Taxol: an antimitotic agent with a new mechanism of action

Pharmac. Ther. Vol. 25, pp. 83 to 125, 1984 Printed in Great Britain. All rights reserved

0163-7258/84 $0.00 + 0.50 Copyright © 1984 Pergamon Press Ltd

Specialist Subject Editor: L. A. DETHLEFSEN

TAXOL:AN

ANTIMITOTIC MECHANISM

AGENT

WITH

A NEW

OF ACTION

JAMES J. MANFREDI a n d SUSAN BAND HORWITZ

Departments of Molecular Pharmacology and Cell Biology, Albert Einstein College of Medicine, Bronx, New York 10461, U.S.A.

ABBREVIATIONS AMP ATP DME EGTA GDP GTP

Adenosine 5'-monophosphate Adenosine 5'-triphosphate Dulbecco's modified Eagle's medium

Ethyleneglycol-bis-(fl-aminoethyl ether)N,N'-tetraacetic acid Guanosine 5'-diphosphate Guanosine 5'-triphosphate

MAPs MES NP-40 PBS PIPES SDS Triton

Microtubule-associated proteins 2(N-morpholino)ethane sulfonic acid Nonidet P-40 Phosphate-buffered saline Piperazine-N,N'-bis-(2-ethane sulfonic acid) Sodium dodecyl sulfate Triton X-100

1. INTRODUCTION Much of what is currently known concerning microtubule biochemistry and physiology has come from studies of the interaction of various plant alkaloids with microtubules. This pharmacological approach has proven to be a powerful one in elucidating the mechanisms which regulate microtubule assembly in vitro and in cells. Such studies have also been useful clinically since drugs such as the Vinca alkaloids have proven to be effective chemotherapeutic agents. The purpose of this review is an examination of the interaction of the novel antimitotic agent, taxol, with cellular microtubules. Since it represents a totally new mechanism of action for an antitumor compound, understanding its cellular activities may provide important information for tumor therapy as well as further our knowledge of the regulation of microtubule assembly and organization in cells. Before discussing experiments involving taxol, a brief introduction to microtubules and microtubule assembly in vitro and in cells is necessary. This is not meant to be a comprehensive review of the literature, but rather to give the reader adequate background to appreciate the data presented on taxol. Microtubules are an integral component of eukaryotic cells, and are implicated in a variety of cellular functions including chromosome movement, regulation of cell form, anchorage of surface receptors in the plasma membrane, and cellular motility associated with cilia and flagella. They are long, hollow cylinders, 25 nm in diameter, consisting of 13 protofilaments aligned longitudinally along the axis of the cylinder, and are assembled from a 100,000 MW protein, tubulin. Tubulin has two subunits: ~ and r, similar but distinct acidic polypeptides, each with a MW of 50,000 (see Dustin, 1978). Tubulin is highly conserved in terms of its molecular weight, antigenic cross-reactivity, ability to bind a variety of small molecules and to copolymerize into microtubules which appear to be ultrastructurally identical regardless of the source of tubulin (see Luduena, 1979). 2. MICROTUBULE ASSEMBLY I N V I T R O A minimum concentration of tubulin dimers are required before microtubule assembly will occur (see Timasheff and Grisham, 1980). This threshold level of tubulin is the critical concentration (Co) for assembly. If microtubules are assembled to steady state, the amount 83

84

J.j. MAr~r'REI~Xand S. B. HORWITZ

of tubulin that remains unassembled is equivalent to the critical concentration. If such steady state microtubules are diluted with buffer, they will disassemble until the critical concentration is reached in solution. If excess tubulin is added to steady state microtubules, assembly will continue until the level of dimers free in solution is equivalent to the critical concentration. A complete discussion of in vitro microtubule assembly is beyond the scope of this review, but several points concerning current thinking of microtubule polymerization should be made (Scheele and Borisy, 1979; Timasheff and Grisham, 1980). There are three phases of microtubule assembly: nucleation, elongation and steady-state. The nucleation phase represents the formation of short microtubule seeds upon which further growth can occur. Elongation of these seeds can occur either by addition of individual tubulin dimers or by addition of oligomeric structures consisting of both tubulin and microtubuleassociated proteins (Weisenberg, 1980). At steady state, there is no net change in the overall microtubule length but there is still a constant addition and loss of tubulin from the microtubule; the polymer is in a dynamic equilibrium with dimers that are free in solution. There is an absolute requirement for G T P in the in vitro microtubule assembly reaction. Tubulin has two binding sites for guanine nucleotides: a nonexchangeable or N site and an exchangeable or E site. The N site binds one molecule of G T P noncovalently; this G T P is not hydrolyzed during microtubule assembly in vitro (Penningroth and Kirschner, 1977) or in cells (Spiegelman et al., 1977) nor does it exchange with nucleotide free in solution. The E site binds one molecule of guanine nucleotide noncovalently which is able to exchange with G T P free in solution. U p o n incorporation of the tubulin dimer into the

MAP 2001(-'150k,..-

........

i~!i~ii~i~!~i~i~

1

<~!~i~%~

2

FIG. 1. One-dimensionalgel analysis of microtubule protein and phosphocellulose purified tubulin. Twice cycled microtubule protein (80#g, lane 1) and phosphocellulose column-purified tubulin (50 p g, lane 2) prepared from calf brains as described in Parness and Horwitz (1981) were subjected to one-dimensional SDS polyacrylamide gel electrophoresis (Laemmli, 1970), and stained with Coomassie blue. Major bands were identified as follows: A, actin (43,000 MW); ~, ~-tubulin; 1~,/3-tubulin; tau, microtubule-associated proteins (58,000-68,000 MW); 68K, 150K, and 200K, intermediate filament proteins; MAPI and MAP2, microtubule-associated proteins 1 and 2 (300,000 and 350,000 MW).

Taxol

85

microtubule, this GTP is hydrolyzed to GDP and becomes nonexchangeable (Weisenberg et al., 1976; Carlier and Pantaloni, 1981). The time course of GTP hydrolysis is not strictly coupled to tubulin incorporation into polymer, but occurs subsequent to it (Carlier and Pantaloni, 1981). This hydrolysis is not necessary for assembly, since nonhydrolyzable (fl, 7) GTP analogs assemble ultrastructurally normal microtubules which depolymerize when treated with calcium or at cold temperatures (Penningroth and Kirschner, 1977). Microtubules can also be assembled using nonhydrolyzable (a, r ) GTP analogs: these microtubules are relatively more stable to depolymerizing conditions (Sandoval et al., 1977; 1978). Since hydrolysis of the (a, r ) high energy bond does not appear to be involved in normal microtubule assembly, these results suggest that more than just the (fl, 7) phosphate bond is involved in the process of microtubule assembly. There are a number of proteins which copurify with tubulin through multiple cycles of assembly and disassembly (Fig. 1). Microtubule-associated proteins have been defined as those which enhance the microtubule assembly reaction. From conventional brain preparations, three proteins have been studied: MAP1 and MAP2 (350,000 and 300,000 MW, respectively; Sloboda et al., 1975), and a group of polypeptides called tau (56,000-68,000 MW; Weingarten et al., 1975). The ability of the exchangeable guanine nucleotide site to become nonexchangeable after the tubulin dimer is incorporated into the microtubule serves as a means of radiolabeling the polymer. Margolis and Wilson (1978) incubated tubulin with tritiated nucleotide and the subsequent incorporation resulted in a time-dependent incorporation of label. If microtubules were allowed to become uniformly labeled, loss of radioactivity from the polymer also occurred in a linear, time-dependent manner. The addition and loss both occurred at equivalent rates. Shearing of the microtubules to shorter lengths increased the rate at which label was released. This suggested that tubulin dimers are added at one end of the microtubule and released from the other end (Margolis and Wilson, 1978). Cote and Borisy (1981) studied this phenomenon in greater detail using a pulse label of [3H]tubulin added to microtubules at steady state. There was a lag time between uptake of label by the polymer and its subsequent release, indicating that tubulin fuxes through the microtubule. This 'treadmilling' is dependent on GTP hydrolysis; microtubules assembled in the presence of nonhydrolyzable guanine nucleotides do not exhibit this behavior (Sandoval and Weber, 1980; Margolis, 1981). A variety of hypothetical models have been proposed for how this treadmilling could be effectively used by the cell (Margolis et al., 1978; Kirschner, 1980; Hill and Kirschner, 1982; Margolis and Wilson, 1981), but recent experiments suggest that treadmilling is probably a very inefficient phenomenon (Caplow et al., 1982; Farrell and Jordan, 1982), and there is no data indicating that microtubules in cells participate in treadmilling. ATP enhances the in vitro treadmilling rate as much as twenty-fold (Margolis and Wilson, 1979). Studies by Jameson and Caplow (1981) have shown that phosphorylation of microtubule-associated proteins enhances the steady state flux of tubulin through the microtubule. This treadmilling of tubulin through the microtubule is sensitivie to particular in vitro conditions, suggesting that tubulin flux may be important for the cellular regulation of microtubule assembly since small changes in intracellular conditions may dramatically affect the tubulin flux rates (Cleveland, 1982). 3. MICROTUBULE ASSEMBLY IN CELLS 3.1. MICROTUBULE ORGANIZING CENTERS

Electron microscopy of thin sections of cells has been helpful in appreciating the ultrastructure of microtubules and their general location and associations with other intracellular organelles, but the utilization of tubulin immunofluorescence (Fuller et al., 1975; Weber et al., 1975) was a major breakthrough in understanding the overall organization of microtubules in cells. Microtubules appear to emanate from the perinuclear region and extend the entire length of the cell (Fig. 2A). In certain cell types, a single focal point can be observed from which all the microtubules appear to grow (Fig.

86

J.J. MANFRED1and S. B. HORW1TZ

FIG. 2. Tubulin immunofluorescence of 3T6 cells which have been treated with colchicine, vinblastine or taxol. 3T6 cells, a mouse fibroblastic cell line, were plated on glass coverslips and incubated for 60 min at 37°C in complete medium (DME containing 10~ heat-inactivated calf serum and supplemented with penicillin, streptomycin, glutamine, and non-essential amino acids) alone (A) or containing 30 pra colchicine (B), 30 #M vinblastine (C), or for 240 min with 30/~M taxol (D). Cells were fixed for 20 min in 3.7~o formaldehyde in PBS, washed in PBS, plunged into absolute acetone at -20'~C for 6 rain, and subsequently further washed in PBS. Incubation with 20 #1 of undiluted sheep anti-tubulin was for 30 min in a humid atmosphere at 37°C. Cover slips were washed extensively in PBS and then incubated with 20/~1 of fluoroscein conjugated rabbit anti-sheep antibody (1:100 dilution) for 30 min at 37'~C.Cover slips were again extensivelywashed in PBS and mounted using Aquamount on clean microscope slides. A Zeiss standard 14 microscope equipped with epifluorescent optics was used to view the preparations. The combination of a × 10 ocular lens and a x 63 oil immersion objective lens (numerical aperture, 1.4) gave a depth of field of 0.8/~m. Photographs were then taken on Kodak Plus-X pan 35 mm film using an ASA of 1000; the meter on the microscope gave exposure times ranging from 15-30 sec. Film was developed in Kodak HC-100 (Dilution B) and printed on Kodabromide F4 grade paper. Bar is 20/~m.

3A). M i c r o t u b u l e s are visualized as fine, wavy, filamentous structures which fill the entire cytoplasm. This cytoplasmic m i c r o t u b u l e complex is seen at interphase. D u r i n g mitosis, the interphase complex breaks d o w n a n d is replaced by the mitotic spindle (Fig. 3 B - F ; Brinkley et al., 1981). Clearly, the cell m u s t have regulatory m a c h i n e r y to c o n t r o l this sophisticated r e a r r a n g e m e n t of the m i c r o t u b u l e cytoskeleton. As cell division nears c o m p l e t i o n , the spindle is r e a r r a n g e d into the m i d b o d y (Fig. 3 G - I ) . Electron microscopic e x a m i n a t i o n o f the m i d b o d y d e m o n s t r a t e s that it consists of a tightly packed b u n d l e o f m i c r o t u b u l e s (Fig. 4), m a i n t a i n i n g a c o n n e c t i o n between the two d a u g h t e r cells. Cellular m i c r o t u b u l e s can be depolymerized with drugs such as colcemid or n o c o d a z o l e and, after w a s h i n g out the drug, the regrowth of cytoplasmic m i c r o t u b u l e s observed. M i c r o t u b u l e s a p p e a r to grow from certain distinct focal points (Brinkley et al., 1981; O s b o r n a n d Weber, 1976; Spiegelman et al., 1979a; D e B r a b a n d e r et al., 1980). These i n i t i a t i o n sites are termed m i c r o t u b u l e o r g a n i z i n g centers ( M T O C s , Pickett-Heaps, 1969). U l t r a s t r u c t u r a l e x a m i n a t i o n of cells recovering from d r u g t r e a t m e n t at early times revealed

Taxol

87

C

8

F

)

E

F

......... ~

t

4

FIG. 3. Tubulin immunofluorescenceof J774.2 cells at various stages of mitosis. J774.2 cells, a mouse macrophage-like cell line, were extracted with 0.1~ NP-40 in 0.1 u PIPES pH 6.9, 1 mM EGTA and 4~ polyethyleneglycol6000, and then processedfor immunofluorescenceas described in Manfredi et al. (1982). Bar is 20#m.

that this microtubule regrowth occurs at the centrosome (Brinkley et al., 1981). The centrosome consists of a pair of centrioles arranged at right angles to one another (Fig. 5A), surrounding osmiophilic pericentriolar material (Fig. 5B), and occasionally associated virus-like particles (Peterson and Berns, 1980). The centriole is a cylindrical structure that is approximately 0.25 # m in diameter and can be up to several microns in length. In cross section in the microscope, it can be seen to consist of nine triplets of microtubules arranged at precise angles to one another (Fig. 5B; Peterson and Berns, 1980). Most cell types contain a single MTOC. Neuroblastoma cells are an exception in that these cells appear to have eight or more sites of microtubule regrowth after drug-induced disassembly (Spiegelman et al., 1979b). Electron microscopic examination demonstrated that each of these multiple sites of reassembly represented a single centriole with surrounding pericentriolar material (Brinkley et al., 1981; Sharp et al., 1981). These cells undergo a normal mitosis, with the multiple centrioles unequally allocated to each of the two poles (Ring et al., 1982). Stimulation of human neutrophilic polymorphonuclear leukocytes with a chemoattractant causes a separation of the centrosome into two solitary centrioles, each

n!a4P "

~

i

/IG. 4. Electron microscopy of the midbody in extracted J774.2 cells. J774.2 cells were extracted with 1~o Triton in 0.1 M PIPES pH 6.9, 1 mM MgSO4, 2 M glycerol, md 2 mM EGTA and processed for electron microscopy as follows: cells were grown on round, carbon-coated, glass coverslips and fixed for 15 min in 2~ glutaraldehyde n 0.1 M cacodylate, pH 7.4. After a 3 min rinse in buffer alone, they were post fixed for 20min in 1~o osmium tetroxide in 0.1 M cacodylate, pH 7.4. Cells were rinsed t min in buffer, followed by two 3 min washes with water, and then stained for 20 min with an aqueous solution of 1~o uranyl acetate. Samples were dehydrated through in ethanol series, and embedded in Epon in inverted BEEM capsules. Thin sections were stained with 4~ uranyl acetate in 40~o ethanol, then with 0.1~o lead citrate, and viewed in a JEOL, HEM-100CX electron microscope at 80 kV. (A) 7500X and (B) 65,000X.

A

q

az

~

0

Na

"~.~

t-rl t o ~'QR

~$~~.o~

0

~.~_~.

