AN AFFINITY REAGENT FOR DEMONSTRATING THE PRESENCE ...

Vol. 262,No. 11,Issue of April 15, pp. 5145-5150,1987 Printed in U.S.A.

THEJOURNAL OF BIOLOGICAL CHEMISTRY

0 1987 by The American Society of Biological Chemists, Inc.

AN AFFINITY REAGENT FOR DEMONSTRATING THE PRESENCEOF Tyr-@311AT THE HYDROLYTIC SITE OF F,-ATPase* (Received for publication, November 10,1986)

Joe C. Wu, Hua Chuan, and JuiH. Wang From the Bioenergetics Laboratory, Acheson Hall, State University of New York, ~ u f fNew ~ ~York , 14214-3094

The compound P1-(5’-adenosyl)-P2-N-(2-mercap-vates the ATPase. It is also possible that Tyr-8’311 is quite toethy1)diphosphoramidate(AMEDA) was synthesized far from the hydrolytic site, but that its labeling induces a as an ATP analogue for in situ reaction with the 4- long-range protein conformation change which inactivates the nitr0-2,1,3-[~~C]benzoxadiazolyl group (NBD) in the enzyme. labeled F1-ATPase (F,). AMEDA was found to reactiIf the phenolic group of Tyr-8‘311 is indeed near the vate O-[14C]NBD-Flvia a dual-path mechanism. The phosphate groups of ATP bound at the hydrolytic site and principal pathinvolves the binding of AMEDA at a site participates directly in thecatalysis (5),it should be possible in F1 with Kd= 14.5 PM and subsequent reaction with to design and synthesize an affinity reactivator for O-NBDthe [14C]NBDlabel. The second slower path involves mitochondrial MF, by replacing the y-phosphate group of the direct biomolecular reaction of AMEDA with the ATP with a suitably positioned m e r c a p ~ ngroup. Such a radioactive label on Fr. The rate of reactivation of 0molecule would be bound in a substrate-like manner to the [14C]NBD-F1byAMEDA was decreased byADP or ATP which competes with theATP analogue for bind- hydrolytic site with its sulfhydryl group placed close to the0ing to the labeled enzyme. The reaction product was NBD label to facilitate the in situ reaction between them. found to contain one adenine group, two phosphate The present article reportsthe synthesis of P1-(5‘-adenosyl)groups, and one [“CINBD label per molecule as ex- P2-N-(2-mercaptoethyl)diphosphoramidate (AMEDA) and pected from the structure of the compound AMEDA- its use in demonstrating the presence of Tyr-8311 near the [“CINBD. Purified AMEDA-[“C]NBD was found to phosphate groups of ATP bound at the hydrolytic site of bind to unlabeled F1 with & = 2 PM. These observations mitochondrial F1. demonstrate the in situreaction of bound AMEDA with EXPERIMENTAL PROCEDURES the nearby [“‘CINBD label attached to Tyr-831 1 and support the assumed presence of Tyr-8311 near the Materiakr phosphate groups of ATP bound at the hydrolytic site ADP (lithium salt), ATP (disodium salt), NADH, NBD-Cl, Pof F,-ATPase. The possible locations of Tyr-8364, His8427, and Tyr-8345 relative to Tyr-8311 in F1 are enolpyruvate, pyruvate kinase, lactic dehydrogenase, morpholine, Hepes, cystamine dihydrochloride, N,~-dicyclohexylcarbodiimide, discussed, and the validity of the previously proposed and were purchased from Sigma. [“CINBD-Cl was obtained model for Fl-ATPase with one hydrolytic site assisted fromDTT Research Chemicals International and