THE JOURNAL OF Vol. 268, No. of 5, pp. for Printed in U. S ... · THE JOURNAL OF BIOLOGICAL CHEMISTRY 0 1993 by The American Society for Biochemistry and Molecular Biology, Inc. Vol.

THE JOURNAL OF BIOLOGICAL CHEMISTRY 0 1993 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 268, No. 34, Issue of December 5, pp. 25320-25328, 1993 Printed in U. S. A.

The Adenine Nucleotide Translocase Modulates Oligomycin-induced Quenching of Pyranine Fluorescence in Submitochondrial Particles*

(Received for publication, April 12, 1993, and in revised form, August 2, 1993)

Mathias Ziegler and Harvey S. Penefsky From the Department of Biochemistry and Molecular Biology, State University of New York, Health Science Center, Syracuse, New York 13210

Incorporation of the fluorescent, nonpermeant pH indicator pyranine into submitochondrial particles (pyranine-SMP) permitted monitoring of intravesic- ular pH changes brought about by proton translocation due to oxidation of respiratory chain substrates or to hydrolysis of ATP.

Addition of oligomycin to beef heart pyranine-SMP was followed by a pH-independent quenching of pyr- anine fluorescence. Quenching was influenced by the presence of adenine nucleotides both inside and outside the submitochondrial particles. The nature of the nu- cleotides required for quenching resembled the speci- ficity of the adenine nucleotide translocase rather than F1-ATPase. Removal of F1 from pyranine-SMP by treatment of the particles with urea did not alter oli- gomycin-induced quenching. Atractyloside, a specific inhibitor of the adenine nucleotide translocase, pre- vented oligomycin-induced quenching when the inhib- itor was coincorporated into submitochondrial parti- cles with pyranine. Bongkrekic acid prevented or re- versed the oligomycin-dependent quenching when added to pyranine-SMP either before or after oligo- mycin, respectively, but only when ATP was present within the particles. A mutant of Saccharomyces cere- uisiae, lacking translocase genes, exhibited oligomy- cin-dependent fluorescence quenching which was not inhibited by bongkrekic acid.

The results support the interpretation that oligomy- cin promotes sequestration of the fluorescent probe in a region of the submitochondrial particle, probably the FoFl complex, that leads to a quenching of fluores- cence. The observed quenching can be modulated in a way that suggests an interaction between the translo- case and Fo.

The mitochondrial ATPase (F1)’ and the adenine nucleotide

* This work was supported in part by Research Grant GM21737 from the National Institutes of Health, United States Public Health Service. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

The abbreviations used are: F,, soluble mitochondrial ATPase; Fo, membrane sector of the ATP synthase complex; DCCD, dicyclo- hexylcarbodiimide; SMP, submitochondrial particles; pyranine-SMP, submitochondrial particles prepared by sonicating mitochondria in the presence of 0.5 mM pyranine; BBBP, butane bis(2,Z”bypyridi- nium); FCCP, carbonyl cyanide p-trifluoromethoxyphenylhydrazone; caged ATP, adenosine-5’triphosphate, 3-0-(1-(2-nitrophenyl)- ethylester); AMP-PNP, 5’-adenylylimidodiphosphate; BCECF, 2’7’- bis(2-~arboxyethy1)-5(and6)-carboxyfluorescein; DAN-ATP, 3’-0-(5- dimethylaminonaphtoyl-1-)-ATP;DTNB, 5,5’-dithiobis-(2-nitroben- zoic acid); Tricine, N-tris(hydroxymethy1)methylglycine.

translocase are among the most abundant of mitochondrial proteins. Each comprises about 10% of the total protein of the inner mitochondrial membrane (1,2). F1 is a highly active ATPase that catalyzes the terminal phosphorylation reaction of oxidative phosphorylation (3). It is an extrinsic, easily solubilized protein consisting of five distinct types of subunits. The ATPase is bound to the membrane via a complex of proteins and phospholipid (Fo). Considerable experimental evidence supports the conclusion that Fo contains the proton channel (4-8), although the pathway of proton translocation through the whole ATPase complex remains to be elucidated. Inhibition of proton translocation, ATP synthesis, and ATP hydrolysis by oligomycin, venturicidin, and low concentra- tions of dicyclohexylcarbodiimide (DCCD) occurs via inter- action of the inhibitors with the Fo component rather than F1

The adenine nucleotide translocase is an apparent dimer, each subunit of which has a molecular mass of about 32 kDa (2, 12, 13). The carrier catalyzes a remarkably specific 1:l exchange of ATP for ADP across the inner mitochondrial membrane (13, 14).

Although FoFl is functionally linked to the translocase in that the carrier facilitates the removal of product ATP to the cytosol and the supply of substrate ADP to the mitochondrial matrix, there has thus far been no direct indication of an interaction between the ATP synthase and the carrier. How- ever, the possibility of an interaction between the ATP syn- thase and the translocase emerged from early kinetic analyses of translocase action and from experiments with inhibitors (reviewed in Ref. 13).

This paper shows that changes in intravesicular proton concentration can be monitored with the fluorescent pH indicator pyranine (15) when the probe is incorporated into submitochondrial particles from beef heart. Addition of oli- gomycin to pyranine-SMP is followed by a specific quenching of pyranine fluorescence that is independent of pH but mod- ified by the presence of adenine nucleotides in a complex manner. Further experiments, including the analysis of an adenine nucleotide translocase-deficient yeast mutant, dem- onstrate that the translocase, and not F1, represents the target for the nucleotide effects. However, the translocase does not interact directly with either oligomycin or pyranine in the observed quench effect.

The evidence supports the conclusion that the interaction of oligomycin with FoFl of submitochondrial particles and the ensuing effects on pyranine fluorescence may be modulated by the mitochondrial adenine nucleotide translocase via con- formational interaction between the two proteins.

(9-11).

EXPERIMENTAL PROCEDURES

Pyranine (8-hydroxypyrene-l,3,6-trisulfonic acid), BBBP (butane bis(2, 2’-bipyridinium), dextran-linked BCECF, 2‘,7’-bis(2-carboxy-

25320

Interactions between Adenine Nucleotide Translocase and F a 1 25321

ethyl)-5(and-6)-carboxyfluorescein, oxonol V, fluorescein maleimide, and caged ATP (adenosine-5’triphosphate, 3-0-(1-(2-nitro- pheny1)ethyl ester) were obtained from Molecular Probes, Eugene, OR. Bongkrekic acid was a gift from Dr. R. L. Cross (original source Dr. W. Berends). DAN-ATP was generously donated by Prof. M. Klingenberg. Atractyloside was purchased from Sigma.

