Modulation by ADP and Mg2+ of the Inactivation of the F1-ATPase from the Thermophilic Bacterium, PS3, with Dicyclohexylcarbodiimide*

The soluble F1-ATPase from the thermophilic bac- terium PS3 (TF1) contains no endogenous adenine nucleotides and contains about 0.2 g ions of Mg2+/mol which resists removal by repeated centrifugation-elu- tion on columns of Sephadex G-50. The isolated enzyme will not bind additional Mg2+ added in the absence of adenine nucleotides nor is the rate of inacti- vation of the isolated enzyme by dicyclohexylcarbodi-imide (DCCD) affected by the addition of Mg2+. When ADP is added to isolated TFl, a 1: 1 TFl .ADP complex is formed which is stable to repeated gel permeation on columns of Sephadex G-50 subjected to centrifuga-tion-elution. On formation of the 1:l TFl.ADP com- plex, the rate of inactivation of the enzyme by DCCD is accelerated 6-fold. The rate of inactivation of the 1: 1 TFl .ADP complex by DCCD is not further stimu- lated in the presence of 2 mM ADP which indicates that the binding of ADP to a single site in the enzyme is sufficient to promote maximal stimulation of the inactivation. Addition of Mg2+ to the 1:l TFl .ADP complex results in the binding of about 1 g ion of Mg2+/mol of enzyme. enzyme were rapidly analyzed in the concentration range of 10 to 500 pmol. Analysis of Enzyme-bound Mg2+-The M e content of the enzyme was determined by atomic absorption spectroscopy using a Varian model 1100 flame atomic absorption analyzer. The buffer in which the enzyme was dissolved, 50 mM TEA-SO4, pH 7.3, containing 1 pM CDTA, contained up to 0.8 p~ contaminating M F . Therefore, due to this background it was necessary to use enzyme concentrations greater than 10 mg/ml (26 PM) to ensure accurate analysis of M e contributed by the enzyme. Standardization was carried out by the method of additions. For a series of solutions with different enzyme concentrations, each was mixed with M2+ solutions of increasing concentration and then subjected to analysis. The MF/TF, ratios were determined from the average of the values obtained for each protein concentration.

The soluble F1-ATPase from the thermophilic bacterium PS3 (TF1) contains no endogenous adenine nucleotides and contains about 0.2 g ions of Mg2+/mol which resists removal by repeated centrifugation-elution on columns of Sephadex G-50. The isolated enzyme will not bind additional Mg2+ added in the absence of adenine nucleotides nor is the rate of inactivation of the isolated enzyme by dicyclohexylcarbodiimide (DCCD) affected by the addition of Mg2+. When ADP is added to isolated TFl, a 1: 1 TFl .ADP complex is formed which is stable to repeated gel permeation on columns of Sephadex G-50 subjected to centrifugation-elution. On formation of the 1:l TFl.ADP complex, the rate of inactivation of the enzyme by DCCD is accelerated 6-fold. The rate of inactivation of the 1: 1 TFl .ADP complex by DCCD is not further stimulated in the presence of 2 mM ADP which indicates that the binding of ADP to a single site in the enzyme is sufficient to promote maximal stimulation of the inactivation. Addition of Mg2+ to the 1:l TFl .ADP complex results in the binding of about 1 g ion of Mg2+/mol of enzyme. The 1:l:l TFl .ADP. Mg2+ complex thus formed is sluggishly inactivated by DCCD. When the Mg2+ is removed from the TFI .ADP.Mg2' complex by treatment with trans-1,2-diaminocyclohexane-N,N,N,N-tetraacetic acid, the rate of inactivation of the enzyme by DCCD is accelerated 4-fold.
Other divalent metal ions protect the 1:l TFl .ADP complex against inactivation by DCCD. Of these, Mn2+, Zn2+, Co2+, and Cd2+, which are about as equally effective as Mg2+ as cofactors for the hydrolytic reaction when present at 0.2 mM, offer about equal protection of the complex against inactivation by DCCD also when present at 0.2 mM. These results indicate that the binding site for ADP in the 1:l TFl .ADP complex is a catalytic site.
TFl, inactivated by 92% with DCCD, has the same capacity to bind ADP as the active enzyme, forming a tight 1: 1 TFl. ADP complex which is stable to repeated centrifugation-elution on columns of Sephadex G-50. The 1: 1 TFl. ADP complex retains its capacity to bind Mg2+ to form the 1:l:l TFl .ADP.Mg2+ complex after it is inactivated by 88% with DCCD.
