Aurovertin , a Fluorescent Probe of Conformational Change in Beef Heart Mitochondrial Adenosine Triphosphatase *

Formation of a complex between aurovertin and soluble mitochondrial adenosine triphosphatase (FJ was accompanied by a 55-fold enhancement of fluorescence, an increase in the polarization of fluorescence from 0.278 for the free form to 0.375 for the bound form and a decrease in absorption at 366 nm. The fluorescence intensity of the complex was partially quenched by addition of ATP or Mg++ and enhanced by ADP. Two binding sites for aurovertin were found on F1 in the presence of ATP and one site in the presence of either ADP, Mg++, or dilute buffer. The dissociation constants of the fluorescent complex were 0.52, 0.07, 0.013, and 0.04 PM in the presence of ATP, ADP, Mg++, or in buffer, respectively. It is proposed that of the two binding sites for aurovertin on F1, only one site participates in inhibition of ATPase activity. Addition of succinate to a fluorescent complex of aurovertin and submitochondrial particles gave rise to an enhancement of fluorescence which depended on maintenance of an energized state. The changes in fluorescence of bound aurovertin were interpreted in terms of changes in the conformational state of the ATPase.

Two binding sites for aurovertin were found on F1 in the presence of ATP and one site in the presence of either ADP, Mg++, or dilute buffer.
The dissociation constants of the fluorescent complex were 0.52, 0.07, 0.013, and 0.04 PM in the presence of ATP, ADP, Mg++, or in buffer, respectively.
It is proposed that of the two binding sites for aurovertin on F1, only one site participates in inhibition of ATPase activity. Addition of succinate to a fluorescent complex of aurovertin and submitochondrial particles gave rise to an enhancement of fluorescence which depended on maintenance of an energized state.
The changes in fluorescence of bound aurovertin were interpreted in terms of changes in the conformational state of the ATPase.
The antibiotic aurovertin was first introduced as a tool for the study of oxidative phosphorylation by Lardy et al. (1) who established that this compound was a potent inhibitor of osidative phosphorylation, the 3'Pi-ATP eschange reaction, and the eschange of I*0 between the oxygens of Pi and ATP catalyzed by rat liver mitochondria.
In addition, aurovertin inhibited the hydrolysis of ATP stimulated by some uncouplers (1). Subsequcntly, it was found by Lenaz (2) and Lee and Ernster (3) that the forward reaction of ATP synthesis in oxidative phosphorylation was much more sensitive to aurovertin inhibition than the reverse energy transfer reactions (supported by either ATP or succinate oxidation) such as the reduction of DPN+ by succinate, the energy-linked reduction of TPN+ by DPNH, and the translocation of Ca++. The adenosine triphosphatase activ-* This research was supported in part by Research Grant Am 12201 of the National Institutes of Health, United Stat,es Public Health Service.
ity of submitochondrial particles also was found to be less zensitive to aurovertin than oxidative phosphorylation (3). Lardy et al. (1) proposed a branched pathway of energy transfer reactions from the respiratory chain to the terminal transphosphorylation step and suggested that aurovertin acted on only one pathway, at a point between the respiratory chain and the site of action of oligomycin.
However, Lee and Ernster (3) and Rob&on et al. (4) have proposed that nurovcrtin acted on the XTP side of the oligomycin-sensitive site. The observation that aurovertin inhibited soluble mitochondrial ATPase (4) and formed stoichiometric fluorescrnt complexes with Fir (5) prompted an examination of the interaction of this fluorophore with both soluble and mcmbranc-bound mitochondrial ATPase . This paper provides evidence in support of the conclusion that aurovertin fluorescence is responsive to conformational changes in mitochondrial ATPase.

MATERIALS AND METHODS
Materials-ATP, Tris-ATP (metal-free), and DPSH were purchased from Sigma. ADP was obtained from Worthington, Pyruvate kinase and lactate dehydrogenase were obtained from Boehringer-Mannheim Corporation. Solvcnt!: and other chemicals were of the highest purity obtainable and were used without further purification.
Aurovertin was a generous gift from Dr. Henry A. Lardy, University of Wisconsin.
Preparation of Fi-Fr was prepared as described previously (6). The preparations used in this study exhibited a specific activity which varied between 100 and 120 units per mg. A molecular weight for F1 of 347,000 (7) was used in binding calculations.
Samples of the enzyme wcrc freed of ammonium sulfate as follows.
i2n aliquot of the ammonium sulfate suspension of the enzyme was centrifuged in the cold, and, after removal of the supernatant, the pellet was dissolved at room temperature by adding 0.5 ml of Buffer I (0.25 M sucrose, 2 mxr EDTA, and 50 m&f Tris sulfate, pH 8.0) containing 4 mu ATP. The sample of enzyme was clarified by low speed centrifugation if necessary and then applied to a column of Sephadex G-50 (1 x 23 cm) which had been equilibrated with the same buffer. The column was eluted with Buffer I-AT]', and the enzyme was collected in a volume of approximately 2 ml. The distribution of ammonium sulfate in the column eluate was monitored with the aid of Nessler's reagent.
