des soixante-sept rCsidus histidine de la cytochrome oxydase sont accessibles a la modification par le diethy1 pyrocarbonate. Les effets sur les bandes Soret et a du spectre de l'htme montrent une perturbation dans l'environnement de l'un ou les deux groupes htmes. Cependant, la modification par le dikthyl pyrocarbonate n'altkre pas la bande d'absorption a 830 nm; cela suggkre que l'environnement du Cu, est inchangt. La modification maximale de la cytochrome oxydase par le ditthyl pyrocarbonate entraine une perte de 85 190% de l'activitt de transfert des Clectrons 11'Ctat d'tquilibre; cette perte peut &re renverste par traitement avec l'hydroxylamine. Cependant, la modification des 20 premitres histidines n'alttre ni l'activitt ni le spectre de l'hkme; il faut modifier 32 rtsidus pour obtenir des changements complets de I'activitC et du spectre de l'htme. Le profil cinttique 1l'ttat dlCquilibre de l'oxydase pleinement modifite est monophasique; la phase correspondant 1la liaison ttroite du cytochrome c et a un turnover faible est retenue alors que la phase du turnover ClevC est abolie. Les prottoliposomes renfermant une oxydase modifite ont un rapport du contr6le respiratoire diminue de 65% et une sto'ichiomCtrie de pompage protonique 40% plus faible que les liposomes contenant une oxydase non modifiCe. Nous discutons de ces rtsultats en fonction d'un modkle de pompage protonique lit a l'ttat rtdox pour le couplage de l'tnergie via la cytochrome oxydase. Mots clis : cytochrome oxydase, modification de l'histidine, transfert des Clectrons, pompage protonique, dikthyl pyrocarbonate. [Traduit par la rCdaction]

Introduction Cytochrome c oxidase is the terminal, membrane-bound complex of the mitochondria1respiratory chain. This enzyme catalyses the transfer of four electrons from ferrocytochrome c to dioxygen according to [l] [l]

4 ferrocytochrome c2+ + 0,

+ 4 H+

-

4 ferrocytochrome c3+ + H,O

The electron transfer reaction supplies the free energy by which cytochrome oxidase generates a proton-motive force used eventually in the synthesis of ATP. At least some of that energy transduction capability has been assigned to the functioning of the oxidase as a redox-linked, proton pump DEPC, d i e t h ~ l pyrocarbonate; TMPD, N,N,Nf,N'-tetramethyl-p-phenylenedimine; CCCP, carbonylcyanide m-chlorophenylhydrazone; DCCD, dicyclohexylcarbodiimide; RCR, respiratory control ratio. 'present address: National Research Council of Canada, Ottawa, Ont.. Canada KIA OR6. 2 ~ u t h oto r whom all correspondence should be addressed. Printed in Canada / lrnprirnt au Canada

(Wikstrom et al. 1981; Malmstrom 1990). According to this concept, electron transport is coupled to the translocation of protons from the matrix to the cytosol side of the mitochondria1 membrane, with some of the free energy of the electron transfer reaction conserved via the generation of a transmembrane ion gradient. Brzezinski and Malmstrom (1986) proposed that for the oxidase to function as a redox-linked proton pump, it is necessary for the enzyme to undergo a conformational transition. This transition allows the oxidase to oscillate between proton input and output states during turnover. The cytochrome oxidase reaction involves proton transfer of two types. Firstly, protons are involved in the reduction of O2to H20and this must involve proton transfer from the media to the site of O2reduction. secondly, protons are transported across the lipid bilayer in the energy ,.onservation process and this involves groups on the protein and at least one of the metal centres. One of the most likely amino acids to be involved in protein-mediated, proton transfer reactions is histidine. There is evidence from both chemical modification and site-directed mutagenesis experiments that

BIOCHEM. CELL BIOL. VOL. 70,

Biochem. Cell Biol. Downloaded from www.nrcresearchpress.com by University of Auckland on 02/08/15 For personal use only.

566

histidine residues are directly involved in proton transfer in the H +-linked, lac-permease reaction (Padan et al. 1979; Kaback 1990). Therefore, we have characterized the reaction of cytochrome oxidase with the chemical modifying reagent DEPC. DEPC reacts with a variety of amino acid side chains, but has considerable specificity for reaction with the imidazole side chain of histidine (Miles 1977). The chemical reaction produces carbethoxyhistidine, a derivative which has lost its capability for reversible protonation. Modification of cytochrome oxidase by DEPC results in an enzyme that has altered spectral, electron transfer, and protonpumping properties. These data are discussed in terms of structural and functional models for cytochrome oxidase, which are aimed at explaining the linkage between electron transfer and energy conservation.

