Characterization and in vitro polymerization of Tetrahymena tubulin

Tetrahymena tubulin was purified from the cell extract using DEAE-Sephadex A-50 ion-exchanger and ammonium sulfate precipitation. About 2.2% of the total protein in the 20,000 X g supernatant was recovered as DEAE-Sephadex-purified tubulin fraction. Applying the temperature-dependent polymerization-depolymerization method to this fraction in the presence of Tetrahymena outer fibers as a seed, almost pure tubulin was obtained. Tetrahymena tubulin dimer showed different behavior on SDS-polyacrylamide gels from porcine brain tubulin, and showed very low affinity for colchicine, amounting to about one-twentieth of the binding to porcine brain tubulin. The tubulin fraction failed to polymerize into microtubules by itself. Addition of a small amount of the ciliary outer fiber fragment induced polymerization as demonstrated by viscometric measurements, but the reconstituted microtubules were very unstable in the absence of glycerol. Microtubule-depolymerizing agents such as Ca2+ ions, low temperature, or colchicine all inhibited in vitro polymerization. Although Tetrahymena tubulin purified by the polymerization-depolymerization method could copolymerize with porcine brain microtubules, the DEAE-Sephadex-purified tubulin fraction suppressed the initial rate of porcine brain microtubule assembly in vitro. There seemed to be no differences between cytoplasmic tubulin and outer fiber tubulin in colchicine binding activity or SDS-gel electrophoretic behavior, or between the fine structure of both reconstituted microtubules observed by electron microscopy.     

In vitro polymerization of flagellar and ciliary outer fiber tubulin into microtubules.

About 10--20% of the total protein in the outer fiber fraction was solubilized by sonication in a solution containing 5 mM MES, 0.5 mM MgSO4, 1.0 mM EGTA, 1.0 mM GTP, and 0 or 50 mM KC1 at pH 6.7. The sonicated extract was shown by analytical centrifugation to consist largely of a 6 S component (tubulin dimer), having a molecular weight of 103,000, as determined by gel filtration, and possessing a colchicine-binding activity of 0.8 mole per tubulin dimer. The tubulin fraction failed to polymerize into microtubules by itself. Addition of a small amount of the ciliary outer fiber fragments or reconstituted short brain microtubules, however, induced polymerization, as demonstrated by viscosity of flow birefringence changes as well as light or electron microscopic observations. The growth of heterogeneous microtubules upon mixing outer fiber tubulin with DEAE-dextran-decorated brain microtubules was observed by electron microscopy. Microtubules were reconstituted from outer fiber tubulin without addition of any nuclei fraction when a concentrated tubulin fraction was warmed at 35degree. A few doublet-like microtubules or pairs of parallel singlet microtubules that were closely aligned longitudinally could be observed among many singlet microtubules. Unlike other fiber microtubules, the reconstituted polymers were depolymerized by exposure to Ca2+ ions, high or low ionic strength, colchicine, low temperature or SH reagents. No microtubules were assembled under these conditions.     

Antimitotic antifungal compound benomyl inhibits brain microtubule polymerization and dynamics and cancer cell proliferation at mitosis, by binding to a novel site in tubulin

