D-Glyceraldehyde 3-phosphate: NADP+ oxidoreductase, nonphosphorylating (GNR; EC 1.2.1.9) purified from spinach leaves was investigated by initial velocity analysis. The hyperbolic saturation curves became nonhyperbolic when NADP+ was varied at elevated D-glyceraldehyde 3-phosphate (G3P) concentrations (sigmoidicity) or when G3P was varied at low NADP+ concentrations (pseudo-substrate inhibition), suggesting a random bi bi mechanism (Scagliarini et al. Plant Physiol. 94, 1337-1344, 1990). Free ATP was a linear competitive inhibitor of both NADP+ with KI 0.5 +/- 0.2 mM (SD) and G3P with KI 3.2 +/- 0.2 mM as determined by data in the hyperbolic range of responses when the nonvaried substrate was saturating. Similarly ADP inhibited competitively with KI 1.9 +/- 0.4 mM (NADP+) and 3.5 +/- 0.5 mM (G3P). Inhibition was mixed-type when the nonvaried substrate was below saturation. ATP, but not ADP, tended to enhance the nonhyperbolic behavior of GNR, resulting in potentiated inhibition at high [G3P]/[NADP+] ratios. The Mg-chelated form of ATP was less effective. The rate equation of a steady state random bi bi reaction mechanism in the presence of a dead-end inhibitor was derived. Suitable values of the rate constants were chosen to fit the kinetic data for the uninhibited enzyme. These values and the measured inhibition constants inserted in the rate equation can satisfactorily account for the nonhyperbolic inhibition patterns of ATP and ADP. The generalized model represents a possible alternative to allosteric models in interpreting nonlinear kinetics and dead-end inhibition of two-substrate enzymes.