was found to have a by two auxiliarysites is examined and compared with specific radioactivity of 93 k 1 mCi/mmol. F,-ATPase was prepared that of the widely accepted alternating sitesmodel. from bovine heart mitochondria (6,7). M e ~ ~ d s Assay of ATPase Activity-The catalytic activities of NBD-labeled Mitochondrial F1-ATPase (FJ1 can be labeled very specif- and control F,-ATPases were determined by coupled oxidation of ically (1)at Tyr-8311 (2,3) by 7-chloro-4-nitro-2,1,3-benzox-NADH in a medium containing 50 mM Hepes/NaOH, pH 8.0,3 mM adiazole (NBD-CI). Each NBD label on Tyr-311 of the hydro- MgC12,50 mM KCl, 2 mM ATP, 2 mM P-enolpyruvate, 0.4 mM NADH, 21 units/ml pyruvate kinase, and 11 units/ml L-lactic-acid dehydrolytic @’ subunit of mitochondrial F1 completely inhibits its genase. The steady-state rate of ATP hydrolysis was computed from ATPase activity (4). The inhibition could be because Tyr- the observed linear decrease of AM due to the coupled oxidation of fi‘311 is at the hydrolytic site, and hence, its labeling inacti- NADH at 30 “C using 6220 as the molar absorbance of NADH. Protein concentrationswere determined by the Coomassie Bluebind* This work was supported in part by Research Grant GM 31463 ing method (8). Radioactivity was assayed by liquid scintillation from the National Institute of General Medical Sciences. The costs counting, with counting efficiency determined by a [“Cltoluene inof publication of this article were defrayed in part by the payment of ternal standard. page charges. This article must therefore be hereby marked ‘‘uduerSynthesis of AMEDA-The synthesis was started in anhydrous tisement” in accordance with 18U.S.C. Section 1734 solely to indicate pyridine solution containing ~,~-dicyclohexylc~bodiimide (9-11) this fact. by reacting the dilithium salt of ADP with cystamine at room temI The abbreviations used are: F1, F,-ATPase; AMEDA, P1-(5’- perature. Cystamine dihydr~hloride was dissolved in absolute adenosyl)-PZ-N-(2-mercaptoethyl)diphosphoramidate; AMP-PNP, ethanol and stirred with an excess of solid anhydrous sodium carbon&-y-imidoadenosine 5”triphosphate; DTT, DL-dithiothreito~FSBA, ate for 8 h. The mixture was filtered, and the clear filtrate was 5’-p-fluorosuifonylbenzoyladenosine;FSBI, 5’-p-fluorosulfonylben- evaporated in vacuum to recover the cystamine-free base, m.p. 78zoylinosine;Hepes, N-2-hydroxyethylpiper~ine-~-2-ethanesulfonic80 “C. acid; n, molar ratio of label to F,; NBD-Cl, 7-chIoro-4-nitro-2,1,3A solution containing 0.4 mmol of N,N‘-dicyclohexylcarbodiimide benzoxadiazole; NBD-FI, F, labeled with NBD-CI at its essential in 5 ml of anhydrous pyridine was added dropwise to a stirred solution tyrosine residue; r, ratio of the specific activity of the labeled enzyme containing 0.08 mmol of the dilithium salt of ADP and 0.8 mmol of to that of the unlabeled enzyme; HPLC, high performance liquid cystamine in 15 ml of anhydrous pyridine at room temperature over chromatography; TNP-ATP, 2‘(3‘)-0-(2,4,6-trinitrophenyl)adeno- a 6-h period. The mixture was stirred for 5 days at room temperature sine 5”triphosphate. (23‘C).The progress of reaction was monitored by taking small