Materials and methods Cytochrome oxidase was prepared from beef heart muscle by the method of Kuboyama et al. (1972), except that cholate was used throughout the entire purification. The purified oxidase was dissolved in phosphate buffer with 1 mg lauryl maltoside/mL. The enzyme was diluted to the desired concentration in 25 mM sodium phosphate buffer (pH 7.4) containing 1 mM EDTA and 1 mg lauryl maltoside/mL. DEPC solutions were prepared fresh in ethanol just before use. The concentration of DEPC was checked by addition of a small aliquot of the stock solution to a 10 mM histidine solution. A molar absorptivity coefficient of 3200 M - .cm - at 240 nm (Miles 1977) was used to calculate the concentration of DEPC. The stock DEPC solution was kept on ice before use and did not show any significant degradation over the time of the experiment. Modification of the oxidase was achieved by adding the desired concentration of DEPC to an enzyme sample previously equilibrated at 14°C. The reaction was monitored at 240 nm. A molar absorptivity coefficient of 3600 M cm (Miles 1977) was used to calculate the number of histidines modified. The oxidase was assayed polarographically with a Yellow Springs 0, electrode system. The reaction medium was thermostatted at 25°C and contained 25 mM sodium phosphate buffer (pH 7.4) with 1 mM EDTA and 1 mg lauryl maltoside/mL. Ascorbate and TMPD were used in the assay at concentrations of 5.25 mM and 350 pM, respectively. The oxidase was added to a final concentration of 10-20 nM and the reaction was initiated by the addition of cytochrome c (0.01-10 pM). Cytochrome oxidase concentration was determined by using a AAE of 27 rnM - .cm - for reductionoxidation at 605-630 nm. Cytochrome oxidase vesicles were prepared by a sonication procedure essentially as described by Proteau et al. (1983). Asolectin was suspended at 50 mg/mL in 25 mM potassium phosphate containing 1 mM EDTA (pH 7.2) and left to sit overnight at 4°C. Twenty nanomoles of cytochrome oxidase was added to 2 mL of lipid, and the mixture was sonicated on ice for 2.5 min at 30% duty cycle in an MSE sonicator. The sonicated solution was centrifuged at 13 000 rpm for 10 min in a minicentrifuge. The concentration of cytochrome oxidase in the proteoliposome preparations and the orientation of the oxidase in the lipid bilayer were determined spectrophotometrically. Cytochrome oxidase with its cytochrome c binding site facing the external medium was reduced by an excess of ascorbate and a catalytic amount of cytochrome c. Following reduction of the externally facing oxidase, TMPD was added and the entire oxidase sample was reduced. The value for the percent of externally facing oxidase in the proteoliposome preparations using either native or DEPC-modified cytochrome oxidase ranged from 73 to 79%. The respiratory control ratio was determined for the proteoliposome preparations by measuring the oxygen consumption rate in the absence and presence of uncouplers. Oxygen consump-

'

'

-'. -'

'

'

1992

tion was measured polarographically as described above. The assay medium was 25 mM potassium phosphate, containing 1 mM EDTA, at pH 7.4 and at 25°C. Ascorbate (5 mM) and TMPD (350 pM) were used as electron donors. The "coupled" rate of oxygen consumption was determined following the addition of ascorbate, TMPD, oxidase-containingvesicles (35 nM cytochrome oxidase), and cytochrome c (5 pM). This was followed by addition of 1 pL of 2 mg valinomycin/mL and 1 pL of 0.4 mg CCCP/mL to measure the "uncoupled" rate of oxygen reduction. The ratio of the uncoupled rate to the coupled rate is reported as the respiratory control ratio. Proton-pumping activity of proteoliposomes containing cytochrome oxidase was measured using a Corning model 250 pH meter. The signal from the pH meter was passed through a tuneable low-pass filter (Frequency Devices model 901 F) to eliminate high frequency noise. The resulting signal was recorded on a Sargent Welch strip chart recorder. The vesicle preparation described above was passed through a Sephadex G-25 column (1 x 40 an)that was equilibrated with 79 mM sucrose - 130 mM KC1 - 1 mM EDTA (pH 7.3) to reduce the external buffering capacity of the vesicle preparation. Reduced, cytochrome c was prepared by treatment with excess sodium ascorbate, followed by gel filtration on Sephadex G-25 (1 x 25 cm) to remove excess reductant. The pH of the reduced cytochrome c was adjusted to match that of the vesicle preparation. The measurements were performed on 1.5 mL of sucrose-KC1 solution containing 1 pL of 2 mg valinomycin/mL and enough of the vesicle preparation to give a final cytochrome oxidase concentration of 0.42 p M . The readings were initiated by the addition of ferrocytochrome c and the pH changes were quantitated by comparison to the changes induced by pulses of standard HCI. All spectrophotometric measurements were made with a Shimadzu UV-160 spectrophotometer. Spectra were stored for later manipulation and printout in the PC160 Personal Spectroscopy program from Shimadzu. Lauryl maltoside was purchased from Boehringer-Mannheim. Diethyl pyrocarbonate, asolectin (L-a-phospatidylcholine type IV-S), ascorbic acid, TMPD, valinomycin. CCCP, and horse heart cytochrome c (type VI) were from Sigma. All other chemicals were from BDH.