Saccharomyces cerevisiae strains used in this study were kindly provided by Drs. D. R. Nelson and M. G . Douglas, University of North Carolina. Wild type strain W303 (MATa ade2-1 leu2-3, 112 his 3-22,15 trpl-1 ura3-1 can 1-100) and DNYl (MATa aac2::HISB aac1::LEUP his 3-11, 15 t r p - 1 ura3-1 canl-100 ade2-1 leu2-3,112) were grown in 1% yeast extract, 2% bactopeptone, and 2% galactose. Yeast cells were harvested and washed twice with 0.4 M sorbitol, 0.1 mM EDTA, and 50 mM Tris acetate, pH 8.2. The cells were disrupted by grinding with glass beads in a DYNO-MIL type KDL cell disruptor (GlenMills, Inc. Maywood, NJ) equipped with a 150-ml batch grinding chamber. The cells were suspended in the same buffer used for washing at a concentration of 0.5 g/ml (wet weight). Grinding was carried out at 12 “C for 3 min. The entire contents of the grinding chamber were removed and centrifuged at 3,000 X g for 15 min at 5 “C. The supernatant was collected and centrifuged again under the same conditions. Yeast mitochondria were collected by centrifugation of the supernatant at 12,000 X g for 20 min and resuspended in 250 mM sucrose, 10 mM Tris-HC1, pH 7.5.

Pyranine was incorporated into submitochondrial particles by sonic oscillation of beef heart or yeast mitochondria in the presence of the probe. A mitochondrial suspension, 4 ml of 15-20 mg/ml, containing 60 mM &SO,, 10 mM MgSO,, 100 mM sucrose, 1 mM Tricine, pH 8.0, and 0.5 mM pyranine was sonicated on ice for 3 min in 20-s intervals at maximum output (microtip limit) with a Heat Systems Ultrasonics sonic oscillator, model 350W. The duty cycle was 30%. ATP (1 mM) and/or atractyloside (100 pM) were added to the sonication medium when it was desired to incorporate ATP into the particle or to inhibit the translocase, respectively. SMP from yeast (with or without pyranine) were always prepared in the presence of 1 mM ATP. When it was desired to incorporate caged ATP into beef heart submitochondrial particles, the sonication medium con- sisted of 100 mM MOPS-KOH, pH 7.5,60 mM KzSO,, 10 mM MgSO,, 0.5 mM pyranine, and 1 mM caged ATP. After sonication, 16 ml of the buffer used for sonication (minus pyranine and caged ATP or ATP) were added, and the suspension was centrifuged for 15 min at 105,000 X g. The supernatant containing submitochondrial particles was centrifuged for 45 min at 144,000 X g. The submitochondrial particles were resuspended in sonication buffer at a concentration of about 30 mg/ml and could be stored at -70 “C for at least 6 months. In order to remove some pyranine that might have been bound to the outside of the particles, an aliquot of frozen particles was thawed and 0.18 ml of the suspension, containing about 5.4 mg of protein was applied to a column of Sephadex G-75 medium formed in a 1-ml tuberculin syringe. The column had been equilibrated with a buffer containing 100 mM MOPS-KOH, pH 7.5, 60 mM KzS04 and 10 mM MgSO,, and the particles were eluted from the column with the same buffer.

The ATPase activity of beef heart pyranine-SMP, 0.2-0.4 units/ mg, was independent of the presence of ATP and/or atractyloside in the sonication medium.

Beef heart submitochondrial particles without incorporated pyr- anine were prepared by sonicating mitochondria in 200 mM sucrose, 1 mM ATP, 10 mM MgCl,, and 10 mM Tris-HC1, pH 7.5 (16). Yeast SMP with no incorporated pyranine were obtained under the same conditions as described above for pyranine-SMP, except the fluores- cent probe was omitted from the sonication medium.

In some experiments, the ATPase activity of beef heart SMP (with or without pyranine) was activated by chromatography of the particles on a column of Sephadex G-50 medium in a buffer containing 2 mM EDTA, 75 mM sucrose, 30 mM Tris sulfate, pH 8.0, and 250 mM KC1 (1). The specific activity of the resultingparticles was about 2.5 units/ mg. When desired, activated SMP were incubated with urea in order to remove Fl from the membrane (1). The specific activity of the resulting particles was about 0.05 units/mg. SDS-polyacrylamide gel electrophoresis of urea-treated pyranine-SMP and subsequent den- sitometric analysis indicated that at least 85% of the (Y and (3 subunits of F1 were removed by the urea treatment.

Fluorescence measurements were carried out at room temperature in a model 8000 spectrofluorometer from SLM Instruments, Inc., Urbana, IL. Pyranine excitation spectra were obtained at an emission

wavelength of 511 nm. Assessment of pH- or oligomycin-dependent pyranine fluorescence changes were carried out at an excitation wavelength of 468 nm (8 nm bandpass). The emitted fluorescent light was monitored by a photomultiplier tube after first removing the exciting light by passage through a 500-nm interference filter and a high pass filter with a cutoff wavelength of 495 nm (Schott Glass Type GG495). Under these instrumental conditions, acidification was detected as a decrease in fluorescence.

The standard reaction mixture for measurement of pyranine fluo- rescence contained 100 mM MOPS-KOH, pH 7.5, 60 mM &So,, 10 mM MgSO,, 12 mM BBBP, and 0.25-0.5 mg of pyranine-SMP in a final volume of 2.5 ml. Additions were made to a 4-ml acrylic cuvette of I-cm path length with four optical surfaces (Elkay Products, Shrewsbury MA) equipped with an efficient magnetic stirrer. In order to accommodate several traces in one figure, the initial base lines were shifted on the ordinate.

Fluorescence measurements with caged ATP were carried out in a quartz cuvette, 1 cm wide, with a 2-mm path length. Light for fluorescence excitation and for release of ATP from the cage entered the cuvette along the 2-mm path. Emitted fluorescent light was monitored at right angles. The reaction mixture, in a final volume of 400 pl, contained the fluorescence assay medium and 0.08 mg of pyranine-SMP into which caged ATP had been incorporated. ATP was released from the cage with a single 2-ms flash of light from a Balcar 3200 watt/s flash lamp (Tekno, Inc., Chicago, IL). Before reaching the sample, light from the flash lamp was passed through a Corning 7-54 glass filter and a Schott WG305 high pass filter. The cuvette was irradiated with approximately 1 Joule of filtered light (measured with a model 2501 energy meter, Scientec, Boulder, CO). A mirror was placed behind the cuvette during the flash in order to reflect activating light back into the sample. Stirring was performed manually after additions were made. In control experiments, the ability of pyranine-SMP to pump protons, associated with the hy- drolysis of externally added ATP, was not influenced by exposure to a single flash from the flash lamp.