5145

Hydrolytic Site of Fl-ATPase

5146

aliquots at intervals and analyzingthe codtent of each by thin layer chromatography on precoated TLCplates (Silica Gel F-254,analytical, EM reagents, or Si500F, preparative, J. T. Baker Chemical Co.). The developing solvent was ethanol, 1 M ammonium acetate at pH 7.5 (100:40). After 5 days, ADP (RF = 0.11) in the mixture was no longer detectable on the TLC plate. The mixture was concentrated to 2 ml by lyophilization and appliedto a 40 X I-cm chromatogmphic column packed with Silica Gel (J. T. Baker Chemical Co., 60-200 mesh), and the condensation product(RF= 0.38) of ADP and cystamine was eluted with the same solvent. Based on the observed Ass due to the adenine group and A m due to the disulfide linkage, the yield of this condensation productwas estimated to be about 70%. This eluted fraction was reduced with sodium borohydride (0.068 mmol) at 0 ‘C for 1h and then stirred at room temperature overnight. The mixture was lyophilized.The residue which containedthe productAMEDA, cysteamine, and salts wasdissolvedin N,-saturated water. The mixture was applied to a 20 X 0.8-cm cation exchanger column(AmberliteIR-120,sulfonatedpolystyrene, H+ form) and eluted with Nz-saturated water. Protonated cysteamine and other cations were retained by the column. The eluate containing AMEDA was neutralized and kept frozen at -20 “C. The overallyield of AMEDA was 6 ~ 9 % . Anulysis of AMEDA-The concentration of adenosine group in the AMEDA sample was determined by measurement of its absorbance at 259 nm, using 15.4 X lo3 as the molar extinction coefficient of adenosine. The concentration of sulfhydryl group in the sample was determined by first reacting it with an excess of 5,5’-dithiobis-(2nitrobenzoic acid) at p H 7.0 and then measuring its absorbance change at 412 nm using 6.5 X lo3 as the molar extinction coefficient of 5 mercapto-2-nitrobenzoate(12). The concentration of total organicphosphatewasdetermined by the method of Fiske and SubbaRow (see Ref. 13). The molar ratios of adenosine to phosphate group to sulfhydryl group determined by these methods were 1:2.00.0.96 for the stored frozensampleofAMEDA and 1:2.001.01 for the freshlyreduced sample. P r e ~ r of~O~- nf ~ 4 C ~ ~ ~ ~ labeling - ~ ~ - of T F, h eby [“CINBDC1 was conducted in Buffer A (50 mM Hepes/NaOH, pH 7.0, 2 mM EDTA,25%glycerol) at room temperaturein the dark (4). The reaction was usuallyterminatedby centrifugalgel filtration (14) when or beforethe label:F1 molarratio reached1.The labeled F, was stored under liquid nitrogen. Measurement of Binding by U~~rufiltru~ion-Equilib~um binding of the reactionproductAMEDA-[14C]NBDbyunlabeled F, was measured by u ~ t r ~ l t r a ~ through jon a Paulus cell (15) fitted with a 13-mmpolysulfonemembrane (MilliporeCorp.,PTGC01310, M, 10,000 limit). Aliquotsof unlabeled F, were incubated with different concentrat~onsof AMEDA-[’~C]NBD before each was addedto the upper chamber. A pressure of about 20 p s i . of Nz was applied to the upperchamber; and after 10min, the concentration of unbound AMEDA-~14C~NBD in the lower chamberwas determinedby injecting 10 pl of the filtrate into a scintillation vial containing 5 ml of Aquasol I1 (New England Nuclear). The concentration of bound AMEDA[‘*C]NBD was obtained as the difference between the total concentration of the radioactive ligand beforeultrafiltration and the concentration of free radioactive ligandin the filtrate. RESULTS

OL

100

ZOO Min

300

40(

FIG. 1. Effect of ADP on removal of NBD label from 0NBD-FI by AMEDA. Each enzyme sample, containing 1.76mg of protein/ml(5 PM) in Buffer A, waspreincubatedat room temperature in the dark with ADP at the concentrations indicated. After 5 min, AMEDA (Fmal concentration 5 phi; r = 0.09) was introduced to the sample, and the mixture was vigorously shaken. A t each indicated time, a lO-~laliquot of the reaction mixturewas injected into 2 ml of ATPase assay medium withoutM$+ to stop the reaction by dilution. The ATPase activity was measured immediately after the addition of MgClz (final concentration4 mM).

-

K = (C-~-Y)(a-x-y)_C(a-x-y) Y

Y

(2)

and theobserved reaction rate is given by dx

= kzC(a - x dt

- y) + kly =

Integration of Equation 3 gives

+

+

where r = x / a and m = (C/( C K ) ) (k,K kl). The observed specific ATPase activities of the labeled enzyme a t different reaction time t over a range of AMEDA concentration in the absenceof added ADP are summarized in Fig. 2 ( A and B ) . The Iinear plots in Fig. 2B support the above dual-path mechanism. For each reagent concentration C, we obtain an observed value of m from the corresponding slope in Fig. 2B. Using any pair of concentrations, say C, and C5, with the correwe obtain the following. sponding slopes m3 and m5, ms= 5 C3 + K ma (c,C, K I c3

-

R - C3Csms - C & m Csm3 - C3ms

(5)