The specific activity of Fr was unchanged by this procedure. F1 was prepared in Buffer I containing ,4DP by the same procedure except that 1 InM ADP was substituted for ATP.
Samples of F1 with a low content of adenine nucleotides were prepared as follows.
About 10 ml of a suspension of Dowex 2-X4 (hydroxide form) in Buffer I was brought to pH 8 by the addition of 1 N HCl.
This material was used to prepare a column (1 X 3 cm) which was then washed with Buffer I. A sample of Fi, freed of ammonium sulfate in Buffer I containing ATP as described above, was applied to the column, and the enzyme was collected in a volume of about 3 ml. The ratio of the absorbance at 280 nm to that at 260 nm ranged between 1.33 and 1.47. The enzyme retained 80 to 90% of the original specific activity and was stable for about 5 hours at room temperature. However, on further standing, the solution lost activity and became turbid, and at higher concentrations (5 to 10 mg per ml) precipitation of protein was observed. It should be emphasized that Domes resin chromatography was the only effective method for removal of adenine nucleotides from F1 which did not cause large, immediate, and irreversible losses in cnzymc activity.
Chromatography on a column of Sephadex G-50 was not always effective in removing adenine nucleotides, as judged by the ratio of absorbance at 280 nm to that at 260 nm, and the emerging enzyme preparation uniformly retained less than 70% of the initial specific activity.
Preparation of Aurovertin Solutions-Crystalline aurovertin was dissolved in methanol at a final concentration of about 220 pM.
The exact concentration of diluted stock solutions in methanol was determined from the absorbance at 367 nm with the use of a molar extinction coefficient at this wave length of 42,700 (9). A molecular weight for aurovertin of 490 was calculated from the molecular formula of Baldwin et al. (9). If the molecular weight of 476 measured by Beechey et al. (10) is correct for the preparation used in this study, the binding data presented here would require a small correction.
Aurovcrtin cshibited a single fluorescent spot when chromatographed, at room temperature, on Whatman No. 1 paper in either water saturated with chloroform (RF = 0.68) or in a solvent consisting of waterdioxane (lO:l, v/v).
The RF in the latter solvent was 0.82. Aliquots of a few microliters were added to reaction mixtures directly from the stock solution of aurovertin in methanol.
No effect of methanol alone was observed on either fluorescence or enzyme activities under these conditions.

EXPERIMENTAL TECHNIQUE
Enzyme Assays-The ATPase activity of F1 was determined in the presence of a regenerating system for ATP either by colorimetric assay of Pi released from ATP (11) or, during kinetic studies, by the spectrophotometric method (11). In the kinetic studies, the reaction was started by the addition of MgATP from an equimolar solution of Mg+f and ITP.
The MgATP stock solution was freed of ADP by incubat,ion for 15 min with 80 pmoles of ATP, pH 7.4, 80 pmoles of XIgSO+ 2 pmoles of phosphoenolpyruvate, pH 7.4, and 100 pg of pyruvate kinase in a final volume of 2 ml. The pH was unchanged at the end of the reaction.
The final concentration of MgATP was 0.04 M. The stock solution contained a small amount of pyruvate which caused a small initial jump in the curve representing ATPase activity, but this did not interfere with accurate measurement of initial rates. All kinetic measurements were carried out at 23" in a Perkin-Elmer model 356 dual wave length spectrophotom-eter. The oxidation of NADH was monitored by recording the difference in absorbance at 340 versus 374 nm. Other conditions of each experiment are given in the legends to figures and tables.
In separate experiments, it was established that aurovertin dicl not inhibit either pyruvate kinase or lactate dehydrogenase activity.
I'luorometric Jieasureme?zfs-Fluorescence emission spectra were obtained in the instruments described previously (12). Aurovertin binding measurements were carried out in eit,her of two instruments as described (12). The filters used to isolate specific portions of the spectrum are mentioned in t,he legends to figures and tables. Fluorescence polarization measurements were made with a modified version of the spectrofluorometer described earlier (12). Right angle optics were used and filters replaced the emission monochromator.
Both polarizer and analyzer consisted of formula 105 UV polarizing filters (Polacoat Incorporated, Cincinnati, Ohio). The polarizer was mounted between the excitation monochromator and the excitation cutoff filter (CS-'i-60, Corning Glass Works).