Results Figure 1 is a typical reaction profile of DEPC and cytochrome oxidase. The time course was measured at 240 nm, near the wavelength maximum for the formation of carbethoxyhistidine. The reaction came t o completion under these conditions in about 600 s. The inset shown in Fig. 1 is a difference spectrum between the DEPC-modified enzyme, at the end of the reaction time course, and the native, oxidized enzyme. A n absorption band centered at 242 nm is consistent with the formation of carbethoxyhistidine. No appreciable change was observed in the 260-300 nm region, as would be expected if there were modification of either tyrosine or tryptophan. However, we could not completely rule out the modification of tyrosine or tryptophan residues by DEPC. The intensity of the 240-nm absorption band was used to count the number of histidines modified and we found on average that 32 k 2, out of the total of 67 histidines found in the primary structure of the oxidase, were modified under the conditions used t o obtain Fig. 1. The inset in Fig. 1 also shows a small, but reproducible, change in the Soret region. This spectral effect is shown at higher sensitivity in Fig. 2a, which also extends the wavelength range t o the near infrared. There was a shift in the oxidized spectrum brought about by DEPC treatment, resulting in a peak at 408 nm and a trough at 430 nm, which was accompanied by a shift in the visible region producing

DORAN AND HILL

--

-

Biochem. Cell Biol. Downloaded from www.nrcresearchpress.com by University of Auckland on 02/08/15 For personal use only.

-

200

3 0

1

200

I

Wavelength (n m) I

400 600 Time (s )

500

-

..

I

800

1000

FIG. 1 . Effect of DEPC treatment on the UV-visible spectrum of cytochrome oxidase. The time course is for the reaction of 0.8 mM DEPC with 7.2 pM cytochrome oxidase, monitored at 240 nm in a 2-mm cuvette. The reaction buffer was 50 mM sodium phosphate (pH 7.4) with 1 mM EDTA and 1 mg lauryl maltoside/mL, thermostatted at 14°C. The inset is the difference spectrum between the DEPC-treated enzyme and the unmodified control measured under the conditions described above.

FIG. 2. The effect of DEPC modification on the spectral parameters of the heme and copper centers of cytochrome oxidase. (a) The difference spectrum between DEPC-treated oxidase and untreated control enzyme. The sample is the same one described in Fig. 1, but the spectrum is extended to the near infrared and the pathlength is increased to 1 cm. ( b ) The difference between the reducedoxidized difference spectra of control and DEPC-treated cytochrome oxidase. A sample of untreated cytochrome oxidase was recuded with 5 mM sodium ascorbate and 200 pM TMPD and the reduced-oxidized difference spectrum was recorded. A sample of DEPCtreated enzyme was reduced by an identical treatment and the reduced-oxidized difference spectrum was recorded. The difference between these two difference spectra is what is shown in b.

a trough at 595 nm and a peak at 640 nm. No difference was observed over the region from 700 to 900 nm. Further evidence of the effect of DEPC modification on the heme groups of the oxidase was observed in comparing reducedoxidized difference spectra of the native and modified enzyme. Initially, our concern was that modification by DEPC might produce a form of the oxidase that was not reducible and this could lead to inactivation of the enzyme. We compared the reduced-oxidized difference spectra, as generated by the reductants ascorbate and TMPD, of the

modified and native enzyme. This was done by taking the modified-native difference between the two reducedoxidized difference spectra. This result is shown in Fig. 2b. The heme spectrum was shifted to shorter wavelength in both the Soret and visible regions, producing a peak at 434 nm and a trough at 449 nm in the Soret region, and a peak at 596 nm and a trough at 608 nm in the visible region. The zero-crossing point in the difference spectrum was close to the peak position in the absolute reduced spectra at 444 nm in the Soret region and 604 nm in the visible region.

BIOCHEM. CELL BIOL. VOL. 70, 1992

Biochem. Cell Biol. Downloaded from www.nrcresearchpress.com by University of Auckland on 02/08/15 For personal use only.

568

"

0

500

1000

1500

2000

1500

2000

Time (s)

0 0

500

1000 Time (s)

FIG. 3. The extent of histidine modification and enzymatic inactivation of cytochrome oxidase at different concentrations of DEPC. The buffer conditions are the same as those outlined in Fig. 1. The oxidase concentration is 5.3 pM. ( a ) The modification time course at three different concentrations of DEPC: ( o ) 0.1 17, (0) 0.588, and (0) 1.17 mM. (b) The results of steadystate activity measurements of samples taken from the modification reactions shown in a. The oxidase activity was determined polarographically in 25 mM sodium phosphate buffer with 1 mM EDTA and 1 mg lauryl maltoside/mL, with 4 pM cytochrorne c as substrate. In addition, the peak and trough absorbances (Fig. 2b) were nearly equal in intensity. These data indicated that the total absorption intensity has not changed quantitatively, but qualitatively. Therefore, we conclude that the reducibility of the hemes has not changed drastically, but one or both of their immediate environments has been altered. We cannot say whether this spectral effect is due to direct modification of a heme group itself or an interacting residue of the protein. In addition, there was no detectable difference observed in the 700-900 nm region, following treatment by DEPC. We conclude that the hemes or their nearby environments are altered as a consequence of DEPC modification, but the CuA site is unchanged. Figure 3 compares time courses for modification and inactivation of cytochrome oxidase at three different concentrations of DEPC. In Fig. 3a the molar ratio of histidine residues modified per oxidase is plotted as a function of