Oxonol V fluorescence was excited at 580 nm. Emitted fluorescent light was isolated with a Schott high pass filter (GG630). Conforma- tional changes in the adenine nucleotide translocase of F1-depleted beef heart SMP were measured in a 2 X 10-mm quartz cuvette by monitoring the binding and fluorescence of the specific probe DAN- ATP (17-19). The fluorescence of DAN-ATP was excited at 360 nm and emitted light was passed through a Schott high pass 455-nm cut- off filter (GG455).

The translocase of yeast mitochondria was specifically labeled with fluorescein-maleimide as follows. Frozen mitochondria (-70 “C) were thawed and washed once in a buffer containing 10 mM Tris-C1, pH 8.0, and 0.25 M sucrose. Washed mitochondria (5 mg in 0.5 ml) were incubated on ice in the same buffer for 20 min with 0.5 mM DTNB, after which 200 PM ADP and 150 PM fluorescein maleimide were added. After an additional 20 min of incubation, 2 pl of 14 M p- mercaptoethanol were added, followed immediately by 5 ml of ice- cold acetone to precipitate the protein (20). After 10 min of incubation on ice, the precipitate was sedimented by centrifugation and the supernatant discarded. Any remaining acetone was evaporated under a stream of argon. The precipitate was dissolved in sample buffer and subjected to gel electrophoresis on 13% gels according to Ref. 21. Fluorescent bands on the gel were photographed with the aid of an ultraviolet transilluminator.

Protein concentration was determined by a biuret procedure in the presence of 0.33% deoxycholate, using bovine serum albumin as standard (22).

ATPase activity was measured by assaying release of inorganic phosphate from ATP in the presence of an ATP regenerating system (22).

RESULTS

Although pyranine-SMP were chromatographed on Seph- adex just prior to use (see “Experimental Procedures”), sig- nificant amounts of pyranine appeared to be extravesicular. This is shown in Fig. 1, using the quencher BBBP. It may be seen that BBBP effectively quenched not only the fluores- cence of pyranine free in solution but also the fluorescence of pyranine-SMP solubilized with 1% Triton X-100. Extrapo- lation of a double-reciprocal plot of the percent fluorescence

25322 Interactions between Adenine Nucteotide Translocase and F p l

o . o o 4 ’ r ’ r ’ ~ . r . ~ ’ r . I 0 . 0 0.1 0 . 2 0.3 0 .4 0.5 0 . 6 0.7

1 / [BBBP] (l/mM)

FIG. 1. Titration of free and membrane-bound pyranine with the quenching agent, BBBP. The reaction mixture contained 100 mM MOPS-KOH, pH 7.5,60 mM K,SO, and 10 mM MgSOa in a final volume of 2.5 ml. Lowermost curue (A) , the reaction mixture contained, in addition, 0.8 p~ pyranine. Middle curue (O), the reaction mixture contained in a final volume of 2.5 ml of the above buffer, 0.5 mg of pyranine-SMP and 1% Triton X-100. Uppermost curue (O), the reaction mixture contained in addition to the above buffer, 0.5 mg of pyranine-SMP. BBBP was added from a 0.5 M stock solution. The fluorescence shown was recorded after stable values for each addition were reached. No corrections were made for dilution by the BBBP additions (maximum dilution was 2.4%).

quenched uersus BBBP concentration indicated that fluores- cence would be completely quenched in both cases at infinite BBBP concentration (0.01 on the ordinate). However, when pyranine was incorporated into submitochondrial particles, only 40-55% of pyranine fluorescence was quenched. Thus, 45-60% of the pyranine was unavailable to the quencher and presumably entrapped within the particles. Since a stable base line of fluorescence was observed, leakage of the BBBP into the vesicles or of pyranine out of the vesicles, was insignificant.

BBBP was routinely added to the assay mixture a t a con- centration of 12 mM. At this concentration, the external pyranine fluorescence was almost entirely quenched, Fig. 1. Consequently, more than 95% of the observed fluorescence signal originated from pyranine inside the submitochondrial particles. The traces in Fig. 2 show the response of pyranine to the internal acidification that resulted when protons were pumped into the vesicles during ATP hydrolysis. Similar results were obtained when oxidation of NADH or succinate was used to generate a ApH via the respiratory chain (not shown). It can be estimated from these experiments that the internal pH was decreased to about pH 6.8 during proton pumping. The slight upward drift following ATP addition is likely due to leakiness of the membrane to protons. In contrast to venturicidin (Fig. 2, upper truce), when oligomycin was added to pyranine-SMP during ATP-dependent proton pumping, only minimal reversal of the fluorescence decrease occurred (Fig. 2, lower curue). Moreover, FCCP was without effect after oligomycin had been added. The experiments described below were designed to explore the nature of this unexpected phenomenon.

As demonstrated in Fig. 3, addition of oligomycin and ADP to pyranine-SMP prepared in the presence of ATP led to a decrease in pyranine fluorescence. Both oligomycin and ADP were necessary for the observed quenching but their order of addition was not important (upper two curues). In contrast, pyranine-SMP prepared in the absence of added ATP exhib- ited oligomycin-induced quenching that did not require the presence of ADP (Fig. 3, lowermost curue). Addition of ADP led to a small additional decrease in fluorescence.

Oligomycin-induced fluorescence quenching of pyranine-

I ATP FCCP

0 1 2 3 -I 4

TIME (rnin)

FIG. 2. Response of pyranine incorporated into submito- chondrial particles from beef heart to ATP-dependent proton pumping and FoFl inhibitors. The reaction was carried out with 0.37 mg of pyranine-SMP prepared in the presence of 1 mM ATP and 0.8 pg/ml valinomycin in the standard reaction mixture of 2.5 ml described under “Experimental Procedures.” A rapid internal acidi- fication occurred immediately after adding 0.4 mM ATP. The break in the curve resulted from removing the cover to the sample com- partment. The second addition was 2 pg/ml venturicidin (uent, upper curue) or 4 pg/ml oligomycin (oligo, lower curue) and the final addition was 1 p~ FCCP. The ATPase activity of the pyranine-SMP had been activated by chromatography on Sephadex columns as described (1).