The polarized fluorescent light passed through a 2-mm column of 2 hr NaNO? and then through a Corning Glass CS-3-71 filter before reaching the analyzer.
Polarization measurements with submitochondrial particles included in addition a Wratten 34.2 filter on the emission side of the instrument.
Measurement of polarization was made by manual rotation of the analyzer and polarizer, and the data were corrected for instrumental error as described by Chen and Bowman (13). The uncertainty in the polarization measurements was within ~~0.003 unit.
The temperature of the sample cell was controlled by circulation of water from a thermostatted water bath through channels in the sample holder.
The temperature of the cell contents was monitored with a small thermistor probe immersed in the liquid column to a point just above the exciting light beam.
The energized fluorescence response of aurovertin in the presence of ETPH was measured as described earlier for other fluorescent probes (12). The filter fluorometer with front surface optics was used with filters given in the legends to appropriate figures.
Measurement of absorption spectra were carried out in a Cary model 14 spectrophotomcter fitted with the 0 to 0.1 absorbance slide-wire.
The possibility that fluorescent light emitted by the sample might be sensed by the photomultiplier tube was minimized because of the distance of the cuvctte from the phototube and the use of low concentrations of aurovertin.
Measurement of Protein Concentration and Definition of Units-The concentration of mitochondrial proteins was measured by a modified biuret reaction (11). The concentration of solutions of F1 was determined by refractometry (14). One unit of ATPase activity is defined as that amount of enzyme which catalyzes the turnover of 1 pmole of substrate per min under the specified conditions of assay. Specific activity is expressed as units per mg of protein.
The fluorescence increment is defined as the fluorescence intensity, in arbitrary units, when 1 nmole of aurovertin is bound to F1.

RESULTS
The absorption spectrum of aurovertin in Buffer I is showy in CuWe 2, fluorescence emission spectrum of aurovertin (2 PM) in glycerol at 6.3". This spectrum was obtained at higher amplifier gain than the spectrum of aurovertin-Fr.
The data were normalized to fit the same axes as Curve 1.
(9). Addition of FI to the aurovertin solution did not affect the position of the maxima or the absorption at 270 nm but led to a 15% decrease in absorption at 365 nm (Fig. IA, Curve 2). It may be calculated from data presented elsewhere in this paper that only 50% of the aurovertin was bound to Fr under the conditions of the measurement.
The decrease in absorption would be 30~~ if all of the inhibitor were bound. Measurements of aurovertin binding to F, described in this paper were based upon fluorometric detection. It would appear, however, that measurement based upon absorption would also be useful but less sensitive.
Although low concentrations of aurovertin were virtually nonfluorescent, addition of the inhibitor to solutions containing F1 resulted in a large enhancement of fluorescence as pointed out by Lardy (5). The fluorescence emission spectrum of the aurovertin-F1 complex is shown in Fig. lB, Curve 1. An emission maximum was found at 469 nm accompanied by a shoulder near 490 nm. A 55-fold enhancement of fluorescence was observed when all of the aurovertin in solution was bound to Fr. The very low fluorescence of micromolar solutions of aurovertin in aqueous buffers was virtually unaffected if the antibiotic was dissolved in solvents such as methanol, ethanol, or dioxane.
Reduction of the oxygen content of the organic solvents by flushing with nitrogen resulted in detectable but minor increases in fluorescence intensit,y.
However, the intensity of fluorescence of aurovertin was increased a-fold in glycerol at room temperature. Cooling of the glycerol solution of aurovertin to 5" resulted in a 16.fold enhancement of fluorescence relative to that in Buffer I. It may be seen in Fig. lB, Curve 2, that the fluorescence emission spectrum of aurovertin in glycerol was similar to that of the aurovertin-F1 complex.
The former was, however, shifted slightly toward the ultraviolet portion of the spectrum with emission maxima occurring at 465 and about 483 nm, respectively.
Because of its large extinction coefficient and low fluores- cence, no attempt was made to obtain the emission spectrum of aurovertin in dilute buffer.
Interaction of F1 with Aurovertin-The nature of the interaction of aurovertin with Fi, monitored by changes in fluorescence intensity, was complex and influenced by adenine nucleotides and Mg++.
In the presence of Buffer I containing ATP, about 80% of the final maximum fluorescence was achieved within the first 4 s after the addition of aurovertin (Fig. 2, Curve C). However, when ATP was replaced by ADP, only about 25% of the final maximum fluorescence was reached during the same time period (Fig. 2, Curve B). In the absence of adenine nucleotides, the rate of the reaction was slower (Fig. 2, Curve A).
No further changes in fluorescence intensity were observed 5 min after the addition of aurovertin.