time. These time courses show that both the rate and extent of modification are dependent upon the DEPC concentration. When the oxidase was exposed to 117 pM DEPC, 20 histidines per oxidase were modified at the completion of the reaction. This represented a near stoichiometric reaction between DEPC and this group of histidine residues. When the oxidase was exposed to 588 pM DEPC, the rate of modification of the oxidase increased as did the number of residues that were modified. At the completion of this time course, 25 histidine residues per oxidase molecule had been modified. Thus, an additional eight residues had been modified when the DEPC concentration was raised from 117 to 588 pM. These additional residues are much less reactive and probably represent a group of histidine residues which are relatively inaccessible to solvent. A further increase in the initial DEPC concentration used for modification further increased the rate and extent of modification. In the presence of 1.17 mM DEPC, 32 histidines per oxidase were modified. Increases of the initial DEPC concentration up to 2 mM did not cause additional histidine modification. At time points in the modification reactions of Fig. 3a, aliquots of enzyme were diluted (1:100) into a prepared assay mixture. Figure 3b is a plot of the steady-state activity at different points in the modification of cytochrome oxidase at different levels of DEPC. At any time point following exposure to 117 pM DEPC, the activity of the oxidase was unchanged relative to the unmodified enzyme. In the presence of 588 pM DEPC, the activity of the oxidase was unchanged after 100 s when about 18 histidines per oxidase had been modified, but then declined to a level of 35070 as the number of histidines modified reached 26. Exposure to 1.17 mM DEPC resulted in significant inactivation of the oxidase at the earliest time point measured. At the completion of modification of the oxidase by 1.17 mM DEPC, the enzyme retained 15% of the activity of the untreated oxidase. These results demonstrate that not all the residues modified by DEPC contributed to the loss in enzyme activity, but there were a subclass of residues that were modified, only late in the reaction or at high DEPC levels that were responsible for inactivation of cytochrome oxidase. Figure 4 compares the progress of cytochrome oxidase modification by DEPC in terms of the number of reacted histidines and the extent of the heme spectral perturbation. The number of histidines modified, shown in curve i, increased nearly in proportion to the concentration of DEPC up to about 0.3 mM. Above 0.6 mM DEPC, the number of histidines modified reached a limit of 32 histidines per oxidase molecule. In this experiment, a series of cytochrome oxidase samples was treated with a different concentration of DEPC for 40 min prior to quantitating the number of modified histidines. The same limiting value for the number of reactive histidines was obtained if a single oxidase sample was subjected to modification by sequential additions of DEPC. The amount of DEPC required to obtain the above level of modification was not in vast excess over the reactive histidine concentration. This was advantageous since keeping the DEPC concentration low reduces the possibility of reactions with other amino acids and because of the double substitution of histidines that can occur when DEPC is used in large excess (Miles 1977). In Fig. 3b, the extent of the heme perturbation (measured as in Fig. lc) is plotted as a function of DEPC concentration. At low levels of

DORAN AND HILL

w 3 0

Q, -

0

E w U)

0.05

30

C3

2

0.04

z5

w

20

Biochem. Cell Biol. Downloaded from www.nrcresearchpress.com by University of Auckland on 02/08/15 For personal use only.

.-

0.03

e

-0

E cn

03

m

E

.C -

-

0.02

10

4 a

u --

0.01

t

Ln .-

-C

2

0.0

0.4

0 .I3

1.2

[DEPC] (mM) FIG. 4. Extent of histidine modification and magnitude of the heme spectral change induced by DEPC treatment as a function of DEPC concentration. The modification reactions were done with 5.0 pM cytochrome oxidase in 50 mM sodium phosphate (pH 7.4) with 1 mM EDTA and 1 mg lauryl maltoside/mL. (i) The number of histidines modified following the complete reaction of cytochrome oxidase with different concentrations of DEPC. Each reaction was allowed to run for 40 min at 14°C. The absorbance difference at 242 nm was used to determine the number of histidines modified and this was divided by the concentration of cytochrome oxidase. (ii) Plot of the extent of the heme spectral modification at different levels of DEPC. Following modification by DEPC, the oxidase samples were reduced as described in the legend to Fig. 1 and reduced-oxidized difference spectra were constructed. A difference spectrum was then measured between the DEPC-treated enzyme and the untreated control. The magnitude of the peak to trough difference at 589 and 609 nm is plotted in curve ii as a function of DEPC concentration. The modification reaction was the same as described for curve i, except the absorbance differences are measured in a I-cm pathlength. DEPC (up to 100 pM) and with 20 modified histidines, no modification of the heme spectrum was detected. However, modification of more than 20 histidines lead to a progressive increase in the heme spectral change. A maximum in the magnitude of the heme perturbation was reached coincident with the maximal histidine modification. Thus it would appear that disturbance of the heme environment of the oxidase happened only after the bulk of the histidine residues were modified. However, it is important to emphasize that we could not exclude the possibility that the heme spectral perturbation was a consequence of modification of residues other than histidine. The generation of carbethoxyhistidine can be reversed by treatment with hydroxylamine and this reversal reaction is specific for modified histidine and tyrosine residues (Miles 1977). When fully modified oxidase was treated with hydroxylamine, the band at 242 nm was abolished, indicating that the histidine modification was reversed. When the enzyme was assayed at this point, it was completely inactive. However, when the hydroxylamine was removed by ultrafiltration, the oxidase recovered steady-state activity (see Table 1). Since we did not see any evidence in the UV spectrum for modification of tyrosine, reversal by hydroxylamine supports the interpretation that it is the chemical modification of histidine residues that leads to enzymatic inactivation. Figure 5 shows the effect that DEPC modification had on the steady-state kinetic profile of cytochrome oxidase activity as a function of cytochrome c concentration. The data are presented in the form of Eadie-Hofstee plots