0

~t

90

J . J . MANFREDI and S. B. HORWITZ

surrounded by an aster of microtubules in approximately half of the cells in the population. In addition, 10~o of these cells have a third, centriole-free aster of microtubules radiating from an accumulation of dense material (Schliwa et al., 1982). Microtubules appear to nucleate at sites other than the centrosome. Indeed, Spiegelman et al. (1979a) suggest that there are three distinct classes of microtubule initiation sites in mammalian cells: a primary site which probably represents the centrosome, secondary sites which are distinct perinuclear sites, and tertiary sites for growth of single microtubules. The latter are also located near the cell nucleus. 3.2. MODELS FOR CELLULARMICROTUBULEORGANIZATION The models presented here relate to the organization of the interphase cytoplasmic microtubule complex in undifferentiated mammalian cells. 3.2.1. Seed or Template Model The centriole appeared to be the focal point for microtubule regrowth in colcemid reversal experiments (see Brinkley et al., 1981). Since the centriole consists of nine triplets of microtubules arranged in a cylinder (Fig. 5B), it represents a potential seed or template upon which cellular microtubules could grow. Careful ultrastructural analysis revealed however, that the microtubules appeared to emanate from the surrounding pericentriolar material (Robbins et al., 1968). Isolated pericentriolar material is capable of nucleating microtubule assembly (Gould and Borisy, 1977). Destruction of the centrioles with a laser beam has no effect on the microtubule organizing activity of the pericentriolar material (Berns and Richardson, 1977). Higher plants, fungi, and diatoms undergo mitosis with an organized microtubule-containing spindle in the absence of centrioles (Peterson and Berns, 1980). Recently, a cultured Drosophila cell line was characterized which grows normally and undergoes normal mitosis, but apparently lacks any centrioles (Debec et al., 1982). Clearly, the centriole is not necessary for the organization of cellular microtubules; microtubules grow from the surrounding pericentriolar material. 3.2.2. Minus-End-Capped Model (Kirschner, 1980) In microtubules, GTP hydrolysis allows for treadmilling of the polymer. Since the net assembly end contains GTP-tubulin whereas the net disassembly end contains GDPtubulin, the two ends have different critical concentrations for assembly. The plus or assembly end has a critical concentration for assembly (C +) which is lower than that of the minus or disassembly end (C~-). The overall steady state critical concentration for assembly (C~) is intermediate between these two values, that is Cc+ < C~ < C~-. There is a monomer concentration, Co, at which one end can add subunits and the other end cannot, that is Cc+ < Co < C~-. Three potential polymer types exist in a cell: plus-end-capped, both-ends-free, and minus-end-capped. The critical concentrations for each of these polymers are C~-, Cc~, Cc+ , respectively. At an intracellular tubulin concentration above C~-, all three types of polymers will grow. As the intracellular tubulin concentration decreases to C~-, the both-ends-free and the minus-end-capped polymers will continue to grow at the expense of the plus-end-capped polymers. As the intracellular tubulin concentration decreases to CSc, the minus-end-capped polymers will continue to grow at the expense of the both-ends-free polymers, until C~+ is reached. Hence, the only stable polymers are those in which the minus end is capped. The function of the MTOC is to cap the minus end of cellular microtubules. Only at the MTOC can the minus end be capped; only microtubules attached to the MTOC will be stable. This model does not allow for any flux of tubulin through cellular microtubules, and requires that there be a capping mechanism at the MTOC for the minus ends (Kirschner, 1980; Cleveland and Kirschner, 1982). It is, of course, also possible that the

Taxol

91

microtubules are capped at both ends and it is this capping mechanism which is responsible for cellular microtubule organization. 3.2.3. Treadmilling Model (DeBrabander, 1982) The microtubules in a cell are constantly treadmilling with both ends free and available for exchange with tubulin dimers. The MTOC functions to focus microtubule assembly by lowering the critical concentration for assembly in its immediate vicinity. Microtubules are nucleated in this region and start to grow outward. This requires that centrosomal microtubules be oriented with their assembly ends distal to the MTOC. Studies by Bergen et al. (1980) suggest that this is the case: centrosomal microtubules elongate at a rate corresponding to the addition of tubulin dimers at the assembly end only. Microtubules are only nucleated in the vicinity of the MTOC. Conditions in the surrounding cytoplasm are not appropriate to nucleate assembly, however, elongation of the MTOC-nucleated microtubules can occur. At steady state, tubulin is added to the distal ends and lost from the proximal ends of the MTOC-associated microtubules. Nucleation of microtubules other than at the MTOC is suppressed. This model allows for flux of tubulin through cellular microtubules and involves the need for assembly-regulating activities in the cell.

4. SMALL MOLECULES THAT INTERACT WITH TUBULIN 4.1. CALCIUMIONS Weisenberg (1972) made a major contribution to the study of tubulin self-assembly by his observation that microtubule polymerization is exquisitely sensitive to calcium. Indeed, only by including a chelator of calcium ions in assembly buffers was he able to demonstrate in vitro tubulin polymerization. The calcium sensitivity of microtubules that have been assembled in vitro is now well-established (Olmsted and Borisy, 1973; Haga et aL, 1974; Kuriyama and Sakai, 1974). Calcium can both inhibit tubulin assembly as well as depolymerize preassembled microtubules. There has been much discussion in the literature over the involvement of microtubule-associated proteins and calmodulin in mediating the sensitivity of microtubules to micromolar calcium concentrations (Nishida and Sakai, 1977; Nishida, 1978; Lee and Wolff, 1982), but it is clear that the ability of millimolar calcium concentrations to depolymerize microtubules refects an intrinsic property of purified tubulin (Berkowitz and Wolff, 1981). There are two classes of calcium binding sites on tubulin: a single high affinity site characterized by a dissociation constant of 3.2 × 10-6M and 16 low affinity sites which have a dissociation constant of 2.8 x 10-aM (Solomon, 1977). Summers and Kirschner (1979) using darkfield microscopy have found that calcium causes disassembly of microtubules with a threefold bias of one end over the other. In vitro studies have demonstrated that calcium induces microtubule disassembly by binding directly to the microtubule and promoting and end-dependent depolymerization (Karr et al., 1980; Weisenberg and Deery, 1981). It has been difficult to examine the calcium sensitivity of microtubules in vivo since the cell membrane serves as an effective barrier to calcium ions. Schliwa (1976) used the ionophore A23187 to demonstrate by electron microscopy that the axonemal microtubule arrays of the heliozoan, Actinosphaerium eichhirni, were degraded by treatment with calcium. Fuller and Brinkley (1976) also observed the calcium sensitivity of microtubules in A23187-treated 3T3 fibroblasts as assayed by tubulin immunofluorescence. Recently, the calcium sensitivity of microtubules in detergent-extracted cultured mammalian cells has been demonstrated by tubulin immunofluorescence and gel electrophoresis (Osborn and Weber, 1977; Solomon et al., 1979; Schliwa et aL, 1981a). 4.2. VINCA ALKALOIDS Vinblastine and vincristine (Fig. 6) are clinically useful antitumor agents isolated from the periwinkle plant, Cantharanthus rosea. Their mechanism of action, in some ways

92

J . J . MANFREDI and S. B. HORW1TZ

CH30~

NH-

R

OCH3

0

ON

OCH3

C HI 3~ II0 ° ~ I

/

~o

c~o

C

OCH3 R : COCH:s

COLCHICINE

R : CH 3

COLCEMID

C H 3 0 ~ C1'.,,O t f y

H

3

0

~

O

CL

'~! CH3 GRISEOFULVlN

NHCOCH3

H 0 " ~

"~'~

, I

CH,O T-To

II

CH)O

CCOCN)

I H

N

H3COC I CH)

O

N"'- NH- ~ - OCH3

R:CHOVgNCItiSrole t R=CH3 VINItASTtNE

II

1.1 OH

NOCODAZOLE

O

OCH~

/"

LUMICOLCHICINE J

0

o

(%C-¢-O

H3CO~OCH3

O

C-N"-C"-C"-C-O-

"'0

"/ q

LO

........

TAXOL

o o

OCH3

PODOPHYLLOTOXIN

H]CO'~OCH) OH VP-16-213

FIG. 6. Structural formulae of drugs which interact with tubulin and some inactive congeners.

similar to that of colchicine, involves the disruption of cellular microtubules. The Vinca alkaloids are unique in that they induce the formation of highly ordered paracrystalline arrays of tubulin in cells (Fig. 2C; see Dustin, 1978). These paracrystals are short rodlike structures which appear hexagonal in cross section. They contain large numbers of macrotubules arranged in parallel along their long axis; each macrotubule consists of two intertwined, helical aggregates of tubulin (Fujiwara and Tilney, 1975). Recently, conditions were determined for the in vitro formation of vinblastine-induced paracrystals. Similar structures can be formed with vincristine and desacetylvinblastine, but not with colchicine or podophyllotoxin (Na and Timasheff, 1982). Triton extraction of cells removes the plasma membrane and most intracellular organelles, leaving behind a cytoskeleton consisting mainly of the nuclear matrix, polysomes, microtubules, actin and intermediate filaments. Treatment of 3T6 cells with 30 #M vinblastine for 60 min causes the formation of numerous, well defined paracrystals in the cytoplasm. One-dimensional gel analysis of vinblastine-treated cells indicates that the tubulin bands are retained in Triton-cytoskeletons of these cells. Examination of these preparations by tubulin immunofluorescence shows that vinblastine-induced paracrystals are retained in the cytoskeletons. If these preparations are then extracted with a calcium-containing buffer, the tubulin bands are retained as seen by gel analysis. Tubulin immunofluorescence of these preparations demonstrates the presence of paracrystals in the cytoskeletons after calcium treatment. If 3T6 cells are first extracted with Triton and allowed to sit for 60 min at 37°C in buffer, the tubulin bands are retained after treatment of these cytoskeletons with 30/~ Mvinblastine for 60 min, but subsequent calcium treatment releases the tubulin. Tubulin immunofluorescence of these samples demonstrates that vinblastine has no effect on microtubules in an extracted cell and that subsequent calcium treatment depolymerizes these microtubules. Vinblastine can neither depolymerize microtubules nor induce paracrystal formation in an extracted cell. Since the unpolymerized tubulin dimer pool has been washed free of the cytoskeletons, these results suggest that

Taxol

93

the Vinca alkaloids may require tubulin dimers to exert their effects; their depolymerizing activity may not be mediated by a direct binding to microtubules (Manfredi and Horwitz, 1984). In cells, vinblastine exerts two effects in a concentration dependent manner. At lower drug concentrations, cellular microtubules are depolymerized. At higher drug concentrations, paracrystals are formed. These paracrystals are associated with the cytoskeleton in that they are retained in Triton extracted preparations (Manfredi and Horwitz, 1984). Further, these paracrystals are stable to the depolymerizing effects of calcium, suggesting that the calcium binding site on tubulin (Solomon, 1977) is either blocked or altered in the paracrystal. This is consistent with recent studies by Na and Timasheff (1982) which demonstrate that paracrystals can form in vitro in the presence of calcium. In vitro, the Vinca alkaloids have been shown to inhibit microtubule assembly substoichiometrically (Wilson et al., 1975), and to depolymerize steady state microtubules in a novel manner by causing spiral coils to form from protofilaments peeling off the ends of the microtubules (Warfield and Bouck, 1974; Himes et al., 1976; Parness and Horwitz, 1981). These spiral coils are stable to depolymerization by cold, calcium ions, or colchicine (Haskins et al., 1981). Vinblastine has two binding sites on the tubulin dimer. Reports on the binding affinities of these two sites have been contradictory (Wilson et al., 1975; Bhattacharyya and Wolff, 1976). Recent in vitro studies suggest that the two binding sites for vinblastine on the tubulin dimer are equivalent; each has a binding constant of 5.6 x 10 -5 M (Na and Timasheff, 1980). The tau proteins are required for vinblastine-induced spiral coil formation (Luduena et al., 1981). Since these coils have the form of a spiraling filament, it is suggested that they consist of tubulin dimers that are joined end to end as in a protofilament and that tau promotes the longitudinal interactions of tubulin dimers. Thus, the lateral interaction sites should be accessible in these structures and not in the assembled microtubule. These structures bind colchicine and podophyllotoxin with high affinity; this suggests that the binding site for these two drugs is at the site of lateral interaction between protofilaments (Palanivelu and Luduena, 1982). [3H]Vinblastine binds to rat platelets in a specific, saturable manner (Secret et al., 1972). Glutamate competitively inhibits the uptake of vinblastine into human leukocytes (Creasey et al., 1971). Vinblastine inhibits DNA, RNA and protein synthesis (Creasey, 1968; Creasey et al., 1971), but enhances the uptake of 3-O-methylglucose into avian erythrocytes (Whitfield and Schworer, 1978). These effects, however, occur at concentrations much greater than those which affect cellular microtubules. 4.3. COLCHICINE AND OTHER SMALL MOLECULES Colchicine (Fig. 6) is isolated from meadow saffron (Colchicum autumnale) and is the drug of choice in the relief of acute gout (Hartung, 1954). The early studies with colchicine demonstrated that the drug depolymerizes microtubules in cells (Fig. 2B). In vitro, colchicine inhibits microtubule assembly at substoichiometric concentrations of colchicine relative to the concentration of tubulin dimers (Olmsted and Borisy, 1973). There is a single binding site for colchicine on the tubulin dimer which is distinct from the binding sites for the Vinca alkaloids and GTP (Bryan, 1972). Dissociation constants for this binding are in the range of 3.0-9.1 × 10 -7 M (see Luduena, 1979). Colchicine binds to the tubulin dimer and it is this colchicine-dimer complex, not the free drug, which mediates the effects of the drug on microtubules (Margolis and Wilson, 1977). The exact mechanism by which colchicine produces microtubule disassembly is, however, not clear. Margolis and Wilson (1978) put forth a 'capping' model for the substoichiometric poisoning of tubulin assembly by cholchicine. A colchicine-tubulin dimer binds to the net assembly end of a treadmilling microtubule and effectively caps that end by preventing further net assembly at that end. Net disassembly continues at the opposite end and the microtubule depolymerizes. Studies by Sternlicht and Ringel (1979) suggest a copolymer model such that colchicine-dimers are incorporated into microtubules and the presence of

94

J. J, MANFREDIand S. B. HORWITZ

these within the polymer reduces the affinity of the ends for additional assembly. Experiments by Deery and Weisenberg (1981) and Caplow and Zeeberg (1981) suggest that colchicine may even cap both ends of the microtubule. Colcemid, podophyllotoxin, and nocodazole (Fig. 6) are all competitive inhibitors of the binding of colchicine to tubulin (Cortese et al., 1977; Hoebeke et al., 1976), and have similar effects on microtubules in vitro and in cells. Colcemid is a synthetic derivative of colchicine; its effects are more readily reversed than the parent compound (Banerjee and Bhattacharyya, 1979; Serpinskaya et al., 1981). Podophyllotoxin is an alkaloid extracted from the root of the May apple (Podophyllum peltatum; Kelly and Hartwell, 1954). The binding site for podophyllotoxin overlaps with that of colchicine but is not identical in that tropolone inhibits the binding of colchicine but not podophyllotoxin (Cortese et al., 1977). The dissociation constant for the binding of podophyllotoxin to tubulin is 8.3 × 10-TM (Cortese et al., 1977). Nocodazole is an indole derivative which inhibits the binding of colchicine to rat brain tubulin with a K~ of 9.5 × 10 6 U (Hoebeke et al., 1976). Its effects in cells are extremely reversible; hence, it has been used extensively in studies of microtubule regrowth in cells. Steganacin is derived from the wood and stems of Steganotaenia araliacea Hochst. It is an antimitotic agent which inhibits microtubule assembly in vitro. The drug is a competitive inhibitor of the binding of colchicine to chicken brain tubulin; the K~ for this inhibition is 3.1 × 10-6M (Schiff et al., 1978). Both colchicine and podophyllotoxin inhibit the transport of nucleosides into cells; this occurs at drug concentrations much greater than those required to depolymerize microtubules. Lumicolchicine and VP-16-213 (Fig. 6), congeners of colchicine and podophyllotoxin, respectively, have no tubulin-related effects, yet retain the ability to inhibit nucleoside transport (Mizel and Wilson, 1972; Loike and Horwitz, 1976) and are often used as controls to verify that certain effects of the parent compounds are microtubulerelated. Colchicine also inhibits the uptake of 3-O-methylglucose by avian erythrocytes (Whitfield and Schworer, 1978). The uptake of [3H]colchicine into human KB cells saturates in a concentration dependent manner (Taylor, 1965). Carlsen et al. (1976), however, report that the concentration dependence of [3H]colchicine binding to Chinese hamster ovary cells is linear and reflects a simple diffusion mechanism for the uptake of the drug. This discrepancy may be due to the fact that Carlsen et al. (1976) completed their binding assay by washing at room temperature with a buffer containing 20/~M unlabeled colchicine (See et al., 1974); such a washing procedure could have caused a loss of [3H]colchicine sufficient to confuse the results. Taylor (1965) used his binding data to demonstrate that cells are unable to form a functional mitotic spindle if only 3-5% of the binding sites for colchicine are complexed. As in vitro, the effects of colchicine in cells occur at substoichiometric concentrations of colchicine to that of its receptor, tubulin. Mutants of Chinese hamster ovary cells have been isolated which are resistant to the cytotoxic effects of colchicine or colcemid. Two-dimensional gel analysis of these resistant cells reveals an altered fl-tubulin. Examination of messenger RNA pools indicates that this altered tubulin is not a result of a posttranslational modification (Cabral et al., 1980). One colcemid-resistant mutant has been shown to be temperature sensitive for growth. Observations of this mutant using tubulin immunofluorescence show that at the nonpermissive temperature, there is a normal organization of cellular microtubules at interphase, but mitosis appears to be abnormal. The altered fl-tubulin apparently does not interfere with interphase microtubule organization but does interfere with mitosis (Cabral et al., 1982). Keates et al. (1981) have isolated colcemid-resistant mutants of Chinese hamster ovary cells which show an altered ~-tubulin. They purified the tubulin from mutant cells on DEAE Sepharose and demonstrated that it had reduced binding affinity for colcemid. Examination of these cells by tubulin immunofluorescence demonstrates normal displays of interphase and mitotic microtubules, however, vinblastine-induced paracrystals in the