A second series of experiments was carried out to test the effects of addition of adenine nucleotides or Mg++ on the fluorescence intensity of the aurovertin-F1 complex. The first arrows in the curves of Fig. 3 indicate the addition of 1.1 pM aurovertin to a solution of F1 in Buffer I. About 2 min later, when the intensity of fluorescence had reached a plateau, 1.6 mM ATP was added (Curve A).
The small transient increase in fluorescence was followed by a slower quenching which reached a new plateau in about 2 to 3 min. Similar observations were reported by Lin (15). In separate experiments, not described here, it was found that the fluorescence emission maxima of the aurovertin-F1 complex in Buffer I containing ATP occurred at the same wave lengths as those shown in Buffer I alone (cf. Fig. 1B). The quenching by ATP was almost completely reversed by subsequent addition of ADP (Fig. 3). Addition of ADP to the aurovertinF1 complex (Fig. 3, Curve B) led to a rapid enhancement of fluorescence that was reversed by addition of ATP. However, the extent of reversal of either ATP-produced quenching or ADP-produced enhancement depended on the ratio of ATP or ADP in the reaction mixture.% The addition of Mg++ in excess of the EDTA in the mixture caused a rapid quenching (Fig. 3, Curve C) which was slowly reversed by the addition of EDTA.
The rate of onset of the changes induced by ADP or Mg++ was too rapid to be resolved by the mixing time of the apparatus used (about 0.5 s). The third set of arrows marks the addition of: Curve A, 2.7 pmoles of ADP; Curve B, 10 pmoles of EDTA; Curve C, 4.8 /*moles of ATP.
Measurement of fluorescence and additions to the cuvette were made as described in the legend to Fig. 2.

ATP-induced
quenching such as that shown in Fig. 3 did not appear to be due to the presence in the ATP preparations of divalent metal ions (16), since similar effects were observed when metal-free preparations of ATP were used. aIoreover, the effects of ATP or ADP were not reversed by EDTA.
The apparent competitive interaction between ATP and ADP may be a reflection of the fact that ATP can displace bound ADP from the enzyme (17). ITP, which is a substrate for Fr, and IDP which does not inhibit the enzyme (II), induced changes in the fluorescence of the aurovertin-Fi complex which were similar to those caused by adenine nucleotides although higher concentrations were required.
The magnitude of the effect was dependent on the ratio of aurovertin to Fi in the solution.
ATP quenching was maximal at low aurovertin to Fr ratios while ADP enhancement was minimal under these conditions (Table I). The brief enhancement of fluorescence which immediately preceded the quenching induced by ATP addition (Fig. 3) was proportional to the measured content of ADP in the ATP preparations used in these experiments.
The amount of ADP (0.05 pmole) added with the usual addition of 4.8 pmoles of ATP would have been enough to account for the observed enhancement.
Prior incubation of the ATP preparations with phosphoenolpyruvate and pyruvate kinase, which converted all free ADP to ATP, eliminated this phenomenon.

Binding of Aurovertin
to Fr-Titrations of Fi with aurovertin were carried out as described under "Materials and Methods." A typical plot of the reciprocal of the observed fluorescence versus the reciprocal of the protein concentration is shown in Fig. 4. Extrapolation to infinite protein concentration of a line fitted to the points by least squares analysis yielded the reciprocal of the fluorescence intensity expected when all of the aurovertin present in solution was bound to Fi. The fluorescence increment is defined as the extrapolated fluorescence intensity divided by the nanomoles of aurovertin present. It may be seen that the fluorescence increment when all aurovertin was bound would be 184 arbitrary units in Buffer I alone and slightly higher (198) in Buffer I containing ADP.
Values of fluorescence intensity obtained in this TABLE I Effect of adenine nucleotides and ratio of aurovertin to FI onjluorescence intensity of aurovertin-F1 complex Fluorescence intensity was measured as described in the legend to Fig. 2. F1, in amounts shown, was prepared in Buffer I. Further additions were made after the fluorescence of the aurovertin-Fi complex had stabilized (about 3 min). Percentage of quenching was obtained by multiplying 100 times the absolute difference between the fluorescence intensities before and after addition of nucleotides divided by the fluorescence before the addition of nucleotides. FIG. 4. The fluorescence intensity of aurovertin completely bound to F1. Each point on the graphs represents a separate experiment in which 0.36 nmole of aurovertin was added to the amount of enzyme protein shown. The value of the fluorescence intensity was recorded after the unchanging plateau value was reached as in Fig. 3. The reciprocal of the fluorescence intensity was then plotted versus the reciprocal of the protein concentration. The resulting data were treated by the method of least squares and plotted as shown. The buffers used were Curve A, buffer I; Curve B, Buffer I containing 4 mM ATP; Curve C, Buffer I containing 1 mM ADP.
way were used to calculate the binding ratios of aurovertin to Fi in the presence and absence of adenine nucleotides. containing Mg++. Lardy and Lin (5) also have observed a low dissociation constant for the aurovertin-Pi complex. In order to test the accuracy of the binding data, the experimentally determined binding ratios from several titrations were plotted versus the log of the free aurovertin concentration and compared to a curve fitted to the equation (18) nKAC fi=- I+ KAC where KA is the association constant for the formation of the inhibitor-F, complex, A is the average number of aurovertin molecules bound per molecule of Fi, n is the number of binding sites on each protein molecule, and C is the concentration of free aurovertin.
