PubMed1998
The conflicting data on the binding of the two molecules of ATP that are involved in the overall reaction catalyzed by carbamoyl-phosphate synthetase (CPS) of Escherichia coli, and a mechanism recently proposed for this reaction, has led us to reexamine ATP binding using pulse/chase techniques. With [gamma-32P]ATP and bicarbonate in the pulse solution, there is a positive intercept at zero time of approximately 1 mol Pi/mol CPS in the plot of 32Pi formation against time, irrespective of whether the incubation is terminated by the addition of acid or by addition of a chase solution containing glutamine, excess unlabeled ATP and bicarbonate. The intercept is decreased to about 50% if the excess unlabeled ATP is added prior to the addition of the glutamine. These are the expected results if the intercept reflects the reversible formation of enzyme-bound ADP and carboxyphosphate. Approximately 0.6 mol carbamoyl [32P]phosphate/mol enzyme is formed in these experiments when the pulse step is terminated by addition to the chase solution. The ATP molecule that provides the phosphoryl group of carbamoyl phosphate, therefore, also binds to the enzyme in the absence of ammonia or glutamine and reacts in the chase to give carbamoyl phosphate before it can dissociate from the enzyme. At 1 mM ATP, the binding of both ATP molecules is essentially complete at 2.5 s, but the dissociation of the ATP that yields carbamoyl phosphate is extremely slow (t(1/2) of about 6 min at 22 degrees C; HCO3-, 40 mM), although it is faster in the absence of bicarbonate. The extreme sequestration from the aqueous environment of this ATP allows the enzyme-ATP complex to be separated from the surrounding ATP by centrifugal gel filtration. After two successive steps of gel filtration through Sephadex G-50 equilibrated with unlabeled ATP and bicarbonate, the majority of the radioactivity remaining in the solution is bound to the enzyme and is released as [gamma-32P]ATP if acid is added, or is converted to carbamoyl [32P]phosphate by addition to chase solution, without concomitant release of 32Pi. K+ is necessary in the pulse solution, but not in the chase solution, to demonstrate this binding. These findings and other confirmatory experiments demonstrate conclusively that, in the presence of K+, both ATP molecules bind to the enzyme in the absence of ammonia or glutamine. The bound ATP that yields Pi in the overall reaction is replaced relatively rapidly by exchange and by hydrolysis in the bicarbonate-dependent ATPase activity of the enzyme, whereas the bound ATP that provides the phosphoryl group of carbamoyl phosphate is replaced very slowly. The temporal pattern of carbamoyl [32P]phosphate formation from [gamma-32P]ATP, in pulse/chase experiments in which a small concentration of ammonia is added to the pulse solution, shows that, in the normal enzyme reaction, this last ATP molecule binds to the enzyme before ammonia. These findings exclude a recently proposed mechanism [Kothe, M., Eroglu, B., Mazza, H., Samudera, H. & Powers-Lee, S. (1997) Proc. Natl Acad. Sci. USA 94, 12348-12353] in which a single molecule of ATP bound at the catalytic center phosphorylates bicarbonate and provides the phosphoryl group of carbamoyl phosphate. A mechanism in which a single ATP molecule binds, followed by the binding of bicarbonate and ammonia (from glutamine) and the release of Pi before the second molecule of ATP is bound is also excluded. We have previously reported very similar findings for carbamoyl-phosphate synthetase (ammonia), strongly suggesting that the different types of CPS share a common mechanism. The virtual sequestration of the ATP that provides the phosphoryl group of carbamoyl phosphate is consistent with a palmate-binding site, with the nucleotide bound within a beta-sheet sandwich, and a loop closure mechanism triggered by the binding of bicarbonate or the formation of carboxyphosphate.