I \ ADP

I - O ATP inside 4 0 2 4 6

TIME (rnin)

FIG. 3. Oligomycin-induced fluorescence quenching of pyr- anine-SMP. The reaction mixtures contained 0.5 mg of beef heart pyranine-SMP in 2.5 ml of the standard fluorescence assay mixture described under “Experimental Procedures.” In addition, FCCP (1 p ~ ) and valinomycin (2 pg) were present. Pyranine-SMP were pre- pared in the presence of 1 mM ATP (two uppermost curues) or in the absence of added nucleotides (lowermost curue). Oligomycin (4 pg/ ml) and ADP (0.2 mM) were added to the medium as shown. Fluores- cence was measured as described under “Experimental Procedures.”

SMP was insensitive to uncouplers of oxidative phosphoryl- ation; since the experiments in Fig. 3 were carried out in the presence of FCCP, valinomycin, and potassium, suggesting that quenching did not reflect alterations in proton concen- tration. Further support for the conclusion that oligomycin- induced quenching was not related to intravesicular changes in proton concentration is shown in Fig. 4. Internal acidifi- cation of the vesicles, resulting from ATP-dependent proton pumping, was accompanied by a wavelength-specific (that is, from about 420 to 480 nm) reduction in the pyranine excita- tion spectrum, Fig. 4A. In contrast, addition of oligomycin and ADP or ATP led to a reduction in the entire pyranine excitation spectrum, Fig. 4B. Fig. 4B also shows that the

Interactions between Adenine Nucleotide Translocase and F81 25323

3 6 0 4 0 0 4 4 0 4 8 0 20 4

3 6 0 4 0 0 4 4 0 4 8 0

nm nm

FIG. 4. Excitation spectra of pyranine-SMP in the presence of oligomycin and during internal acidification. The reaction mixtures contained 0.25 mg of beef heart pyranine-SMP prepared in the presence of 1 mM ATP in 2.5 ml of the standard fluorescence assay mixture (“Experimental Procedures”). Emitted fluorescent light was monitored via the emission monochromator set at 511 nm with a 16 nm bandpass. The sample was excited by light from the excitation monochromator (4 nm bandpass) in the wavelength range 360-480 nm. Panel A, the effect of internal acidification due to ATP-supported proton pumping on the excitation spectrum of pyranine-SMP. Solid curue, pyranine-SMP. Dotted curue, the excitation spectrum was recorded during proton pumping in the presence of 2 mM ATP. B, the effect of oligomycin on the excitation spectrum of pyranine-SMP. The uppermost, solid curue is the control spectrum of pyranine-SMP. The dotted lowermost curue: after a 5-min incubation of pyranine-SMP with 4 pg/ml oligomycin (oligo), ATP (2 mM) was added and the spectrum recorded. The central (dashed line) curue was recorded 10 min after adding 10 p~ bongkrekic acid (BKA) to the reaction mixture used to obtain the lowermost curve.

c ‘S 0 1 2 3 4

TIME (min)

OLIGO

0 2 4 6 0 1 0

TIME (mln)

FIG. 5. Effects of the translocase inhibitors atractyloside and bongkrekic acid on the oligomycin-induced fluorescence quenching of pyranine-SMP. Panel A, the standard fluorescence assay mixture contained 0.37 mg of beef heart pyranine-SMP that had been prepared in the absence of added adenine nucleotides, 0.8 pg/ml valinomycin, and 1 PM FCCP. Pyranine-SMP used in the experiment shown by the lowermost curue were prepared in the presence of 100 p~ atractyloside (ATR). Panel B, pyranine-SMP were prepared in the presence of 1 mM ATP with (+ATR) or without (no ATR) 100 p~ atractyloside. After isolation, the ATPase activity of the pyranine-SMP was activated by chromatography on columns of Sephadex, and the obtained particles were treated with urea to remove F, (1). Using untreated pyranine-SMP the same results were obtained. The cuvette contained in a final volume of 2.5 ml, the standard fluorescence assay mixture, 0.37 mg of submitochondrial particles, 0.8 pg/ml valinomycin, 1 p~ FCCP, and 0.2 mM ADP. Where shown in both panels, 4 pg of oligomycin (oligo), 10 p~ bongkrekic acid (BKA), and 0.2 mM ADP were added.

reduced excitation spectrum in the presence of oligomycin and nucleotides was largely reversed following addition of the adenine nucleotide translocase inhibitor, bongkrekic acid. Valinomycin, potassium, and FCCP were routinely added to reaction mixtures in this study in order to eliminate possible contributions from pH changes to observed oligomycin-de- pendent quenching.

Because adenine nucleotides on both sides of the vesicle membrane influenced oligomycin-dependent quenching of pyranine fluorescence (Fig. 3), it seemed likely that the target for nucleotide binding was the adenine nucleotide translocase. Fig. 5 shows the effects of the highly specific translocase inhibitors, atractyloside and bongkrekic acid on oligomycin- induced fluorescence quenching. In Fig. 5A, the two upper- most curves, it may be seen that bongkrekic acid was without effect on oligomycin-induced quenching of pyranine fluores-

cence (whether added before or after oligomycin) when the submitochondrial particles were prepared in the absence of ATP. Submitochondrial particles prepared in the absence of ATP but in the presence of atractyloside exhibited minimal responses to oligomycin, Fig. 5A, lowermost curve. If, however, the submitochondrial particles were prepared in the presence of ATP and the medium contained ADP, addition of bong- krekic acid substantially reversed oligomycin-induced quenching of fluorescence, Fig. 5B, upper curve. Addition of bongkrekic acid before oligomycin prevented oligomycin-in- duced quenching (not shown). The reversal of oligomycin- induced quenching of pyranine fluorescence by bongkrekic acid was also reflected in the almost complete restoration of the pyranine-SMP excitation spectrum following addition of the inhibitor (Fig. 4B) . Submitochondrial particles prepared in the presence of ATP and atractyloside exhibited minimal

25324 Interactions between Adenine Nucleotide Translocase and F$,

quenching by oligomycin and no reversal by bongkrekic acid, Fig. 5B, lower curve.

The experiments shown in Fig. 5B were carried out with submitochondrial particles that were activated by chromatog- raphy on Sephadex columns and then treated with urea to remove membrane-bound F1 (see “Experimental Proce- dures”). The same results were obtained with submitochon- drial particles that had not been treated to remove the ATPase (not shown).

