Numerical modelling of the action of acetolactate synthase isozyme II using simplex optimization

Abstract

This stimulation investigates the dominant mode of action of acetolactate synthase isozyme II. Several models are considered. A variable step size, fourth-order Runge-Kutta-Fehlberg method integrated the coupled sets of stiff nonlinear rate equations using an NEC PowerMate SX personal computer fitted with a 16-MHz, 32-bit 80386/387 Intel co-processor. The Nelder-Mead simplex method optimized rate constants through a nonlinear least squares procedure. An adapted Larsson-Pardue, technique initialized rate parameter sets. In all cases, the simplex procedure converged by minimizing the sum of the squared errors. In addition, several other convergence criteria were simultaneously applied to ensure model integrity and allow valid cross model comparison, the most stringent being that the absolute relative error of each point be less than the experimental error. Models considered included: the classic case proposed by Schloss et al.; a single path concerted mechanism with complete substrate coverage: asymmetric dual path action with an without concerted substrate conversion to desorbed product; and multiple site action of the A2B2 dimer with different efficiencies at each. This model optimized all convergence criteria better than any other but subsequent work showed poor inhibitory action. The simplest model which optimized both inhibition and convergence criteria reasonably well had an asymmetric independent dual path action, with the possibility of either direct or indirect production of the desorbed product from the activated substrate complex.