The antifungal agent benomyl [methyl-1-(butylcarbamoyl)-2-benzimidazolecarbamate] is used throughout the world against a wide range of agricultural fungal diseases. In this paper, we investigated the interaction of benomyl with mammalian brain tubulin and microtubules. Using the hydrophobic fluorescent probe 1-anilinonaphthalene-8-sulfonic acid, benomyl was found to bind to brain tubulin with a dissociation constant of 11.9 +/- 1.2 microM. Further, benomyl bound to at a novel site, distinct from the well-characterized colchicine and vinblastine binding sites. Benomyl altered the far-UV circular dichroism spectrum of tubulin and reduced the accessibility of its cysteine residues to modification by 5,5'-dithiobis-2-nitrobenzoic acid, indicating that benomyl binding to tubulin induces a conformational change in the tubulin. Benomyl inhibited the polymerization of brain tubulin into microtubules, with 50% inhibition occurring at a concentration of 70-75 microM. Furthermore, it strongly suppressed the dynamic instability behavior of individual brain microtubules in vitro as determined by video microscopy. It reduced the growing and shortening rates of the microtubules but did not alter the catastrophe or rescue frequencies. The unexpected potency of benomyl against mammalian microtubule polymerization and dynamics prompted us to investigate the effects of benomyl on HeLa cell proliferation and mitosis. Benomyl inhibited proliferation of the cells with an IC(50) of 5 microM, and it blocked mitotic spindle function by perturbing microtubule and chromosome organization. The greater than expected actions of benomyl on mammalian microtubules and mitosis together with its relatively low toxicity suggest that it might be useful as an adjuvant in cancer chemotherapy.     

Autopalmitoylation of tubulin

Pure rat brain tubulin is readily palmitoylated in vitro using [3H]palmitoyl CoA but no added enzymes. A maximum of approximately six palmitic acids are added per dimer in 2-3 h at 36-37 degrees C under native conditions. Both alpha and beta tubulin are labeled, and 63-73% of the label was hydroxylamine-labile, presumed thioesters. Labeling increases with increasing pH and temperature, and with low concentrations of guanidine HCl or KCl (but not with urea) to a maximum of approximately 13 palmitates/dimer. High SDS and guanidine HCl concentrations are inhibitory. At no time could all 20 cysteine residues of the dimer be palmitoylated. Polymerization to microtubules, or use of tubulin S, markedly decreases the accessibility of the palmitoylation sites. Palmitoylation increases the electrophoretic mobility of a portion of alpha tubulin toward the beta band. Palmitoylated tubulin binds a colchicine analogue normally, but during three warm/cold polymerization/depolymerization cycles there is a progressive loss of palmitoylated tubulin, indicating decreased polymerization competence. We postulate that local electrostatic factors are major regulators of reactivity of tubulin cysteine residues toward palmitoyl CoA, and that the negative charges surrounding a number of the cysteines are sensitive to negative charges on palmitoyl CoA.     

The binding of isocolchicine to tubulin. Mechanisms of ligand association with tubulin

Isocolchicine is a structurally related isomer of colchicine altered in the methoxytropone C ring. In spite of virtual structural homology of colchicine and isocolchicine, isocolchicine is commonly believed to be inactive in binding to tubulin and inhibiting microtubule assembly. We have found that isocolchicine does indeed bind to the colchicine site on tubulin, as demonstrated by its ability to competitively inhibit [3H]colchicine binding to tubulin with a KI approximately 400 microM. Isocolchicine inhibits tubulin assembly into microtubules with an I50 of about 1 mM, but the affinity of isocolchicine for the colchicine receptor site, 5.5 +/- 0.9 x 10(3) M-1 at 23 degrees C, is much less (approximately 500-fold) than that of colchicine. Unlike colchicine, isocolchicine binds rapidly, and the absorption and fluorescence properties of the complex are only modestly altered compared to free ligand. It is proposed that the binding of isocolchicine to tubulin may be rationalized either in terms of conformational states of colchicinoids when liganded to tubulin or by the structural requirements for C-10 substituents for high affinity binding to the colchicine receptor.     