Taxol

95

mutant line differ markedly in size and shape from those in the parental line (Connolly et al., 1981). Podophyllotoxin-resistant mutants of Chinese hamster ovary cells have been

shown to have an altered peptide of 66,000 MW which by several criteria appears to be a microtubule-associated protein (Gupta et al., 1982). Thus, resistance to colchicine and colchicine-like molecules manifests itself as alterations in both ~- and fl-tubulin as well as in microtubule-associated proteins. Chinese hamster ovary cells resistant to the cytotoxic effects of VP-16-213 have been isolated which are not cross-resistant to podophyllotoxin (Gupta, 1983) indicating that the growth inhibitory effects of these two congeners can be genetically separated. Treatment of cells with either colchicine or nocodazole causes an inhibition of tubulin synthesis (Ben-Ze'ev et al., 1979). This inhibition reflects a decrease in the levels of tubulin messenger RNA (Cleveland et al., 1981). Treatment with vinblastine, on the other hand, does not cause such inhibition. A model for regulation' of tubulin synthesis has been proposed in which colchicine, by depolymerizing cellular microtubules, increases the level of unpolymerized tubulin; this unpolymerized tubulin, in a manner not understood, is able to decrease the level of tubulin messenger RNA for tubulin. Vinblastine, by sequestering tubulin in the form of paracrystals, does not substantially alter unpolymerized tubulin levels (Cleveland et al., 1981). 5. TAXOL Taxol (Fig. 6), a novel diterpenoid, was originally isolated from the stem bark of the western yew, T a x u s brevifolia, and has been found also in the leaves, stems and roots of a variety of other T a x u s species (Miller et al., 1981; Wani et al., 1971). The drug is a complex ester, shown to be a taxane derivative containing a rare oxetan ring, and is the first compound of this type to have antileukemic and tumor inhibitory properties. Early work in our laboratory demonstrated that the drug inhibits replication of HeLa cells (Schiff et al., 1978; 1979). Studies with P-388 cells taken from taxol-treated mice identified taxol as a mitotic spindle poison (Fuchs and Johnson, 1978). 5.1. EFFECTS OF TAXOL IN VITRO Studies in vitro have shown that in marked contrast to other antimitotic drugs, taxol enhances both the rate and yield of microtubule assembly. The critical concentration of microtubule protein required for assembly is reduced, and the microtubules formed are stable to depolymerization by calcium or cold (Schiff et al., 1979). Taxol also assembles tubulin under conditions in which polymerization would not normally occur; these include the absence of microtubule-associated proteins, exogenously added GTP, or organic buffer (Kumar, 1981; Schiff and Horwitz, 1981b). Taxol induces microtubule assembly even at low temperatures (Hamel et al., 1981; Schiff et al., 1979; Thompson et al., 1981b). This taxol-induced assembly at low temperature requires the presence of either GTP or microtubule-associated proteins (Hamel et aL, 1981). Microtubules assembled to steady state and then incubated with taxol become resistant to depolymerization by calcium, suggesting that there is a taxol binding site on the microtubule. This taxol binding site is distinct from the exchangeable GTP binding site and the binding sites for colchicine or podophyllotoxin and vinblastine (Schiff and Horwitz, 1981b; Kumar, 1981). Taxol does not inhibit the GTP hydrolysis reaction that normally accompanies assembly (Schiff and Horwitz, 1981b). Maximal effects in vitro are seen at taxol concentrations stoichiometric with the tubulin dimer concentration (Schiff and Horwitz, 1981a). Treadmilling of in vitro microtubules is substantially reduced by treatment with taxol (Schiff and Horwitz, 1981b; Kumar, 1981; Thompson et al., 1981a; Caplow and Zeeberg, 1982). [3H]Taxol binds to assembled microtubules in vitro with a stoichiometry approaching one mole of taxol bound/mole of tubulin dimer in the polymerized form. When added prior to assembly, podophyllotoxin and vinblastine inhibit the binding of [3H]taxol in a complex manner which probably reflects a competition between these drugs, not for a single binding

96

J . J . MANFREDI and S. B. HORWITZ

site, but for different polymeric forms (dimer or microtubule) of tubulin (Parness and Horwitz, 1981; Horwitz et al., 1982). In vitro, taxol does not influence actin polymerization nor does it bind to intermediate filaments or DNA (Parness and Horwitz, 1981). The drug also does not affect the interaction of microtubules with their associated proteins in vitro (Kumar, 1981; Schiff and Horwitz, 1981b; Vallee, 1982) or in cells (DeBrabander et al., 1981c). Taxol assembles highly purified tubulin, and these microtubules bind [3H]taxol and are more stable to podophyllotoxin-induced disassembly than control microtubules (Parness and Horwitz, 1981; Schiff and Horwitz, 1981b).

5.2. EFFECTS OF TAXOL IN CELLS 5.2.1. Taxol Induces the Formation of Parallel Arrays of Cellular Microtubules HeLa cells incubated with taxol accumulate in the (32 and M phases of the cell cycle. In contrast to cells treated with colchicine or vinblastine, these taxol-treated cells exhibit an unusual interphase microtubule cytoskeleton as seen by tubulin immunofluorescence and electron microscopy. These taxol-cytoskeletons are characterized by the presence of bundles of microtubules (Fig. 2D; Schiff and Horwitz, 1980; DeBrabander et al., 1981a; DeBrabander, 1982; Albertini and Clark, 1981; Brenner and Brinkley, 1982). These bundles tend not to be associated with the microtubule organizing center. Electron microscopic examination of the centrosome of taxol-treated PtK~ cells show few, if any, microtubules associated with it (DeBrabander et al., 1981a). Mouse splenic lymphocytes are an exception to this. In these cells, taxol induces the formation of a single bundle of microtubules. The centrosome is displaced towards the cell periphery and the bundle of microtubules extends from the centrosome into the cytoplasm (Paatero and Brown, 1982). Roberts et al. (1982) observed taxol-induced microtubule bundles associated with the centrosomes in human neutrophils, but their treatment was with 1/tM taxol for 45 min; longer incubations may be necessary to observe bundle formation which is independent of an MTOC. Brenner and Brinkley (1982) noted electron dense material at one end of taxol-induced bundles that form in colcemid-reversed 3T3 cells. Treatment of J774.2 cells with 30 #M taxol for 60 min at 37°C induces the formation of microtubule bundles that are not associated with the microtubule organizing center (Fig. 14b). These bundles tend to form in the peripheral cytoplasm and contain microtubules that appear to be shorter in length than microtubules in control cells which sometimes can span the entire length of the cell. Further, in control cells, it is unusual to observe two or more microtubules in a parallel array. Saturation of [1H]taxol binding is reached within 60min at 0.3 #M taxol (Fig. 9), however, no effect is seen on cellular microtubule organization at that time (Fig. 13D). If cells are treated with 0.3 #M taxol for 24 hr, the characteristic bundle formation is observed (Fig. 13A). Thus, this novel reorganization of cellular microtubules is both a concentration-dependent and time-dependent phenomenon. Examination of these taxol-induced bundles by electron microscopy demonstrates that the bundles are composed of parallel arrays of microtubules (Fig. 7). Thin sections of control cells show that microtubules remain in the plane of section only for short distances; this reflects the fact that these microtubules curve and bend. Thin sections of taxol-treated cells show microtubules remaining in the plane of section for fairly long distances; this probably reflects the relative straightness of the microtubules in the taxol-induced parallel arrays. Extraction of cells with the non-ionic detergent Triton, releases a considerable amount of soluble cellular material. Cytoskeletal elements are retained and examination of thin sections of these preparations provides much greater clarity and resolution of these elements, particularly microtubules. Extracted preparations of taxol-induced bundles consist of parallel arrays of cellular microtubules. These parallel arrays can be observed both in longitudinal- and cross-section. These microtubules do not have extensive longitudinal contacts nor are they as tightly packed as actin filaments in the microvillus

Taxol

97

FIG. 7. Electron microscopyof taxol-treated J774.2 cells. J774.2 cells were incubated with complete medium (DME supplemented with 20~ heat-inactivated horse serum, penicillin, streptomycin, glutamine and non-essential amino acids) containing 30 pu taxol for 60 min at 37°C, and processed for electron microscopyas described in the legend to Fig. 4.44,800X. The arrows point to the limits of the most prominent parallel array of microtubules.

core (Matsudaira and Burgess, 1979), but cross-bridges can be observed between adjacent microtubules within these parallel arrays (Fig. 8). Treatment of mitotic J774.2 cells with 30/~M taxol for 60 min at 37°C induces the formation of numerous fluorescent foci asters scattered uniformly throughout the cytoplasm. These cells are identified as mitotic because no nuclei are observed by phase contrast microscopy. Similar results have been observed by Brenner and Brinkley (1982), DeBrabander et al. (1981 a) and DeBrabander (1982). Electron microscopic examination of these structures indicated that they are asters of microtubules which radiate out from a center that does not contain any centrioles; these asters are not associated in any way with condensed chromosomes (DeBrabander et al., 1981a). Thus, taxol treatment prevents the formation of the normal mitotic spindle. Brenner and Brinkley (1982) also observed electron dense material at the center of these asters that does not react with an antisera J.P.T. 2 5 / 1 ~

98

J . J . MANFREDI and S. B. HORW1TZ

A

FIG. 8. Electron microscopy of extracted, taxol-treated J774.2 cells. J774.2 cells were incubated with complete medium (see legend to Fig. 7) containing 30/~M taxol for 60 min at 37°C, extracted with 1~o Triton in 0.1 M PIPES pH 6.9, 1 mM MgSO4, 2 M glycerol and 2 mM EGTA, and processed for electron microscopy as described in the legend to Fig. 4. 99,000X. The arrowheads point to sites at which adjacent microtubules appear to be linked by cross-bridges.

specific for centrioles. Treatment of mitotic Haemanthus endosperm cells with taxol causes cell division to be retarded, but not arrested. Anaphase appears to be temporarily reversed in that chromosomes move back towards the spindle equator before resuming their poleward migration (Bajer et al., 1982). There is a promotion of microtubule assembly, particularly of polar, aster-like microtubules which are not associated with kinetochores, and a marked enhancement of lateral associations between the microtubules of the spindle (Mole-Bajer and Bajer, 1983). Cande et al. (1981) found that taxol added to PtK1 cells at anaphase retards chromosome movement and blocks spindle elongation. Treatment with taxol stabilizes cellular microtubules to depolymerization by cold, steganacin (Schiff and Horwitz, 1980; 1981a) and colchicine (Crossin and Carney, 1981a). Taxol injected into unfertilized Xenopus eggs induces aster-like tubulin structures (Heidemann and Gallas, 1980). In unfertilized sea urchin eggs, taxol likewise induces the formation of punctate microtubule-containing asters which do not contain centrioles (Bestor and Schatten, 1982; Schatten et al., 1982). In fertilized eggs, the sperm aster is incompetent to complete migration of the egg nucleus (Schatten et al., 1982). Treatment with taxol blocks amoebae of Physarum polycephalum in mitosis causing the accumulation of cells with monoasters surrounded by condensed chromosomes. Escape from this block involves the formation of a long microtubule bundle which interacts with a small chromosomal mass outside the monoaster (Wright et al., 1982). Centriole replication is not affected by taxol, and the monoasters do not necessarily have centrioles at their center (Wright and Moisand, 1982). Treatment of the microtubule-containing axonemes of the heliozoan Actinophyrs sol with taxol causes a disruption of the normal double spiral pattern due to a large increase in the length and number of microtubules and a loss or inactivation of intermicrotubule linkages (Hausmann et al., 1983). The drug induces abnormal microtubule arrays in the various cell types of primary mouse dorsal root ganglion-spinal cord cultures; these arrays

Taxol

99

include microtubules aligned along the endoplasmic reticulum and hexagonal groupings of microtubules in the cytoplasm (Masurovsky et al., 1981; 1983). Microtubules that were aligned along the endoplasmic reticulum were also observed in HeLa cells by Schiff and Horwitz (1980). 5.2.2. [3H] Taxol Binds to Cellular Microtubules in a Specific, Saturable, and Reversible Manner Specific binding of [3H]taxol to macrophage-like J774.2 cells saturates in a concentration-dependent manner. Maximal binding occurs at 21 pmol bound/mg total cellular protein; half-maximal binding occurs at 0.08 #M. As to be expected, non-specific binding increases linearly in a concentration-dependent manner, ranging from 25~ of total binding at 0.01 # u to 45~o at 0.4 #M (Fig. 9A). Scatchard analysis of the specific binding data produces a linear plot, indicating a single set of binding sites (Fig. 9B). The binding of 0.3 #M [3H]taxol is completely reversible after 90 min. Unlabeled taxol readily displaces the [3H]taxol. Colchicine also facilitates the release of [3H]taxol, but not as well as unlabeled taxol and the kinetics of the colchicine-facilitated release are more complex than a simple displacement curve (Fig. 10B). The efltux of [3H]taxol from J774.2 cells is inhibited at 6°C regardless of the presence of ten-fold excess of unlabeled taxol or colchicine (Fig. 10B). Simultaneous addition of [3H]taxol and a 100-fold excess of either colchicine, podophyllotoxin, vinblastine, nocodazole, or unlabeled taxol completely inhibits specific binding to J774.2 cells. Griseofulvin (Fig. 6), an antimitotic agent whose mechanism of action is still controversial, but whose cellular activity may not be mediated by microtubule depolymerization (Cox et al., 1979; Grisham et al., 1973), does not inhibit the binding of [3H]taxol. Lumicolchicine and VP-16-213, congeners of colchicine and podophyllotoxin, respectively, which do not interact with tubulin (Loike and Horwitz, 1976; Wilson and Friedkin, 1967) are also without effect. The binding results are compatible with observations made with tubulin immunofluorescence. Simultaneous treatment with 3 0 # M vinblastine and 0.3 #M [3H]taxol results in a disruption of cytoskeletal structure and the formation of numerous well-defined paracrystals. Since the binding of [3H]taxol is inhibited in these preparations, vinblastine-induced paracrystals must not bind taxol.

v

~20 I

!

CI Z I0

v

i

20

3o

x 150

/

i ,°°

/9,..( ''~"

x ,,o,,'6"

o'.~

<>'.2

o13

3H- TAXOL (~.M)

o'.4

I

10

3H-TAXOLBOUND

FIG. 9. [3H]Taxol binds to J774.2 cells in a saturable manner. (A) Confluent 35 mm dishes of J774.2 were incubated for 60 min at 37°C in 2 ml of complete medium (see legend to Fig. 7) containing various concentrations of [3H]taxol. Cells were washed five times with 2 ml of ice-cold PBS, and lysed in 1 ml of 0.1 N NaOH. Radioactivity was determined by liquid scintillation counting and protein concentrations were determined by the method of Lowry et al. (1951). Specific binding (O) is calculated as the difference between binding of [3H]taxol in the presence and absence of a 100-fold excess of unlabeled taxol. Non-specific binding (O) is binding of [3H]taxol in the presence of a 100-fold excess of unlabeled taxol. Values are the average of duplicate determinations. (B) Scatchard analysis reveals a single class of specific binding sites for [3H]taxol in J774.2 cells. The specific binding data from (A) have been plotted in the appropriate Scatchard analysis form. The line was determined by a least squares linear regression. Units on the abscissa are pmol [3H]taxol/mg protein. Units on the ordinate are pmol [~H]taxol/mg protein/#r,l. (From Manfredi et al., 1982.)

100

J . J . MANFREDI a n d S. B. HORW1TZ r A

c

i

l

i

~

--i

i

i _

30

"-6

2o ~

2o Z

8

o io

e~

0

30

60 TIME

;

3b

~o

9'o

e~

(rain)

FIG. 10. The binding of [3H]taxol to J774.2 cells is reversible. (A) Confluent 35 mm dishes of J774.2 cells were incubated with 2ml of complete medium (see legend to Fig. 7) containing 3/~M nocadazole for 120rain at 37°C. Cells were washed three times with 2ml of DEM at room temperature, incubated with 2ml of complete medium containing 0.3pM [3H]taxol for the appropriate time interval at 37°C, washed five times with 2 ml of ice-cold PBS and lysed in 1 ml of 0.1 N NaOH. Radioactivity and protein concentration were then determined. Values are the average of triplicate determinations. (B) Confluent 35 mm dishes of J774.2 cells were incubated with 2 ml of complete medium containing 3 # Mnocodazole for 120 min at 37°C. Cells were washed three times with 2 ml of DME at room temperature, and further incubated with 2 ml of complete medium containing 0.3/~M [3H]taxol for 60 min at 37°C. They were then washed five times with 2 ml of ice-cold PBS and incubated with 2ml of complete medium alone (0), or complete medium containing 3 #M taxol (A), or 3/~M colchicine (111) for the appropriate time interval at 37°C, or complete medium alone (C)), or complete medium containing 3/tM taxol (A), or 3 #M colchicine (D) for 90 min at 6°C. Cells were washed five times with 2 ml of ice-cold PBS and lysed in 1 ml of 0.1 N NaOH. Radioactivity and protein concentration were determined. Values are the average of triplicate determinations. Similar results were obtained in the absence of pretreatment with nocodazole.