A summary of a series of experiments similar to those presented in Fig. 5 is shown in Table I. Oligomycin quenched the fluorescence of pyranine-SMP prepared in the presence of ATP, “ATP inside,” only if nucleotide also was added to the medium, compare lines 1 and 2. Under the latter condi- tions (line 2), oligomycin-induced quenching was inhibited by both atractyloside and bongkrekic acid. When ATP was ab- sent from the sonication medium (lines 3 and 4), oligomycin- induced quenching was independent of the presence of me- dium nucleotides. Quenching was inhibited by atractyloside but insensitive to bongkrekic acid whether or not ADP was present in the medium.

It has been reported that binding of the translocase inhib- itors, atractyloside and bongkrekic acid, to the adenine nu- cleotide translocase resulted in an altered conformation of the carrier and complete loss of transport activity (12,13). It was of interest to determine if the reversal of oligomycin-induced fluorescence quenching by bongkrekic acid could take place only after formation of a translocase-inhibitor complex. Al- ternatively, reversal of quenching might be achieved simply by permitting a physiological effector such as ATP to bind to the inner (cytosolic face) of the translocase. In order to test these possibilities, pyranine-SMP were prepared in the pres- ence of 1 mM caged ATP. The behavior of these particles was similar to that prepared in the absence of added adenine nucleotides, Le. addition of external nucleotide was not re- quired to observe oligomycin-induced fluorescence quenching (Fig. 6). When the entrapped caged ATP was released by flash photolysis, fluorescence quenching was reversed (Fig. 6, upper curve). Subsequent addition of ATP to the medium led to quenching which was reversed by bongkrekic acid. The lower curve of Fig. 6 shows that when ATP was added to the medium before the light flash, release of ATP within the vesicle did not reverse oligomycin-induced fluorescence quenching. However, the reversal by bongkrekic acid indicates that ATP had indeed been released within the vesicle (cf. Table I). ATP and ADP were equally effective as external nucleotides, Fig. 6 and Table 11.

It would appear that adenine nucleotide released within the

TABLE I Dependence of oligomycin-induced quenching of pyranine

fluorescence on the presence of nucleotides The data were obtained from experiments as described in Fig. 5.

ATP was incorporated into pyranine-SMP by sonication of mito- chondria in the presence of pyranine and 1 mM ATP. ADP (0.2 mM) was added to the medium. Atractyloside (ATR) was present during sonication of the mitochondria. Bongkrekic acid (BKA) was added to

1 + 2 + + + + + 3 + + + 4 + +

a - -

- - - - -

~ ~~

a No quenching.

C”

0 2 4 6 0 TIME (min)

FIG. 6. Reversal of oligomycin-induced fluorescence quenching of pyranine following release of caged ATP within pyranine-SMP. The standard fluorescence assay mixture contained 0.08 mg of pyranine-SMP, 1 p~ FCCP, and 0.32 pg of valinomycin in a final volume of 0.4 ml in a 10 mm wide by 2 mm deep fluorescence cell. Pyranine-SMP were prepared from beef heart mitochondria in the presence of 1 mM caged ATP as described under “Experimental Procedures.” Light for excitation of pyranine fluorescence and release of caged ATP entered along the 2-mm path of the cuvette. Emitted fluorescent light was monitored at right angles. The flash lamp apparatus is described under “Experimental Procedures.” Oligomycin (oligo), 4 pg/ml, ATP, 1 mM, and bongkrekic acid (BKA), 10 pM, were added as shown.

TABLE I1 Substrate specificity of oligomycin-induced fluorescence quenching of

pyranine-SMP Pyranine-SMP were prepared by sonicating beef heart mitochon-

dria in the presence of pyranine and 1 mM ATP as described under “Experimental Procedures.” The standard fluorescence assay mixture contained 0.5 mg of pyranine-SMP. The particles were incubated for 5 min with 0.8 pg/ml valinomycin, 1 p~ FCCP, and 4 pg/ml oligo- mycin. The observed fluorescence level was set to 100%. The nucle- otides shown were added from a 100 mM stock solution, except for AMP-PNP which was 20 mM. The extent of fluorescence quenched was determined after about 2 min when a stable level was reached. The level of quenching observed with ATP (about 40%) was arbitrar- ily set to 100%. Levels of quenching reached by other additions are shown as the percent of ATP-induced quenching.

Addition Quenching

% ATP (1 mM) 100 ADP (1 mM) 95

GTP (3 mM) 0 GDP (3 mM) 0 ITP (3 mM) 0 IDP (3 mM) 0

AMP-PNP (0.6 mM) 74

vesicle by photolysis of caged ATP interacted with the trans- locase rather than the ATPase, which faces the external medium. However, as shown above, external nucleotide also modulated oligomycin-dependent quenching. Consequently, the target for adenine nucleotide added to the medium could be both the translocase and F1-ATPase. Two types of exper- iments were conducted to address this point. First, removal of F1 from the membrane by treatment with urea did not alter the oligomycin-induced fluorescence quenching (Fig. 5B) . Second, the nature of the medium nucleotide effective in promoting oligomycin-dependent fluorescence quenching was determined. The results are shown in Table 11. It may be seen that the effective nucleotides (ATP, ADP, and AMP-PNP) were those that are recognized ligands of the adenine nucleo- tide translocase (13,14,19). GTP and ITP, which were readily

Interactions between Adenine Nucleotide Translocase and Fal 25325

hydrolyzed by FoF1-ATPase (22), supported proton translo- cation in the same manner as shown for ATP in Fig. 2. However, guanine and inosine nucleotides, which do not bind to the translocase (13, 14, 19) did not promote oligomycin- dependent quenching of pyranine fluorescence, Table 11.

The interactions of pyranine with submitochondrial parti- cles could be distinguished from recognized conformational changes in the adenine nucleotide translocase using DAN- ATP, a highly sensitive probe of conformational change in the carrier (17-19). The experiments shown in Fig. 7 were carried out with submitochondrial particles treated with urea in order to remove F1 from the membrane, because DAN- ADP may bind to F, (17). Addition of DAN-ATP to F,- depleted submitochondrial particles from beef heart was fol- lowed by the increase in fluorescence that results when DAN- ATP binds to the translocase (18, 19). Addition of a translo- case substrate such as ATP, Fig. 7, upper curve, resulted in a decrease in DAN-ATP fluorescence, presumably due to com- petition at the binding site on the translocase (19). Oligomycin had no influence on the fluorescence of DAN-ATP bound to these particles, whether or not ATP was present. However, as expected for this probe of the adenine nucleotide translo- case (17-19), bongkrekic acid effectively quenched the fluo- rescence of DAN-ATP, Fig. 7. It is important to note that the conformational change in the carrier, monitored by DAN- ATP when bongkrekic acid was added, was completed within a few seconds whether or not oligomycin was present. In contrast, the complete reversal of oligomycin-induced quench- ing of pyranine fluorescence in beef heart submitochondrial particles by bongkrekic acid required several minutes, Fig. 5B.