Human brain tubulin purification: Decrease in soluble tubulin with age

The soluble tubulin of human cerebral cortex, as assessed by [³H]colchicine binding of the 100,000g supernatant fraction, decreases drastically with age, 75 percent from age 0 to age 90. There is also a considerably lower concentration of high molecular weight proteins in the soluble fraction of postmortem human cerebral cortex than in that of nonhuman species. Human brain tubulin can be polymerized into microtubules with DEAE-dextran. The DEAE-dextran induced microtubules are stable to cold temperature (4°) and calcium. However, in the presence of 1 M glutamate, the microtubules become cold labile and depolymerize at 4°. Thus we have developed a novel method for purifying polymerization competent tubulin from fresh or frozen human cerebral cortex. Human brain tubulin purified by our novel method is very similar to tubulin from the brains of other mammals in molecular weight, amino acid composition, polymerization-depolymerization parameters, and structural dimensions of the microtubules formed.Some aspects of this work have been published as an abstract in 1981. Fed. Proc. 40:1548.     

The natural naphthoquinone plumbagin exhibits antiproliferative activity and disrupts the microtubule network through tubulin binding

Plumbagin (5-hydroxy-2-methyl-1,4-naphthoquinone), a naphthoquinone isolated from the roots of Plumbaginaceae plants, has potential antiproliferative activity against several tumor types. We have examined the effects of plumbagin on cellular microtubules ex vivo as well as its binding with purified tubulin and microtubules in vitro. Cell viability experiments using human non-small lung epithelium carcinoma cells (A549) indicated that the IC 50 value for plumbagin is 14.6 microM. Immunofluorescence studies using an antitubulin FITC conjugated antibody showed a significant perturbation of the interphase microtubule network in a dose dependent manner. In vitro polymerization of purified tubulin into microtubules is inhibited by plumbagin with an IC 50 value of 38 +/- 0.5 microM. Its binding to tubulin quenches protein tryptophan fluorescence in a time and concentration dependent manner. Binding of plumbagin to tubulin is slow, taking 60 min for equilibration at 25 degrees C. The association reaction kinetics is biphasic in nature, and the association rate constants for fast and slow phases are 235.12 +/- 36 M (-1) s (-1) and 11.63 +/- 11 M (-1) s (-1) at 25 degrees C respectively. The stoichiometry of plumbagin binding to tubulin is 1:1 (mole:mole) with a dissociation constant of 0.936 +/- 0.71 microM at 25 degrees C. Plumbagin competes for the colchicine binding site with a K i of 7.5 microM as determined from a modified Dixon plot. Based on these data we conclude that plumbagin recognizes the colchicine binding site to tubulin. Further study is necessary to locate the pharmacophoric point of attachment of the inhibitor to the colchicine binding site of tubulin.     

Aging of tubulin at neutral pH: stabilization by colchicine and its analogues.

The effect of colchicine and its analogues, allocolchicine, 2,3,4-trimethoxy-4'-carbomethoxy-1,1'biphenyl, 2,3,4,4'-tetramethoxy-1,1'-biphenyl, 2,3,4-trimethoxy-4'-acetyl-1,1'-biphenyl, and tropolone methyl ether, on the aging process of tubulin has been examined. In contrast to the vinca alkaloid drugs which accelerate the formation of the paucidisperse 9 S polymers by a factor of 3.5, the colchicine class of ligands stabilize alpha,beta-tubulin. Less than 10% of the protein is transformed into the aggregates after 50 h of incubation in the presence of 1 x 10(-3) M colchicine, as compared to nearly 70-75% transformation in its absence. These results are supported by fluorescence examination of the retention of colchicine binding ability, as well as circular dichroism spectroscopy. In the presence of colchicine, the rate determining step is a conformational change, just as in its absence. The colchicine analogues which bind to tubulin in a rapidly reversible equilibrium were almost as effective in tubulin stabilization. Addition of vincristine to the system reduced the stability of the tubulin-colchicine complex. Furthermore, vincristine was found to have the same effects on the fresh complex as it does on pure tubulin; i.e., it induced the isodesmic linear polymerization and inhibited assembly into the microtubule-mimicking large polymers. This inhibition, however, was stoichiometric, whereas it is substoichiometric in the case of microtubules.     

Mechanism of inhibition of microtubule polymerization by colchicine: inhibitory potencies of unliganded colchicine and tubulin-colchicine complexes.

