Effect of reversible reactions on isotope label redistribution

Analysis of the pentose phosphate pathway


  • Correspondence to G. Stephanopoulos, Department of Chemical Engineering, Massachusetts Institute of Technology, Cambridge, MA 02139, USA

  • Fax: +1 617 253 3122.

  • E-mail:gregstep@mit.edu

  • Abbreviations. Ery4P, erythrose 4-phosphate; Gra3P, glyceraldehyde 3-phosphate; Hex6P, hexose pool consisting of glucose 6-phosphate and fructose 6-phosphate; Pen5P, pentose pool consisting of ribulose 5-phosphate, ribose 5-phosphate, and xylulose 5-phosphate; Sed7P, sedoheptulose 7-phosphate.

  • Enzymes. Glucose-6-phosphate 1-dehydrogenase ( EC1.1.1.49); ribose-5-phosphate isomerase ( EC5.3.1.6); ribulose-phosphate 3-epimerase ( EC5.1.3.1); transketolase ( EC2.2.1.1); transaldolase ( EC2.2.1.2); glucose-6-phosphate isomerase ( EC5.3.1.9); 6-phosphofructokinase ( EC2.7.1.11); fructose-bisphosphate aldolase ( EC4.1.2.13); triose-phosphate isomerase ( EC5.3.1.1).


The pentose phosphate pathway plays several key roles in metabolism including supply of biosynthetic carbon skeletons and reducing power. Previous research has focused on determining the fluxes through the reactions of this pathway using carbon-labeled substrates and models that make certain assumptions about the reversibility of the transketolase and transaldolase reactions in the nonoxidative pathway. These assumptions, however, have resulted in inconsistencies between the predicted carbon label distributions using these models and those determined experimentally. A general metabolic reaction network model developed in this paper and applied to the pentose phosphate pathway not only incorporates reaction reversibility but also accounts for the effect of individually varying extents of reaction reversibility on labeled carbon fractional enrichment values for intermediate metabolites. In addition, an algorithm is presented that can be used to calculate the three individual transaldolase and transketolase extents of reversibility. The results of this method show that varying extents of reaction reversibility have an observable effect on the metabolite carbon label distributions which can in turn affect flux calculation for other parts of the metabolic network such as the tricarboxylic acid cycle. In addition, the observability of reversibility extent and accuracy of flux calculations depend on the particular choice of metabolite carbon enrichments measured. In particular, [6-13C]hexose 6-phosphate and [4-13C]erythrose 4-phosphate carbon enrichment values resulting from [1-13C]glucose feeding contained more information as compared to those from ribose 5-phosphate. This analysis was applied to literature data of metabolite carbon labeling that resulted from supplying either 13C- or 14C-enriched substrates to several cell types growing under various conditions. The specific activities of metabolite carbon atoms taken from rat epididymal adipose tissue, goosefish islet cells, Corynebacterium glutamicum, and Escherichia coli supplied with either [2-14C]glucose or [1-13C]glucose demonstrate how reversibility is present in the pentose phosphate pathway and the extents of reversibility can be estimated from labeled carbon data sets.