TABLE 1. Reversal of DEPC-induced inactivation of cyrochrome oxidase by incubation with hydroxylamine Conditions Native oxidase 1.07 mM DEPCa 0.48 M NH,OH~ NH,OH removedc by ultrafiltration

Turnover, s-' (%) 35(100) 5.4(15) o(0) 26(74)

'The native oxidase was incubated for 30 min at 14OC with DEPC. treatment was for 30 min at 14°C prior to assay. %H~OH '%ltrafiltration was performed over a period of 20 h at 4°C and involved three changes of the modification solution with buffer.

(TN/[S] against TN, where TN is the turnover number of the enzyme in s - '). The unmodified oxidase exhibited the standard biphasic curve in which there was a phase of low Kmfor cytochrome c with low turnover and a phase of high Km associated with high turnover (Coooper 1990). When the oxidase was modified with DEPC, the kinetics became monophasic. Figure 5 documents this effect at two levels of DEPC: 0.25 and 0.75 mM. The curves that fit to the data are from a two-component Michaelis-Menten equation. The unmodified enzyme fits with a high turnover of 220 s - ' and a high Km of 3.0 pM, and a low turnover of 33 s - ' and a low Km of 0.62 pM. Satisfactory fits for the two modified oxidase samples were obtained if the two Kmsand the low turnover were fixed to those of the native enzyme

BIOCHEM. CELL BIOL. VOL. 70, 1992

2. Effect of DEPC treatment of cytochrome TABLE oxidase on respiratory control and H electron ratios +

in reconstituted proteoliposomes Oxidase sample

Biochem. Cell Biol. Downloaded from www.nrcresearchpress.com by University of Auckland on 02/08/15 For personal use only.

Native DEPC modifiedC

RCR'

H /electron

5.70 2.05

0.44 0.26

+

'The values reported here are averages of two determinations. The range of values was 0.40 for the native and 0.15 for the modified vesicles. 9hese values are the means of eight determinations. The standard deviation is 0.06 for the native vesicles and 0.02 for the modified vesicles. T h e oxidase was modified by incubation with 0.53 mM DEPC as described in Materials and methods.

FIG. 5. Steady-state kinetics of cytochrome oxidase modified with DEPC. Enzyme activity was determined polarographically as outlined in Materials and methods. ( a ) Untreated enzyme; (A) enzyme treated with 0.25 mM DEPC; (w) enzyme treated with 0.75 mM DEPC. The enzyme samples were modified at a concentration of 5 pM cytochrome oxidase for 30 min prior to assay. and the high turnover was allowed to vary. At 0.25 mM DEPC the high turnover was reduced to 100 s-', and at 0.75 mM DEPC the high turnover was about 10 s-'. Thus we concluded that DEPC modification of cytochrome oxidase converted the enzyme to a form which exhibited only a single steady-state kinetic phase of low turnover and high affinity for cytochrome c. Maximal modification did not completely inhibit the oxidase, but the enzyme turned over at a rate of about 10-15% of the high turnover maximum. DEPC modification did not alter the tight binding of cytochrome c as inferred from the low Km seen in the steady-state kinetics of the modified enzyme. We had also looked for substrate protection from DEPC modification of the oxidase by the binding of cytochrome c. Under conditions where cytochrome c formed a high affinity complex with cytochrome oxidase, the extent of DEPC modification of the oxidase was unchanged. Therefore, we concluded that the binding of cytochrome c did not make any of the histidine residues of the oxidase less accessible to DEPC. The effect of DEPC treatment on the oxidase's protonpumping capacity is shown in Table 2. The oxidase was first modified by treatment with DEPC as in other experiments and then it was incorporated into asolectin vesicles by a sonication method. Table 2 shows the respiratory control ratios obtained as the rate of cytochrome c oxidation in the presence of valinomycin and CCCP divided by the rate seen in the absence of uncouplers. The respiratory control ratio was diminished by 65% in the vesicles containing treated enzyme. Furthermore, the difference in this ratio was found in the difference between the uncoupled rates for the native

and modified oxidase samples. The controlled rate of cytochrome c oxidation was the same for the native and modified oxidase, implying that the intrinsic proton permeability of the two classes of vesicles was the same. Table 2 also shows the observed H+/electron stoichiometries for native and DEPC-treated oxidase vesicles. The protonpumping capability was decreased by 40% as a result of DEPC treatment. Therefore, DEPC-treated cytochrome oxidase was less effective, but was not completely impaired in its energy conservation function.