* This work was supported by the Ministry of Education, Japan and by United States Public Health Service Grant GM-16974 and National Science Foundation Grant 81-12467. 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.
Evidence is accumulating that the F1-ATPases contain one or more carboxylic acid side chains which are essential for the hydrolytic activity catalyzed by these enzymes (1)(2)(3)(4)(5)(6)(7)(8)(9)(10)(11)(12)(13). The F1-ATPases from a variety of organisms have been inactivated by DCCD' (1, 3-5, 8, 11). With the use of [14C]DCCD it has been demonstrated that the inactivations of the F,-ATPases, TF1, MF1, and EFI, are due to the modification of a single glutamic acid side chain in the fi subunit of each enzyme (8)(9)(10). Surprisingly, a different glutamic acid residue, designated by an asterisk in the sequence shown below, is labeled by ["C] DCCD in the fi subunit of TF1 than is labeled by [14C]DCCD in the fi subunits of MFI and EFI, designated by a dagger in the sequence shown below. The highly conserved amino acid sequence in the fi subunits of these enzymes which contain both the DCCD-reactive glutamic acid residue of TFI (E*) and the DCCD-reactive glutamic acid residue of MFl and EF1 The function or functions of the DCCD-reactive glutamic acid residue of TF1 and the DCCD-reactive glutamic acid residue common to MF1 and EF1 are not known. The rate of inactivation of the F1-ATPases by DCCD is affected by the presence of adenine nucleotides and by M$+. The rate of inactivation of TF1 by DCCD is accelerated about 7-fold in the presence of ADP while it is only slightly stimulated by ATP (8).
The stimulatory effect of ADP on the inactivation of TF, by DCCD is abolished by the addition of M e . It was also observed that the addition of MgZ+ to native TF, affords slight protection of the enzyme against inactivation by DCCD (8).
The effects of adenine nucleotides and Mg2f on the rate of inactivation by DCCD of MF1 depleted of loosely bound endogenous nucleotides are very similar to those described for the inactivation of TF1 by DCCD (2, 14). In the presence of ADP or ATP the rate of inactivation of MF, by DCCD is accelerated by about 50% while the rate of inactivation of the enzyme by DCCD is decreased by about 60% in the presence of Mg2+. The addition of Mg.ADP to MFI depleted of loosely bound endogenous adenine nucleotides has very little effect on the rate of inactivation of the enzyme when compared to a control with no additions (14). The rate of inactivation of EF, (3) is affected differently by DCCD than is TF1 and MF, in the presence of adenine nucleotides and M e . In the presence of ATP o r ADP the rate of inactivation of EF, by DCCD is slowed by a factor of two. It is slowed 3-to 4-fold in the presence of M$+ (3). Therefore, in terms of the characteristics of inactivation by DCCD, MF, resembles TF, more than it does EF,, although MF, and EF, have a common glutamic acid residue that reacts with DCCD which is essential for the hydrolytic reaction catalyzed by the enzymes.
In order to provide clues to elucidate the function or functions of the carboxyl groups in the F1-ATPases which, when modified by DCCD abolish hydrolytic activity, an investigation was initiated to explore in more detail the characteristics of ADP and M$+ binding to TF, which modulate the rate of inactivation of the enzyme by DCCD. These characteristics are presented and discussed in this report.

EXPERIMENTAL PROCEDURES
Materials-The adenine nucleotides used were of the highest purity available from Sigma. MES was purchased from Calbiochem-Behring. CDTA was purchased from Sigma. Stock solutions of DCCD (Sigma) were prepared in absolute ethanol. The divalent metal salts were of the highest purity available from Wako Chemical Company. The MgSO, used was a special grade prepared for use in atomic absorption spectroscopy.
Preparation and Assay of TF,-TF, was prepared from plasma membranes of PS3 as described in detail previously (15). The enzyme was stored as a lyophilized powder at -20 "C after exhaustive dialysis of the purified enzyme against distilled water. TF, is stable for years when stored in this manner. The lyophilized enzyme was dissolved in the appropriate buffer just before each of the experiments described.