By choosing the five well-spaced pairs of straight lines from

Reactivation of O-NBD-FI by AMEDA-In a solution conFig. 2B, i.e. (m7, md, (m7, m4), (me, m d , (ms, m d , and (m5, and 5 p~ AMEDA, the mz), taining 5 p~ O-([14C]NBD)0.~1-FI we obtained anaverage valueof I( = 14.5 rt 2.7 p ~Thus, . at 25 “C. AMEDA is bound to the NBD-labeled site in F, with fairly inhibited enzyme was rapidly reactivated in the dark Fig. 1 shows that this reactivation process was retarded by high affinity and can presumably reactin situ with thelabel. micromolar concentrations of ADP. In the absence of ADP, On the other hand, there isalso the bimolecular path ( k z ) the reaction between NBD-F, and AMEDA may be reprewhich does not require prior binding of the reactant. The sented by thefollowing dual-path scheme. b i m o l e c u l ~ p a t his directly demonstrated by the effect of ATP on the reaction rate shown in Fig. 3. The Iabeled site in NBD-F1 F1 was obviously already saturated by ATP at concentrations K k from 5 to 20 mM since the inhibited reaction ratewas found + AMEDANBD-F,-AMEDA 1 , F1 + AMEDA-NBD to be independent of ATP concentration. But the observed kz reactivation rate due to the bimolecular path is still quite of 8 p~ O-NBDLet ususe the following notation: at t = 0, [NBD-F1] = u and appreciable. As a control, duplicate solutions [AMEDA] = C; at time t, IF,] = [AMEDA-NBDJ = x and F1were treated with mercaptoethanol(HOCH&H2SH) either of 5 mM ATP. The results inFig. [NBD-F,-AMEDA] = y. The dissociation equilibrium con- in the absence or presence 4 show that the rate of removal of NBD labelby 8 p~ stant may be written as

Hydrolytic Site of Fl-ATPase

5147

c

l 20

0

1

2

3

4

I

Min I

40

Min

60

80

100

FIG. 4. Removal of NBD label from 0-NBD-F, by mercaptoethanol. Enzyme samples containing 2.78 mg of protein/ml(8 p ~ ) in Buffer A without (0,0, and A) orwith (0)5 mM ATP were incubated in the dark at 26 "C with micromolar concentrations of mercaptoethanol as indicated. At the times indicated,2-pl aliquots of each reaction mixture were assayed for specific ATPase activity both in the absence and presence of 2.5 mM D m . I

FIG. 2. Reaction of AMEDA with 0-NBD-F,. A , the reaction inBuffer A wasfollowedby monitoring the increase in ATPase activity. The final concentrations ofAMEDA for the numbered experimental curves are 2 p M (curue I ) , 4 p M (curue 2), 8 p M (curue 3), 12 p M (curue 4 ) , 19 p~ (curue 51, 39 p M (curue 6 ) , and 68 p~ (curue 7). For each experiment,a measured volume of AMEDAstock solution was injected into a 0.86 p~ enzyme solution in BufferA and at once vigorouslyshaken witha Vortex mixer.The recorded reaction time was started at 5 s after the beginning of the shaking. A t 30-5 intervals, 2 0 - 4 aliquots ofeach reactionmixture were taken for ATPase assay. Each 20-gl aliquot was diluted 100-fold by injection into 2 mlof assay medium without M F . The assay was started 5 Min min later by the addition of 4 mM MgCl,. B, plot of kinetic data in A according to Equation 4. The straight lines intersect at -0.5 min on FIG. 5. Isolation of reaction product AMEDA-["C]NBD by the time axis due to chemical change before the start of recorded HPLC. AMEDA-["C]NBD was prepared by reacting 46.7 p~ 0-["C] reaction time. The relative rate r represents the ratio of ATPase NBD-F, with 38.6 PM AMEDA in Buffer A at room temperature in activity of the enzyme in the absence of DTT to that in the presence the dark for 5 h. After the removal of the protein by centrifugal gel of 2.5 mM DTT. filtration through SephadexG-50-80, the product was eluted from the Sephadex with Hz0 at a flow rate of 0.1 ml/min. The eluate was concentrated by lyophilization. A 50-p1 sample of the concentrated solution was applied to a Radial-PAK NOVAPAKClacolumn (8 mm X 10 cm) which had been pre-equilibrated with 4% methanol, 96% H20 and eluted with the same solvent at a flow rate of 1 ml/ min. r

