Can yeast glycolysis be understood in terms of in vitro kinetics of the constituent enzymes? Testing biochemistry.

Model Status

This CellML runs in both PCEnv and COR to recreate the published results. The units have been checked and they are consistent. This particular CellML model is based on a Matlab file which, in turn, was exported from the working SBML file in the BioModels database. There are several subtle differences between the published paper and the working Matlab code which have been previously documented by the BioModels curators. In brief these are: in the kinetic law for ADH the species a should denote NAD and b Ethanol the last term in the equation should read bpq/(Kib Kiq Kp) and the rate equation for the PFK contains two errors: the R term should read: R = 1 + lambda1 + lambda2 + gR lambda1 lambda2 and the last term in the L equation should be squared. Also there are slight differences in the parameterisation of the model depending on whether it comes from the JWS or the BioModels database and, for Vmax values, please note that there is a conversion factor of approxinately 270 to convert from U/mg-protein as shown in Table 1 of the paper to mmol/(min*L_cytosol).

Model Structure

In April 2003 the human genome sequence was completed. One of the greatest challenges currently facing biologists is the elucidation of the molecular function of the proteins encoded by the genes. Traditionally, the approach has been to isolate these proteins and then characterise them in vitro. More recently, scientists have begun to question whether this will generate an accurate picture of how these proteins behave within a functional cell (in vivo). It is highly likely that the environment within a living cell is very different from the conditions within a test tube. Interacting metabolites, intracellular compartmentation, and enzyme concentrations are all likely to differ. The question to be asked is whether these differences between in vitro and in vivo conditions seriously affect scientists' understanding of the functional behaviour of the living cell.

The paper described here by Teusink et al. (2000) asks the question: can the behaviour of biochemical pathways be described by simply combining the properties of the components in isolation? In other words, can the in vivo behaviour of yeast glycolysis be understood in terms of the in vitro kinetic properties of the constituent enzymes? Based on in vitro, experimentally determined enzyme kinetics data, they developed a mathematical model of the glycolysis pathway in the yeast Saccharomyces cerevisiae. The simulation results from this model were then compared with recorded fluxes and metabolite levels in living cells under similar conditions. They employed mathematical modelling to test whether knowledge of the in vitro kinetic properties of the glycolytic enzymes allows a good understanding of the overall pathway behaviour in vivo.

The model includes most enzymes in the glycolysis pathway (see the figure below). A set of ordinary differential equations was used to represent the time-dependence of the metabolite concentrations. Reversible Michaelis-Menten and irreversible Hill kinetics were amongst the types of enzymes kinetics included in the model equations.

The complete original paper reference is cited below:

Can yeast glycolysis be understood in terms of in vitro kinetics of the constituent enzymes? Testing biochemistry, Bas Teusink, Jutta Passarge, Corinne A. Reijenga, Eugenia Esgalhado, Coen C. van der Weijden, Mike Schepper, Michael C. Walsh, Barbara M. Bakker, Karel van Dam, Hans V. Westerhoff, and Jacky L. Snoep, 2000, European Journal of Biochemistry , 267, 5313-5329. Full text and PDF versions of the article are available to journal subscribers.) PubMed ID: 10951190

A schematic diagram of the reactions included in the Teusink et al. 2000 model of the glycolysis pathway in the yeast Saccharomyces cerevisiae.

Model simulations revealed that the orginal question posed does not have a simple answer. For half of the enzymes studied, the in vitro kinetics did describe the in vivo activity. However, the other half of the enzymes analysed, they were larger deviations between their in vitro and in vivo characteristics.