ATPase activity was determined a t 25 "C by a phosphate release method with an ATP-regenerating system (16) unless specified otherwise. Protein concentrations were determined with Coomassie brilliant blue as described by Bradford (17) using a standard curve prepared from lyophilized TF, for each experiment. The average of triplicate measurements for the standard curve and the unknowns was used. The centrifugation-elution method of Penefsky (18) was used to separate enzyme-bound ligands from free ligands using columns equilibrated with 50 mM TEA-SO4, pH 7.5, containing 1 p M CDTA unless specified otherwise. When the volume of the TF1-ligand solutions was 0.1 ml or less, 1.0-ml syringes were used for centrifugation-elution. For larger volumes up to 0.5 ml, centrifugation-elution was performed on 5-ml syringes.
Analysis of Enzyme-bound Adenine Nucleotides-Enzyme-bound adenine nucleotides were determined by anion exchange high performance liquid chromatography. Bound nucleotides were released from TF, by adding 2 pl of 60% perchloric acid to 50 pl of the enzymeadenine nucleotide complexes. After adding perchloric acid the mixtures were incubated at 0 "C for 30 min at which time denatured enzyme was removed by centrifugation. Then 2 p1 of 9.2 M KOH was added to the supernatants. The pH of the supernatants was adjusted to neutrality as indicated by pH test paper by the addition of small volumes of 9.2 M KOH or 85% H3P04. The precipitated KC1O4 in the samples was then removed by centrifugation a t 4 "C. The supernatant was applied to a Toya Soda IEX-540 column (4 X 300 mm) which was equilibrated with 400 mM sodium phosphate, pH 6.0. The column was eluted isocratically with the same buffer at a flow rate of 0.8 ml/ min using a Waters model 204 liquid chromatograph. Nucleotides were monitored at 254 nm. Peak areas were determined by automatic integration. Using this method adenine nucleotides released from the enzyme were rapidly analyzed in the concentration range of 10 to 500 pmol.
Analysis of Enzyme-bound Mg2+-The M e content of the enzyme was determined by atomic absorption spectroscopy using a Varian model 1100 flame atomic absorption analyzer. The buffer in which the enzyme was dissolved, 50 mM TEA-SO4, pH 7.3, containing 1 p M CDTA, contained up to 0.8 p~ contaminating M F . Therefore, due to this background it was necessary to use enzyme concentrations greater than 10 mg/ml (26 PM) to ensure accurate analysis of M e contributed by the enzyme. Standardization was carried out by the method of additions. For a series of solutions with different enzyme concentrations, each was mixed with M 2 + solutions of increasing concentration and then subjected to analysis. T h e M F / T F , ratios were determined from the average of the values obtained for each protein concentration. Table I shows that treatment of TF, with ADP or ATP followed by centrifugation elution on Sephadex (2-50 (18) leaves the enzyme in a form more sensitive to inactivation by DCCD. ADP is more effective in this capacity than is ATP. Sensitization of TF1 to DCCD inactivation is not promoted by AMP or Pi. When TF,, sensitized to DCCD inactivation in the manner described above, is subsequently treated with M e , inactivation by DCCD becomes sluggish as shown in Table I. The TF,. ADP complex responds rapidly to the addition of M$+ as is illustrated by the lower curve in Fig. 1. The addition of DCCD to the TF, .ADP complex prepared by centrifugation-elution led to a rapid rate of inactivation. When M$+ was added 25 min after the inactivation was initiated, the rapid inactivation ceased abruptly. The upper curve in Fig. 1 shows that the addition of ADP to a reaction mixture containing DCCD and TF, leads to an abrupt increase in the rate of inactivation of the enzyme.