The role of the adenine nucleotide carrier in the oligomycin- induced fluorescence quenching of pyranine in submitochon- drial particles was further investigated with a yeast mutant (DNY1) lacking the entire coding regions of the genes AACl and AAC2 (23). These genes encode isoforms of the translo- case. The mutant has the same genotype as a previously described mutant JLY1053 (24), in which the two genes had been disrupted. Mitochondria isolated from JLY1053 cells not only contained no immunologically detectable translocase (24) but in addition, binding of the tight binding translocase inhibitor, carboxyatractyloside, was not detected (25). Fig. 8

I .

0 ; 2 3 4

TIME (min) FIG. 7. Effects of oligomycin and bongkrekic acid on the

fluorescence of DAN-ATP in submitochondrial particles from beef heart. The reaction mixture contained 0.8 mg of submitochon- drial particles, treated with urea so as to remove F1 (l), in 0.4 ml of a buffer containing 10 mM Tris-HC1, pH 7.5 and 250 mM sucrose. The sample was placed in a 10 mm wide by 2 mm deep fluorescence cuvette. Exciting light (see “Experimental Procedures”) entered along the 2-mm path. DAN-ATP (3 p ~ ) , ATP (1 mM), oligomycin (oligo) (4.8 pg), and bongkrekic acid (10 p ~ ) were added as indicated.

w1’ DKYl

FIG. 8. SDS-polyacrylamide gel electrophoresisof yeast mi- tochondria from wild type ( W T ) and the translocase-deficient mutant (DNYZ). Samples of mitochondria were prepared for stain- ing with fluorescein-maleimide and for electrophoresis on 13% poly- acrylamide gels as described under “Experimental Procedures.” Ap- proximately 60 pg of protein were added to each lane of the slab gels. The arrow points to a fluorescein-maleimide-stained fluorescent pro- tein band of apparent molecular weight 32,000. The second band in each lnne has not been identified. The location of the two uppermost bands in each lane coincides with the top of the separating gel. The bands at the bottom of the gel represent free dye.

shows an electrophoretic separation on SDS-polyacrylamide gels of the proteins of mitochondria obtained from wild type and mutant cells and incubated with fluorescein-maleimide as described under “Experimental Procedures.” It has been shown that the translocase of heart and liver mitochondria could be specifically labeled with eosin-maleimide (26, 27). After pretreatment with N-ethylmaleimide (26) or DTNB (27), only a single fluorescent band corresponding to the translocase was observed. As shown in Fig. 8, left l a n e , labeling of wild type yeast mitochondria with fluorescein-maleimide resulted in the appearance of two fluorescent protein bands. The upper band (arrow) with an apparent molecular weight of 32,000 is close to that expected for the AAC2 gene product. It is clear that submitochondrial particles from the DNYl mutant, Fig. 8, right l a n e , lacked this protein. The more rapidly migrating second band in the left lane of Fig. 8 is also present in the mutant and represents an unknown protein. Of the two genes coding for the translocase, the AACl gene product is produced in amounts too small to be visualized on these gels (24). It might be expected that the translocase would be readily visualized on gel electrophoresis since the enzyme constitutes about 10% of the protein of the inner mitochondrial membrane (2).

Submitochondrial particles prepared from both wild type and DNYl mutant cells exhibited similar levels of ATPase activity which were equally sensitive to oligomycin, Fig. 9. Both types of submitochondrial particles were capable of generating a membrane potential as reported by oxonol V. Addition of ATP to submitochondrial particles previously incubated with oxonol V led to a decrease in oxonol V fluo- rescence in both wild type and mutant particles. The ATP-

25326 Interactions between Adenine Nucleotide Translocase and F g 1

2.0 t

0 1 0 2 0 3 0 1 7 0 [OLIGOMYCIN] ( pg I mg OF PROTEIN)

FIG. 9. Oligomycin inhibition of ATPase activity of submi- tochondrial particles from wild type (WT) and translocase deficient (DNYI) cells. Submitochondrial particles (1 mg) were incubated with oligomycin as shown in a buffer containing 10 mM Tris-CI, pH 7.5, and 250 mM sucrose. The final volume was 1.0 ml. After 15 min of incubation, 100 p1 were transferred to 0.9 ml of assay medium (see “Experimental Procedures”) for determination of ATP-

DNYl

0 2 4 6 8

TIME (min)

FIG. 10. Oligomycin-induced quenching of the fluorescence of pyranine incorporated into submitochondrial particles from wild type (WT) and mutant (DNYZ) yeast cells. The fluorescence of pyranine-SMP (0.25 mg) was monitored in 2.5 ml of fluorescence medium as described in Fig. 3, except 0.2 mM ADP was also present. Oligomycin (oZigo) at 10 pg and bongkrekic acid ( B K A ) at 10 p~ were added as shown. The truces were corrected for a decreasing base line of about 1-2%/min.

dependent decrease in fluorescence of both types of particles was sensitive to the ATPase inhibitor venturicidin and the uncoupler FCCP. Valinomycin collapsed the potential, since potassium was present at high concentration (120 mM) both inside and outside the submitochondrial particles. When the fluorescent pH indicator pyranine was incorporated into mu- tant and wild type submitochondrial particles (pyranine- SMP), acidification of the interior, as indicated by a decrease in pyranine fluorescence, was observed upon addition of ATP in a similar manner as shown for beef heart pyranine-SMP in Fig. 2. As expected, fluorescence quenching was reversed by subsequent addition of venturicidin or FCCP.’

An oligomycin-dependent quenching of pyranine fluores- cence was observed in pyranine-SMP prepared from both wild type and mutant particles, Fig. 10. As shown above for beef heart pyranine-SMP, fluorescence quenching due to a de- crease in pH could be easily distinguished from an oligomycin- dependent effect. The experiments shown in Fig. 10 were in fact carried out in the presence of FCCP, valinomycin, and potassium in order to ensure an absence of pH effects on ~~

M. Ziegler and H. S. Penefsky, unpublished observations.

pyranine fluorescence. I t may be seen that the extent of fluorescence quenching was about equal in wild type and mutant submitochondrial particles. However, oligomycin-de- pendent fluorescence quenching was reversed upon addition of bongkrekic acid to wild type particles, Fig. 10, upper curve, but not to mutant particles, Fig. 10, lower curve. Additional oligomycin was without effect on fluorescence quenching, suggesting that quenching was maximal following the first addition. I t is clear from the experiments shown in Fig. 10 that the presence of the adenine nucleotide translocase is not required for oligomycin-dependent quenching of pyranine flu- orescence.

