This is the first in a series of three papers [see also Mulquiney and Kuchel (1999) Biochem. J. 342, 579-594; Mulquiney and Kuchel (1999) Biochem. J. 342, 595-602] that present a detailed mathematical model of erythrocyte metabolism which explains the regulation and control of 2,3-bisphosphoglycerate (2,3-BPG) metabolism. 2,3-BPG is a modulator of haemoglobin oxygen affinity and hence plays an important role in blood oxygen transport and delivery. This paper presents an in vivo kinetic characterization of 2,3-BPG synthase/phosphatase (BPGS/P), the enzyme that catalyses both the synthesis and degradation of 2,3-BPG. Much previous work had indicated that the behaviour of this enzyme in vitro is markedly different from that in vivo. (13)C and (31)P NMR were used to monitor the time courses of selected metabolites when erythrocytes were incubated with or without [U-(13)C]glucose. Simulations of the experimental time courses were then made. By iteratively changing the parameters of the BPGS/P part of the model until a good match between the NMR-derived data and simulations were achieved, it was possible to characterize BPGS/P kinetically in vivo. This work revealed that: (1) the pH-dependence of the synthase activity results largely from a strong co-operative inhibition of the synthase activity by protons; (2) 3-phosphoglycerate and 2-phosphoglycerate are much weaker inhibitors of 2,3-BPG phosphatase in vivo than in vitro; (3) the K(m) of BPGS/P for 2,3-BPG is significantly higher than that measured in vitro; (4) the maximal activity of the phosphatase in vivo is approximately twice that in vitro, when P(i) is the sole activator (second substrate); and (5) 2-phosphoglycollate appears to play no role in the activation of the phosphatase in vivo. Using the newly determined kinetic parameters, the percentage of glycolytic carbon flux that passes through the 2, 3-BPG shunt in the normal in vivo steady state was estimated to be 19%.