Sensitization of TF, to DCCD Inactivation by Pretreatment with ADP or ATP-
Binding of ADP and Mg2+ to TF, Correlated with Sensitivity to Inactivation by DCCD-The amount of ADP and M$+

TABLE I
The effect of pretreatment of TF, with adenine nucleotides and P; on its rate of inactivation with DCCD TF, (600 pg) was incubated with 2 mM ADP, ATP, or Pi in 50 mM TEA-SO4, pH 7.3, for 1 h at 23 "C. The solutions were then subjected to centrifugation-elution on columns of Sephadex G-50 equilibrated with 50 mM TEA-S04, pH 7.3. To 50-pl samples of each effluent the following additions were made: none; 1 pl of 100 mM ADP; and 1 p1 of 100 mM MgSO,. After an incubation of 1 h at 23 "C, the inactivations were initiated by the addition of 0.5 p1 of 15 mM DCCD to each reaction mixture and the rate of inactivation was determined. bound to TF, under various conditions of pretreatment followed by centrifugation-elution is presented in Table I1 which also shows the rate of inactivation of the enzyme by 0.15 mM DCCD in the presence and absence of 2 mM ADP. The isolated enzyme used in this study did not contain bound ADP nor did it contain bound ATP. The amount of M$+ detected in the isolated enzyme by atomic absorption spectroscopy was not stoichiometric as shown in the control of Table 11. Pretreatment of the enzyme with CDTA or M$+ did not change the M e content of TFI nor did it affect its inactivation by DCCD as shown by Experiments 1 and 2 of Table 11. Incubation of TF, with ADP in the absence of M e leads to the binding of 1.5 mol of ADP/mol of TF, after a single centrifugation-elution as shown in Experiment 3a of Table 11. This value decreased to 1.1 mol of ADP/mol of TF, when the TFI.
ADP complex was subjected to a second centrifugation-elution step as shown by Experiment 3b. Both complexes are equally sensitive to inactivation by DCCD in the presence and absence of 2 mM ADP. This shows that the binding of 1 mol of ADP per mol of TF1 is sufficient to cause maximum stimulation of inactivation by DCCD. The TF,. ADP complex obtained after the second centrifugation elution step is stable. It has been stored for 1 week at 4 "C with no change in the amount of bound ADP when examined by centrifugationelution and with no change in its sensitivity to inactivation by DCCD. However, a single ammonium sulfate precipitation removed 20% of the ADP bound to the TF1. ADP complex which may explain the absence of endogenous ADP in the preparation of TF, used in this study and the very low content of endogenous adenine nucleotides usually found in TF, preparations (19). Fig. 2 shows the effect of increasing Mg2' concentrations on the rate of inactivation of the TFI. ADP complex by DCCD. Saturation occurs at 5 2 0 PM M e showing that the TF1-ADP complex has a high affinity for Mg2+ which does not exist in the apoenzyme. When 2 mM ADP is present in the inactivation medium, higher concentrations of Mg2' are required to saturate the complex for inhibition of DCCD inactivation. Medium ADP acts like medium EDTA by chelating Mg2+, thus reducing the concentration of free Mg2+ available to bind the ADP moiety of the TF, .ADP complex. M$+ binds in amounts stoichiometric with bound ADP when it is added to the TF, . ADP complex as shown by Experiment 3c of Table   11. When TFI was pretreated with ADP and M e simultaneously and then subjected to a single centrifugation-elution step, 3 mol of both ADP and M e were bound per mol of TFI  After 1 h at 25 "C the solution was subjected to centrifugation-elution to produce effluent 3a from which samples were removed for analyses. The remainder of effluent 3a was incubated for 12 h at which time it was subjected to centrifugationelution to produce effluent 3b from which samples were removed for the analyses listed. To the remainder of effluent 3b (0.12 ml) was added 2.4 pl of 0.10 M MgSO4. The resulting solution was subjected to centrifugationelution after 1 h to produce effluent 3c from which samples were removed for analyses. Experiment 4, TFI (13 mg) was dissolved in 0.60 ml of 50 mM TEA-SO4, pH 7.3, containing 1 p M CDTA, 5 mM MgSO,, and 2 mM ADP. After 1 h at 25 "C the solution was subjected to centrifugation-elution to produce effluent 4a from which samples were removed for analyses. The remainder of effluent 4a was incubated for 12 h at which time it was subjected to centrifugation-elution, and the effluent was incubated an additional 12 h at which time it was subjected to centrifugation-elution to produce effluent 4b from which samples were removed for analyses. To the remainder of effluent 4b (0.19 ml) was added 48 pl of 0.10 M CDTA. The resulting solution was incubated at 25 'C for an additional 12 h at which time it was subjected to the third centrifugation-elution to produce effluent 4c from which samples were removed for analyses.