The tubulin-colchicine binding reaction appears to involve a number of intermediate steps beginning with rapid formation of a transient preequilibrium complex that is followed by one or more slow steps in which conformational changes in tubulin and colchicine lead to formation of a poorly reversible final-state complex. In the present study, we investigated the relative ability of unliganded colchicine and preformed final-stage tubulin-colchicine complex to incorporate at microtubule ends and to inhibit addition of tubulin at the net assembly ends of bovine brain microtubules in vitro. Addition of 0.1 microM final-stage tubulin-colchicine complex to suspensions of microtubules at polymer-mass steady-state resulted in rapid incorporation of one to two molecules of tubulin-colchicine complex per microtubule net assembly end concomitant with approximately 50-60% inhibition of tubulin addition. Incorporation of colchicine-tubulin complex continued slowly with time, without significant additional change in the rate of tubulin addition. In contrast, addition of unliganded colchicine to microtubule suspensions resulted in incorporation of small numbers of colchicine molecules at microtubule ends and inhibition of tubulin addition only after periods of time that varied from several minutes to approximately 20 min depending upon the concentration of colchicine. Inhibition of tubulin addition beginning with unliganded colchicine increased slowly with time, concomitant with increases in the concentration of final-state tubulin-colchicine complex and the amount of colchicine bound per microtubule end. The results indicate that inhibition of tubulin incorporation at microtubule ends is caused by colchicine-liganded tubulin in the form of a final-state complex.(ABSTRACT TRUNCATED AT 250 WORDS)     

Polymerization of tubulin in the presence of colchicine or podophyllotoxin. Formation of a ribbon structure induced by guanylyl-5'-methylene diphosphonate

GTP-dependent in vitro polymerization of rat brain microtubular protein is inhibited to 50% by substoichiometric concentrations of the antimitotic drugs colchicine (0.12 mol/mol of tubulin) and podophyllotoxin (0.14 mol/mol of tubulin). Substitution of pp(CH₂)pG² for GTP, however, results in an extensive microtubular protein polymerization at such concentrations. In the presence of pp(CH₂)pG, suprastoichiometric concentrations of podophyllotoxin (19 mol/mol of tubulin) are required to inhibit the polymerization process by 50%. Colchicine is very ineffective since 3 × 10⁵ moles/mole of tubulin are required to give a 50% inhibition. Electron microscopical analysis shows that the polymers formed by microtubular protein in the presence of suprastoichiometric concentrations of drugs are not the normal short microtubules typical of pp(CH₂)pG-driven polymerization, but are ribbons with three or four protofilaments. The colchicine content of the harvested ribbons has been measured directly and found to be approximately 0.8 moles colchicine/mole of tubulin. Treatment of microtubular protein with substoichiometric concentrations of drugs results in an increase in the number of protofilaments forming the ribbons. Many of the ribbons can close into morphologically normal microtubules when microtubular protein is treated with only 0.05 moles of either colchicine or podophyllotoxin per mole of tubulin.     

Resistance of Rosa microtubule polymerization to colchicine results from a low-affinity interaction of colchicine and tubulin