The theoretical curve was obtained by calculating fi (with the use of values of n and KA obtained in Plots A, C, and D of Fig. 5) at different concentrations, C, of free aurovert)in.
The close correspondence of the points obtained from three titrations in the presence of ATP with the theoretical curve is shown in Plof B of Fig. 5. Similar observations were made from the data for tit'rations with ADP and in Buffer I without additions.
The fluorescence increment and the dissociation constant calculated for t,he aurovertin-Fi complex in these experiments was influenced by the manner in which the enzyme was prepared. The dissociation constant was lowest (0.015 pM) and the fluorescence increment, highest when F1 was prepared in Buffer I and ADP was added just before the titration with aurovertin was begun. If the enzyme was prepared in the presence of ADP, It has been reported by Weber (19) that the study of polarization phenomena may provide an insight into the environment of a fluorophore, whether this be the solvent surrounding a molecule free in solution or a protein to which the fluorophore may be bound.
If the lifetime of the excited state is long relative to the time required for rotational diffusion (due to Brownian motion) then the resulting emitted fluorescent light will exhibit little or no polarization.
However, if the fluorophore is in a viscous solvent or attached to a large macromolecule which rotates more slowly in solution, appreciable polarization of fluorescence may be observed.
A summary of some measurements of the polarization of fluorescence of aurovertin, along with computed values of the viscosity of the solvent used, is given in Table II. The polarization of the free inhibitor in methanol or in Buffer I containing ATP was appreciable, 0.234 and 0.278, respectively. However, the observed polarization was increased considerably in solvents of high viscosity such as 2 M sucrose or 100% glycerol at low temperatures.
The polarization values recorded were affected little if at all over the range of aurovertin concentration tested. A considerable increase in polarization also was detected when aurovertin was bound to Fr. The average value of 0.375 at 25" was similar to that observed in solvents of high viscosity.
The possibility was investigated that changes in polarization of fluorescence might accompany changes in fluorescence intensity following the addition of adenine nucleotides or Mg++.
The results shown in Table III indicate that both ATP and ADP caused a small increase in polarization whereas addition of Mg+f caused a decrease. The magnitude of the observed changes was 3 to 8 times greater than the uncertainty in the measurements (kO.003).
In each case measurements were made immediately before and after the additions shown, and the effects were observed repeatedly with a number of different enzyme preparations.
Interaction oj Burovertin with Submitochondrial Particles-As by guest on September 22, 2017 http://www.jbc.org/ Downloaded from shown in Fig. 6, addition of aurovertin to a suspension of ETPH resulted in a large fluorescence enhakement. Subsequent addition of ATP or Mg++ led to a quenching of fluorescence whereas addition of ADP caused an enhancement of fluorescence.
The quenching of ATP was reversed by addition of ADP and the observed enhancement by ADP was reversed by addition of ATP in a manner similar to that described for soluble Fr (Fig. 3). Although the responses to adenine nucleotides and Mg++ are thus qualitatively similar to that of the aurovertinF1 complex, the rate of the response was more rapid for the soluble enzyme (cf. Fig. 3).
It should be pointed out that quenching by Mg++ appeared to depend on the Mg++ content of the particles.
It was necessary to preincubate the particles with a chelating agent for 10 min before addition of aurovertin in order to demonstrate clearly The uarticles were suspended in 3.0 ml of Buffer I at a concentration of 0.37 mg per r&l and incubated for 10 min. A large increase in fluorescence followed the addition of 0.88 nmole of aurovertin (j&t arrow). After the increase in fluorescence intensity had slowed or stopped, 3.3 mu ATP, 4 mu MgSO4, or 1.5 mu ADP was added as shown. Conditions of the measurement were the same as used in Fig. 2 except that the trace was interrupted when additions were made to the cuvette. the quenching observed in Fig. 6. EDTA could be replaced by sodium tripolyphosphate but not by ethylene glycol bis@-aminoethyl ether)-N,N'-tetraacetate which forms a complex with Mg++ characterized by a significantly lower stability constant. An examination of the polarization of fluorescence of the complex formed between aurovertin and ETPH was undertaken in an effort to compare the values found with soluble F1 with those for the membrane-bound enzyme.