TABLE 111
The divalent metal ion specificity of the ATPase reaction and for protection against inactivation by DCCD The 200-pl reaction mixtures for the determination of ATPase activity contained 2 mM ATP and 0.2 mM metal ions (Cu" and Fe" present as sulfates and the others present as chlorides) in 50 mM TEA-SO,, pH 7.3. The enzyme assays were carried out as described under "Experimental Procedures." The TF, .ADP complex was prepared as described in the legend of Table 11. The rate of inactivation by 0.15 mM DCCD was determined a t 20 "C. The inactivations were initiated after incubating the enzyme with the metal ions for 40 min at 20 "C by adding 2 pl of 3.75 mM DCCD to 20 pg of TF, in 50 pl of 50 mM TEA-SO,, pH 7.3, containing the divalent metal ions at 0.2 mM. Divalent Metal Ions That Protect TF1 against Inuctiuation by DCCD-The relative effectiveness of various divalent metal ions which protect the TFl. ADP complex against inactivation by DCCD when present at a concentration of 0.2 mM is shown in Table 111. Also shown in Table I11 is the relative effectiveness of the same metal ions as cofactors in the ATPase reaction. It is clear that those divalent metal ions (Mn2+, Zn2+, Co2+, and Mg2+) which are the most active catalytically are also the most effective in protecting the TF, .
ADP complex against inactivation by DCCD. This suggests that the divalent metal ions protect TFI against inactivation by DCCD when they are bound to the catalytic site of the enzyme. We have no explanation for the observation that Cd", Ni", and Fez+ appear to be more active as cofactors when assayed with TFl than they do when they are assayed with the TF, .ADP complex. The ATPase activity is very low in the presence of 0.2 mM Ca2+ when compared to its activity in the presence of 0.2 mM M$+. This appears to contradict results previously reported which showed that 5 mM Ca2+ is as effective as 5 mM Mg2+ as cofactor for the hydrolytic reaction catalyzed by TF, (15). However, a t 5 mM both divalent metal ions saturate the enzyme and, therefore, the differences observed between Ca2+ and Mg2+ at low nonsaturating concentrations were not exhibited.
The Binding of ADP to TF, Inactivated with DCCD-When TFI was inactivated by 92% with DCCD the enzyme retained its capacity to bind ADP at pH 7.3 as shown by comparison of Experiment 3 of Table I1 with Experiment 2 of Table IV. The DCCD-inactivated enzyme, like native TF,, binds 1 mol of ADP tightly which is not removed by repeated centrifugation-elutions. When Mg2+ and ADP are added together to TFI The DCCD-inactivated enzyme was prepared as follows. To 15 mg of TF1 in 1.0 ml of 50 mM Na-MES, pH 5.5, containing 1 p~ CDTA was added 20 pl of 10 mM DCCD. 150 min after initiating the inactivation a second 20-pl dose of DCCD was added. After 10 h 92% inactivation was attained at which time the reaction mixture was subjected to centrifugation-elution (CE) on Sephadex G-50 equilibrated with 50 mM TEA-SOI, pH 7.3, containing 1 p~ CDTA. Experiment 1, samples of the DCCD-inactivated enzyme were removed for the determination of ADP and protein. Experiment 2, to 0.4 ml of the DCCD-inactivated enzyme was added 16 pl of 100 mM ADP. This mixture was incubated for 1 h at 25 "C at which time it was subjected to centrifugation to produce effluent 2a. Samples of effluent 2a were removed for analyses and the remainder was incubated an additional 12 h at 25 'C at which time it was subjected to centrifugation-elution to produce effluent 2b. Samples of effluent 2b were removed for analyses. Experiment 3, to 0.2 ml of DCCDinactivated TF, was added 8 pl of 100 mM ADP and 10 pl of 100 mM MgS04. This mixture was incubated for 1 h at 25 "C at which time it was subjected to centrifugation to produce effluent 2a. Samples of effluent 2a were removed for analyses and the remainder was incubated an additional 12 h at 25 "C at which time it was subjected to centrifugation-elution. The resulting effluent was incubated an additional 4 h at 25 "C at which time it was subjected to another centrifugation-elution to produce effluent 2b. Samples of effluent 2b were then removed for analyses.   Table 11. To 4.6 mg of the TF1. ADP complex in 0.40 ml of TEA-SO,, pH 7.3, containing 1 p~ CDTA was added 8 p1 of 10 mM DCCD. After 150 min at 25 "C another 8 pl of 10 mM DCCD were added to the reaction mixture. After another 150 min 88% of the ATPase was tion-elution (CE) on Sephadex G-50 equilibrated with TEA-SO,, pH inactivated at which time excess DCCD was removed by centrifuga-7.3, containing 1 p~ CDTA. Samples were removed from the effluent to determine enzyme-bound ADP and M e as well as the protein concentration as described under "Experimental Procedures." To 0.20 ml of the remaining solution was added 4 pl of 100 mM MgSO, which was followed by an incubation of 1 h at 25 "C. The incubated mixture was then subjected to centrifugation-elution on Sepbadex G-50 equilibrated with TEA-SO,, pH 7.3, containing 1 p~ CDTA. Samples were removed from the effluent for M%+, ADP, and protein determinations.  Table I1 and Experiment 3a of Table IV.