The inhibition of the polymerization of tubulin from cultured cells of rose (Rosa. sp. cv. Paul's scarlet) by colchicine and the binding of colchicine to tubulin were examined in vitro and compared with data obtained in parallel experiments with bovine brain tubulin. Turbidimetric measurements of taxol-induced polymerization of rose microtubules were found to be sensitive and semiquantitative at low tubulin concentrations, and to conform to some of the characteristics of a nucleation and condensation-polymerization mechanism for assembly of filamentous helical polymers. Colchicine inhibited the rapid phase of polymerization at 24°C with an apparent inhibition constant (K _i) of 1.4·10^-4 M for rose tubulin and an apparent K _i=8.8·10^-7 M for brain tubulin. The binding of [³H]colchicine to rose tubulin to form tubulin-colchicine complex was mildly temperature-dependent and slow, taking 2–3 h to reach equilibrium at 24°C, and was not affected by vinblastine sulfate. The binding of [³H]colchicine to rose tubulin was saturable and Scatchard analysis indicated a single class of low-affinity binding sites having an apparent affinity constant (K) of 9.7·10² M^-1 and an estimated molar binding stoichiometry (r) of 0.47 at 24°C. The values for brain tubulin were K=2.46·10⁶ M^-1 and r=0.45 at 37°C. The binding of [³H]colchicine to rose tubulin was inhibited by excess unlabeled colchicine, but not by podophyllotoxin or tropolone. The data demonstrate divergence of the colchicine-binding sites on plant and animal tubulins and indicate that the relative resistance of plant microtubule polymerization to colchicine results from a low-affinity interaction of colchicine and tubulin.Abbreviations MT microtubule - TC tubulin-colchicine complex     

Interaction of oncodazole (R 17934), a new antitumoral drug, with rat brain tubulin.

Oncodazole (R 17934), methyl [5-(2-thienylcarbonyl)-1H-benzimidazol-2-yl] carbamate (I), a new synthetic drug with anti-tumoral activity, inhibits the polymerization of rat brain tubulin $\text{in vitro}$. It has no depolymerizing effect on preformed microtubules $\text{in vitro}$. Binding studies by means of molecular sieving and equilibrium dialysis indicates that the drug binds to purified rat brain tubulin in a mole to mole ratio. Finally the drug competitively inhibits colchicine binding to purified rat brain tubulin. From these results the conclusion may be drawn that oncodazole is a true microtubule inhibitor.     

Selenium: inhibition of microtubule formation and interaction with tubulin.

We have studied the interaction of Na2SeO3 with microtubule proteins and tubulin. This selenium compound inhibits the polymerization of MTP (half-inhibition occurred for Na2SeO3 10 microM), and to a lesser that of tubulin. This effect of selenite is related to the formation of disulfide bridges between tubulin sulfhydryl groups, inducing a conformational change of the protein. This is corroborated by the modified binding of colchicine and vinblastine in presence of selenium. The selenite inhibitory concentrations are similar to the toxic blood levels of selenium (40 microM).     

Theory for modeling the copolymerization of tubulin and tubulin-colchicine complex

Substoichiometric concentrations of tubulin-colchicine complex (TC) inhibits microtubule assembly through a copolymerization reaction between tubulin and TC. We have determined the rates and extent of TC incorporation into bovine brain microtubules and developed a theory that models copolymerization. Our analysis suggests that while the apparent association rate constants for tubulin and TC are similar, the apparent dissociation rate constants for TC are a factor of five or more larger than those of tubulin. Copolymer composition showed only slight changes during assembly despite changes in the solution phase and showed little dependence at high TC upon the initial tubulin concentration. The theory was based on coupled Oosawa-Kasai equations that allow for the co-assembly of two components, tubulin and TC. An expression was derived that relates copolymer composition to reaction mixture composition and to the affinity of microtubule ends for tubulin and TC. This expression predicts copolymer composition at TC concentrations less than 10 microM and correlates composition with assembly inhibition. We perceive copolymerization as a facilitated incorporation of TC requiring the presence of tubulin. TC incorporation was dependent on the ratio of total tubulin to the dissociation constant for TC bound to microtubule ends. The copolymerization reaction is thus characterized by an interplay of two effects (a) where tubulin facilitates the incorporation of TC into the microtubule, and (b) where TC inhibits the assembly of tubulin into microtubules.     