DISCUSSION

The main observation of this paper is that addition of oligomycin to pyranine-SMP led to an apparent sequestration or binding of the probe and quenching of fluorescence. Fluo- rescence quenching appeared to be modulated by the adenine nucleotide translocase. Quenching of pyranine fluorescence by oligomycin could be distinguished from pH-dependent quenching because the excitation spectrum of the latter was wavelength specific whereas oligomycin caused a uniform decrease in fluorescence over the entire excitation spectrum, Fig. 4. In addition, it is unlikely that oligomycin-dependent quenching was due to a pH effect since these experiments were carried out in the presence of FCCP, valinomycin, and potassium.

The requirements for nucleotides on both sides of the membrane as well as the nucleotide specificity of oligomycin- induced quenching point to a role for the translocase in these observations. F1 itself does not seem to participate since removal of F1 by extraction of the particles with urea was without effect on quenching (Fig. 5B) . Thus, the translocase is likely to be the target for adenine nucleotides added to either side of the membrane. Table I shows that quenching was not observed with pyranine-SMP prepared in the pres- ence of nucleotides but tested without ADP in the medium. Quenching was restored by adding to the medium nucleotides that are known to interact with the translocase (ADP, ATP, AMP-PNP). ITP and GTP, which were hydrolyzed by Fl(22) and supported proton pumping by FoF1, did not bind to the translocase (13, 14, 19) and did not replace adenine nucleo- tides in quenching, Table 11. A similar nucleotide specificity was observed in studies of the modulation of fluorescence of the analog DAN-ADP which was specifically bound by the translocase (19).

Further support for a role for the adenine nucleotide trans- locase was provided by the effects of specific inhibitors of the carrier on oligomycin-induced quenching of pyranine fluores- cence. Since atractyloside binds to the translocase only on the cytosolic side of the membrane (12, 13) it could be used to prevent oligomycin-induced quenching, Fig. 5, but not for a reversal of quenching. However, bongkrekic acid, which binds exclusively to the matrix side of the translocase (12,13) substantially, if not completely, reversed oligomycin-induced quenching, Figs. 4-6, and 10. Bongkrekic acid prevented oli- gomycin quenching if added before the antibiotic.

I t is of interest that reversal of oligomycin-induced quench- ing of pyranine fluorescence by bongkrekic acid was effective only if an adenine nucleotide was present inside the vesicle, Table I. The implication of this finding is that an adenine nucleotide must be present on the cytosolic side (interior of SMP) of the translocase if bongkrekic acid is to bind to the matrix side. The presence of ATP inside the particles might stabilize a conformation which renders the translocase SUS-

Interactions between Adenine Nucleotide Translocase and FoF, 25327

ceptible to bongkrekic acid but does not permit oligomycin- induced fluorescence quenching (Figs. 3, 5, and 6). Under these conditions, the addition of external (matrix side) sub- strate could initiate a nucleotide exchange leading to cyclic alterations in conformation, one of which might be trapped by the addition of oligomycin.

An alternative explanation of these translocase interactions is based on suggestions that atractyloside and bongkrekic acid induce the c or m states, respectively, in the translocase (12, 13, 19). If each functional translocase molecule contains only one binding site for either substrates or inhibitors, as has been proposed (12), ATP bound to the translocase within the submitochondrial particle would have to dissociate before the binding center could relocate to the exterior where it could bind bongkrekic acid. The experiments using photolysis of incorporated caged ATP suggest that structural reorientation rather than the formation of an inactive inhibitor-translocase complex is required for the reversal of oligomycin-induced quenching. Since the deletion mutant (DNY1) exhibits oli- gomycin-induced fluorescence quenching, neither pyranine nor oligomycin is likely to bind to the translocase during these quenching experiments.

Evidence presented in this paper (Fig. 8) that SDS gels of submitochondrial particles prepared from the mutant lack the 32-kDa protein characteristic of the translocase is consistent with the recent report of Nelson et al. (23) that antibodies prepared against the translocase failed to detect any of the carrier in mitochondria from mutant cells. I t is important to point out that no functional differences were detected in energy-dependent reactions catalyzed by the FoFl-ATPase of wild type and mutant submitochondrial particles. The levels of ATPase activity and sensitivity of ATPase to oligomycin were similar (Fig. 9) and both types of particles were capable of establishing a A,J(H+ during ATP hydrolysis as indicated by membrane potentials and internal acidification.' Thus, dele- tion of the translocase genes did not affect either expression or assembly of the FoFl-ATPase. These experiments also show that a binding site on the translocase for either oligomycin or venturicidin is not needed for the inhibition of proton trans- location through the FoF1-ATPase.

No effect of oligomycin could be detected when conforma- tional changes in the translocase were monitored using DAN- ATP in beef heart submitochondrial particles (Fig. 7). Addi- tion of bongkrekic acid to the fluorescent complex of DAN- ATP and the translocase was followed by a decrease in fluo- rescence signal that was within the mixing time of the instru- ment (less than 1 s, Fig. 7). In contrast, complete reversal of oligomycin-induced quenching of pyranine fluorescence by this translocase inhibitor took 1-2 min (Fig. 5B).

Taken together, a reasonable interpretation of the obser- vations presented is that binding of oligomycin to Fo causes a conformational change in Fo creating a binding site for pyr- anine. Binding of pyranine to Fo results in a quenching of fluorescence. Subsequent conformational rearrangements of the adenine nucleotide translocase induced, for example, by binding of bongkrekic acid to the carrier (2, 12, 13, 19), is transmitted to Fo and reverses the quenching of pyranine fluorescence. It is relevant to this model to point out that pyranine-SMP were prepared by sonication of mitochondria in 0.5 mM pyranine. Since the internal volume of submito- chondrial particles from beef heart is about 1 pl/mg protein (28, 291, approximately 0.5 nmol of pyranine/mg protein might be within the intravesicular space. In addition, each milligram of particle protein also contains about 0.4 nmol of F1 (30). Thus, the ratio of intravesicular pyranine to F1 in

these experiments is not far from unity. Gupte et al. (27) concluded from their studies that a specific

interaction between FoFl and the translocase was not needed to explain the character of the lateral diffusion of the trans- locase in the inner mitochondrial membrane. These experi- ments were carried out with preparations in which the trans- locase had been labeled with eosin-maleimide (27). However, it has been reported that N-ethylmaleimide inhibits the action of the translocase (13), and it was found in the current study that N-ethylmaleimide, like bongkrekic acid, reverses the oligomycin-dependent quenching of pyranine fluorescence.' Thus, the eosin-maleimide label may have interfered with detection of a specific interaction between FoFl and the trans- locase in the lateral diffusion studies (27).