Discussion Malmstrom and his co-workers (Brzezinski and Malmstrom 1986; Thornstrom et al. 1988; Malmstrom 1990) have proposed that the non-Michaelis-Menten, biphasic steady-state kinetics exhibited by cytochrome oxidase are a consequence of the oxidase functioning as a proton pump, in which the enzyme oscillates between proton input and output states via a conformational transition. Models based on these proposals have been elaborated and simulate the experimental biphasic kinetic results observed for the oxidase (e.g., Michel and Bosshard 1989). Malmstrom's model is attractive, because it is an attempt to reconcile the nonconventional, steady-state kinetic behaviour of the oxidase with the mechanistic requirements of a proton pump. According to such a model, the oxidase undergoes a conformational transition from a state, E,, to another conformational state, E2, and this transition is strictly dependent upon protonation of the enzyme. The two conformations of the oxidase differ in their intrinsic affinity for cytochrome c and their electron transfer capability. Thus, the El conformation of the oxidase binds cytochrome c with high affinity and has a low maximal electron transfer rate, whereas conformation E2 has low affinity for cytochrome c and a high electron transfer rate. Since this model stipulates that the enzyme undergoes rapid conversion from El to E2 only when the enzyme is protonated, it follows that if the ability of the enzyme to bind protons is impaired, then the steady-state kinetics should be those of the E, conformation: high affinity for cytochrome c and low turnover. Modification of cytochrome oxidase by DEPC produces a change in the form of the steady-state kinetics of the oxidase, which can be interpreted to arise from blockage of the conformational transition, such that the oxidase is locked into the El state. The change in the pattern of the steady-state kinetic profile of DEPC-modified

Biochem. Cell Biol. Downloaded from www.nrcresearchpress.com by University of Auckland on 02/08/15 For personal use only.

DORAN AND HILL

cytochrome oxidase indicates that modification has produced a homogenous kinetic effect and does not arise simply because 85% of the enzyme is blocked completely, while 15% remains completely active. The obstruction of the oxidase's conformational change by DEPC is postulated to arise from the modification of particular residues, most likely histidines or tyrosines, that are directly involved in one of the proton binding steps in the enzyme's catalytic cycle. These protonation steps are obligatorily coupled to the conformation of the oxidase and, thereby, to its steadystate electron transfer activity. However, it should be borne in mind that Malmstrom's two-state conformational model is not the only one which explains the oxidase's nonhyperbolic kinetics (see Cooper 1990). For example, DEPC modification could be destroying a second site on the oxidase for cytochrome c which binds the substrate weakly, but gives rise to a high turnover rate. However, we do not observe any effect of cytochrome c binding on the rate or extent of modification of the oxidase by DEPC. The primary structure of bovine heart cytochrome oxidase shows that there are 67 histidine residues distributed amongst the 13 subunits of the oxidase (Buse et al. 1985). We have shown that 32 of these residues are accessible to DEPC modification. Presumably, the remaining 35 histidine residues which are nonreactive are buried in the threedimensional structure of the protein, such that DEPC cannot penetrate their immediate environment or they are involved in other interactions which preclude their reaction with DEPC. Papa et al. (1988) have measured the effects of a variety of amino acid modifying reagents on electron transfer and proton-pumping activities, as well as the RCR, of membrane-reconstituted cytochrome oxidase. They report a similar decrease, as seen here, in H+-pumping capacity for liposomal oxidase treated with 2 mM DEPC, but smaller declines in activity and RCR. However, their results are difficult to compare with those reported here, since there is no quantitation of the extent of protein modification. In addition, since their modification was performed on liposomeincorporated oxidase, there is a good chance of chemical interference by the lipid in the modification reaction and physical interference by the membrane in not allowing access of the modifying reagent to the entire oxidase molecule. The last possibility may provide a means for getting a clearer understanding of the relationship between the effects caused by DEPC and will be of interest to explore further. Our activity and spectroscopic measurements show that the first 20 histidines that are modified do not appear to have any consequence for the steady-state activity or the heme spectral properties of the oxidase. However, the reaction of DEPC with the more inaccessible and difficult to modify residues leads to enzymatic inactivation. Current models of the oxidase assign catalytic activity for electron transfer and proton translocation to the three largest subunits; I, 11, and I11 (Holm et al. 1987; Malmstrom 1990). Of the 67 histidine residues in bovine heart oxidase, 41 are found in these core subunits. Of the 41 histidine residues in subunits I, 11, and 111, 15 are conserved between animal, plant, and bacterial species (Saraste 1990). We suggest that these conserved histidine residues are good candidates for those residues that are modified by DEPC and lead to enzymatic inactivation. According to the structural model of Holm et al. (1987), 8 of the 15 conserved histidines are involved in metal ligation.