Enhancement of tubulin assembly as monitored by a rapid filtration assay

The early kinetics of microtubule formation from lamb brain tubulin isolated by affinity chromatography can be followed by a newly developed filter assay. The rapid collection of microtubules on glass fiber filters permits the calculation of the moles of tubulin polymerized. The filter assay gives both a rate and extent of polymerization that are identical to those obtained by turbidity or sedimentation analysis, respectively. The microtubules trapped by the filter are readily depolymerized by cold (t12= 3 min) and slowly by colchicine (t1/2= 32min). Tubulin purified by affinity chromatography requires a high protein concentration (>4 mg/ml) for polymerization. Although 5m glycerol allows polymerization to occur at tubulin concentrations below 2 mg/ml, the maximum amount of microtubule formation is observed at low tubulin concentration when microtubule-associated proteins are present. These proteins are not retained by the affinity resin; however, they can be eluted from diethylaminoethyl-Sephadex by solutions containing 0.3m KCl. Microtubule-associated proteins enhance both the rate of polymerization and the total amount of tubulin polymerized as assessed by the filter assay, suggesting that they are involved in both initiation and elongation of microtubules.     

Interaction of tubulin with octyl glucoside and deoxycholate. 2. Protein conformation, binding of colchicine ligands, and microtubule assembly

The structural change induced by binding of mild detergents to cytoplasmic calf brain tubulin and the effects on the functional properties of this protein have been characterized. Massive binding of octyl glucoside or deoxycholate monomers induces circular dichroism changes indicating a partial alpha-helix to disordered structure transition of tubulin. The protein also becomes more accessible to controlled proteolysis by trypsin, thermolysin, or V8 protease. This is consistent with the looser protein structure proposed in previous binding and hydrodynamic studies [Andreu, J. M., & Mu?oz, J. A. (1986) Biochemistry (preceding paper in this issue)]. Micelles of octyl glucoside and deoxycholate bind colchicine and its analogue 2-methoxy-5-(2,3,4-trimethoxyphenyl)-2,4,6-cycloheptatrien-1-one (MTC). This impedes the determination of colchicine binding in the presence of detergents. Both detergents cause a reduction in the number of tubulin equilibrium binding sites for the colchicine site probe MTC. Deoxycholate monomers bind poorly to the tubulin-colchicine complex, but deoxycholate above the critical micelle concentration effectively dissociates the complex. Microtubule assembly in glycerol-containing buffer is inhibited by octyl glucoside, which raises the critical protein concentration. Low concentrations of deoxycholate enhance tubulin polymerization, allowing it to proceed without glycerol. The polymers formed are microtubules, pairwise associated open microtubular sheets, and macrotubules possibly generated by helical folding of the sheets, as indicated by the optical diffraction patterns. Saturation of tubulin with octyl glucoside, followed by full dissociation of the detergent, allowed the recovery of binding to the colchicine site and microtubule assembly, indicating the reversibility of the protein structural change.     

Light-microscopic observations of individual microtubules reconstituted from brain tubulin.

The course of polymerization of individual brain microtubules could be observed with a light microscope employing dark-field illumination. Statistical analysis of the increase in microtubule length during the polymerization was in accordance with the time course of viscosity change of the tubulin solution. After a plateau level in viscosity was attained, there was no significant change in histograms showing length distribution. These observations were confirmed with fixed and stained microtubules, using a phase-contrast microscope. Observations with dark-field illumination revealed that reconstituted microtubules depolymerized and disappeared immediately upon exposure to buffer containing CaCl2 or sulphydryl reagents such as p-chloromercuriphenyl sulphonic acid (PCMPS) and p-chloromercuribenzoic acid (PCMB). They were also cold-labile. The growth of heterogeneous microtubules which were assembled by mixing purified tubulin dimers with ciliary outer fibres could also be followed with these optical systems.     