Hence, the taxol binding site present on one polymeric form, the microtubule, is absent from another, the vinblastine-induced paracrystal (Manfredi et al., 1982). Microtubule disassembly can cause a rapid, characteristic change in cell shape. Treatment with agents such as colchicine or podophyllotoxin induce the formation of a large bulge or protuberance in the cytoplasm of J774.2 (Fig. 13E; Manfredi et al., 1982; Melmed et al., 1981). Vinblastine-induced paracrystals appear to be excluded from the protuberance. This is presumably due to their displacement by actin filaments which are concentrated in these regions (Albertini et al., 1977). Pretreatment with 3.0/~M nocodazole for two hours followed by washing and incubation with 0.3 ]AM [3H]taxot results in binding similar to that of control cells (Manfredi et al., 1982). When these cells are examined by tubulin immunofluorescence, a microtubule cytoskeleton is observed (Fig. 13F). This is probably true because the effects of nocodazole are readily reversible (DeBrabander et al., 1981b; Solomon, 1980). The effects of other antimitotic drugs, however, are less reversible; for example, the binding of colchicine to tubulin in vitro is essentially irreversible (Banerjee and Bhattacharyya, 1979; Wilson and Friedkin, 1967). Pretreatment with colchicine, colcemid or vinblastine, followed by washing, results in inhibition of [3H]taxol binding. Immunofluorescence of these cells reveals the complete loss of all microtubule structure, and formation of the characteristic protuberance (Manfredi et al., 1982). Cells pretreated with nocodazole, washed, and then incubated with 0.3 #M [3H]taxol for 60 rain form bundles of microtubules which are not associated with an organizing center (Fig. 13F). This differs from what is seen when 0.3 ]AM taxol is added directly to a control cell, where at 60 min no alteration in the tubulin cytoskeleton is seen (Fig. 13D). Similar results have been reported by others (DeBrabander et al., 1981a; DeBrabander, 1982; Brenner and Brinkley, 1982). Extraction of cells with non-ionic detergents such as Nonidet P-40 (NP-40) or Triton X-100 releases most cellular lipid and soluble protein (Lenk et al., 1977) including

Taxol

101

unassembled tubulin (Duerr et al., 1981). The remaining cytoskeletal structure consists mainly of various filaments and microtubules, and their associated proteins (Schliwa and van Blerkom, 1981; Schliwa et al., 1981b). If [3H]taxol is allowed to bind directly to such cytoskeletons, a major fraction of the binding seen with unextracted cells is retained (Table 1). Since these cytoskeletons have been washed free of their dimer pool, unassembled tubulin does not appear to contribute to the binding of [3H]taxol, and thus it is unlikely that the addition of 0.3/AM [3H]taxol to intact cells is affecting the extent of microtubule polymerization. This also supports the in vitro demonstration that taxol binding is not dependent on a free tubulin dimer pool (Parness and Horwitz, 1981); there is a taxol binding site on the microtubule. If cytoskeletons are extracted in the presence of 5 mu CaC12, 70~o of the specific binding is lost without an appreciable change in non-specific binding (Table 1). Calcium extraction disrupts the microtubule cytoskeleton in these preparations (Schliwa et al., 1981 a; Osborn and Weber, 1977). SDS-polyacrylamide gel analysis of these extracts reveal a band comigrating with bovine brain tubulin in the calcium extract but not in the control; the calcium treatment facilitates the release of tubulin (Solomon et al., 1979). Thus, when the calcium concentration is increased, the resulting release of tubulin causes a disruption of the microtubule cytoskeleton, and a concomitant decrease in the binding of [3H]taxol (Manfredi et al.,~ 1982). The specific binding of [3H]taxol remaining after calcium extraction (Table 1) is probably due either to cells which are not completely lysed or to extracted cells in which the microtubules are not totally depolymerized. In contrast to intact cells, treatment of cytoskeletons with 0.3/AM [3H]taxol in the presence of 30/AM unlabeled taxol for 60 min does not alter the tubulin cytoskeletal organization as seen by tubulin immunofluorescence. The organizing center remains visible and no bundle formation occurs. An intact cell appears to be necessary for the cytoskeletal rearrangement to occur that results in microtubule boundle formation. The total tubulin content of J774.2 cells was determined by a [3H]colchicine binding assay in which a correction is made for the time-dependent decay of colchicine binding to tubulin at 37°C (Wilson et al., 1974). J774.2 cells contain 57.6 pmol tubulin/mg total cellular protein. This represents 0.6~o of the total cellular protein or 1.2~ of the soluble TABLE 1. Binding o f [3H]Taxol to Extracted J774.2 Cells

[3H]Taxol Bound (pmol/mg protein) Non-Specific Unextracted cell Ca 2+-treated unextracted cell Cytoskeleton Ca2+-treated cytoskeleton

5.1 7.5 4.6 4.6

_ + + +

1.1 0.4 1.1 0.4

Specific 18.7_ 18.2 _ 17.0 _ 6.2 +

2.0 2.6 1.3 1.2

Subconfluent 35 mm dishes of J774.2 were washed twice at room temperature with 2 ml of 0.1 u PIPES pH 6.9, 1 mM E G T A , 4 ~ polyethylene glycol 6000 and then incubated for 5 rain at 37°C with 1 ml of this buffer alone (unextracted cells), buffer containing 5 m~l CaC12 (Ca2+-treated unextracted cells), buffer containing 0.17o NP-40 (cytoskeletons) or buffer containing 0.1~o NP-40 and 5ram CaCI2 (Ca2+-treated cytoskeletons). Cells were subsequently washed twice at room temperature with 2 ml of buffer and the extraction procedure was repeated. After two additional washes with buffer at room temperature, the dishes were further incubated with 2 ml of buffer containing 0 . 3 / ~ [SH]taxol _+ 3 0 # u unlabeled taxol for 45 rain at 37°C. Cells were washed four times with 2 ml of buffer at room temperature, and lysed in 1 ml of 0.1 N N a O H . Radioactivity was determined by liquid scintillation counting and protein concentration was determined by the method of Lowry et al. (1951). Since detergent extraction releases two-thirds of the cellular protein (Lenk et al., 1977) the bound radioactivity in the extracted samples was related to the amount of protein present without NP-40 treatment, that is, to the amount of protein in the unextracted cell samples. Values are the average of six determinations + one standard deviation. See Manfredi et al., 1982 for details.

102

J . J . MANFREDI and S. B. HORW1TZ

protein pool. If the binding of [3H]taxol reflects the amount of polymerized tubulin in these cells, then approximately 36% of the total tubulin pool is in the polymerized form in J774.2 cells. Values in the literature for the amount of tubulin in the polymerized form in a variety of cell types range from 2-5% in sea urchin eggs to 90% in human platelets (Fulton and Simpson, 1979). Hiller and Weber (1978) estimate that approximately 40% of the total tubulin is polymerized in mouse 3T3 cells. Since conditions which depolymerize microtubules in cells inhibit binding of [3H]taxol, it is suggested that the single set of saturable, specific binding sites represents cellular tubulin in its polymerized form. As noted above, binding of taxol to the J774.2 cell line saturates at 21 pmol taxol/mg total cellular protein. Assuming one mole of taxol bound per mole of polymerized tubulin (Parness and Horwitz, 1981), simple calculations reveal that 0.3% of the total cellular protein is tubulin in the polymer form. Such a value is well within the range of total tubulin content reported here for J774.2 cells (0.6%) and in the literature for human peripheral unseparated monocytes (0.6%) and purified lymphocytes (0.4%, Sherline and Munday, 1977), as well as guinea pig peritoneal macrophages (0.3%, Pick et al., 1979). If cells are pretreated for two hours at low temperature, subsequent binding of [3H]taxol at low temperature is inhibited (Fig. 11). This lack of binding correlates with a loss of microtubles as seen by tubulin immunofluorescence. Unlike drug-induced microtubule disassembly in J774.2, cold treatment does not cause the unusual change in cell shape described above. This may be due to incomplete depolymerization of cellular microtubules, as the organizing center and occassionally small stubs of microtubules can still be visualized in these cells. Several laboratories have shown that taxol promotes microtubule polymerization in vitro at 4°C (Hamel et al., 1981; Schiff et al., 1979; Thompson et al., 1981 b); this suggests that taxol binds to tubulin even at low temperatures. Thus, it is likely that the inhibition of [3H]taxol binding in cells at low temperatures may not be totally microtubule related; low temperature may be inducing non-specific membrane changes which result in inhibition of the transport of [3H]taxol across the plasma membrane. Specific binding of [3H]taxol is completely inhibited at both 6°C and 10°C. Non-specific binding of the drug is completely inhibited at 6°C but is recovered at 10°C (Fig. 11). These results are consistent with the notion that the specific binding reflects the polymerization state of cellular microtubules; cellular microtubules are depolymerized at both 6°C and 10°C. The non-specific component represents simple diffusion across the plasma membrane; inhibition of non-specific binding at 6°C but not 10°C may reflect differences in the fluidity of the membrane at these temperatures (Inesi et al., 1973). This correlates with the observation that the efflux of [3H]taxol from J774.2 cells is inhibited at 6°C (Fig. 10B). Mammalian red blood cells do not contain tubulin (Dustin, 1978). The binding of [3H]taxol to sheep erythrocytes does not saturate in a concentration dependent manner, but is linear. There is no specific binding of the drug to these cells; the same binding of

O~

81-

/,/,A

m ~- 61-~, /

.~" /

x... 4[--z ~ ~

_------

-L.T.-~4~'-~ ~ 0

" - --

0.1

9 0.2

"

r 0.3

aH- TAXOL (/4M )

FIG. 11. Effect o f low temperature on the concentration dependence of [3H]taxol binding to J774.2 cells. Confluent 35 mm dishes of J774.2 cells were pretreated in complete medium for 120 min at either 6°C ( 0 , ©) or 10°C (A, A). This medium was then aspirated and the cells incubated for 60min at the same respective temperature in 2 ml of complete medium containing various concentrations of [3H]taxol. Cells were washed five times with 2 ml of ice-cold PBS, and lysed in 1 ml o f 0.1 N NaOH. Radioactivity and protein concentration were then determined. Specific binding ( 0 , A ) is calculated as the difference between binding of [3H]taxol in the presence and absence of a 100-fold excess of unlabeled taxol. Non-specific binding (O, A) is binding of [3H]taxol in the presence o f a 100-fold excess of unlabeled taxol. Values are the average of duplicate determinations.

Taxol

103

[3H]taxol is observed in the presence and absence of a 100-fold excess of unlabeled taxol (Fig. 12). This lack of specific binding reflects the lack of a receptor for [3H]taxol in these cells. The limited solubility of the drug in aqueous solution and its molecular structure suggest that taxol is hydrophobic. Depolymerization of cellular microtubules with drugs such as colchicine removes the receptor for [3H]taxol. Saturable binding is no longer seen; the concentration dependent binding of [3H]taxol under these conditions is linear. Since mammalian red blood cells do not contain tubulin (Dustin, 1978), the concentration dependence of [3H]taxol binding to these cells is also linear. Both the uptake and efflux of [3H]taxol is inhibited at low temperature. All these observations suggest that the drug is probably transported across the plasma membrane by passive diffusion, and it is the specific binding of taxol to the microtubule, its cytoplasmic receptor, which results in a saturating binding curve for the drug. Since the binding of [3H]taxol saturates, is diluted by an excess of unlabeled taxol, and is inhibited by conditions which depolymerize microtubules, it is unlikely that the specific binding of the drug reflects its partitioning into the cell membrane or cellular interior. Thus, taxol binds directly to the pre-existing cytoplasmic microtubule complex. It does not bind vinblastine-induced paracrystals but rather, specifically recognizes the microtubule. In the original studies with taxol, Schiff and Horwitz, (1980) used primary Balb/c fibroblasts in which it is difficult to identify a microtubule organizing center. Bundle formation was demonstrated but it was difficult to discern whether these were MTOCassociated. DeBrabander et al., (1981a), using a cell line with a more prominent MTOC (PtK2), were able to definitively observe bundles independent of the organizing center. Using 3T3 cells and colcemid reversal, Brenner and Brinkley (1982) corroborated this finding. The studies with the J774.2 cell line concur. The clearly visualized organization of the macrophage cytoskeleton makes it an ideal tool for such studies. The mechanism of microtubule bundle formation is not clear. This taxol-induced process requires an intact, unextracted cell. Since conditions for in vitro microtubule bundle formation have yet to be identified, it is likely that there are specific cellular conditions which are necessary for this unusual cytoskeletal reorganization. DeBrabander and associates (DeBrabander et al., 1981a) have studied the time course of taxol-induced cytoskeletal rearrangements in control and nocodazole-reversed PtK2 cells visualized by peroxidase-antiperoxidase immunostaining. They suggest that the pre-existing tubulin cytoskeleton depolymerizes as bundles form. This idea is supported by our observation that depolymerization of cytoplasmic microtubules appears to facilitate bundle formation. Treatment with 0.3/~M taxol for 60min does not induce bundles in J774.2 (Fig. 13D). However, if the cytoplasmic microtubules are depolymerized by nocodazole, the cells washed and 0.3 # u taxol is then added, bundles do form within 60 min (Fig. 13F). Brenner and Brinkley (1982) using colcemid reversal have reported similar results.

I

I

I

I

I

2.0 o

CO

"r

0.1 0.2 aH. TAXOL {~M )

I

0.3

FIG. 12. Binding of [3H]taxol to sheep erythrocytes. Aliquots of sheep erythrocytes were incubated for 60 rain at 37°C in 0.1 M PIPES pH 6.9, 1 mu EGTA, and 4~o polyethylene glycol 6000 containing various concentrations of [3H]taxol. Cells were washed three times with ice-cold PBS, and lysed in 0.1 N NaOH. Values are the average of duplicate determinations.

104

J.J. MANFREDIand S. B. HORWITZ

FIG. 13. Evidence that taxol-induced bundle formation involves depolymerization of preexisting microtubules. (A and B) J774.2 cells were treated with 0.3 #M taxol for 24 hr. (C and D) J774.2 cells were treated with 0.3~M taxol for l hr. (E and F) J774.2 cells were treated with 3~M nocodazole for 2 hr (E), washed and further incubated with 0.3,UM taxol for l hr (F). Cells were processed for immunofluorescence as described in Manfredi et al., 1982. Bar is 20 #m.