( n = 1-00, r = 0.03) for 5 h in the dark at pH 7.0, the protein was removed by centrifugal gel filtration through Sephadex G-50-80, andthereactionproducts of smaller molecular weights were eluted from thegel with water and lyophilized. T h e solid residue wasredissolved in a smaller volume of water, applied to a C18-HPLC column, and eluted with water containing 4% methanol. Fig. 5 shows that the elution profile 0 20 40 60 contained essentially only oneradioactive fraction. Min This radioactive fraction, which was eluted asa single peak FIG. 3. Effect of ATP on removal of NBD label from 0= 0.181 and a NBD-F, by AMEDA. Enzyme samples containing2.78 mg protein/ inthechromatogram, was foundtohave ml (8 p ~ in) Buffer A were preincubated without (0) or with 5, 10, radioactivity concentration of 6143 cpm/3 111, which corre15,and 20 mM ATP (0,m,A, and A,respectively) for 15 min at room sponds to [adenosine group] = 11.7 p~ and [NBD group] = temperature. The reaction was started by the addition of AMEDA 11.2 PM, respectively. These estimated concentrations give (final concentration 8 p M ) to each sample. At the times indicated, [NBD group]:[adenosine group] = 0.96, which is consistent aliquots of eachreactionmixturewerecentrifugallygel-filtered of the through Sephadex G-50-80, and the specific ATPase activity of the withoneNBDgroupattachedtothesulfuratom AMEDA moiety in each productmolecule as expected. filtrate was determined as described in the legend to Fig. 2. There is a small error in theabove computation due to the mercaptoethanol is unaffected by 5 mM ATP and that ther absorbance at 259 nm by the NBD group attached to the atom. Thiswas determined in the following way. A 70 uersus time plot is similar to that for the bimolecular path sulfur in the presence of saturating concentrations of ATP, as shown ~ L Mmitochondrial F1, 70 g M NBD-Cl, 2.5 mM D T T solution in BufferA was allowed to react in the dark 25at"Cuntil its in Fig. 3. A,,s reached the maximumvalue in about 3 h. Measurement Separation and Characterization of the Reaction ProductAfter AMEDA had reacted witha slight excess of 0-NBD-F, of absorbance a t 259 nm gave Az,, = 1.474, which was only

Hydrolytic Site of Fl-ATPase

5148

2.7% higher than the absorbance in the absence of NBD-Cl. After this 2.7% contribution by the NBD group was subtracted from the observed Azss of the isolated reaction product, the NBD:adenine molar ratio in the product molecule was found to be 0.98. Binding of the Reaction Product AMEDA-[T]NBD to Unlabeled K-ATPase-The equilibrium binding of the above purified reaction product was measured by ultrafiltration as described under "Methods." In order to avoid the unnecessary complication due to thepresence of multiple binding sites in F1, the total concentrations of F1 and AMEDA-['*C]NBD, respectively, were so chosen in these experiments that the molar ratio of bound AMEDA-[WJNBD to F, was always less than 1. In thisway, the average value of the dissociation constant for AMEDA-[l*C]NBDbound to thehydrolytic site of F, was found from four experiments to be Kd = 1.8 f 0.3 PM. DISCUSSION

Since NBD-Fl still binds AMP-PNP (23) and the NBD label has no effect on the furtherlabeling of F1 by FSBA(19) and FSBI(ZO),it has been pointed out repeatedly that perhaps the NBD-labeled Tyr-311 is quite far away from the nucleotide-binding site of the subunit and that the labeling could have induced a long-range protein conformation change which inactivated the enzyme. This possibility is clearly ruled out by the present results for the following reasons. AMEDA reactivates 0-NBD-F, by removing the NBD label via two reaction paths. The principal path involves the binding of AMEDA at a sitewith Kd = 14.5 p~ and subsequent reaction in situ with the NBD label to form AMEDA-NBD which is still bound to thesite with Kd = 2 pM, as illustrated in Fig. 6. Kinetic data show that this reaction path can be blocked by either ATP or ADP which competes with AMEDA for binding to 0-NBD-F,. The second path involves a simple bimolecular reaction of AMEDA with the NBD label at a rate which remained constant when the concentration of ATP in the medium was raised from 5 to 20 mM (Fig. 3). The nonaffinity reagent mercaptoethanol removes NBD label from O-NBDF1 only via the simple bimolecular path at a rate which is unaffected by the presence of 5 mM ATP (Fig. 4). In view of the structural similarity between AMEDA and ATP, the above observations may be regarded as compelling evidence for the presence of NBD-labeled Tyr-0311 near the phosphate groups of ATP bound at the hydrolytic site of F,-ATPase. It KM

T i 6 (AMEDA)

.:; ;.

iii_.

*.-..x '

lir..:

FIG. 6. I n situ removal of NBD label by bound AMEDA in FI-ATP-.

