How does GAP catalyze the GTPase reaction of Ras?: A computer simulation study

The formation of a complex between p21(ras) and GAP accelerates the GTPase reaction of p21(ras) and terminates the signal for cell proliferation. The understanding of this rate acceleration is important for the elucidation of the role of Pas mutants in tumor formation. In principle there are two mai n options for the origin of the effect of GAP. One is a direct electrostati c interaction between the residues of GAP and the transition state of the P as-GAP complex and the other is a GAP-induced shift of the structure of Pas to a configuration that increases the stabilization of the transition stat e. This work examines the relative importance of these options by computer simulations of the catalytic effect of Ras. The simulations use the empiric al Valence bond (EVB) method to study the GTPase reaction along the alterna tive associative and dissociative paths. This approach reproduces the trend in the overall experimentally observed catalytic effect of GAP: the calcul ated effect is 7 +/- 3 kcal/mol as compared to the observed effect of simil ar to 6.6 kcal/mol. Furthermore, the calculated effect of mutating Arg789 t o a nonpolar residue is 3-4 kcal/mol as compared to the observed effect of 4.5 kcal/mol for the Arg789Ala mutation. It is concluded, in agreement with previous proposals, that the effect of Arg789 is associated with its direc t interaction with the transition state charge distribution. However, calcu lations that use the coordinates of Pas from the Pas-GAP complex (referred to here as Pas') reproduce a significant catalytic effect relative to the P as coordinates. This indicates that part of the effect of GAP involves a st abilization of a catalytic configuration of Pas. This configuration increas es the positive electrostatic potential on the beta-phosphate (relative to the corresponding situation in the free Pas). In other words, GAP stabilize s the GDP bound configuration of Ras relative to that of the GTP-bound conf ormation. The elusive oncogenic effect of mutating Gln61 is also explored. The calculated effect of such mutations in the Ras-GAP complex are found to be small, while the observed effect is very large (8.7 kcal/mol). Since th e Pas is locked in its Pas-GAP configuration in our simulations, we conclud e that the oncogenic effect of mutation of Gln61 is indirect and is associa ted most probably with the structural changes of Pas upon forming the Pas-G AP complex. In view of these and the results for the Pas' we conclude that GAP activates Pas by both direct electrostatic stabilization of the transit ion state and an indirect allosteric effect that stabilizes the GDP-bound f orm. The present study also explored the feasibility of the associative and dissociative mechanism in the GTPase reaction of Pas. It is concluded that the reaction is most likely to involve an associative mechanism.