We have shown that taxol binds specifically to cellular microtubules. The ability of [3H]taxol to stoichiometrically bind to cellular microtubules m a y prove to be an ideal assay for quantitating tubulin in its polymer form in cells. Such an assay is sorely lacking in the current arsenal of microtubule technology and is necessary if we are to satisfactorily answer key questions concerning the role of the tubulin polymer-dimer equilibrium in a variety of cellular functions. For example, the effect of cycloheximide or actinomycin on the colchicine-induced depolymerization of cellular microtubules can be assayed by the binding of [3H]taxol. Treatment of J774.2 cells with cycloheximide or actinomycin inhibits the incorporation of [3H]leucine or [3H]uridine, respectively, into trichloroacetic acid insoluble material but has no effect on the binding of [3H]taxol or the ability of 30/tM colchicine to inhibit the binding of 0.3 #N [3H]taxol (Table 2). Thus, protein or R N A synthesis does not appear to be necessary for either binding of [3H]taxol or colchicineinduced depolymerization of cellular microtubules. Pretreatment of cells with 10 mM NaN3 for 60 min at 37°C depletes cellular A T P levels but has no effect on either the specific or non-specific binding of [3H]taxol (Manfredi et al., 1982). Such azide treatment likewise has no effect on the organization of cellular microtubules as seen by tubulin immunofluorescence (DeBrabander et al., 1980). However, under conditions in which taxol induces microtubule bundles that are independent of an organizing center in control cells (Fig. 14B), no such alteration is seen in the tubulin cytoskeleton of azide-treated cells (Fig. 14D). Pretreatment of cells with 3/~M nocodazole depolymerizes all cellular microtubules. The effects of nocodazole are readily reversible; washing of the cells allows for complete

105

Taxol TABLE 2. Effect o f Cycloheximide and Actinomycin on Binding o f [ 3H]Taxol to J774.2 Cells

[3H]Taxol

Precursor Incorporation (nmol/plate)

Bound (pmol/mg protein)

Control +Cycloheximide +Actinomycin

[3H]Taxol Alone

+ 30 ~M Taxol

+ 30 ~M Colchicine

[3H]Leucine

[3H]Uridine

46.6 + 0.6 47.6 + 2.2 51.7 + 1.4

19.2 + 2.2 19.4 + 0.3 23.9 + 0.4

14.7 + 1.0 13.4 + 0.7 13.0 + 0.8

376.0 + 23.0 38.2 + 1.2 288.0 + 1.6

28.9 + 2.2 27.5 + 1.6 1.9 + 0.3

Confluent 35mm dishes of J774.2 cells were incubated with 0.1 u PIPES pH 6.9, 1 mM EGTA, and 4~o polyethylene glycol 6000 alone (control), buffer containing 20/~ M cycloheximide ( + cycloheximide), or containing 10/~M actinomycin ( + actinomycin) for 30 min at 37°C. Cells were then incubated in buffer with the appropriate addition plus 0.3 #M [3H]taxol or plus 0.3 # u [3H]taxol and 30 pM unlabeled taxol or 30 #M colchicine for 60 rain at 37°C. They were then washed five times with ice-cold PBS and lysed in 1 ml of 0.1 N NaOH. Radioactivity was determined by liquid scintillation counting and protein concentration was determined by the method of Lowry et al. (1951). Precursor incorporation was determined for confluent 35 mm dishes of J774.2 cells. Values are the average of triplicate determinations + one standard deviation.

reassembly of a normal cytoplasmic microtubule complex. Depletion of cellular ATP by treatment with sodium azide does not inhibit microtubule reassembly, but it does prevent the formation of an organized microtubule complex (Fig. 14F). What is observed is an unusual tubulin fluorescent display which DeBrabander and colleagues, using peroxidase-

FIG. 14. Taxol-induced bundle formation is an ATP dependent process. (A and B) J774.2 cells were treated with 30 #M taxol for 1 hr. (C and D) J774.2 cells were treated with 10 mM NaN 3 for 1 hr (C), and filrther incubated with 10 mM NaN 3 and 30 pM taxol for 1 hr (D). (E and F) J774.2 cells were treated with 3 #M nocodazole for 1 hr, 3/IM nocodazole and 10 mM N a N 3 for 1 hr (E), washed and further incubated with 10 mM NaN 3 and 30 #M taxol (F). (G and H) J774.2 cells were treated with 3 pM nocodazole for 1 hr, 3 pM nocodazole and 10 mM NaN3 for 1 hr (G), washed, and further incubation with 10 mM NaN 3 (H). Cells were processed for immunofluorescence as described in Manfredi et al., 1982. Bar is 20/~m.

106

J.J. MANFREDIand S. B. HORWITZ

antiperoxidase staining have resolved as numerous, randomly-oriented microtubules scattered throughout the cytoplasm and not associated with the microtubule organizing center (DeBrabander, 1982). Cellular microtubules are depolymerized by treatment with nocodazole, followed by treatment with sodium azide to deplete cellular ATP levels. If cells are then washed to remove the nocodazole and reassembly is observed in the presence of 30 #M taxol, no effect of taxol is seen (Fig. 14H). The fluorescent display is identical in the control recovery and taxol-induced recovery (Fig. 14F and H). Under such conditions, the uptake of [3H]taxol is unaffected; the drug is entering the cells. Clearly, taxol-induced bundle formation after reversal from the effects of nocodazole is energy dependent. Taxol-induced bundle formation requires ATP regardless of whether the taxol is added to cells with a fully assembled microtubule complex or to cells with completely disassembled microtubules. These nocodazole-reversed, ATP-depleted preparations were examined by electron microscopy in an attempt to better resolve the ultrastructural basis for this unusual fluorescent display. Treatment with sodium azide, however, has sufficiently affected the cellular ultrastructure that it is impossible to identify any cytoskeletal elements in the cytoplasm. The cytoplasm has collapsed on the nucleus and what little cytoplasm can be seen is obscured by the presence of numerous electron-dense granules, preventing any further resolution of the cytoskeletal ultrastructure. Depletion of cellular ATP can inhibit depolymerization by cotchicine or vinblastine (Bershadsky and Gelfand, 1981; Moskalewski et al., 1980). Since ATP is also required for taxol-induced bundle formation, it is interesting to speculate that microtubule depolymerization may be an important part of the cytoskeletal reorganization caused by taxol. In vitro data demonstrating that taxol suppresses microtubule treadmilling (Schiff and Horwitz, 1981b; Kumar, 1981; Thompson et al., 1981a; Caplow and Zeeberg, 1982) would tend to discredit depolymerization as an intermediate step in bundle formation. As yet, however, treadmilling of microtubules in cells has not been demonstrated, nor is it unreasonable to envisage unique cellular mechanisms that could overcome taxol-induced microtubule stability. The inhibition of taxol-induced bundle formation in the extracted cell may reflect the need for ATP in this process. Bershadsky and Gelfand (1981) found, for example, that addition of ATP in the absence of any microtubule depolymerizing drugs causes rapid depolymerization of the microtubules of the cytoskeleton. Treatment of J774.2 cells with sodium azide causes a variety of ultrastructural alterations. The observation that sodium azide inhibits the effects of taxol and colchicine may not be directly related to ATP depletion. There are a number of consequences of ATP depletion which could be responsible for the inability of taxol or colchicine to exert their effects. If the ATP dependence of taxol-induced bundle formation is due to the need for microtubule depolymerization prior to assembly of the microtubule bundles, then bundle formation from cells in which microtubules are already depolymerized should not be energy dependent. This is clearly not the case. Taxol-induced bundle formation from nocodazole-reversed cells is inhibited by treatment with sodium azide (Fig. 14F). This suggests that there are, in fact, two energy dependent steps in taxol-induced bundle formation. The first is depolymerization of pre-existing cellular microtubules and the second is the actual formation of microtubule bundles. 5.2.3. Taxol Stabilizes Cellular Microtubules against Calcium Extraction but not ColchicineInduced Depolymerization The plasma membrane forms an impermeable barrier to calcium ions making the cell relatively insensitive to changes in the calcium ion concentration of the external medium. The calcium ionophore A23187, has been used to demonstrate the calcium sensitivity of cellular microtubules in heliozoan axopodia (Schliwa, 1976) and 3T3 fibroblasts (Fuller and Brinkley, 1976). Although a variety of experimental conditions were tried, no effect of A23187 and calcium ion was demonstrable in J774.2 cells. We therefore used the

Taxol

107

detergent extraction method of Solomon et al. (1979) to show the calcium sensitivity of cellular microtubules. If J774.2 cells are extracted with 1~o Triton, electron microscopic examination of thin sections of embedded samples shows that the plasma membrane is removed, as well as all membrane-bound intracellular organelles. What remains includes the nuclear matrix, polysomes, and a cytoskeletal network consisting of actin filaments, intermediate filaments, and microtubules. Similar observations have been made in other cell types by whole mount high voltage electron microscopy (Schliwa et al., 1981b; Schliwa and van Blerkom, 1981). Such detergent extraction has proven useful in improving tubulin immunofluorescence images (Osborn and Weber, 1977) and has been successfully exploited by Solomon and colleagues in identifying and characterizing microtubule-associated proteins (MAPs) in cultured cell lines (Solomon et al., 1979; Duerr et al., 1981; Zieve and Solomon, 1982; Pallas and Solomon, 1982). J774.2 cells were metabolically labeled with [35S]methionine and then extracted with 1~o Triton. This extraction releases a large number of soluble proteins including presumptive unassembled tubulin, leaving cytoskeletally-associated peptides as seen by twodimensional gel electrophoresis (Fig. 15A). Treatment with a calcium-containing buffer causes subsequent release of tubulin from these cytoskeletons (Fig. 15B). Treatment of J774.2 cells with 30 #M podophyllotoxin for 60 min causes complete depolymerization of cellular microtubules as seen by tubulin immunofluorescence (Manfredi et al., 1982). Examination of [35S]labeled Triton-extracted preparations by two-dimensional gel analysis shows that indeed ~- and /~-tubulin are specifically released (Fig. 15C). The ~- and /~-tubulin spots are thus identified by the criteria of Solomon et al. (1979): comigration with unlabeled purified calf brain tubulin, retention in a Triton-extracted cytoskeleton, release upon extraction with a calcium-containing buffer, and absence in a Triton-extracted cytoskeleton in which there was a prior treatment of the whole cell with a microtubule depolymerizing drug such as podophyllotoxin. Treatment of J774.2 cells with 30/~ M taxol for 60 min induces the formulation of discrete microtubule bundles in the peripheral cytoplasm (Manfredi et al., 1982). Examination of these cells by gel electrophoresis confirms the retention of ~- and/~-tubulin (Fig. 15E). Extraction of these preparations with a calcium-containing buffer does not release tubulin as seen by two-dimensional gel electrophoresis (Fig. 15F). The activity of a number of natural and semisynthetic congeners of taxol have been studied (Fig. 16; Parness et al., 1982). Both cytotoxicity toward the macrophage-like cell line, J774.2, and the ability to promote microtubule assembly in vitro in the absence of exogenous GTP require an intact taxane ring and ester side chain at position C13. Structures A - F (Fig. 16) were cytotoxic to J774.2 cells; structures G - H (Fig. 16) were not. Only structures A, B, E and F (Fig. 16) were able to assemble calcium stable microtubules in vitro in the absence of exogenous GTP. Thus, addition of acetyl moieties at both positions 2' and 7 results in loss of in vitro activity but not cytotoxicity (Fig. 16, structures C and D). Some of these congeners of taxol were examined for their ability to stabilize cellular microtubules against calcium extraction in J774.2 cells as determined by two-dimensional gel electrophoresis. 10-Deacetylcephalomannine (Fig. 16, structure F) enhances microtubule assembly in vitro in the absence of added guanine nucleotide and stabilizes those microtubules against calcium-induced depolymerization (Parness et al., 1982). Treatment with 30#M 10-deacetylcephalomannine for three hours stabilizes the microtubules in J774.2 cells against calcium extraction (Fig. 17B). Three other congeners of taxol, 2',7-diacetyltaxol, 2',7-diacetyl,10-deacetyltaxol, and baccatin III (Fig. 16, structures C, D and G), which do not promote in vitro microtubule assembly in the absence of added GTP, likewise do not stabilize against calcium extraction after a three hour incubation with cells at a drug concentration of 30#M (Fig. 17C-E). A longer incubation could result in stabilization. The stabilization of microtubules against calcium extraction in J774.2 cells occurs in a concentration dependent manner over a range which is consistent with the saturable

108

J . J . MANFREDIand S. B. HORW1TZ

Q

A

t

o

B

-till m

,.J

C

F FIG. 15. Two-dimensional gel analysis demonstrating microtubule stabilization by taxol in intact J774.2 cells. Confluent 16mm wells of J774.2 cells were labeled for 8hr at 37°C in 250pl of methionine-free DME containing [3SS]methionine (0.25 Ci/ml). At 7 hr, 50 #1 of methionine-free DME alone (A and B), or methionine-free DME containing podophyllotoxin for a final concentration of 30 pm (C and D), or containing taxol for a final concentration of 30 pM (E and F) was added. After 1 hr, cells were then extracted with 1~o Triton in 0.1 M PIPES pH 6.9, 1 mM MgSO4, 2M glycerol and 2mM EGTA (A, C a n d E) or 0.1M PIPES pH 6.9 and 1 mM CaC12 (B, D and F), lysed in isoelectric focusing sample buffer, and subjected to two-dimensional gel analysis (O'Farrell et al., 1977). Each gel was loaded with 1.2 × 106cpm in a volume of 5-25#1. The migration of marker proteins is shown by the arrows, ct, fl and A refer to ct-tubulin, fl-tubulin, and actin, respectively.

b i n d i n g c u r v e f o r [3H]taxol in this cell line. H a l f - m a x i m a l b i n d i n g at 60 m i n is seen at 0.05/AM t a x o l w i t h s a t u r a t i o n o c c u r r i n g b y 0.3/AM t a x o l (Fig. 9). S i m i l a r l y , the s t a b i l i z a t i o n a g a i n s t c a l c i u m - i n d u c e d d e p o l y m e r i z a t i o n a p p e a r s to s a t u r a t e by 0.3 #M as seen b y t w o - d i m e n s i o n a l gel e l e c t r o p h o r e s i s (Fig. 18). A t 60 m i n , 0.3 #r,l t a x o l d o e s n o t i n d u c e the

Taxol

O

O

109

laC

II 4' ~' 2' II ,,-c-N.-c.-c.-c-o-4 s'

I

I R3

L~JJ

COMPOUND A. Toxol

\..

r

~

,~°°' J' l ~r~_

~

:I

~Jk . A

I

O--C-CH 3

RI

R2

R3

OCOCH 3

OH

OH

R4

B. IO-deacetyhoxol

OH

OH

OH

C 2', 7-diocetyltaxol

OCOCH 3

OCOCH 3

OCOCH 3

OH

OCOCH 3

OCOCH 3

D. 2', 7-diacetyl, 10-deocetyltoxol E. Cepholomonnine

~)

OCOCH 3

OH

OH

C(CH3)=CHCH 3

OH

OH

OH

C(CH3)=CHCH 3

F. 10- deacetylce pholomannine O

H3C--C--O OI--C--CH3

o:,,,

,x,~..,L A

"~F--CS~i Y

,o

/~x

"O-C-C"=CH-'t(, )?

O - ICI--CH3 0 0

COMPOUND

R

G. Baccatin TIT H. 19-hydro~ybaccatin

CH 3

TIT

CH2OH

COMPOUND

R

0-cinnamoyltoxicin-T

triacetate

OH

J O-cinnamoyltaxicin-~

triacetate

H

I

FIG. 16. Structural formulae of taxol congeners.

characteristic formation of microtubule bundles (Fig. 13D). Hence, binding of [3H]taxol and stabilization against calcium-induced depolymerization are distinct and separate from taxol-induced bundle formation. J774.2 cells are first extracted with Triton to release unassembled tubulin, and then incubated with 0.3 #M taxol for 60 min. Even in the absence of an unassembled tubulin pool, taxol stabilizes cellular microtubules against calcium-induced depolymerization (Fig. 19C-D). This observation is consistent with our in vitro studies demonstrating that the addition of taxol to preassembled microtubules confers calcium stability (Schiff and Horwitz, 1981a). It also confirms the data demonstrating that [3H]taxol binds directly to polymerized microtubules thereby indicating that there is a taxol binding site on the microtubule (Parness and Horwitz, 1981). Microtubules assembled in vitro in the presence of taxol are resistant to depolymerization by calcium. This stabilization occurs at stoichiometric concentrations of taxol to polymerized tubulin. Addition of substoichiometric concentrations of taxol confers only partial resistance to calcium-induced depolymerization; this partial resistance reflects the fact that taxol acts in a stoichiometric manner (Schiff et al., 1979). Treatment of preassembled in vitro microtubules with taxol also confers this stability against the effects of calcium; taxol binds directly to the microtubule thereby conferring stability against calcium-induced depolymerization (Schiff and Horwitz, 1981a). Indeed, the binding of [3H]taxol has a stoichiometry of one taxol bound per tubulin dimer in the polymerized from (Parness and Horwitz, 1981). The presence of microtubule-associated proteins (MAPs) has no effect on the ability of taxol to confer stability against the effects of calcium (Schiff and Horwitz, 1981b; Kumar, 1981). The stabilizing effects of taxol in cells correlate well with these in vitro studies. Incubation of intact cells with taxol confers stability against calcium extraction in both J774.2 and 3T6 cells. This occurs in a concentration dependent manner over a range consistent with the saturable binding curve for [3H]taxol in the J774.2 cell line. Direct addition of taxol to cytoskeletons confers stability against calcium extraction; this further corroborates our findings both in vitro and in cells that taxol binds directly to the microtubule thereby conferring stability to the effects of calcium.

110

J.J. MANFREDIand S. B. HORWITZ

FIG. 17. Two-dimensional gel analysis demonstrating microtubule stabilization by various congeners of taxol in intact J774.2 cells. Confluent 16 mm wells of J774.2 cells were labeled for 8 hr in 250#1 of methionine-free DME containing [35S]methionine(0.25Ci/ml). At 5 hr, 50/11 of methionine-free DME containing taxol (A), 10-deacetylcephalomannine(B), 2',7-diacetyltaxol (C), 2',7-diacetyl, 10-deacetyltaxol(D), or baccatin III (E) for a final concentration of 30 #M in all cases was added. Cells were then extracted with 1~o Triton in 0.1 M PIPES and 1mM CaC12 lysed in isoelectric focusing sample buffer, and subjected to two-dimensional gel analysis (O'Farrell et al., 1977). Each gel was loaded with 1 × l 0 6 cpm in a volume of 12-22 #1. The migration of marker proteins is shown by the arrows. ~, 13, and A refer to ~-tubulin, 13-tubulin,and actin, respectively.

