Comparison of computer simulations of the F-type and L-type non-oxidative hexose monophosphate shunts with 31P-NMR experimental data from human erythrocytes.

Mathematical modelling was used to predict the behaviour of the two most favoured schemes for the operation of the non-oxidative hexose monophosphate shunt (HMS), the F-type and the L-type pathways. The models simulate the time courses of sugar-phosphate concentrations when various substrates are metabolized via each pathway. A 31P-NMR technique, with which to observe time courses of concentrations of sugar phosphates in a human red cell lysate, was developed. The accuracy of each hypothesised scheme was then evaluated by comparing predicted with observed data. The results were more consistent with time courses of sugar-phosphate levels predicted by the F-type (classical) pathway than those predicted by the L-type model. However, the accumulation of sedoheptulose 1,7-bisphosphate when a haemolysate was incubated with ribose 5-phosphated showed that the F-type pathway is not a complete description of the system of reactions. Transaldolase was demonstrated to be essential for the normal metabolism of sugar phosphates by haemolysates. The effects of the heat-inactivation of transaldolase on the metabolism of sugar phosphates were accurately predicted by the F-type model. The relevance of attempting to describe the reaction of the non-oxidative HMS as a distinct 'pathway' or 'cycle' is discussed.