wouldhave been inconceivable for an AMEDAmolecule bound at the hydrolytic site to react with a faraway NBD label to form the product AMEDA-NBD and to be again bound to thesite with even higher affinity. Recent energy transfer measurements show that the fluorescent N-NBD label attached to Lys-0162 is about 25.6 A from the chromophore group of TNP-ATP bound at anactive site for ATP hydrolysis (16). Since the distance from the chromophore of bound TNP-ATP to its farthest oxygen atom is about 16 A, it seems quite possible for the €-amino group of Lys-0162 to be hydrogen-bonded to they-phosphate group of the bound substrate (17). Since N-NBD-Fl was prepared through adirect transfer of NBD label from Tyr-0311 to Lys162 in the dark at pH 9, it seems very likely that Tyr-B311 is also near the y-phosphate end of the bound ATP, as illustrated in Fig. 6. This could explain why NBD-Fl still binds ATP, why ATP does not completely protect F1 from labeling by NBD-Cl (18),and why the NBD label does not affect the labeling of F1 by 5'-FSBA (19) or FSBI (20). Bullough and Allison (20) showed that FSBA labels Tyr8368 or His-0427 (21), whereas FSBI labels Tyr-0345 (20). Also, the complete inactivation of ATPase activity of F, by FSBA requires modification of Tyr-0368 or His-0427 in all three copies of the 0 subunit, whereas the complete inactivation of ITPase activity by FSBA is accomplished by modification of these residues in only one copy of the 0 subunit. On the other hand, the complete inactivation of both ATPase and ITPase activity of F1 by FSBI requires modification of Tyr-0345 in only a single copy of the ,O subunit. The structural interpretation of these exciting discoveriesposes a challenging problem to investigators of this important enzyme. The observed specific labeling of Tyr-0311 by NBD-Cl, specific labeling of Tyr-0368 or His-0427 by FSBA, and specific labeling of Tyr-0345 by FSBI could be explained by assuming that these labeled residues were located at different nucleotide-binding sites or domains in the/3 subunit (20). But such an assumption would require the binding of more than one nucleotide analogue per 0 subunit, which is difficult to reconcile with the well-established fact that there are only five or six nucleotide-binding sites in the undissociated F1 molecule, and three of them are unavailable to affinity reagents because they are still occupied by tightly bound ATP or ADP even after gel filtration (22-26). A simpler and more realistic interpretation is possible if it can be assumed that each (3 subunit has only one nucleotidebinding site (with Kd in the submillimolar range) which is surrounded by all of the above labeled residues and that the mode of binding or labeling depends on the nature of the reagent. For NBD-Cl, which is not an affinity reagent, the highly specificlabeling of Tyr-/3311is probably due to greatly enhanced reactivity of this residue in the hydrolytic subunit toward certain reagents. The present resultsshow that the0NBD label is near the y-phosphate group of bound ATP. Although the affinity reagents FSBA and FSBI areboth expected to be bound at this nucleotide-binding site, the absence of atriphosphateordiphosphate group in these substrate analogues could make them bind in different conformations and orientationsfrom that of bound ATP or ADP. Furthermore, the replacement of adenine with its one primary amino group and adjacent tertiary amino group by inosine with its one carbonyl group and adjacent secondary amino group could also cause FSBA to be bound at the site in a different conformation and orientation from that of FSBI. In this way, FSBA and FSBI could label different residues at or near the same site. The observed variation in the number of covalent label per

Site

Hydrolytic

F, required for total inactivation of ATPase or ITPase activity by FSBA or FSBI could also be interpreted from the assumption of one nucleotide-binding site per B subunit, if we accept the notion of functional differentiation of p subunits due to subunit interaction in F,. Recent studies (27) show that the intrinsically identical @ subunits can be functionally distinct in F, due to subunit interaction: one p subunit, represented by p’ in Fig. 7, catalyzes the hydrolysis of ATP directly; whereas the other two, represented by p”, play auxiliary roles through interaction between the subunits. By attaching the NBD label exclusively to either Tyr-/3’311 or Tyr-P”311, two geometric isomers of labeled F, with contrasting different properties have been prepared (4).The difference in their properties has been shown by relabeling experiments to be indeed due to thelabeling of Tyr-311 in functionally distinct p subunits in F,, not due to different orientations of the NBD label attached to Tyr-311 of the same /3 subunit. According to the alternating sitesmodel of F,-ATPase, the rapid hydrolysis of an ATP molecule bound to an active site of F1 requires aprotein conformation change at the site coupled to ligand change at two other equivalent sites which alternate with the first (28-31). Labeling the essential Tyr311 at any one of the three alternating sites would block one of the consecutive steps in the catalytic cycle and thereby stop the steady-state hydrolysis of ATP. From the standpoint of this model, the observed inactivity of O-#?’-NBD-F,with a label:F1 molarratio of n = 1 is expected, but the observed full ATPase activity (relative to that of the control unlabeled enzyme) of O-@”-NBD-F,for steady-state ATP hydrolysis should not be possible. According to another previously proposed model (27), the rapid hydrolysis of an ATP molecule bound to thehydrolytic site in p’ also requires a protein conformation which is maintained by and coupled to the nature of ligands bound at the auxiliary p” sites. But unlike in the alternating sites model, the nucleotides bound at p‘ and p” sites in this model are not required to turnover at thesame rate during the steady-state hydrolysis of ATP. Direct labeling of Fl by NBD-Cl takes place predominantly at the8‘ site on account of the enhanced reactivity of its Tyr-311 phenolic group. Thus, it would also give an inactive O-P’-NBD-F, with n = 1. However, from the standpoint of this model, the observed full ATPase activity of O-P”-NBD-F1is also possible because the NBD label at p” site does not preventthe binding of adenine nucleotide at this auxiliary site for promoting steady-state hydrolysis of ATP at thehydrolytic site. With affinity reagents such as FSBA and FSBI, labeling at the p’ site and p” sites in F, could take place at comparable rates. But in this case, the SBA or SBI label will prevent the binding of nucleotides at the same site. In the case of inacti-

