Reversible binding of Pi by beef heart mitochondrial adenosine triphosphatase.

Beef heart mitochondrial ATPase (F1) exhibited a single binding site for Pi. The interaction with Pi was reversible, partially dependent on the presence of divalent metal ions, and characterized by a dissociation constant at pH 7.5 of 80 micronM. A variety of substances known to influence oxidative phosphorylation or the activity of the soluble ATPase (F1) also influenced Pi binding by the enzyme. Thus aurovertin, an inhibitor of oxidative phosphorylation, which was bound tightly by F1 and inhibited ATPase activity, enhanced Pi binding via a 4-fold increase in the affinity of the enzyme for Pi (KD = 20 micronM) but did not alter binding stoichiometry. Anions such as SO4(2-), SO3(2-), chromate, and 2,4-dinitrophenolate, which stimulated ATPase activity of F1, also enhanced Pi binding. Inhibitors of ATPase activity such as nickel/bathophenanthroline and the protein ATPase inhibitor of Pullman and Monroy (Pullman, M. E., and Monroy, G. C. (1963) J. Biol. Chem. 238, 3762-3769) inhibited Pi binding. The adenine nucleotides ADP, ATP, and the ATP analog adenylyl imidodiphosphate as well as the Pi analog arsenate, also inhibited Pi binding. The observations suggest that the Pi binding site was located in or near an adenine nucleotide binding site on the molecule.

The cold-labile ATPase of oxidative phosphorylation was first described about 15 years ago (1). It was proposed at that time that the ATPase activity of the soluble beef heart enzyme was an artifact of isolation and that when the enzyme was bound to the mitochondrial membrane, it served as a phosphate transfer catalyst in the terminal transphosphorylation step of oxidative phosphorylation (2). In the intervening years, it also has been suggested that F,' functioned as a component part of a system which synthesized ATP via reversal of the hydrolytic reaction (3,4) and that energy-dependent conformational changes which altered the affinity of adenine nucleotide binding sites on F, were critical features of ATP synthesis in oxidative phosphorylation (5,6 As isolated in this laboratory, three of the five sites on beef heart F, were occupied by very tightly bound adenine nucleotides (8) and two sites were engaged in readily reversible binding of adenine nucleotides (8). The function of all of the sites has yet to be elucidated. However, one or more may be control sites (8-10) and at least one site is the hydrolytic site (11,12  It may be seen that essentially the same results were obtained when binding of AMP-P(NH)P by nucleotide-depleted F, (8) was examined by the two methods. Three observations regarding the properties of the centrifuge column are relevant to a discussion of the mode of separation of free ligand from protein.
Direct measurement of the time of transit of F, through the column, carried out with the aid of a device for automatic layering of 100 ~1 of sample on the column while the rotor was revolving, indicated that the bulk of the protein emerged from the column in a volume of 10 to 20 ~1 about 30 s after application of the sample. An additional 80 to 90 ~1 of buffer followed during the next 60 to 90 s of centrifugation.
Experiments in which the reaction mixture applied to the column in Step 2 contained a colored ion such as chromate also indicated that unbound ions were sequestered in the upper one-third of the column. The third observation was obtained from experiments in which the Sephadex used to fill the centrifuge column was pre-equilibrated with 32Pi of known concentration.
Application in Step 2 of 100 ~1 of water to such a column resulted, after centrifugation, in the appearance of"2P, in the column effluent of the same concentration as that originally equilibrated with the Sephadex.
These observations indicate that a considerable dehydration of the protein sample, accompanied by loss of most if not all unbound ions probably occurred in the upper one-third of the column. A concentrated protein sample emerged from the column followed by a volume of buffer which may have been held within the gel beads in the lower portion of the column. It would thus appear that the centrifuge column procedure differs in important ways from gel permeation chromatography as normally used. In addition to its sensitivity and rapidity in ligand binding studies, the centrifuge column procedure thus exhibits two further features of interest.
First, it may be used to exchange buffers in small protein samples with virtually no loss of protein.
Second, it may be used to achieve a 5-to lo-fold concentration of protein in small samples and is particularly useful for purposes of gel electrophoresis. The concentration step was, however, accompanied by a loss of protein of 10 to 30%. A method of desalting proteins, very similar to the one presented here, has been described by Neal  It may also be seen in Fig. 3    At the end of the incubation period, loo-p1 aliquots were withdrawn and transferred to the tops of centrifuge columns for measurement of P, binding as described under "Experimental Procedures." The pH was determined in a separate series of tubes which were identical except that "*Pi was replaced by water. FIG. 6 ( (Fig. 2), in the presence of aurovertin the K, was 20 PM.
Adenine nucleotides partially inhibited Pi binding by F,. In Fig. 9 it is shown that ATP was considerably less effective than the ATP analog AMP-P(NH)P.
Approximately 50% of the total observable inhibition occurred at an ATP concentration of 45 PM, whereas only 4 PM AMP-P(NH)P was sufficient to produce 50% inhibition. Moreover, the extent of the inhibition also was greater with the analog (Fig. 9). ADP appeared to be about as effective an inhibitor as ATP under these experimental conditions.
The inhibition of Pi binding by ADP was prevented when aurovertin was included in the reaction mixture (Table IV)