Perturbation of microtubule polymerization by quercetin through tubulin binding: a novel mechanism of its antiproliferative activity

The dietary flavonoid quercetin has a broad range of biological activities, including potent antitumor activity against several types of tumors. Recently, it has been shown that quercetin inhibits cancer cells proliferation by depleting cellular microtubules and perturbing cellular microtubule functions. However, the direct interactions of quercetin with tubulin and microtubules have not been examined so far. Here, we found that quercetin inhibited polymerization of microtubules and depolymerized microtubules made from purified tubulin in vitro. The binding of quercetin with tubulin was studied using quercetin fluorescence and intrinsic tryptophan fluorescence of tubulin. Quercetin bound to tubulin at a single site with a dissociation constant of 5-7 microM, and it specifically inhibited colchicine binding to tubulin but did not bind at the vinblastine site. In addition, quercetin perturbed the secondary structure of tubulin, and the binding of quercetin stimulated the intrinsic GTPase activity of soluble tubulin. Further, quercetin stabilized tubulin against decay and protected two cysteine residues of tubulin toward chemical modification by 5,5'-dithiobis-2-nitrobenzoic acid. Our data demonstrated that the binding of quercetin to tubulin induces conformational changes in tubulin and a mechanism through which quercetin could perturb microtubule polymerization dynamics has been proposed. The data suggest that quercetin inhibits cancer cells proliferation at least in part by perturbing microtubule functions through tubulin binding.     

Insulin release and the microtubular system of the islets of Langerhans. Identification and characterization of tubulin-like protein.

1. Incubation of islets of Langerhans in vitro in the presence of colchicine produced a progressive inhibition of the insulin-secretory response to glucose, which was dependent on the time of incubation. 2. The uptake of [3-H]colchicine by islet cells was a rapid process, equilibrium being reached in less than 30 min. Part of the colchicine taken up was bound to protein material, which was recovered largely in a post-microsomal supernatant fraction prepared from the islets. In contrast with this rapid uptake, the binding of colchicine by islet-cell proteins in intact islets or in islet homogenates was a slow process, and equilibrium was not reached for 60-90 min. After an initial 30 min delay, the time-course of the binding of [3-H]colchicine to islet-cell proteins paralleled that for the inhibitory effect of colchicine on insulin release. 3. Some purification of the colchicine-binding material present in islet homogenates could be achieved by precipitation of the protein with 2mM-CaCl2 (2.8-fold). However, ion-exchange chromatography on DEAE-Sephadex produced a further 27-fold purification on elution with 0.6M-NaCl. 4. Colchicine-binding protein prepared from islets by ion-exchange chromatography showed an intrinsic association constant for colchicine of 1.4muM and an apparent molecular weight on gel filtration of 110000. 5. These results suggest that colchicine-binding protein in islet cells closely resembles tubulin extracted from the other tissues. The delayed effectiveness of colchicine in inhibiting insulin secretion is not due to poor penetration of colchicine into the cells but rather to slow binding of the alkaloid to islet-cell tubulin. It seems likely that, as in other tissues, this binding prevents polymerization of the tubulin into microtubules, and thus interferes with the release process.     

33K protein--an inhibitory factor of tubulin polymerization in porcine brain

A factor (33K protein) that modulates tubulin polymerization in vitro has been purified to homogeneity from porcine brain by ammonium sulfate fractionation and Whatman DE52, Toyo-pearl HW65C and Bio-Gel A 0.5 m column chromatographies. The purified fraction was free of nucleic acids and sugars. The activity of the purified 33K protein is pronase E sensitive but apparently heat- and trypsin-resistant though it undergoes tryptic digestion. The 33K protein inhibits polymerization of brain microtubule proteins in a dose-dependent manner and partially depolymerizes preformed microtubules. It also inhibits polymerization of purified starfish tubulin and microtubule elongation involving fragellar outer doublet microtubules and purified porcine brain tubulin. This suggests that the target of the 33K protein is tubulin rather than microtubule-associated proteins. The 33K protein causes incomplete depolymerization of microtubules and a new steady state is quickly attained which is apparently independent of microtubule mass concentration. Divalent cations such as calcium and magnesium do not modulate the inhibitory activity of the 33K protein.