FIG.I . Functional differentiation of i.9 subunits in F, for catalytic hydrolysis due to contact interaction between specific amino acid residues of adjacent subunits. The gross (low resolution) 2-fold molecular axis of symmetry is represented by the horizontal line. b’ designates the hydrolytic j3 subunit; 6” designates the auxiliary3!, subunits. The circles represent the equivalent spheres of the respective subunits.

of Fl-ATPase

5149

vation of the ATPase activity of F, by FSBA, the labeling of hydrolytic p’ site should give a totally inactive O-p’-NBD-F,. The labeling of an auxiliary 8’’ site also might not reduce the observed rate of ATP hydrolysis because the structure of the SBA label at this site might not be sufficiently different from that of a bound ATP to cause a damagingly large structural perturbation. Consequently, when an effector ATP molecule from the assay medium is bound to the unlabeled 8’’ site, it could still convert the hydrolytic p’ site from inert “uni-site” conformation to promoted conformation for rapid steadystate hydrolysis of ATP. If all three p sites were labeled by FSBA at similar rates under the experimental conditions of Bullough and Allison, total inactivation of the ATPase would be expected to require 3 mol of SBA label per mol of F, as observed (21). But in the case of inactivation of the ITPase activity of F, by FSBA, the labeling of an auxiliary p” site could also lead to an essentially inactive enzyme because, when an effector ITP molecule from the assay medium is bound to the unlabeled p” site, structurally it might not fit the auxiliary site aswell as ATP, andhence its binding might not be sufficient to prevail over the perturbing effect of the SBA label at theother p” site for converting the /3’ site from uni-site to promoted conformation. Consequently, the steadystate hydrolysis of ITP could be inhibited by labeling either at the8‘ or p” site. If this was indeed the case, the inactivation of ITPase by FSBA would be faster than thatof ATPase, and the observed total inactivation of ITPase would require only 1 mol of SBA label per mol of F,. For the inactivation of F, by FSBI, the perturbing effect of the SBI label could be so damagingly large that the enzyme could be inactivated by labeling either at the 6’ or 8’’ site irrespective of whether ATP or ITP was used as the substrate. Consequently, the observed total inactivation of either ATPase or ITPase by FSBI would require only 1 mol of SBI label per mol of F, as observed (20). The above discussions are for catalytic hydrolysis only. Similar studies of the effect of selective labeling of the p’ and p” sites in F, on the efficiency of phosphorylation coupled to proton flux, which will be reported in a separate publication, show that the catalytic mechanism of F, for oxidative phosphorylation might not be the exact reversal of that for hydrolysis. The arrangement of the subunits of F, illustrated in Fig. 7 was suggested by the x-ray diffraction data andelectron microscopy results (32-35). By symmetry, the a‘ subunit in Fig. 7, the counterpart of p’ for a subunits, would beexpected to have some properties different from those of either a” subunits. Wise et al. (24) demonstrated convincingly that adenine nucleotides bound to the a subunits of F,do not participate in either catalytic hydrolysis or oxidative phosphorylation. Kironde and Cross (26) have recently discovered that the adenine nucleotide bound to one of the three noncatalytic sites in beef heart mitochondrial Fl exchanged rapidly with nucleotides in the medium, whereas those bound to the other two noncatalytic sites did not. Energy transfer studies of chloroplast F1 by Nalin et al. also indicate that the three a subunits in chloroplast Fl are not structurally equivalent (36). Acknowledgments-The beef heart mitochondrial F1-ATPase was prepared by Betty Stone in our laboratory. J. H. W. wishes to thank Dr. Joyce J. Diwan of Rensselaer Polytechnic Institute for helpful conversation concerning the arrangement of subunits in F,. REFERENCES 1. Ferguson, S. J., Lloyd, M. H., Lyons, M. H., and Radda, G. K. (1975) Eur. J. Bwchem. 54, 117-126 2. Andrews, W. M., Hill, F. C., and Allison, W. S. (1984) J. Bwl.

