Chemistry And Function Of Microtubules

Wolff, Jan

Abstract

In an attempt to understand post-translational modifications of tubulin, particularly lipid modifications on cysteines, it became necessary to investigate the 20 thiols of the tubulin heterodimer. Alkylation reactions with iodoacetamide or N-ethyl maleimide revealed that only 5 of 12 cysteines of a tubulin and 2 of 8 cysteines of ? tubulin reacted in the native protein. This was shown to be due to the presence, within 6.5?, of positively charged lysines, arginines or aromatic ring edges. Negatively charged environments prevented alkylation suggesting an apparent lack of accessibility except for certain unreactive surface cysteines. The electrostatic environment of cysteines is an important rate-determining factor. With reagents forming disulfide bonds (i.e. disulfide interchange), all 20 thiols react in the native protein and reaction rates are much faster than with ioidoacetamide. Although the reaction seems to be biphasic we show that it is much more complex and required stopped flow instrumentation. For Ellman?s reagent (DTNB) under pseudo first-order conditions curve could not resolve the kinetics with sums of exponentials despite correlation coefficients of 0.99. We looked for and found numerous side reactions resulting in protein/protein disulfide bond formation ? both intra- and inter- monomer - seen in non-reducing gels. These side reactions a sensitive to G nucleotides, pH, concentration and the nature of the reagent. By contrast, the related reagent, n-octyl mercaptan thionitrobenzoate showed no side reactions (because the mercaptan is a poor leaving group) and this permitted satisfactory kinetic analysis as the sum of three exponential terms. Except for possible dependence on structural fluctuations these disulfide reagents are less sensitive to local protein factors but more dependent on reagent properties such as the nature of the leaving group and hydrophobicity. We have been less successful in the study of thioester formation by palmitoyl CoA. Attempts using S. cerevisiae were unsuccessful because of low yields and massive hydrolysis. An HPLC method for determining highly hydrophobic peptides has been developed in collaboration with Dr. Jin Kim of the Mass Spec section. This is being used for auto- palmitoylated tubulin and hopefully later for in situ palmitoylated tubulin. Preliminary findings suggest that substitution occurs in cysteines different from alkylation. We have started to investigate the mysterious role of tubulin in apoptosis with a summer student, Tim Panosian. Antimitotic drugs, both polymerizing and depolymerzing, induce apoptosis in various cells. This effect has been blamed on changes in microtubule dynamicity, drug actions on tubulin bound to mitochondria, or direct action of the drugs on Bcl2 (taxol). Using monoclonal antibodies. We find that tubuilin is present in mitochondria fo HeLa and PC-12 cells but is hard to detect on liver mitochondria. Bcl2 shows a similar distribution. More examples of mitochondria are needed. A Bcl2 fusion protein is slightly incorporated into microtubules. Other interactions between these two proteins are under study as is the nature of the attachment of tubulin to mitochondria.