57 1

Site-directed mutagenesis studies on the homologous cytochrome oxidase from Rhodobacter sphaeroides strongly suggests that six of the eight conserved histidines in subunit I are involved in ligation to cytochrome a and the cytochrome a3 - CUB center (Shapleigh et al. 1992). Cytochrome a has two histidine residues ligated at the 5th and 6th axial positions, the CuA site contains two histidines as ligands, CUB may have as many as three histidine ligands, and cytochrome a3, has a histidine in one of the axial positions. We suggest that we have not modified the axial ligands to cytochrome a nor the proximal histidine ligand of cytochrome a3, because we would anticipate a larger spectral perturbation than the one we observe. In addition we have not modified the ligands to CuA, since the 830-nm band is unaltered by DEPC modification. Histidine has also been considered as a possible bridging ligand between CUBand cytochrome a3, (Palmer et al. 1976). This bridging ligand is present in the resting enzyme and renders this form of the oxidase slowly reactive to extraneous ligands, such as cyanide. When the enzyme is reduced, the cytochrome a3 - CUBsite becomes more reactive and this is assigned to the realignment of the bridging group to allow for access of an extraneous ligand. Thus, ligand exchange at the binuclear center is possible, might involve a histidine residue, and would be mechanistically compatible with a role in proton translocation. It is possible that we have modified this putative bridging histidine residue and this modification alters the electron transfer properties of the enzyme and the heme spectral properties. Since this histidine is not tightly bound to cytochrome a3, as evidenced by its ready exchange and its weak-field ligand properties, we anticipate much less of a spectral perturbation than if we had modified an inner sphere histidine of cytochrome a. An additional role for histidine might be in an H-bond to the formyl side chain of heme A. Callahan and Babcock (1983) proposed that an H-bonding interaction between the formyl side chain of heme A and an amino acid residue accounts for the resonance Raman properties of cytochrome a and its formyl side chain. Histidine is a possible candidate to fulfil this H-bonding function to cytochrome a. Such a residue would be relatively buried and, owing to its H-bonding interaction with heme A, relatively nonreactive towards a reagent such as DEPC. Conversely, DEPC modification of this histidine would render the residue incapable of H-bonding to heme A. Such a change in the environment of cytochrome a would raise the energy of the main electronic transition of the heme group and result in a shift of its spectrum to lower wavelength, as we observed. In addition, if this residue is involved in the redox-linked proton pumping of cytochrome oxidase, it could lead to reduced electron transfer and proton-pumping activity. It is possible that the effect of histidine on the electron transfer rate of cytochrome oxidase could be due to a nonspecific inactivation such as might occur if the structure of the protein is generally denatured. However, that we find specific and reversible changes in the oxidase's steady-state kinetics that are correlated with specific spectroscopic changes argues against the effects of DEPC being due to nonspecific protein structural alterations. We suggest that DEPC may represent a new type of inhibitor of cytochrome oxidase. DEPC inhibition of electron transfer arises from a blockage of a proton binding amino acid residue and, owing to the mechanistic linkage between proton pumping

BIOCHEM. CELL BIOL.

Biochem. Cell Biol. Downloaded from www.nrcresearchpress.com by University of Auckland on 02/08/15 For personal use only.

572

and electron transfer, electron transfer is inhibited. This indirect inhibition by DEPC is different from an oxidase inhibitor such as the respiratory poison cyanide, which derives its inhibitory properties by binding directly t o the Fe atom of cytochrome a3. Moreover, DEPC produces a modification that is different in form from a reagent such as DCCD, which reacts preferentially with carboxylic acid residues in hydrophobic environments. Glutamate 90 of subunit I11 is particularly reactive with DCCD (Prochaska et al. 1981) and this modified form of the oxidase is substantially inhibited (i.e., r 50%) in its proton-pumping ability without much change in its steady-state electron transfer activity (Casey et al. 1980). In contrast, cytochrome oxidase that is fully modified with DEPC is impaired in its proton-pumping ability and electron transfer activity. The view that a histidine residue adjacent t o the cytochrome a o r a3 heme is the key functional site of DEPC modification is consistent with results that demonstrate transmembrane potential sensitive electron transfer t o or from the heme groups (e.g., Morgan and Wikstrom 1991; Gregory and Ferguson-Miller 1989). In contrast, our results neither support nor can they exclude a direct role in proton pumping of one of the Cu, ligands as advocated by Chan (1988). The reactions catalysed by cytochrome oxidase involve the interaction of the enzyme with protons. There could be direct involvement of the protein matrix in the uptake of protons during the reduction of oxygen t o water. In addition, there is almost certainly a direct role for amino acids of the oxidase in transmembrane proton translocation. Histidine residues are ideal candidates for involvement in proton transport by the oxidase and, therefore, we have studied the effect of the histidine-modifying reagent DEPC on spectroscopic and steady-state kinetic properties of cytochrome oxidase. Much work still needs t o be done t o identify which, if any, of the metal centers are directly involved in proton translocation and how their involvement is coupled t o any amino acid residues. Our results suggest that histidine may be positioned with respect to the heme prosthetic groups of the oxidase, t o fulfil a role in coupling the processes of intramolecular electron transfer and transmembrane proton translocation.

