The maximum achievable tyrosination level of neurotubulin, in vitro, is about 50%. We have developed a method to obtain a complete separation of the tyrosinatable and nontyrosinatable species. We use an immunoaffinity column, with coupled YL 1/2 monoclonal antibody (anti-Tyr-tubulin) and rapid desalting methods. Both subpopulations can be obtained in a polymerizable, apparently native, form. We find that about 35% of the brain tubulin is truly nontyrosinatable, despite the fact that it is assembly competent. Using a polyclonal antibody directed against nontyrosinatable tubulin, we find that it recognizes a specific epitope on the alpha-subunit of the dimer. The existence of an abundant tubulin subspecies, structurally different from tyrosinatable tubulin, should obviously be kept in mind in immunofluorescence studies of the distribution of nontyrosinated tubulin in brain tissues. Furthermore, we have extensively investigated the effect of tubulin tyrosination on microtubule dynamics. Despite the homogeneity of the populations under comparison, we find no significant effect of tyrosination on microtubule dynamics. Similarly, the stabilizing effects of microtubule associated proteins and of STOP protein were identical in both subpopulations. The drug taxol seems more efficient in stabilizing detyrosinated microtubules, but the difference is moderate. Taken together, these findings suggest that tubulin tyrosination does not effect microtubule stabilization, neither through modifications of the intrinsic tubulin properties nor through a differential binding of stabilizing proteins. Finally, the complete separation of two tubulin species (tyrosinated or detyrosinated) with similar kinetic properties, but immunologically different, should be of value in many kinetic studies of microtubule assembly.