After repeated centrifugation-elutions the ADP content of both the native and DCCD-inactivated TF, decreased to about 1 mol per mol of enzyme when the initial loading with ADP took place in the presence of Mg2+ as shown in Experiment 3 of Table IV.
The Binding of Mg2+ to the TF, -ADP Complex Inactivated with DCCD-When the 1:l TF1 .ADP complex was inactivated by 88% with DCCD, the bound ADP in the complex resisted removal by centrifugation-elution as shown in Table  V. After inactivation with DCCD the 1:l TFl. ADP complex retained its capacity to bind M$+ as shown by comparison of Experiment 3c of Table I1 with Experiment C of Table V. Therefore, the DCCD-reactive glutamic acid residue in TFl which, when modified leads to inactivation of the enzyme, does not appear to function in binding the Mg2+ moiety of M$+-adenine nucleotide complexes as was suggested by us previously (8,10).

DISCUSSION
From the results presented it is clear that Mg2+ does not bind tightly to TFl in the absence of bound adenine nucleotides nor does the addition of Mg2+ to TFl free of adenine nucleotides protect the enzyme against inactivation by DCCD. Although the TF1 used in the experiments reported here was completely free of endogenous adenine nucleotides, some preparations of the enzyme have been found to contain small amounts of endogenous ADP. The presence of a small amount of ADP in the TF, preparation examined by us previously (8) might have been responsible for the slight protective effect that was observed when DCCD inactivation was carried out in the presence of M$+ but in the absence of exogenous adenine nucleotides. The observation that TF, contains very little endogenous Mg2+ differs considerably from the results of Senior and his colleagues who have reported that isolated MF, contains 1 g ion of tightly bound M$+ per mol which remains bound even after the tightly bound adenine nucleotides were removed from the enzyme (20, 21). Isolated EF, has been reported to contain about 2 g ions of M$+ per mol (21). T o explain the fact that MF1 depolymerizes when the tightly bound Mg2+ is removed from it, Senior has postulated that the tightly bound M e is an integral component of the active enzyme complex which might function to stabilize the quaternary structure of the complex (20). From the results presented here, it is clear that TF, does not require endogenous Mg2+ for stabilization of its quaternary structure. Scheme 1 summarizes the sensitivity of different TF, complexes to DCCD inactivation which have been Solated after incubating the enzyme with ADP or ADP plus M$+ and then subjecting it to centrifugation-elution in buffers not containing the ligands. When TF, binds only ADP the rate of inactivation of the enzyme is stimulated as designated by an asterisk in Scheme 1. This increased sensitivity of TF, to inactivation by DCCD is probably due to the binding of ADP to a single catalytic site for the following reasons. 1) The 1:l TF, .ATP complex has been isolated and shows maximal stimulation of DCCD inactivation as indicated in Scheme 1.
2) Only those divalent metal ions which are active catalytically protect the 1:l TF, .ADP complex against inactivation by DCCD. 3) It has been shown that the ADP in the 1:l TF,. ADP complex can be converted to ATP when the complex is incubated with M e and Pi in 50% dimethyl sulfoxide under slightly acidic conditions (22).