In view of the likelihood that the rotational mobility of membrane-bound F, would be hindered relative to that of the soluble enzyme, it might be anticipated that an increase in the polarization of fluorescence of aurovertin would occur in this instance.
However, the protein concentrations required t.o produce a fluorescence signal of sufficient intensity also introduced light-scattering artifacts which led to an apparent depolarization.
A plot of the reciprocal of the observed polarization versus the concentration of particle protein is shown in Fig. 7. The increased polarization values found at lower protein concentrations would be expected if light scattering were in fact interfering with the measurements.
Extrapolation yielded a polarization of 0.356 at zero protein concentration.
Effect of Energization on Fluorescence of Aurovertin-The addition of succinate to an incubation mixture containing ETPH and aurovertin led to an enhancement of fluorescence (Fig. 8, Curve A). Energy conservation apparently was not affected by the concentrations of aurovertin used since the energized fluorescence of l-anilinonaphthalene-8-sulfonate (12, 22) was not influenced under these conditions.2 Quenching of the fluorescence of aurovertin followed the onset of anaerobiosis (Fig. 8, Curves A  and B) or the addition of carbonyl cyanide p-trifluoromethoxyphenylhydrazone (not shown). Whereas the energized fluorescence response of l-anilinonaphthalene-8sulfonate was enhanced by rutamycin (22), that of aurovertin was prevented by prior incubation with rutamycin or carbonyl cyanide p-trifluoromethoxyphenylhydrazone (Fig. 8, Curve C). A quantitative difference in aurovertin fluorescence was observed when EDTA replaced Mg+f in the reaction mixture (Curve B). It may be seen that in this case both the fluorescence enhancement following the addition of aurovertin and the response to succinate was markedly increased in the presence of EDTA.
Moreover, only a small quenching was observed following anaerobiosis. Inhibition of ATPase Activity by Aurovertin-In order to correlate binding of aurovertin with inhibition, the time course of ATPase activity was compared with the time course of the development of fluorescence as illustrated in Fig. 2. In Fig. 9 Curve A is a control showing the rate of ATP hydrolysis. Fluorescence measurements were made in the filter ffuorometer with front surface optics and filters described in the legend to Fig. 2. In addition, a Wratten number 55 alter was placed in the emitted fluorescent light path. The reaction mixture contained, in a final volume of 3.0 ml, 1.2 mg of ETPH and 20 mu potassium phosphate buffer, pH 8.0. The experiment represented by Curve A contained in addition 2 mM MgSOh while that of Curve B contained 2 mu EDTA.
In Curve C, the particles were pretreated with either 8 rg of rutamycin or 2 PM carbonyl cyanide p-trifluoromethoxyphenylhydrazone.
At the$rst arrows (Point I), 2.2 nmoles of aurovertin were added. When the fluorescent signal reached a final value, the position of the recorder pen was readjusted with the zero control to establish a new base-line (Point 2). Succinate (8.3 mu) was added as shown.
nmole of aurovertin, respectively, were added to the reaction mixture 15 s after the addition of substrate.
It may be seen that there was a lag in the onset of inhibition unless the inhibitor was incubated with F1 for 2 min prior to the addition of substrate.
Lineweaver-Burk plots (23) indicated that the inhibition was uncompetitive with ATP ( Fig. 10). Whereas the inhibition by aurovertin of the ATPase activity of submitochondrial particles was reported to be different in Tris versus phosphate buffer (24), the inhibition curves obtained with the soluble enzyme were the same in either buffer. DISCUSSION A variety of observations suggest that the conformation of F1 in solution may be modified by interaction with adenine nucleotides or Mg++.
Thus these substances alter the fluorescence intensity, the dissociation constant, and the polarization of fluorescence of the aurovertin-F1 complex. Direct evidence for an ATP-induced structural change in FL was obtained from measurement of intrinsic viscosity (12) and the polarization of fluorescence of pyrenebutyric acid covalently linked to the enzyme. The rotational relaxation time of the conjugate was 750 ns in Buffer I which contained ATP and 2000 ns in Buffer I alone. 2 As prepared in this study, the ATPase activity of the Dowextreated enzyme preparations remained stable at room temperature for about 5 hours. The subsequent losses in activity were accompanied by the development of turbidity and, at high concentrations (5 to 10 mg per ml), precipitation of denatured protein. These time- of the velocity of ATP hydrolysis were carried out as described in the legend to Fig. 9. The experiments with the inhibitor were carried out by incubating the enzyme in the reaction mixture with 0.221 PM aurovertin for 2 min prior to the addition of substrate.