The b i n d i n g o f [3H]taxol reflects the extent o f t u b u l i n p o l y m e r i z a t i o n in b o t h intact a n d extracted cells ( M a n f r e d i et al., 1982). T r e a t m e n t of intact J774.2 cells with 30#M colchicine depolymerizes all cellular m i c r o t u b u l e s a n d inhibits the b i n d i n g o f 0.3 #M [3H]taxol to the same extent as s i m u l t a n e o u s a d d i t i o n o f 30 pM u n l a b e l e d taxol ( M a n f r e d i et al., 1982). If J774.2 cells are first extracted with a n o n - i o n i c detergent such as N P - 4 0 or Brij 58, 30 #M colchicine n o longer has a n y effect o n the b i n d i n g of 0.3/aM [3H]taxol. The ability of u n l a b e l e d taxol to dilute the b i n d i n g of radiolabeled d r u g is, however, unaffected by prior extraction (J. J. M a n f r e d i a n d S. B. Horwitz, u n p u b l i s h e d data). Thus,

Taxol

111

FIG. 18. Two-dimensional gel analysis demonstrating the concentration dependence of microtubule stabilization by taxol in intact J774.2 cells. Confluent 16 mm wells of J774.2 cells were labeled for 8 hr in 250/~1 of methionine-free medium containing [35S]methionine(0.25 Ci/ml). At 7 hr, 50 FI of methionine-free DME containing taxol (for a final concentration of 0.01 # u (A), 0.05 p~ (B), or 0.3/~u (C) was added. After 60 min, cells were then extracted with 1~o Triton in 0.1 i PIPES pH 6.9 and 1 mu CaC12, lysed in isoelectric focusing sample buffer, and subjected to twodimensional gel analysis (O'Farrell et al., 1977). Each gel was loaded with 1 × 106cpm in a volume of 10-20/~1.The migration of marker proteins is shown by the arrows, ct, fl and A refer to ct-tubulin, fl-tubulin, and actin, respectively.

p r i o r detergent extraction prevents the colchicine-induced d e p o l y m e r i z a t i o n o f cellular m i c r o t u b u l e s as j u d g e d by b i n d i n g o f [3H]taxol. Similar results have been reported for p r i m a r y m o u s e e m b r y o fibroblasts by Bershadsky e t al. (1978). F u r t h e r i n c u b a t i o n o f extracted J774.2 cells with 30 pM p o d o p h y l l o t o x i n for 60 m i n does n o t depolymerize

112

J. J. MANFREDI and S. B. HORWITZ

A

C

D

FIG. 19. Two-dimensional gel analysis demonstrating microtubule stabilization by taxol in J774.2 cytoskeletons. Confluent 16 mm wells of J774.2 cells were labeled for 8 hr in 250/~1 of medium containing [35Slmethionine(0.25 Ci/ml). Cells were then extracted with 17o Triton in 0.1 MPIPES pH 6.9, 1 mM MgSO4, 2 N glycerol and 2 mM EGTA, washed once with this buffer, and incubated for 60 min at 37°C in buffer alone (A and B), or buffer containing 0.3 ~M taxol (C and D), or buffer containing 30/2Mpodophyllotoxin (E). Cells were then treated for 10 min at room temperature with either 0.1N PIPES pH 6.9, l mM MgSO4, 2M glycerol and 2mM EGTA (A, C and E) or 0.1N PIPES pH 6.9 and 1mN CaCI2 (B and D), washed twice with the appropriate buffer, lysed in isoelectric focusing sample buffer, and subjected to two-dimensional gel analysis (O'Farrell et al., 1977). Each gel was loaded with 5 x 105cpm in a volume of 25 50 ktl. The migration of marker proteins is shown by the arrows. ~, fl and A refer to ~-tubulin, [:~-tubulin,and actin, respectively.

cellular m i c r o t u b u l e s as seen by t w o - d i m e n s i o n a l gel electrophoresis (Fig. 19E). A n intact cell is required for b o t h p o d o p h y l l o t o x i n - a n d colchicine-induced d e p o l y m e r i z a t i o n o f cellular m i c r o t u b u l e s . T r e a t m e n t of J774.2 cells with 20 mM NaN3 for 60 m i n decreases the cellular A T P level to less t h a n 3~o o f its c o n t r o l level. Such a depletion o f the cellular energy pool prevents 30/~M colchicine from i n h i b i t i n g the b i n d i n g of 0.3/~M [3H]taxol w i t h o u t affecting the ability o f 3 0 / ~ u n l a b e l e d taxol to do so. This azide d e p e n d e n t effect can be reversed by the a d d i t i o n o f glucose to the i n c u b a t i o n buffer, allowing cells to recover their A T P levels t h r o u g h energy p r o d u c t i o n by glycolysis (Table 3). The inability o f colchicine to affect [3H]taxol b i n d i n g in A T P - d e p l e t e d cells is n o t due to reduced u p t a k e o f colchicine after

Taxol

113

azide treatment. In fact, pretreatment with sodium azide enhances the ability of colchicine to enter J774.2 cells on the order of three- to five-fold. See et al. (1974) have previously demonstrated a similar stimulatory effect on colchicine uptake in CHO cells by treatment with the metabolic inhibitors, sodium azide, potassium cyanide, or dinitrophenol. This binding data reflects the energy dependence of colchicine-induced depolymerization of cellular microtubules. J774.2 cells were metabolically labeled with [35S]methionine, their energy levels were depleted by treatment with sodium azide, and they were subsequently incubated in the presence and absence of 30 ~tM colchicine. One-dimensional gel analysis after Triton extraction demonstrated that azide pretreatment indeed prevented the colchicine-induced release of tubulin from the cytoskeletons. Pretreatment with both azide and glucose allowed such colchicine-induced release. These observations correlate with the [3H]taxol binding data and verify results found by others working with a variety of cultured cell types (Moskalewski et al., 1980; Bershadsky and Gelfand, 1981; DeBrabander et al., 1981b). Thus, colchicine-induced depolymerization of cellular microtubules requires ATP. Treatment of J774.2 cells with 0.3/~M taxol for 60 min stabilizes cellular microtubules against depolymerization by calcium extraction as seen by gel electrophoresis and tubulin immunofluorescence. If cells are first treated for 60min with 0.3/~M taxol and then subsequently treated for an additional 60 min with both 0.3 ~M taxol and 90/~M colchicine, cellular microtubules depolymerize as seen by tubulin immunofluorescence (J. J. Manfredi and S. B. Horwitz, unpublished data). Treatment with taxol can make cellular microtubules less sensitive to the depolymerizing effects of drugs such as colchicine, steganacin or nocodazole (Crossin and Carney, 1981a; Schiff and Horwitz, 1980; DeBrabander et al., 1981 a). If, however, a high enough concentration of colchicine is used, even taxol-saturated cellular microtubules can be depolymerized. The effects of calcium and colchicine on cellular microtubules are different and are summarized in Table 4. Although both calcium and colchicine can depolymerize microtubules in intact cells, calcium does not require ATP or an unpolymerized tubulin pool to exert its effects whereas colchicine does. One potential explanation is that calcium binds directly to the microtubule and facilitates depolymerization; its mechanism of action appears to be the same both in vitro and in cells. Colchicine must first bind a free tubulin dimer and be incorporated into the microtubule. Its depolymerizing effect is then a function of an intrinsic biochemical property of the microtubule itself such as, for example, tubulin dimer flux through the microtubule. Calcium is a direct antagonist of microtubule assembly; colchicine exerts its effects in an indirect and considerably more complex manner.

TABLE 3. Effect o f Sodium Azide Treatment on Colchicine Inhibition o f [ 3H]Taxol Binding to J774.2 Cells [3H]Taxol Bound (pmol/mg protein) [3H]Taxol Alone +30/~M Taxol Control +Glucose +Azide + G l u c o s e and azide

41.5 _ 4.1 45.2 _ 0.8 38.6 + 1.5 41.9___2.1

18.0 + 1.5 20.3 + 0.9 14.0 -4- 0.4 18.9+2.1

+30btM Colchicine 11.8 _ 0.8 13.9 ___1.3 31.3 _ 2.2 16.0+0.7

[3H]Colchicine A T P Level Uptake (pmol/plate) (pmol/mg protein) 61.5 _ 62.0 + 1.5 _ 61.0+

1.5 1.7 0.3 1.7

5.7 + 0.3 8.5 ___0.4 30.7 _ 1.3 9 . 0 + 1.0

Confluent 35 m m dishes of J774.2 cells were washed once with 2 ml of 0.1 M PIPES pH 6.9, 1 mM EGTA, and 4 ~ polyethylene glycol 6000 and incubated with 2 ml of buffer alone (control), or buffer plus 30 mM glucose ( + glucose), or 20 mM sodium azide ( + azide), or both ( + glucose and azide) for 60 min at 37°C. Cells were further incubated with 2 m l o f buffer, the appropriate additions, and 0.3 # u [3H]taxol alone or containing 0.3 #M [3H]taxol plus 30 #M unlabeled taxol or 30 #r,I colchicine for 60 min at 37°C. They were then washed five times with 2 ml of ice-cold PBS and lysed in 1 ml of 0.1 N N a O H . Radioactivity was determined by liquid scintillation counting and protein concentration was determined by the method o f Lowry et al. (1951). A T P levels were determined for confluent 35 m m dishes of J774.2 cells according to Loike et al. (1979). A standard curve in the range of 10 -7 to 10-9M A T P was used; dilutions were made from a 10-SM stock solution in 0.04 M Tris-borate, p H 9.2, and calibrated by absorption measurements (e259= 1.54 x 104/M/Cm). All values are the average of triplicate determinations _ one standard deviation. [3H]Colchicine uptake was determined using a concentration of 1 #M. These values are the average of five determinations _+ one standard deviation. J.P.T. 25/I--H

114

J.J. MANFRED!and S. B. HORWITZ TABLE4. Compar&on of Calcium and Colchicine-lnduced Depolymerization of Cellular Microtubules Calcium Colchicine Depolymerizesmicrotubules in intact cells +* + Depolymerizesmicrotubules in cytoskeletons + Requires ATP + Depolymerizes taxol-treated microtubules + *Intact cells are impermeable to calcium ions, however, the use of the calcium ionophore, A23187, has been used to demonstrate the calcium sensitivity of cellular microtubules (Schliwa, 1976; Fuller and Brinkley, 1976). No effect of A23187 and calcium was demonstrable in the J774.2 cell line. Treatment with 2#M A23187 and 5 mM calcium chloride for 120min had no effect on the one-dimensional tubulin gel patterns of J774.2 cytoskeletons.

If taxol-microtubules are depolymerized by the cell in order to form microtubule bundles, the mechanism of depolymerization must not involve direct interaction of calcium ions since these microtubules are stable to such treatment. Job et al. (1981) have reported that cold-stable brain microtubules are resistant to depolymerization by millimolar concentrations of calcium ions but are sensitive to treatment with micromolar concentrations of calcium in the presence of calmodulin. Thus, calmodulin may mediate the depolymerization process. However, if cells are pretreated with the calmodulin antagonist, trifluoperazine (Tfp, 30 ~M) for 60 min and then further incubated with 30 pM Tfp and 30/~M taxol for another 60 min, Tfp has no effect on microtubule bundle formation in J774.2 cells. This particular concentration of Tfp has been shown to affect both cell growth and Fc-mediated phagocytosis in the closely related cell line, J774.16 (Horwitz et al., 1981). Another potential mechanism for this disassembly involving microtubule treadmilling has been suggested by DeBrabander (1982). Consider that microtubules associated with the microtubule organizing center (MTOC) are constantly in flux, and dimers are continuously added and lost because a lower critical concentration of tubulin is necessary for assembly at the M T O C than in the rest of the cell. Treatment with taxol lowers the critical concentration in the peripheral cytoplasm allowing the tubulin dimer pool to assemble there rather than at the MTOC. Eventually all MTOC-associated tubulin will be shuttled into bundles. This model requires that taxol-treated microtubules undergo flux. In vitro data (Schiff and Horwitz, 1981b; Kumar, 1981; Thompson et al., 1981a; Caplow and Zeeberg, 1982) show the tubulin flux through taxol-treated microtubules is substantially reduced, but the relevance of these studies to the situation in intact cells is not clear. Since colchicine is capable of depolymerizing taxol-treated microtubules in cells, the more attractive possibility is that the depolymerization prior to bundle formation is mediated by some endogenous colchicine-like molecule. There are suggestions that such endogenous molecules do, in fact, exist (Sherline et al., 1979; Lockwood, 1979). The energy dependence of taxol-induced bundle formation (Manfredi et al., 1982) may reflect the energy dependence of cellular microtubule depolymerization (Moskalewski et al., 1980; Bershadsky and Gelfand, 1981; DeBrabander et al., 1981b). 5.2.4. All o f the Effects o f Taxol Appear to be Related to the Tubulin-Microtubule System The effects of taxol have been studied in a number of different systems, and in all cases examined the effects of taxol appear to be related specifically to the tubulin-microtubule system. The migration of 3T3 cells through gold particles (Schiff and Horwitz, 1980) as well as chemotaxis in rabbit leukocytes (Nath et al., 1981) is inhibited by taxol. In the trypanosome, T. cruzi, taxol inhibits cytokinesis, yet duplication of cellular organelles continues (Baum et al., 1981). Taxol enhances the in vitro detyrosinolation of ~-tubulin (Kumar and Flavin, 1981). Intact microtubules are necessary for this detyrosinolation reaction (Thompson et al., 1979); taxol is presumably acting by shifting the equilibrium in favor of the polymer.

Taxol

115

Taxol inhibits the growth factor-induced deciliation of primary cilia in 3T3 cells (Tucker, 1980). Separation of daughter centrioles precedes DNA synthesis in HeLa cells; it is not clear whether these two events are related. Epidermal growth factor (EGF) stimulates both DNA synthesis and centriole separation. Taxol inhibits the centriole separation, while colchicine and nocodazole stimulate it; this suggests that microtubule disassembly is necessary for centriole movement (Sherline and Mascardo, 1982). Treatment of cells with nocodazole causes the formation of bundles of intermediate filaments around the nucleus. In cells pretreated with nocodazole and then treated with taxol, the preexisting intermediate filament coils disappear and the intermediate filaments appear to associate with the taxol-induced bundles of microtubules (Geuens et al., 1983). Green and Goldman (1983) find that in taxol-treated cells, microtubules form bundles within which intermediate filaments interdigitate; this confirms previous observations that there is a close association between microtubules and intermediate filaments. Treatment with taxol also induces post-mitotic myoblasts and cells undergoing myofibrillogenesis to assemble interdigitating microtubule-myosin arrays that exclude actin filaments (Antin et al., 1981; Toyama et al., 1982). Taxol has no effect on the uptake of nucleotides, glucose or amino acids in primary mouse embryo cells (Crossin and Carney, 1981b), nor on the synthesis of DNA, RNA, or protein in exponentially growing HeLa cells (Schiff et al., 1979) or J774.16 macrophagelike cells (Horwitz et al., 1981). The drug has no effect on [,25I]thrombin and [~25I]EGF binding to specific surface receptors and their subsequent internalization in primary mouse embryo cells (Crossin and Carney, 198 l b). No alteration of Fc-mediated phagocytosis in J774.16 cells occurs (Horwitz et al., 1981). Taxol does, however, inhibit the secretion of plasma proteins in cultured rat hepatocytes, suggesting a role for microtubules in this process (Oda and Ikehora, 1982). Taxol has been useful in stabilizing cellular microtubules prior to extraction and electron microscopic examination (Schliwa and van Blerkom, 1981). Self-assembly of tubulin from cells of a higher plant, Paul's Scarlet rose cells, was achieved through the use of taxol (Morejohn and Fosket, 1982). By assembling tubulin in the absence of GTP with taxol, Gaskin (1981) showed that Zn2÷-induced structures are not due to a Zn-GTP complex; there is presumably a binding site for Zn 2÷ on tubulin. Zieve and Solomon (1982) treated mitotic baby hamster kidney cells (BHK) with taxol, extracted with detergent to remove soluble proteins, treated with DNAase to remove chromatin, and extracted further at low ionic strength to remove the cage of intermediate filaments which surrounds the spindle. The resulting preparations were analyzed by gel electrophoresis to identify mitosis-specific MAPs. Vallee (1982) established a procedure for isolating microtubule-associated proteins in vitro by stabilizing microtubules with taxol and selectively extracting MAPs from the assembled microtubules by treatment with high concentrations of salt. Cleveland et al. (1981) were able to expand their studies on the regulation of tubulin messenger RNA levels by observing that treatment with taxol leaves unchanged or slightly increases the rate of tubulin protein synthesis by affecting the level of translatable messenger RNA for tubulin. A fluorescent conjugate of desacetylcolchicine has been synthesized which appears to bind to assembled microtubules in vitro (Zimmermann et al., 1982) and in PtK~ cells, particularly after treatment with taxol (Moll et al., 1982). This suggests that colchicine may bind to assembled microtubules. It is difficult to relate these results to studies done with colchicine and podophyllotoxin since it is not clear that the conjugated congener acts by an identical or even similar mechanism. Taxol-resistant Chinese hamster ovary (CHO) cells have altered ct-tubulin (Cabral et al., 1981), while taxol requiring CHO cells (Cabral, 1983) have a normal cytoplasmic microtubule complex but have an impaired ability to form a mitotic spindle in the absence of taxol. Electron microscopic analysis shows a reduced number of kinetochore to pole microtubules and an apparent absence of pole to pole microtubules when the mutant cells are cultured in the absence of taxol (Cabral et al., 1983). Warr et al. (1982) have reported on a mutant of Chinese hamster ovary cells which is resistant to the cytotoxic effects of

