The stabilization of microtubules in isolated spindles by tubulin‐colchicine complex

We have analyzed the effect of colchicine and tubulin dimer‐colchicine complex (T‐C) on microtubule assembly in mitotic spindles. Cold‐ and calcium‐labile mitotic spindles were isolated from embryos of the sea urchin Lytechinus variegatus employing EGTA/glycerol stabilization buffers. Polarization microscopy and measurements of spindle birefringent retardation (BR) were used to record the kinetics of microtubule assembly‐disassembly in single spindles. When isolated spindles were perfused out of glycerol stabilizing buffer into a standard in vitro microtubule reassembly buffer (0.1 M Pipes, pH 6.8, 1 mM EGTA, 0.5 mM MgCl₂, and 0.5 mM GTP) lacking glycerol, spindle BR decreased with a halftime of 120 s. Colchicine at 1 mM in this buffer had no effect on the rate of spindle microtubule disassembly. Inclusion of 20 μM tubulin or microtubule protein, purified from porcine brain, in this buffer resulted in an augmentation of spindle BR. Interestingly, in the presence of 20 μM T‐C, spindle BR did not increase, but was reversibly stabilized; subsequent perfusion with reassembly buffer without T‐C resulted in depolymerization. This behavior is striking in contrast to the rapid depolymerization of spindle microtubules induced by colchicine and T‐C in vivo. These results support the current view that colchicine does not directly promote microlubule depolymerization. Rather, it is T‐C complex that alters microtubule assembly, by reversibly binding to microtubules and inhibiting elongation. In vivo, colchicine can induce depolymerization of nonkinetochore spindle microtubules within 20 s. In vitro, colchicine blocks further microtubule assembly, but does not induce rapid disassembly. The rate of tubulin dissociation from spindle microtubules in vitro in reassembly buffer without soluble tubulin is about 20 times slower than the rate of dissociation in vivo when assembly is blocked abruptly by T‐C. The rate of tubulin dissociation from the spindle microtubules may determine their response to T‐C, since the tubulin dissociation rate in vivo is about 12 times faster than the rate measured here for spindle microtubules in standard microtubule reassembly buffer at physiological temperature.