Acknowledgement This work was supported by an operating grant t o B.C.H. from the Natural Sciences and Engineering Research Council of Canada. Buse, G., Meinecke, L., and Bruch, B. 1985. The protein formula of beef heart cytochrome c oxidase. J. Inorg. Biochem. 23: 149-153. Brzezinski, P., and Malmstrom, B.G. 1986. Electron-transportdriven proton pumps display nonhyperbolic kinetics: simulation of the steady-state kinetics of cytochrome c oxidase. Proc. Natl. Acad. Sci. U.S.A. 83: 4282-4286. Callahan, P.M., and Babcock, G.T. 1983. Origin of the cytochrome a absorption red shift: a pH-dependent interaction between its heme a formyl and protein in cytochrome oxidase. Biochemistry, 22: 452-461. Casey, R.P., Thelen, M., and Azzi, A. 1980. Dicyclohexylcarbodiimide binds specifically and covalently to cytochrome c

VOL. 70, 1992

oxidase while inhibiting its H +-translocatingactivity. J. Biol. Chem. 255: 3994-4000. Chan, S.I. 1988. The coppers in cytochrome c oxidase and the possible role of Cu, in proton pumping. Ann. N.Y. Acad. Sci. 550: 207-222. Cooper, C.E. 1990. The steady-state kinetics of cytochrome c oxidation by cytochrome oxidase. Biochim. Biophys. Acta, 1017: 187-203. Gregory, L., and Ferguson-Miller, S. 1989. Independent control of respiration in cytochrome c oxidase vesicles by pH and electrical gradients. Biochemistry, 28: 2655-2662. Holm, L., Saraste, M., and Wikstrom, M. 1987. Structural models of the redox centres in cytochrome oxidase. EMBO J. 6: 2819-2823. Kaback, H.R. 1990. Active transport: membrane vesicles, bioenergetics, molecules and mechanisms. In Bacterial energetics. Edited by T.A. Krulwich. Academic Press, San Diego. pp. 151-202. Kuboyama, M., Yong, F.C., and King, T.E. 1972. Studies on cytochrome oxidase. VIII. Preparation and some properties of cardiac cytochrome oxidase. J. Biol. Chem. 247: 6375-6383. Malmstrom, B.G. 1990. Cytochrome oxidase: some unresolved problems and controversal issues. Arch. Biochem. Biophys. 280: 233-241. Michel, B., and Bosshard, H.R. 1989. Oxidation of cytochrome c by cytochrome oxidase: spectroscopic binding studies and steadystate kinetics support a conformational transition mechanism. Biochemistry, 28: 244-252. Miles, E.W. 1977. Modification of histidyl residues in proteins by diethylpyrocarbonate. Methods Enzymol. 47: 431-442. Morgan, J.E., and Wikstrom, M. 1991. Steady-stateredox behaviour of cytochrome c, cytochrome a, and Cu, of cytochrome oxidase in intact rat liver mitochondria. Biochemistry, 30: 948-958. Padan, E., Patel, L., and Kaback, H.R. 1979. Effect of diethylpyrocarbonate on lactose/proton symport in Escherichia coli membrane vesicles. Proc. Natl. Acad. Sci. U.S.A. 76: 6221-6225. Palmer, G., Babcock, G.T., and Vickery, L.E. 1976. A model for cytochrome oxidase. Proc. Natl. Acad. Sci. U.S.A. 73: 2206-2210. Papa, S., Capitanio, N., and Steverding, D. 1988. Characteristics of the protonmotive activity of mammalian cytochrome c oxidase and their modification by amino acid reagents. Ann. N.Y. Acad. Sci. 550: 238-253. Prochaska, L.J., Bisson, R., Capaldi, R.A., et al. 1981. Inhibition of cytochrome c oxidase function by dicyclohexylcarbodiimide. Biochim. Biophys. Acta, 637: 360-373. Proteau, G., Wrigglesworth, J.M., and Nicholls, P. 1983. Protonmotive functions of cytochrome c oxidase in reconstituted vesicles. Biochem. J. 210: 199-205. Saraste, M. 1990. Structural features of cytochrome oxidase. Q. Rev. Biophys. 23: 331-366. Shapleigh, J.P., Hosler, J.P., Tecklenburg, M.M.J., et al. 1992. Definition of the catalytic site of cytochrome c oxidase: the specific ligands of heme a and the heme a3 - CuB center. Proc. Natl. Acad. Sci. U.S.A. 89: 4786-4790. Thornstrom, P.-E., Brzezinski, P., Fredriksson, P.-0.. and Malmstrom, B.G. 1988. Cytochrome c oxidase as an electrontransport-driven proton pump: pH dependence of the reduction levels of the redox centres during turnover. Biochemistry, 27: 5441-5447. Wikstrom, M., Krab, K., and Saraste, M. 1981. Cytochrome oxidase: a synthesis. Academic Press, New York.