Recent high-resolution analysis of tubulin's structure has led to the prediction that the taxol binding site and a tubulin acetylation site are on the interior of microtubules, suggesting that diffusion inside microtubules is potentially a biologically and clinically important process. To assess the rates of transport inside microtubules, predictions of diffusion time scales and concentration profiles were made using a model for diffusion with parameters estimated from experiments reported in the literature. Three specific cases were considered: 1) diffusion of αβ-tubulin dimer, 2) diffusion/binding of taxol, and 3) diffusion/binding of an antibody specific for an epitope on the microtubule's interior surface. In the first case tubulin is predicted to require only ~1 min to reach half the equilibrium concentration in the center of a 40 μm microtubule open at both ends. This relatively rapid transport occurs because of a lack of appreciable affinity between tubulin and the microtubule inner surface and occurs in spite of a three-fold reduction in diffusivity due to hindrance. By contrast the transport of taxol is much slower, requiring days (at nM concentrations) to reach half the equilibrium concentration in the center of a 40 μm microtubule having both ends open. This slow transport is the result of fast, reversible taxol binding to the microtubule's interior surface and the large capacity for taxol (~12 mM based on interior volume of the microtubule). An antibody directed toward an epitope in the microtubule's interior is predicted to require years to approach equilibrium. These results are difficult to reconcile with previous experimental results where substantial taxol and antibody binding is achieved in minutes, suggesting that these binding sites are on the microtubule exterior. The slow transport rates also suggest that microtubules might be able to serve as vehicles for controlled- release of drugs.
Bibliographical noteFunding Information:
Acknowledgements The author wishes to thank Ken Downing for stimulating this work. The author is also grateful to Michael Caplow for sharing unpublished data and to Ted Salmon for helpful comments. The work was supported by a Biomedical Engineering Research Grant from the Whitaker Foundation.