The molar ratio of Mg+* to ATP was kept fixed throughout the series of measurements. The reciprocal of the rate of DPNH oxidation (nanomoles per min) was plotted versus the reciprocal of the substrate concentration according to the procedure of Lineweaver and Burk (23). The final concentration of aurovertin was 0.22 PM in Curve A. Curve B, no inhibitor present.
observed in the polarization of fluorescence of the aurovertin-F1 complex in Buffer I as well as the finding that addition of ATP to the inhibitor-F1 complex in Buffer I frequently did not restore the polarization to values observed with enzyme samples which were prepared in the presence of ATP.
It would appear, therefore, that the effectiveness of ATP and ADP in restoring the original conformation of F1 decreased with the age of the preparation.
The changes in structure which did occur in the enzyme following removal or addition of adenine nucleotides were not accompanied by significant uncoiling of the protein since estimates of the molar ellipticity of the enzyme, based on measurements of circular dichroism at 220 nm, were identical in Buffer I alone and in Buffer I which contained adenine nucleotides or MgS04.2 The molar ellipticity of Fi also was unaffected by aurovertin. 2 Xechanism of Enhancement and Quenching of Fluorescence of Burovertin-It has been observed that a variety of dyes such as l-anilinonaphthalene-8-sulfonate and 2-p-toluidinylnaphthalene-6-sulfonate exhibit little or no fluorescence in water but fluoresce strongly when dissolved in nonpolar solvents (25,26) and when bound in apparently hydrophobic regions of some proteins (27). It seems unlikely that the enhanced fluorescence of the aurovertin-Fr complex was a reflection of "polarity" effects since the inhibitor exhibited about the same low fluorescence in nonpolar solvents as in aqueous buffers. It seems more likely that the fluorescence of the aurovertin-Fl complex reflected an increased rigidity imposed upon the fluorophore in the protein-binding site. This conclusion is supported by the finding that aurovertin fluorescence ( Fig. 1) and its fluorescence polarization (Table II) were enhanced considerably in solutions of high viscosit'y. The increased polarization of fluorescence of bound versus free aurovertin, 0.375 and 0.279, respectively, strengthens the suggestion of reduced mobility in the enzyme-binding site. Thus a reduction of collisional quenching by solvent molecules would appear to make an important contribution to the fluorescence enhancement of aurovertin bound to F1.
An alternate explanation is suggested by the work of Oster and Nishijima (28) who observed an increased quantum yield of fluorescence of auramine 0, a substituted diphenylmethane dye, when the compound was dissolved in solutions of high viscosity.
The author's interpretation of their data, that the fluorescence intensity of auramine 0 depended on the rotational diffusion constant of the rotating phenyl groups of the molecule, might equally well apply to aurovertin if this compound also contained light-emitting groups capable of internal rotation. The quenching of fluorescence brought about by addition of ATP to the aurovertin-FL complex was, at least in some instances, partly due to the release of aurovertin from the enzyme.
It may be calculated from Table I, line 5, that 0.194 nmole of aurovert.in was bound to F1 in the absence and 0.127 nmole in the presence of ATP.
In this case the molar ratio of aurovertin to F1 was 0.2. When the molar ratio was 6 ( Table I,  However, most of the fluorescence quenching by ATP may be ascribed to the decrease in the fluorescence increment (Fig. 4). The enhancement of fluorescence by ADP may be explained by the observation that the fluorescence increment increased without change in the stoichiometry of binding. Similarly, the remarkable quenching by Mg++, which also occurred in the absence of a change in the amount of aurovertin bound to Fi, appeared to be due to a a-fold decrease in the fluorescence increment.
In separate experiments, not shown here, it was found that Mg++ did not quench the fluorescence of aurovertin in viscous solvents such as 2 M sucrose. Direct interaction of Mg++ with aurovertin on the enzyme would thus seem to be an unlikely explanation of the quenching. However, it is not ruled out that such an interaction may be promoted in the enzyme-binding site. It seems reasonable to conclude that the conformational changes which occur in Fi alter the aurovertin-binding sites and give rise to the observed changes in fluorescence.
Thus auro-vertin would appear to be a sensitive probe of conformational changes in the enzyme. Mechanism of Inhibition of ATPase Activity by Aurovertin-The finding that aurovertin inhibition of ATP hydrolysis catalyzed by F1 was uncompetitive agreed with previous observations of Lin (15). Lardy et al. (1) also found that the inhibition of mitochondrial ATPase by aurovertin was uncompetitive. Although two binding sites for aurovertin were available on Fi in the presence of ATP (Fig. 5A), only one of these sites appeared to be the inhibitory site since the reciprocal of the inhibited rate of the ATPase was linearly related to the aurovertin concentration?
Although other interpretations are possible, it may be expected that a linear relation would be obtained only with the square of the aurovertin concentration if inhibition required occupation of both sites (29). It is not entirely clear which of the two sites might be inhibitory.