116

J.J. MANFREDIand S. B. HORWITZ

taxol but sensitive to colchicine, and a second mutant which is resistant to colcemid, but hypersensitive to the effects of taxol. 5.3. PROPOSED MECHANISM OF ACTION OF TAXOL The results presented here suggest the following pathway for the cellular action of taxol: Taxol crosses the plasma membrane, probably by passive diffusion. The drug binds directly to preexisting cytoplasmic microtubules conferring stability to calcium-induced depolymerization. These calcium-stable microtubules could be depolymerized in an energy dependent manner by an endogenous colchicine-like activity. Parallel arrays of microtubules (bundles) form in an organized manner at discrete sites; this bundle formation is energy dependent. With this scheme as a basis, a model for the mechanism of action of taxol in vitro and in cells will now be discussed. 5.3.1. A Model f o r the Action o f Taxol In Vitro In vitro, taxol enhances the nucleation phase of microtubule assembly and microtubules assembled with taxol are more stable to low temperature, podophyllotoxin or calcium ions than untreated microtubules (Schiff et al., 1979; Schiff and Horwitz, 1981a; 1981b). In addition, taxol substantially inhibits treadmilling by reducing the flux rate of tubulin dimers through steady state microtubules (Schiff and Horwitz, 1981b; Kumar, 1981; Thompson, 1981a; Caplow and Zeeberg, 1982). Sandoval et al. (1977) have studied the effect of the (~,fl)-nonhydrolyzable analog of GTP, guanosine 5'-(~,fl-methylene) triphosphate, pp(CHz)pG on microtubule assembly and stabilization. The GTP analog shares a number of its effects with taxol, pp(CH2)pG enhances microtubule nucleation in vitro and such microtubules are more resistant to depolymerization by cold, colchicine, podophyllotoxin, or calcium ions (Sandoval et al., 1977). pp(CHz)pG also inhibits treadmilling of steady state microtubules (Sandoval and Weber, 1980). These similarities suggest that taxol and pp(CH 2)pG may act by a similar mechanism. One possibility is that taxol may be acting as a GTP analog. Taxol, however, has no effect on the binding of GTP to tubulin nor does it inhibit the rate of hydrolysis of GTP during the microtubule assembly reaction (Schiff and Horwitz, 1981b). Because microtubule assembly occurs in the presence of nonhydrolyzable analogs (Penningroth and Kirschner, 1977), GTP hydrolysis is not necessary for the assembly reaction to occur. Since these analogs assemble microtubules which are more stable to depolymerizing conditions than microtubules that are assembled with GTP (Penningroth and Kirschner, 1977; Sandoval et al., 1977; 1978), the role of GTP hydrolysis may be to facilitate disassembly. One can therefore postulate the following model. There are two conformational states of tubulin: a state capable of assembly, state 1, and a state capable of disassembly, state 2. During normal microtubule assembly, GTP binds to the E site of tubulin and this GTP-tubulin complex assembles into microtubules. This GTP-tubulin is in a state 1 conformation. Subsequent to assembly, the GTP is hydrolyzed to GDP, converting the tubulin-nucleotide complex to a state 2 conformation. Tubulin in state 2 is capable of flux through the microtubules whereas tubulin in state 1 is not. The role of pp(CH2)pG and taxol is to stabilize tubulin in the conformation equivalent to state 1. pp(CH2)pG does this presumably by binding at the E site for guanine nucleotides. Taxol does this by binding at a distinct site; this binding then maintains the tubulin in the state 1 conformation in spite of hydrolysis of GTP at the E site. It has yet to be shown, however, that pp(CH2)pG does in fact bind to the E site; one explanation for the similarities of its effects with taxol may be that they both bind to the same site on tubulin. The E site is present on the tubulin dimer. Preliminary evidence suggests that the taxol binding site does not exist on the unpolymerized tubulin dimer (J. Parness and S. B. Horwitz, unpublished data) but it may be formed when two dimers come together. Dimers are constantly becoming associated and dissociated in solution. When two dimers transiently associate in this manner, the binding site for taxol may be created, and the dimer-dimer interaction stabilized by the drug (Horwitz et al., 1982). Because of the

Taxol

117

reversible exchange of GTP and GDP at the E site, tubulin in solution may be constantly changing conformation between state 1 and state 2. In the presence of pp(CH2)pG, tubulin may remain in state 1 longer and when two dimers randomly collide in the presence of pp(CH2)pG, their respective conformations are such that a nucleation site is formed. Thus, a role of taxol may be to stabilize tubulin in the state 1 conformation. State 1 tubulin has a reduced flux rate and an enhanced nucleation rate; this accounts for the assembly-promoting and microtubule-stabilizing activity of taxol, pp(CHE)pG binds to the E site of tubulin, and like taxol, maintains tubulin in the state 1 conformation; as with taxol, nucleation is enhanced and treadmilling is inhibited. Wehland and Sandoval (1983) recently reported that microinjection of pp(CH2)pG into PtKz cells causes the formation of bundles of microtubules that are independent of the microtubule organizing center. There are similarities between the effects of taxol and pp(CHE)pG in cells. 5.3.2. A Model for the Action of Taxol in Cells The model presented here is for the effects of taxol on the cytoplasmic microtubule complex of undifferentiated mammalian cells, such as J774.2 cells. The data presented in this review demonstrate that taxol binds directly to preexisting cytoplasmic microtubules, thereby conferring stability to calcium-induced depolymerization. Formation of parallel arrays of microtubules that are independent of the microtubule organizing center occurs subsequent to this binding. Although the binding of 0.3 #M taxol saturates within 60 rain, bundle formation does not occur until much later (Fig. 13, A-D). Since depolymerization of cellular microtubules facilitates taxol-induced bundle formation (Fig. 13, E-F), the first step in bundle formation after saturable binding of taxol to cellular microtubules may be depolyrnerization of those microtubules. DeBrabander (1982) has suggested that the microtubules in cells are constantly in flux, and the depolymerization of taxol-treated microtubules occurs as part of this normal process. In vitro studies suggest that this may be unlikely since taxol inhibits the treadmilling of in vitro microtubules (Schiff and Horwitz, 1981b; Kumar, 1981; Thompson et al., 1981a; Caplow and Zeeberg, 1982). Since taxol-treated cellular microtubules are sensitive to colchicine depolymerization, albeit at a higher concentration than untreated microtubules, another possibility is that an endogenous molecule which is similar to colchicine in its mechanism of action is responsible for depolymerization. The phenomenon of taxol-induced microtubule bundle formation requires energy (Fig. 14, A-D). If cellular microtubules are first depolymerized and taxol-induced microtubule bundle formation is then attempted, energy is still required (Fig. 14, E-H). Therefore two energy dependent steps are postulated. The first is the depolymerization of taxol-treated microtubules and the second is the actual formation of parallel arrays of microtubules. The ability of actin filaments to form bundles in vitro and in cells appears to depend on the presence of actin-associated proteins. In the intestinal microvillus core, these 'bundling' proteins are fimbrin and villin (Matsudaira and Burgess, 1979). It is not unreasonable to expect that similar proteins exist for mediating the lateral associations of microtubules. Taxol-induced microtubule bundles are discrete, organized structures. The requirement for an intact cell to initiate bundle formation may reflect the need for a soluble 'bundling' factor. The second energy requiring step in taxol-induced bundle formation may be due to the energy dependence of such a 'bundling' factor. The ability of plant alkaloids such as colchicine and taxol to interact with tubulin suggests that there are endogenous molecules in animal cells which have similar activities. Any complete model for cellular microtubule organization, therefore, must postulate their existence and assign them a role. One such role is to regulate the critical concentration for microtubule assembly at various sites in the cell. By balancing taxol-like and colchicine-like activities, the cell would have a mechanism by which it could alter the critical concentration at various intracellular locations in a cell cycle specific manner. There have been attempts to identify colchicine-like and taxol-like endogenous molecules (Lockwood, 1979; Sherline et al., 1979; Parness et al., 1983).

# 118

J . J . MArqFREDI and S. B. HORWlTZ

The Kirschner (1980) minus-end-capped model for cellular microtubule organization can be modified to provide a reasonable explanation for taxol-induced bundle formation. Kirschner proposed that the microtubule organizing center (MTOC) functions by capping the minus end of cytoplasmic microtubules and that these are more stable than microtubules with their plus end capped or with both ends free. Although the data demonstrating that taxol-treated microtubules do not treadmill tend to weaken the arguments of the DeBrabander (1982) model, certain aspects of this model remain relevant. The primary assembly site, the MTOC, has a certain intrinsic critical concentration, call this, C~, and secondary sites in the cytoplasm have another intrinsic critical concentration, call this, C 2. In control cells, conditions are such that microtubules assemble preferentially at the primary MTOC. Since C~ >/C 2, only microtubules growing at the primary MTOC will be stable. This need not be an absolute situation: most microtubules may be MTOC-associated, with a few associated with certain secondary sites. These MTOCassociated microtubules are stabilized because the MTOC caps their minus ends (Kirschner, 1980). In the presence of taxol, the critical concentration for cellular microtubule assembly is lowered (DeBrabander et al., 1981a). If endogenous colchicinelike molecules are present, the overall critical co,acentration for microtubule assembly in the cell does not have to approach zero, as in vitro (Schiff et al., 1979), but can be some finite value. Thus, in the presence of taxol, the critical concentration for assembly at the secondary sites is lowered such that C~ = C~; the primary MTOC is no longer the only site for microtubule nucleation. The tubulin dimers derived from the depolymerization of the preexisting microtubule cytoskeleton can thus reassemble at various secondary sites present in the cell periphery. Since there are many of these sites, but only a single primary MTOC, these secondary sites dominate microtubule assembly, and little nucleation occurs at the primary MTOC. By Kirschner's arguments, these secondary sites must act by capping the minus ends of the newly assembled microtubules. Thus, a colchicine-like activity may depolymerize taxol-treated microtubules associated with the primary MTOC in an energy dependent manner. Reassembly occurs at secondary sites in the cell periphery. These newly assembled microtubules are organized into parallel arrays by endogenous 'bundling' factors in an energy dependent manner. The signal for depolymerization by the colchicine-like activity may be the binding of taxol to cytoplasmic microtubules. One can envisage that during the cell cycle, the signal for depolymerization of the interphase microtubule complex at the starting of mitosis may be the binding of taxol-like molecules thereby triggering a reaction similar to the one seen when exogenous taxol is added to cells. Thus, understanding the mechanism of action of taxol may give clues to understanding how the cell orchestrates the major reorganization of its microtubules during mitosis. What is proposed here is an extremely hypothetical model, but it should serve to suggest further experimentation to understand the mechanism of action of taxol in cells. Taxol can be used as a powerful pharmacological tool for studying the regulation of microtubule assembly and organization in cells. 5.3.3. The Potential of Taxol in Cancer Therapy It is impossible at the present time to anticipate the role of taxol, either as a single drug or in drug combinations, in the treatment of malignancies, since clinical trials in humans are just beginning. Such trials will encourage the development of methods for detecting taxol or its metabolites in biological fluids and for following the pharmacokinetics of the drug. Hamel et al. (1982) working with rabbits demonstrated that taxol is rapidly cleared from the serum, even though it circulates almost totally as a protein-bound complex. The concentration of taxol in serum was measured utilizing the drug's unique ability to form tubulin polymers in 1.0 M glutamate at 37°C in the absence of exogenous GTP. In addition to antitumor activity, taxol may be helpful as a tumor cell synchronizing agent. The usefulness of taxol as a cell synchronizing agent depends on the drug's ability to block cells in metaphase and the cell's ability to reverse this block. The tumor cells

Taxol

119

w o u l d need to proceed t h r o u g h the cell cycle m a i n t a i n i n g their s y n c h r o n y a n d be susceptible to drugs that i n h i b i t D N A synthesis as they enter the S-phase o f the cell cycle. Some i n f o r m a t i o n m u s t be available c o n c e r n i n g the cell kinetics a n d h o m o g e n e i t y of the specific t u m o r being treated. O u r knowledge o f taxol indicates that the d r u g will block cells in m e t a p h a s e (Schiff a n d Horwitz, 1980) b u t the reversibility o f this effect seems to vary with the p a r t i c u l a r cell type. Vincristine has been studied as a t u m o r cell s y n c h r o n i z i n g a g e n t ( C a m p l e j o h n , 1980) a n d there is little evidence to indicate that cells can be s y n c h r o n i z e d in a clinical situation. U n f o r t u n a t e l y , there is n o reason to expect that t u m o r cells w o u l d be s y n c h r o n i z e d m o r e efficiently t h a n n o r m a l proliferating cells by either taxol or vincristine. The potential o f taxol in cancer t h e r a p y m u s t await further e x p e r i m e n t a t i o n a n d clinical trials. Acknowledgements--The authors would like to thank Jane Fant and Frank Macaluso for their expertise in

electron microscopy, Jerome Parness, Peter Schiff and George Orr for many helpful discussions, and Mary Rutigliano for typing the manuscript. This investigation was supported in part by National Institutes of Health (NIH) grants GM 29042, CA 15714, P30-CA-13330, training grant 2T32-GM-07260 and American Cancer Society grant CH-86. J.J.M. is a recipient of a fellowship from the Pharmaceutical Manufacturers' Association, and the data presented in this review are from a thesis submitted in partial fulfillment for the degree of Doctor of Philosophy from the Sue Golding Graduate Division of the Albert Einstein College of Medicine, Bronx, New York. The authors also gratefully acknowledge the critical reviews of this manuscript by Drs. B. R. Brinkley (Baylor College of Medicine) and J. B. Olmsted (University of Rochester).

REFERENCES

ALBERT1NI,O. F. and CLARK,J. I. (1981) Visualization of assembled and disassembled microtubule protein by double label fluorescence microscopy. Cell Biol. int. Rep. 5: 387-397.