Oligomycin-dependent quenching was specific for pyranine. Signals from the fluorescent pH probes acridine orange, car- boxyfluorescein, or dextran-linked BCECF did not respond to oligomycin. In addition, alteration of the fluorescence assay medium, for example, use of a buffer containing 250 mM sucrose, 1 mM EDTA, and 10 mM Tris-HC1, pH 7.5, did not alter the oligomycin-induced quenching. The likely presence of a specific binding site for oligomycin in these experiments is suggested by the absence of a quench effect when the inhibitor was added to everted Escherichia coli membrane vesicles or to phospholipid vesicles containing 0.5 mM pyr- anine.' The suppression of oligomycin-induced quenching via inhibition of the translocase further suggests that oligomycin acted via its generally recognized binding site on Fo and that other, nonspecific, binding sites for oligomycin were not involved.

Preincubation of pyranine-SMP with DCCD, which acts similarly to oligomycin in that it binds to Fo and causes inhibition of ATPase activity, proton pumping, and ATP synthesis (10, l l ) , was without effect on oligomycin-induced fluorescence quenching of pyranine.' Whereas the site of action of DCCD is well defined (it modifies a single carboxyl group of an amino acid residue in position 61 of subunit 9 of mitochondrial Fo or subunit c of the E. coli enzyme (11, 31)), oligomycin may interact with two components of yeast Fo, subunits 6 and 9 (11, 31). Thus, these pyranine quenching experiments point to an operational distinction between the modes of action of oligomycin and DCCD. Similarly, in spite of functional and even structural similarities between oligo- mycin and venturicidin (lo), venturicidin did not quench the fluorescence of pyranine in submitochondrial particles (Fig. 2). Genetic analyses indicated (reviewed in Ref. 31) that venturicidin resistance in yeast mapped exclusively to a region close to the DCCD binding site in subunit 9. It appears that the interaction of oligomycin with subunit 6 is specific for this inhibitor. This information may serve as a starting point in the search for the site of interaction of Fo with pyranine and, perhaps, the translocase.

Acknowledgments-We thank Drs. D. R. Nelson and M. G. Douglas for kindly providing us with samples of DNY 1 and Dr. M. Klingenberg for DAN-ATP. W e are grateful to Marcus Hutcheon for expert technical assistance.

REFERENCES 1. Racker, E., and Horstman, L. L. (1967) J. Biol. Chem. 242 , 2547-2551 2. Hackenberg, H., and Klingenberg, M. (1980) Biochemistry 19.548-555 3. Penefsky, H. S., Pullman, M. E., Datta, A,, and Racker, E. (1960) J . Biol.

4. Penefsky, H. S., and Cross, R. L. (1991) Adu. Enzymol. Rel. Areas Mol.

5. Futai, M., Noumi, T., and Maeda, M. (1989) Annu. Reu. Biochem. 58,111-

Chem. 235,3330-3336

Biol. 6 4 , 173-214

1 RG 6. Senior, A. E. (1990) Annu. Reu. Biophys. Biophys. Chem. 19 , 7-41

8. Issartel, J. P., Dupuis, A., Garin, J., Lunardi, J., Michel, L., and Vignais, 7. Fillingame, R. H. (1990) Bacteria 12 , 345-391

_"

25328 Interactions between Adenine Nucleotide Translocase and Fal

9. Lardy, H. A., Connelly, J. L., and Johnson, D. (1964) J . Bid. Chem. 12, P. V. (1992) Experientia 48,351-362

10. Linnett, P. E., and Beechey, R. B. (1979) Methods Enzymol. 55,472-518 11. Sehald, W., and Hoppe, J. (1981) Curr. Topics Bioenerg. 12, l -64 12. Klingenberg, M. (1989) Arch. Biochem. Biophys. 270, l -14 13. Vignais, P. V. (1976) Biochim. Biophys. Acta 456, 1-38 14. Pfaff, E., and Klingenberg, M. (1968) Eur. J. Biochem. 6,66-79 15. Kano. K., and Fender, J. H. (1978) Biochim. Biophys. Acta 5 0 9 , 289-299 16. Beyer, R. E. (1967) Methods Enzymol. 10,186-194 17. Schafer, G., and Onur, G. (1980) FEBS Lett. 109,197-201 18. Mayer, I., Dahms, S., Riezler, S., and Klingenberg, M. (1984) Biochemistry

19. Klingenberg, M., Mayer, I., and Dahms, A. S. (1984) Biochemlstry 23,

20. Lester, R. L. (1967) Methods Enzymol. 10, 264-266 21. Laemmli, U. K. (1970) Nature 227,680-685

1961-1968

23,2436-2442

2442-2449

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

Pullman, M. E., Penefsky, H. S., Datta, A,, and Racker, E. (1960) J. Biol.

Nelson, D. R., Lawson, J. E., Klingenberg, M., and Douglas, M. G. (1993)

Lawson, J. E., Gawaz, M., Klingenberg, M., and Douglas, M. G. (1990) J.

Gawaz, M., Douglas, M. G., and Klingenberg, M. (1990) J. Biol. Chem.

Muller M. Krebs J. J. R., Cherry, R. J., and Kawato, S. (1984) J . Biol.

Guite, S . S., Chazotte B., Leesnitzer, M. A,, and Hackenbrock, C. R. (1991)

Sorgato, M. C., Galiazzo, F., Panato, L., and Ferguson, S. J. (1982) Biochim.

Sorgato, M. C., Lippe, G., Seren, S., and Ferguson, S. J. (1985) FEBS Lett.

Harris, D. A,, Radda, G. K., and Slater, E. C. (1977) Biochim. Biophys. Acta

Nagley, P. (1988) Trends Genet. 4,46-52

Chem. 235,3322-3329

J. Mol. Biol. 230, 1159-1170

Biol. Chem 265,14195-14201

265,14202-14208

Cheh. 269,3033-3043

~ochm. Bmphys. Acta 1069,131-138

BLophys. Acta 682, 184-188

181,323-327

459,560-572