The observations that the occupancy of a single catalytic site of TF, by ADP is sufficient to promote full acceleration of inactivation of the enzyme by DCCD raises interesting questions concerning the stoichiometry of modification of the enzyme by the reagent. It has been shown that the complete inactivation of TF, with [14C]DCCD in the presence of 1 mM ADP is accompanied by the modification of 1.6-1.8 copies of the /3 subunit to form a single N-y-glutamyl derivative of dicy~lohexyl['~C]urea with each /3 subunit modified (8). This suggests that selective modification of two of the three p subunits of the enzyme with DCCD is sufficient to abolish all ATPase activity. However, as has been discussed in detail elsewhere (lo), it is not certain that the 0-acylisourea formed on the initial reaction of the essential carboxyl group with DCCD rearranges quantitatively to form the N-acylurea which is isolated. Thus firm experimental evidence is not available to distinguish between "two-thirds of the sites reactivity" and "all of the sites reactivity" when TF, is inactivated by DCCD. Cross and Nalin have suggested on the basis of diminished binding of [3H]adenyl 5"yl imidodiphosphate to MFl inactivated with DCCD that inactivation of the mitochondrial enzyme proceeds with "two-thirds of the sites reactivity" (23). However, since DCCD modifies different glutamic acid side chains when it inactivates TF, and MF, (8), the observations of Cross and Nalin with MF, may not bear on the stoichiometry of carboxyl group modification which occurs when TF1 is inactivated by DCCD. Feldman and Sigman have shown that isolated CF, will synthesize enzyme-bound ATP under slightly acidic conditions (24). Isolated MF, (25) and isolated TF, (22) will also synthesize enzyme-bound ATP under certain conditions. Moreover, the studies with the latter two enzymes have provided evidence which suggests that the DCCD-reactive glutamic acid residue of TF, has a different functional role than the DCCD-reactive glutamic acid residue to MF,. Sakamoto and Tonomura have shown that MFI will synthesize up to 0.6 mol of bound ["PIATP per mol of the MF, in the presence of 30% dimethyl sulfoxide, ADP, 32Pi, and MgZ+ at slightly acidic pH (25). They also showed that MFI inactivated with DCCD retained full capacity to synthesize enzyme-bound ATP under the same conditions. The differential effect of DCCD on the hydrolytic reaction opposed to the artificial synthetic reaction catalyzed by MF, is similar to the well documented observation that isolated MF1 modified with 7chloro-4-nitrobenzo-2-oxa-l,3-diazole will not catalyze ATP hydrolysis, but will catalyze ATP synthesis when added to depleted submitochondrial particles (26, 27). Enzyme-bound ATP is also formed when Pi and M$+ are added to the TFI. ADP complex in 50% dimethyl sulfoxide under slightly acidic conditions (22). In contrast to what is observed with MF1, enzyme-bound ATP is no longer formed under these conditions when the TFI. ADP complex is inactivated with DCCD.' Therefore, the DCCD-reactive glutamic acid side chain of TF, appears to be essential for both ATP synthesis and ATP hydrolysis while the DCCD-reactive glutamic acid side chain of MF1 is essential for ATP hydrolysis but apparently not for ATP synthesis. These results suggest that the DCCD-reactive glutamic acid residue of TF, has a direct role in the reversible reaction catalyzed by the enzyme while the DCCD-reactive glutamic acid side chain of MF, does not.
If the DCCD-reactive glutamic acid residue of TF, does indeed have a direct functional role in catalysis then why are the corresponding glutamic acid residues in the ( 3 subunits of MF, and EF, unreactive with DCCD? Very little information is available which bears directly on this dilemma. However, several reports have appeared which indicate that individual F,-ATPases contain more than one essential carboxyl group which can be distinguished by their reactivities with different electrophilic reagents. For example, Pougeois and his colleagues have presented evidence that N-ethoxycarbonyl-2ethoxy-1,2-dihydroquinoline and DCCD modify different carboxyl groups when MF, is inactivated by these reagents (2,3,12). Vallejos and his colleagues (6,7,13) have provided evidence that Woodward's reagent K inactivates CFI (6), TF1 (7), and the F1-ATPase from Rhodospirillum rubrum chromatophores (13) by modification of a different carboxyl group in each of these enzymes than is modified during inactivation with DCCD. It is possible that the essential carboxyl groups in these ATPases which react with N-ethoxycarbonyl-2ethoxy-1,2-dihydroquinoline or Woodward's reagent K might correspond to the DCCD-reactive glutamic acid residue of TF, which appears to have a direct catalytic role. Evidence that a carboxyl group participates directly int the hydrolytic M. Yoshida, unpublished experiments. reaction catalyzed by the mitochondrial F1-ATPase has been obtained from kinetic analyses (28). Godinot and Penin have demonstrated the participation of a carboxyl group in the hydrolytic reaction catalyzed by MF1 from pig heart by careful examination of the pH rate profiles of the enzyme as a function of buffer ion species, organic solvents, and temperature (28). From this examination they concluded that V,,, and V,,,/K,,, of ATP hydrolysis depend on the presence of the conjugate base of a carboxyl group with a pK, of 5.4-5.9.