Research

Page 3

Suggestions for Calculations

The free energy per molecule stored in the proposed concave asymmetric brush of polymer loops needs to be compared to the energy stored in the intertubulin bonds (total ~15 kBT/molecule or ~10 kcal/mol), 4 out of 6 of which are broken upon disassembly. A crude estimate, counting 1 kBT for every "Pincus blob" in a flat, linear array of self-avoiding 42-mers (0.3 nm monomers) 2 nm apart, is ~5 kBT/molecule (~4 kcal/mol). The implication is that only a modest energy ~3 kBT/molecule (~2 kcal/mol) must be overcome to initiate disassembly.
A serious calculation would also clarify the quantitative dependence of the brush free energy on experimental parameters, such as temperature, solvation, microtubule radius (i.e. protofilament number) and domain size. This information should be most useful in predicting characteristics of dynamic instability.
Finally, a more elaborate calculation is called for to interpret the effects of such parameters on microtubule bending stiffness.

Suggestions for Experiment

The most direct test of the hypothetical random-coil domain would involve genetic modification of the putative domain itself. For example, deleting a stretch of 10 amino acids should reduce the effective radius of the domain from 2.7 nm to 2.6 nm, enough to change the radius of protofilaments from 19 to 23 nm (easily detectable in the electron microscope) but probably not enough to interfere with the overall folding/expression of the protein. Reducing the domain size may also noticeably diminish microtubule dynamics...
Another direct approach would involve visualization of the domain by atomic force microscopy. AFM is uniquely suited to detect the soft signature of a random-coil domain. We have already begun to image the large 2-dimensional crystalline sheets of tubulin that form in the presence of Zinc. Given the anti-parallel arrangement of protofilaments in the Zinc sheet structure, we expect to see soft regions in an 8 nm lattice superimposed on the 4 nm tubulin monomer lattice.
A somewhat less direct test of the hypothesis would measure the size of the change in the radius of gyration of tubulin dimers in solution with either GTP or GDP using, for example, dynamic light scattering.
One particularly counter-intuitive prediction of the random-coil hypothesis involves the relation between nucleation rate and size of the critical aggregate. A standard kinetic analysis of microtubule nucleation predicts that the lower the tubulin concentration, for example, the smaller the typical sub-critical aggregate. Therefore, microtubules should be more likely to nucleate with fewer protofilaments. However, if random-coil domains are released on the sheet-like sub-critical aggregates, they would work against its natural curvature, making a greater number of protofilaments necessary for closure. Thus, the random-coil hypothesis predicts that the slower assembly rate, the more random-coil domains released per sub-critical sheet, and the larger a sheet will have to be before it closes to form a microtubule.

Suggestions for Dynamic Instability

Knowledge of the conformational change in tubulin should lend insight into microtubule disassembly and suggest criteria for transitions to and from the disassembling state, the hallmark of dynamic instability. Current speculations are summarized below.

Disassembly is... driven by the free energy stored in the brush. Perturbations which do nothing more than lower the free energy of the brush should reduce the disassembly rate and have no affect on assembly. We have already observed such an effect with increasing concentrations of glycerol, a poor solvent for polypeptides. Others have seen a corresponding sensitivity on proportion of GTP hydrolyzed. It remains to test a larger variety of poor solvents (e.g., DMSO, D2O) and develop a quantitative interpretation of the results.

Catastrope is... a crack. If a pair of protofilaments at the growing end of a microtubule extend beyond the rest AND if random coil domains are released along their length, they may separate and begin curling away fromthe microtubule. If this "crack" propagates into the bulk of the microtubule, weakening lateral bonds below the energy required to contain the brush, a cascade of protofilament separation should follow.

Rescue is... a blunt end. As protofilaments peel apart, they sometimes break. If, by chance, all protofilaments break at the same location, creating a blunt end, their full strength lateral bonds will continue to confine the brush and disassembly will halt, giving opportunity for growth to resume.

These speculations are in the process of being refined and tested with the help of precise measurements on induced and spontaneous disassembly events in microtubules.

The Big Picture.....

Dynamic instability is an intriguing phenomenon that begs to be understood. If the hypothesis presented here proves true, the physics of polymer brushes may find new application in understanding the biology of proteins and chemotherapeutics which influence dynamic instability.

Tubulin is a "primitive" protein: present in all nucleated cells, derived from a prokaryote homologue, and relatively unchanged by evolution. Its design may well reflect a time of relatively limited biochemical complexity when simplicity of conformational change would be a plausible prerequisite. Unfolding is arguably the simplest conformational change.
A biological instance of unfolding as conformational change would validate a new scientific approach to protein folding and function. The vast array of molecular machinations might be projected onto a small set of such "simple" conformational changes. Then, design constraints on those conformational changes might illuminate the evolutionary history of protein diversification and guide human creativity in protein engineering and other molecular designs.

Page 3

Return to Main Research Page

RESEARCH