To examine the behavior of microtubule-associated proteins (MAPs) in living cells, MAP 4 and MAP 2 have been derivatized with 6-iodoacetamido-fluorescein, and the distribution of microinjected MAP has been analyzed using a low light level video system and fluorescence redistribution after photobleaching. Within 1 min following microinjection of fluoresceinated MAP 4 or MAP 2, fluorescent microtubule arrays were visible in interphase or mitotic PtK1 cells. After cold treatment of fluorescent MAP 2-containing cells (3 h, 4 degrees C), microtubule fluorescence disappeared, and the only fluorescence above background was located at the centrosomes; microtubule patterns returned upon warming. Loss of microtubule immunofluorescence after nocodozole treatment was similar in MAP-injected and control cells, suggesting that injected fluorescein-labeled MAP 2 did not stabilize microtubules. The dynamics of the MAPs were examined further by FRAP. FRAP analysis of interphase cells demonstrated that MAP 2 redistributed with half-times slightly longer (60 +/- 25 s) than those for MAP 4 (44 +/- 20 s), but both types of MAPs bound to microtubules in vivo exchanged with soluble MAPs at rates exceeding the rate of tubulin turnover. These data imply that microtubules in interphase cells are assembled with constantly exchanging populations of MAP. Metaphase cells at 37 degrees C or 26 degrees C showed similar mean redistribution half-times for both MAP 2 and MAP 4; these were 3-4 fold faster than the interphase rates (MAP 2, t1/2 = 14 +/- 6 s; MAP 4, t1/2 = 17 +/- 5 s). The extent of recovery of spindle fluorescence in MAP-injected cells was to 84-94% at either 26 or 37 degrees C. Although most metaphase tubulin, like the MAPs, turns over rapidly and completely under physiologic conditions, published work shows either reduced rates or extents of turnover at 26 degrees C, suggesting that the fast mitotic MAP exchange is not simply because of fast tubulin turnover. Exchange of MAP 4 bound to telophase midbodies occurred with dynamics comparable to those seen in metaphase spindles (t1/2 = approximately 27 s) whereas midbody tubulin exchange was slow (greater than 300 s). These data demonstrate that the rate of MAP exchange on microtubules is a function of time in the cell cycle.