During
the past decade, genome-scale stoichiometric models have enabled the
exploration and quantification of metabolic phenotypes by using constraint-based
modeling methods. Such methods have proved to be extremely useful in
characterizing metabolic behaviours under different environmental conditions as
well as predicting the result of particular genetic modifications. Although
stoichiometric models enable a diversity of analysis, their predictions are
limited by lack of enzyme kinetics. Kinetics is frequently represented by simplified
approximate formulas (e.g. mass action, reversible Michaelis-Menten, etc.).
Many real kinetic behaviours are ignored by these approaches. Moreover, the
full thermodynamic relationship between parameters is lost which complicates
sampling feasible parameter sets. Here, we present a General Reaction Assembly
and Sampling Platform (GRASP), which enables parameterizing and sampling
kinetic models by integrating the generalized MCW model with the elementary reaction formalism. By formulating the appropriate thermodynamic constraints, we developed
a framework that enables parameterization of simple as well as complex
reaction mechanisms, such as random-order and allosteric mechanisms, without
sacrificing complexity or using simplifying assumptions. This thermodynamically
consistent parameterization incurs no loss in generality and can be efficiently
sampled employing standard probability distributions. We used the sampled models to explore the full kinetic space of reactions and to assess the impact of
thermodynamics and reaction mechanisms on kinetics. Moreover, we applied this
approach to a tightly-regulated small network comprising the methionine cycle
in human hepatocytes. Our results reinforce the notion that biological
meaningful kinetic parameter sets are constrained to a small portion of the
kinetic space allowable by thermodynamics.