In 1963, improved fixation methods led to the definitive identification of microtubules (see “Microtubules get a name” JCB. 168:852). Just one year later, Gary Borisy embarked on a daring project to isolate the main component of those microtubules.
The effort was initiated by Edwin Taylor at the University of Chicago. Taylor was interested in using colchicine to study mitosis. Unfortunately, “the literature on the effects of colchicine was very confused,” says Borisy. Colchicine was known to destroy the mitotic spindle but could also inhibit a disparate collection of other processes including distracting oddities such as cellulose deposition in plants.
Colchicine did, however, bind with high affinity and simple kinetics to cells, suggesting that isolation of a complex of colchicine with its binding protein(s) should be possible (Taylor, 1965). As a new graduate student in Taylor's lab, Borisy “got very excited” by the prospect of finding the colchicine-binding protein. “I begged to do this as a thesis project,” says Borisy. “He said no, no, no, it's too risky. But I begged to do it.” Others told him, “ ‘You’ll have nonspecific binding and it'll be a mess.' But what did I know—I was a student.”
Colchicine did turn out to be specific, and Borisy did succeed in isolating a colchicine-binding activity from extracts of tissue culture cells (Borisy and Taylor, 1967a). The highest binding activity came from dividing cells, the isolated mitotic apparatus (Borisy and Taylor, 1967b), cilia, sperm tails (Shelanski and Taylor, 1967), and brain tissue. The brain tissue was a temporary scare: “It seemed like, oh my goodness, this is a totally nonspecific binding reaction and it's a mess,” says Borisy. But there was a common denominator in that all the sources had an abundance of microtubules.
Further correlation came when the colchicine-binding activity was extracted under low salt conditions that led to the disappearance of microtubules (Borisy and Taylor, 1967b). The group took pains to measure detailed in vitro binding kinetics and show that they matched those seen for intact cells, where colchicine disassembled microtubules. “The results,” concluded the authors, “are consistent with the hypothesis that the binding site is the subunit protein of microtubules.”
For now, the protein was nameless. “We did not give it a name, which was a blunder,” says Borisy. Although tubulin was the obvious candidate given the existing name of microtubules, “it sounded so jarring to our ears.”
But soon enough Mohri (1968) “gave it the obvious name—the name we considered and rejected.” The term “tubulin” was now official, although “spactin,” “flactin,” and “tektin” stuck around as alternative monikers for a little while (Satir, 1968).
As purification from sperm tails (Shelanski and Taylor, 1968) and then brain (Weisenberg et al., 1968) continued, “there were many red herring findings and inconsistent findings and blind alleys,” says Borisy. Initially there were candidate microtubule proteins of very different sizes from both Daniel Mazia's study of the mitotic apparatus and Ian Gibbon's study of Tetrahymena cilia. The Mazia candidate turned out to be a yolk protein contaminant, and the Gibbons group at Harvard was probably looking at a monomeric version of what Taylor's group was isolating under less-denaturing conditions (Gibbons, 1963; Renaud et al., 1968).
For Gibbons, cilia had the advantage of having a vast excess of tubulin over other proteins, although early on it was far from certain that the same main component would be the basis of “microtubules” from flagella, cilia, the cytoplasm, and spindles. Not only did this turn out to be the case, but the Gibbons group also spotted that the microtubules consisted of two closely related proteins (Renaud et al., 1968).
By 1966, the Taylor group had a protein that was pure to homogeneity