We have previously shown that rat brain tubulin, a heterodimer consisting of an a and fi monomer, can be covalently labeled with [3H]colchicine by near UV irradiation. Most of the label appears in (&tubulin. We show here that (3-tubulin can be separated and purified from SDS preparative gels and analyzed by proteolysis. Chymotrypsin yielded a labeled -4-kDa band that contained two peptides. Tryptic digestion also yielded an =4-kDa band containing two peptides. Sequence analysis revealed a peptide ofresidues 1-36 and 213-242 for chymotrypsin and a peptide of residues 1-46 and 214-241 for trypsin. To identify which peptide carried the label, limited hydrolysis of -tubulin was done with trypsin; this procedure yielded a labeled 16-kDa N-terminal peptide and a 35-kDa C-terminal peptide, as identified by antibodies. Isolation of these peptides and extensive digestion with trypsin yielded two labeled peptides corresponding to residues 1-46 from the 16-kDa N-terminal fragment and residues 214-241 from the 35-kDa C-terminal fragment. These results show that at least two regions in P-tubulin are specifically involved in colchicine binding and that the span of the colchicine molecule, s11 A, bridges these two regions in the native /3 monomer.Colchicine is the most studied of all antimitotic agents. It binds to tubulin with a stoichiometry of one and inhibits microtubule assembly substoichiometrically. Colchicine binding to tubulin exhibits pseudoirreversible kinetics; it displays a fast step followed by a slow step involving conformational changes of both ligand and tubulin. The latter, in turn, promote fluorescence characteristic of the tropolone moiety of colchicine. Many analogs have been studied, resulting in the proposal that the A and C rings of colchicine both bind to specific loci, whereas the B ring is primarily a regulator of binding kinetics (see refs. 1 and 2 and the references therein).Attempts to localize the binding site(s) in the tubulin dimer have led to conflicting results. Studies with bromacetylcolchicine indicated binding to a-tubulin (3), but these results have been criticized on the basis of the nonspecificity of the alkylation reaction. Proteolytic studies of the colchicinetubulin complex yielded a-tubulin-derived, colchicinebearing peptides of 16-18 kDa; because z90% of the label bound to the protein was lost during processing, conclusions are uncertain (4). Others (5) have also shown colchicine localization to a-tubulin by using a photoaffinity label with a long spacer arm. A shorter spacer arm led to colchicine binding on both a-and ,B-tubulins (6), suggesting that localization was a function ofthe spacer length. On the other hand, Luduena and Roach (7) [3H]colchicine that the label is localized almost exclusively on 3-tubulin. In late stages of irradiation, or with "damaged" tubulin, label also appeared in a-tubulin (9). These results suggested the possibility that the colchicine-binding domain on 3-tubulin might be near a-tubulin. That proposal was strengthened by the finding that co...