Chem. 259,8219-8225

5150

Site

Hydrolytic

3. Sutton, R., and Ferguson, S. J. (1985) Eur. J. Biochem. 148, 551-554 4. Wang, J. H., Joshi, V., and Wu, J. C. (1986) Biochemistry 25, 7996-8001 5. Ting, L. P., and Wang, J. H. (1980) Biochemistry 19,5665-5670 6. Low, H., and Vallin, I. (1963) Biochim. Biophys. Acta 69, 361374 7. Knowles, A. F., and Penefsky, H. S. (1972) J. Biol. Chem. 247, 6617-6623 8. Bradford, N.M. (1976) Anal. Biochern. 72, 248-257 9. Chambers, R. W., Shapiro, P., and Kurkov, V. (1960) J. Am. Chem. SOC.82,970-975 83, 10. Moffatt, J. G., and Khorana, H. G . (1961) J. Am. Chem. SOC. 649-658 11. Roseman. S.. Distler, J. J.. Moffatt, J. G., and Khorana, H. G. (1961) i.Am. C h i . Soc: 83,659-663 12. Ellman, G. L. (1959) Arch. Biochem. Biophys. 82, 70-77 13. Leloir, L. F., and Cardini, C. E. (1957) Methods Enzymol. 3,843844 14. Penefsky, H. S. (1977) J. Biol. Chem. 252, 2891-2899 15. Paulus, H. (1969) Anal. Biochem. 32,91-100 16. Wu, J. C., and Wang, J. H. (1986) Biochemistry 25,7991-7995 17. Ting, L. P., and Wang, J. H. (1980) Biochemistry 19,5665-5670 18. Ting, L. P., and Wang, J . H. (1980) J. Bioenerg. Biomembr. 12, 79-93 19. Esch, F. S., and Allison, W. S. (1979) J. Biol. Chem. 254, 1074010746 20. Bullough, D.A., and Allison, W. S. (1986) J. Biol. Chem. 261, 14171-14177

of Fl-ATPase 21. Bullough, D. A., and Allison, W. S. (1986) J. Biol. Chem. 261, 5772-5730 22. Wagenvoord, R. J., Kemp, A., and Slater, E. C. (1980) Biochim. Biophys. Acta 593,204-211 23. Cross, R. L., and Nalin, C. M. (1982) J. Bid. Chem. 267,28742881 24. Wise, J. G., Duncan, T. M., Latchney, L. R.,Cox,D. N., and Senior, A. E. (1983) Biochem. J. 215,343-350 25. Weber, J., Lucken, U., and Schafer, G. (1985) Eur. J. Biochem. 148,41-47 26. Kironde, F. A. S., and Cross, R. L. (1986) J. Biol. Chem. 261, 12544-12549 27. Wang, J. H. (1985) J. Biol. Chem. 260, 1374-1377 28. Gresser, M. J., Myers, J. A., and Boyer, P. D. (1982) J. Biol. Chem. 257,12030-12038 29. O’Neal, C. C., and Boyer, P.D. (1984) J. Bwl. Chem. 259,57615767 30. Cross, R. L., Grubmeyer, C., and Penefsky, H. S. (1982) J. Bwl. Chem. 257,12101-12105 31. Grubmeyer, C., Cross, R. L., and Penefsky, H. S. (1982) J. Bwl. Chem. 267,12092-12100 32. Akey, C.W., Spitsberg, V., and Edelstein, S. J. (1983) J. Biol. Chem. 258,3222-3229 33. Amzel, L. M., McKinney, M., Narayanan, P., and Pedersen, P. L. (1982) Proc. Natl. Acud. Sci. U. S. A. 79,5852-5856 34. Ehrig, K., Lunsdorf, H., Fried, P., and Schairer, H. U. (1986) Methods Enzymol. 126, 770-775 35. Boekma, E. J., Berden, J. A., andvan Heel, M. G . (1986) Biochim. Biophy~.Acta 851,353-360 36. Nalin, C. M., Snyder, B., and McCarty, R. E. (1985) Biochemistry 24,2318-2324