The experiment illustrated in Fig. 9 suggested that the single site available for aurovertin in Buffer I was the inhibitory site since preincubation of the enzyme with aurovertin prior to the addition of substrate led to an inhibited initial reaction rate whereas a lag was observed in the onset of inhibition when aurovertin was added to an ongoing reaction.
This lag appeared to be of about the same order as the time required for formation of the aurovertin-Fi fluorescent complex in Buffer I containing ATP.
However, a rapid rearrangement of bound aurovertin to an inhibitory site upon the addition of ATP cannot at present be ruled out. The possibility that of the two binding sites for aurovertin on Fr only one site participates in the inhibition of ATPase activity may serve to explain observations that the forward reactions of oxidative phosphorylation (for example, ATP synthesis) are much more sensitive to aurovertin inhibition than the reverse reactions such as the ATPsupported reduction of DPN+ by succinate and the ATPase of submitochondrial particles (2,3). It may be that the second binding site for aurovertin on F1 participates more directly in the inhibition of forward reactions of oxidative phosphorylation. The possibility of two apparently unidirectional catalytic sites on Fi is consistent with the findings of Asami et al. (30) that the protein inhibitor of Fi (31) inhibited the reverse reactions of oxidative phosphorylation but, as shown by Pullman and Monroy (31), did not inhibit the forward reaction.
Interaction of Aurovertin with Submitochondrial Particles-The fluorescence responses of aurovertin on the mitochondrial membrane may possibly be more readily interpreted than those of lanilinonaphthalene-8sulfonate and other dyes since it seems probable that aurovertin is localized exclusively on Fi. Lin has noted little enhancement of fluorescence in the presence of submitochondrial particles treated so as to remove most or all of the membrane-bound F1 (15). Lardy and Lin (5) have noted that F1 was the only protein in mitochondria that enhanced the fluorescence of aurovertin.
It was also reported in this paper that ATP, XDP, and Mg+f caused changes in the fluorescence intensity of aurovertin bound to ETPH which were similar to those wit,h F1. Thus the available evidence supports the suggestion that aurovertin forms a specific complex with Fi on the mitochondrial membrane.
The inhibition by rutamycin of the succinate-induced energized response of aurovertin was consistent with the suggestion of Lee and Ernster (3) and Roberton et al. (4) that aurovertin acts on the ATP side of the rutamycin-sensitive site. Chance et al. (32) have reported no change in the fluorescence intensity of aurovertin bound to submitochondrial particles that was related to turnover of the respiratory chain. However, since these experiments were carried out in the presence of ruta-by guest on September 22, 2017 http://www.jbc.org/ Downloaded from mycin, energized responses of the kind reported in this paper would not have been observed.
The changes in the fluorescence intensit,y of the aurovertin-ETPH corn& which accompany the appearance of the energized state may well reflect an altered conformational state of membrane-bound Fr in a manner analogous to the alterations suggested above for the soluble Fl-inhibitor complex.
The ambiguities introduced into the measurement of polarization by scattered light (33), which was considerable in suspensions of submitochondrial particles (Fig. 7) precluded, in our view, a search for the effects of adenine nucleotides or Mg++ on the polarization of fluorescence of the membrane-bound complex which might parallel the effects observed with soluble Fi. It may well be, however, that measurement of pulsed anisotropy of emission (34,35) will provide a useful approach to this type of study of conformational change in F1 on the mitochondrial membrane.
That a change in the conformation of a coupling factor may occur during the energized state is suggested by the tritium exchange experiments of Ryrie and Jagendorf (36) with the coupling factor from spinach chloroplasts.
In addition, the suggestions of Boyer (37) and the work of Hackenbrock (38), Green and Ji (39) and others, predict conformational changes in t,he mitochondrial membrane in the energized state. It was pointed out above, however, that at least part of the observed fluorescence quenching, following for example the addition of ATP, may be ascribed to dissociation of the fluorescent aurovertin-Fi complex. Moreover, the experiment illustrated in Fig. 8 suggests that Mg++ is intimately involved in the energized fluorescence response of aurovertin on submitochondrial particles. It may be that Mg++ is released from membrane-bound Fi during energization, thus permitting an enhancement of fluorescence by the mechanism suggested above. Following anaerobiosis, Mg++ reattaches to F1 if the metal is available in solution and quenching ensues. If Mg++ is not available, for example in the presence of a chelator, little quenching occurs.
Based upon studies carried out with soluble Fi, the changes in aurovertin fluorescence during energization of ETPH would thus appear to be due to a change in conformation of the bound enzyme which alters the aurovertin-binding sites on the molecule. The resulting alterations might then cause quenching or enhancement as discussed above, or alternatively, might promote dissociation of the fluorescent complex.