Measurement of the affinity of microtubules for the anti-cancer drug taxol is problematic, because microtubules are not stable at the very low concentrations required to detect taxol dissociation. We have circumvented this problem by using the GTP analogue GMP-CPP (guanylyl alpha, beta-methylenediphosphonate), which renders microtubules sufficiently stable to allow binding studies with nonsaturating concentrations of taxol. AKd value equal to about 10 nM was estimated from the effect of taxol concentration on the dilution-induced disassembly rate and on the binding of [3H]taxol. With GTP-microtubules the Kd value for taxol binding by tubulin-GDP subunits in the core of the microtubule appears to be comparable with that of GMPCPP-microtubules. However, the stabilizing effect of the drug bound to tubulin subunits that arrive at ends of disassembling microtubules is attenuated by a two-step reaction sequence in which taxol dissociates (k = 30 s-1), followed by rapid (k = 1000 s-1) loss of the taxol-free tubulin subunit. This sequential reaction can be disrupted by high (micromolar) concentrations of taxol, which react rapidly with tubulin subunits at the ends of microtubules (k = 2 x 10(9) M-1 s-1). The inhibitory effect of taxol on microtubule disassembly at concentrations a thousand-fold greater than the Kd value suggests the desirability of using high taxol concentrations in chemotherapy with this compound.