ALBERTINI,D. F., BERLIN,R. D. and OLIVER,J. M. (1977) The mechanism of concanavalin A cap formation in leukocytes, d. Cell Sci. 26: 57-75. ANTXN,P. B., FORRY-SC~uDmS,S., FRIEDMAN,T. M., TAPSCOTT,S. J. and HOLrZER,H. (1981) Taxol induces postmitotic myoblasts to assemble interdigitating microtubule-myosin arrays that exclude actin filaments. J. Cell Biol. 90: 300-308. BAh'ERJEE,A. C. and BriATTACHARYYA,B. (1979) Colcemid and colchicine binding to tubulin. FEBS Lett. 99: 333-336. BALER,A. S., CYPHER,C., MOLE-BAlER,J. and HOWARD,H. M. (1982) Taxol-induced anaphase reversal: evidence that elongating microtubules can exert a pushing force in living cells. Proc. natn. Acad. Sci. U.S.A. 79: 6569-6573. BAUM,S. G., WITTNER,M., NADLER,J. P., HORWXTZ,S. B., DENNlS,J. E., SCruFF,P. B. and TANOWlTZ,H. B. (1981) Taxol, a microtubule stabilizing agent, blocks the replication of Trypanosoma cruzi. Proc. natn. Acad. Sci. U.S.A. 78: 4571-4575. BEN-ZE'EV,A., FARMER,S. R. and PENMAN,S. (1979) Mechanisms of regulating tubulin synthesis in cultured mammalian cells. Cell 17: 319-325. BERGEN,L. G., KURIYAMA,R. and BORISY,G. G. (1980) Polarity of microtubules nucleated by centrosomes and chromosomes of Chinese hamster ovary cells in vitro. J. Cell Biol. 84: 151-159. BERKOWIrZ,S. A. and WOLFF,J. (1981) Intrinsic calcium sensitivity of tubulin polymerization. J. Biol. Chem. 256: 11216-11223. BERNS,M. W. and RICrL~ROSON,S. M. (1977) Continuationof mitosis after selectivelaser microbeam destruction of the centriolar region. J. Cell Biol. 75: 977-982. BERSr~DSKY, A. D. and GELFAt~D, V. I. (1981) ATP-dependent regulation of cytoplasmic microtubule disassembly. Proc. natn. Acad. Sci. U.S.A. 78: 3610-3613. BERSHADSKY,A. D., GELFAND,V. I., SWTKINA,T. M. and TINT, I. S. (1978) Microtubules in mouse embryo fibroblasts extracted with Triton X-100. Cell Biol. int. Rep. 2: 425-432. BESTOR,T. H. and SCHATTEN,G. (1982) Configurations of microtubules in artificially activated eggs of the sea urchin Lytechinus variegatus. Exp. Cell Res. 141: 71-78. BrIArTACr~RYYA,B. and WOLFF,J. (1976) Tubulin aggregation and disaggregation: mediation by two distinct vinblastine-binding sites. Proc. natn. Acad. Sci. U.S.A. 73: 2375-2378. BP,ENN~, S. L. and BRINKLEY,B. R. (1982) Tubulin assembly sites and the organization of microtubule arrays in mammalian cells. Cold Spring Harb. Syrup. quant. BioL 46: 241-254. BRINKLEY,B. R., Cox, S. M., PEPPER,D. A., WIBLE,L., BRENNER,S. L. and PARDUE,R. L. (1981) Tubulin assembly sites and the organization of cytoplasmic microtubules in cultured mammalian cells. J. Cell Biol. 90: 554-562. BRYAN, J. (1972) Definition of three classes of binding sites in isolated microtubule crystals. Biochemistry 11: 2611-2616. CABRAL,F. R. (1983)Isolation of Chinese hamster ovary cell mutants requiting the continuous presence of taxol for cell division. J. Cell Biol. 97: 22-29. CABRAL,F., SOBEL,M. E. and GOTTESMAN,M. M. (1980) CHO mutants resistant to colchicine, colcemid or griseofulvin have an altered fl-tubulin. Cell 20: 29-36. CABRAL,F., ABRAHAM,I. and GOTTESMAN,M. M. (1981) Isolation of a taxol-resistant Chinese hamster ovary cell mutant that has an alteration in g-tubulin. Proc. natn. Acad. Sci. U.S.A. 78: 4388-4391.

120

J.J. MANFREDI and S. B. HORW1TZ

CABRAL,F., ABRAHAM,I. and GOTTESMAN,M. M. (1982) Revertants of a Chinese hamster ovary cell mutant with an altered fl-tubulin: Evidence that the altered tubulin confers both colcemid resistance and temperature sensitivity on the cell. Molec. cell. Biol. 2: 720-729. CABRAL, F., WIBLE, L., BRENNER, S. and BR1NKLEY, B. R. (1983) Taxol-requiring mutant of Chinese hamster ovary cells with impaired mitotic spindle assembly. J. Cell Biol. 97: 30-39. CAMPLF.JOHN,R. S. (1980) A critical review of the use of vincristine (VCR) as a tumor cell synchronizing agent in cancer therapy. Cell Tissue Kinet. 13:327 335. CANOE, W. Z., MCDONALD, K. and MEEUSEN, R. L. (1981) A permeabilized cell model for studying cell division: a comparison of anaphase chromosome movement and cleavage furrow constriction in lysed PtK 1 cells. J. Cell Biol. 88: 618-629. CAPLOW, M. and ZEEaERG, B. (1981) Incorporation of radioactive tubulin into microtubules at steady state. Experimental and theoretical analysis of the effect of podophyllotoxin. J. biol. Chem. 256:5608-5611. CAPLOW, M. and ZEEBERG, B. (1982) Dynamic properties of microtubules at steady state in the presence of taxol. Fur. J. Biochem. 127: 319-324. CAPLOW, M., LANGEORD, G. M. and ZEEBERG, B. (1982) Concerning the efficiency of the treadmilling phenomenon with microtubules. J. biol. Chem. 257: 15012-15021. CARLIER, M. and PANTALONI, D. (1981) Kinetic analysis of guanosine 5'-triphosphate hydrolysis associated with tubulin polymerization. Biochemistry 20: 1918-1924. CARLSEN, S. A., TILL, J. E. and LING, V. (1976) Modulation of membrane drug permeability in Chinese hamster ovary cells. Biochim. biophys. Acta 455: 900-912. CLEVELAND, D. W. (1982) Treadmilling of tubulin and actin. Cell 28:689 691. CLEVELAND, D. W. and K1RSCHNER, M. W. (1982) Autoregulatory control of the expression of ct- and/3-tubulins: implications for microtubule assembly. Cold Spring Harb. Syrup. quant. Biol. 46: 171-183. CLEVELAND, D. W., LOPATA, M. A., SHERLINE,P. and KIRSCHNER,M. W. (1981) Unpolymerized tubulin modulates the level of tubulin mRNAs. Cell. 25: 537-546. CONNOLLY,J. A., KALNINS,V. I. and LING, V. (1981) Microtubules in colcemid-resistant mutants of CHO cells. Exp. Cell Res. 132: 147-155. CORTESE,F., BHATTACHARYYA,B. and WOLFF,J. (1977) Podophyllotoxin as a probe for the colchicine binding site of tubulin. J. biol. Chem. 252: 1134-1140. COTE, R. H. and BORISY, G. G. (1981) Head-to-tail polymerization of microtubules in vitro. J. Molec. Biol. 150: 577-602. Cox, S. M., RAO, P. N. and BRINKLEY, B. R. (1979) Action of nitrous oxide and griseofulvin on microtubules and chromosome movement in dividing cells. In: Effects o f Drugs on the Cell Nucleus, pp. 521-549, BUSCH, H., CROOKE, S. T. and DASKAI, Y. (eds) Bristol-Myers Cancer Symposia, Vol. l, Academic Press, New York. CREASEY, W. A. (1968) Modifications in biochemical pathways produced by the Vinca alkaloids. Cancer Chemother. Rep. 52: 501-507. CREASEY, W. A., BENSCEI, K. G. and MALAWISTA, S. E. (1971) Colchicine, vinblastine and grieseofulvin. Pharmacological studies with human leuokocytes. Biochem. Pharmac. 20: 1579-1588. CROSSIN, K. L. and CARNEY,D. H. (1981a) Evidence that microtubule depolymerization early in the cell cycle is sufficient to initiate DNA synthesis. Cell 23:61 71. CROSS1N,K. L. and CARNEY,D. H. (1981 b) Microtubule stabilization by taxol inhibits initiation of DNA synthesis by thrombin and by epidermal growth factor. Cell 27: 341-350. DEBEC, A., SZOLLOSI,A. and SZOLLOSI,D. (1982) A Drosophila melanogaster cell line lacking centriole. Biol. Cell 44: 133-138. DEBRABANDER, M. (1982) A model for the microtubule organizing activity of the centrosomes and kinetochores in mammalian cells. Cell Biol. int. Rep. 6: 901-915. DEBRABANDER, M., GEUENS, G., NUYDENS, R., WILLEBRORDS, R. and DEMEY, J. (1980) The mierotubule nucleating and organizing activity of kinetochores and centrosomes in living PtK 2 cells. In: Microtubules and Microtubule Inhibitors, pp. 255-268, DEBRABANDER M. and DEMEY J. (eds) Elsevier, Amsterdam. DEBRABANDER, M., GEUENS, G., NUYDENS, R., WILLEBRORDS, R. and DEMEY, J. (1981a) Taxol induces the assembly of free microtubules in living cells and blocks the organizing capacity of the centrosomes and kinetochores. Proc. natn. Acad. Sci. U.S.A. 78: 5608-5612. DEBRABANDER, M., GEUENS, G., NUYDENS, R., WILLEBRORDS, R. and DEMEY, J. (1981b) Microtubule assembly in living cells after release from nocodazole block: the effects of metabolic inhibitors, taxol, and pH. Cell Biol. int. Rep. 9: 913-920. DEBRABANDER, M., BULINSKI, J. C., GEUENS, G., DEMEY, J. and BORISY, G. G. (1981c) Immunoelectron microscopic localization of the 210,000-mol wt microtubule-associated protein in cultured cells of primates. J. Cell Biol. 91: 438~445. DEERY, W. J. and WEISENBERG, R. C. (1981) Kinetic and steady-state analysis of microtubules in the presence of colchicine. Biochemistry 20:2316-2324. DUERR, A., PALLAS, O. and SOLOMON,F. (1981) Molecular analysis of cytoplasmic microtubules in situ: identification of both widespread and specific proteins. Cell 24:203-211. DUSTIN, P. (1978) Microtubules, Springer Verlag, Berlin. FARRELL, K. W. and JORDAN, M. A. (1982) A kinetic analysis of assembly~zlisassembly at opposite microtubule ends. J. biol. Chem. 257: 3131-3138. FUCHS, D. A. and JOHNSON, R. K. (1978) Cytologic evidence that taxol, an antineoplastic agent from Taxus brevifolia, acts as a mitotic spindle poison. Cancer Treat. Rep. 62: 1219-1220. FUJIWARA, K. and TILNEY, L. G. (1975) Substructural analysis of the microtubule and its polymorphic forms. Ann. N.Y. Acad. Sci. 253: 27-50. FULLER, G. M. and BRINKLEY, B. R. (1976) Structure and control of assembly of cytoplasmic microtubules in normal and transformed cells. J. Supramolec. Struct. 5: 497-514. FULLER, G. M., BRINKLEY, B. R. and BOUGHTER,J. M. (1975) lmmunofluorescence of mitotic spindles by using monospecific antibody against bovine brain tubulin. Science 187: 948-950.

Taxol

121

FULTON,C. and SIMPSON,P. A. (1979) Tubulin pools, synthesis, and utilization. In: Microtubules, pp. 117-174, ROBERTS, K. and HYAMS, J. S. (eds) Academic Press, London. GASKIN, F. (1981) In vitro microtubule assembly regulation by divalent cations and nucleotides. Biochemistry 20: 1318-1322. GEUENS, G., DEBRABANDER,M., NUYDENS, R. and DEMEY, J. (1983) The interaction between microtubules and intermediate filaments in cultured cells treated with taxol and nocodazole. Cell. Biol. int. Rep. 7: 35-47. GOULD, R. R. and BORISV, G. G. (1977) The pericentriolar material in Chinese hamster ovary cells nucleates microtubule formation. J. Cell Biol. 73: 601-615. GREEN, K. J. and GOLDMAN,R. D. (1983) The effects of taxol on cytoskeletal components in cultured fibroblasts and epithelial cells. Cell Motility 3: 283-305. GRISHAM, L. M., WILSON, L. and BENSCH, K. G. (1973) Antimitotic action of griseofulvin does not involve disruption of microtubules. Nature 244: 294-296. GUPTA,R. S. (1983) Genetic, biochemical, and cross-resistance studies with mutants of Chinese hamster ovary cells resistant to the anticancer drugs, VM-26 and VP-16-213. Cancer Res. 43: 1568-1574. GUPTA, R. S., HO, T. K. W., MOFEAT, M. R. K. and GUPTA, R. (1982) Podophyllotoxin-resistant mutants of Chinese hamster ovary ceils. Alteration in a microtubule-associated protein. J. biol. Chem. 257: 1071-1078. HAGA,T., ABE,T. and KUROKAWA, M. (1974) Polymerization and depolymerization of microtubules in vitro as studied by flow birefringence. F E B S Lett. 39: 291-295. HAMEL,E., DELCAMPO,A. A., LOWE,M. C. and LIN, C. M. (1981) Interactions of taxol, microtubule-associated proteins, and guanine nucleotides in tubulin polymerization. J. biol. Chem. 256: 11887-11894. HAMEL,E., LIN, C. M. and JOHNS, D. G. (1982) Tubulin-dependent biochemical assay for the antineoplastic agent taxol and application to measurement of the drug in serum. Cancer Treat. Rep. 66: 1381-1386. HARTUNG, E. F. (1954) History of the use of colchicum and related medicaments in gout. Ann. Rheum. Dis. 13: 190-200. HASKINS, K. M., DONOSO, J. A. and HIMES, R. H. (1981) Spirals and paracrystals induced by Vinca alkaloids: evidence that microtubule-associated proteins act as polycations. J. Cell Sci. 47: 237-247. HAUSMANN, K., LINNENBACH, M. and PATTERSON, D. J. (1983) The effects of taxol on microtubular arrays: in vivo effects on heliozoan axonemes. J. Ultrastruct. Res. 82: 212-220. HEIDEMANN,S. R. and GALLAS,P. T. (1980) The effect of taxol on living eggs of Xenopus laevis. Devl Biol. 80: 489-494. HILL, T. L. and KIRSCHNER, M. W. (1982) Subunit treadmilling of microtubules or actin in the presence of cellular barriers: Possible conversion of chemical free energy into mechanical work. Proc. natn. Acad. Sci. U.S.A. 79: 490-494. HILLER, G. and WEBER,K. (1978) Radioimmunoassay for tubulin: a quantitative comparison of the tubulin content of different established tissue culture cells and tissues. Cell 14: 795-804. HIMES, R. H., KERSEY,R. N., HELLER-BETTINGER,I. and SAMSON, F. E. (1976) Action of the Vinca alkaloids vincristine, vinblastine, and desacetylvinblastine amide on microtubules in vitro. Cancer Res. 36: 3798-3802. HOEBEKE, J., VANNIJEN, G. and DEBRABANDER,M. (1976) Interaction of oncodazole (R17934), a new antitumoral drug, with rat brain tubulin. Biochem. biophys. Res. Commun. 69: 319-324. HORWITZ, S. B., CH1A, G. H., HARRACKSINGH,C., ORLOW, S., PIFKO-HIRST,S., SCHNECK,J., SORBARA,L., SPEAKER, M., WILK, E. W. and ROSEN, O. M. (1981) Trifluoperazine inhibits phagocytosis in a macrophagelike cell line. J. Cell Biol. 91: 798-802. HORWITZ, S. B., PARNESS,J., SCHIFF,P. B. arid MANFREDI,J. J. (1982) Taxol: A new probe for studying the structure and function of microtubules. Cold Spring Hath. Syrup. Quant. Biol. 46: 219-226. INESI, G., MILLMAN,M. and ELETR, S. (1973) Temperature-induced transitions of function and structure in sarcoplasmic reticulum membranes. J. molec. Biol. 81: 483-504. JAMESON,L. and CAPLOW, M. (1981) Modification of microtubule steady-state dynamics by phosphorylation of the microtubule-associated proteins. Proc. natn. Acad. Sci. U.S.A. 78: 3413-3417. JOB, D., FISCHER, E. H. and MARGOLIS, R. L. (1981) Rapid disassembly of cold-stable microtubules by calmodulin. Proc. natn. Acad Sci. U.S.A. 78: 4679-4682. KARR, T. L., KRISTOFFERSON,D. and PURICH, D. L. (1980) Calcium ion induces endwise depolymerization of bovine brain microtubules. J. biol. Chem. 255: 11853-11856. KEATES, R. A. B., SARANGI,F. and LING,V. (1981) Structural and functional alterations in microtubule protein from Chinese hamster ovary cell mutants. Proc. natn. Acad. Sci. U.S.A. 78: 5638-5642. KELLY, M. G. and HARTWELL,J. L. (1954) The biological effects and the chemical composition of podophyllin. A review. J. natn. Cancer Inst. 14: 967-1010. KIRSCHNER, M. W. (1980) Implications of treatmilling for the stability and polarity of actin and tubulin polymers in vivo. J. Cell Biol. 86: 330-334. KUMAR,N. (1981) Taxol-induced polymerization of purified tubulin. J. biol. Chem. 256: 10435-10441. KUMAR, N. and FLAVIN, M. (1981) Preferential action of a brain detyrosinolating carboxypeptidase on polymerized tubulin. J. biol. Chem. 256: 7678-7686. KUR1YAMA,R. and SAKAI, H. (1974) Viscometric demonstration of tubulin polymerization. J. Biochem. 75: 463-471. LAEMMLI, U. K. (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227: 680-685. LEE, Y. C. and WOLFF,J. (1982) Two opposing effects of calmodulin on microtubule assembly depend on the presence of microtubule associated proteins. J. biol. Chem. 257: 6306--6310. LENK, R., RANSOM, L., KAUFMANN,Y. and PENMAN, S. (1977) A cytoskeletal structure with associated polyribosomes obtained from HeLa cells. Cell 10: 67-78. LOCKWOOD,A. H. (1979) Molecules in mammalian brain that interact with the colchicine site on tubulin. Proc. natn. Acad. Sci. U.S.A. 76: 1184-1188. LOIKE,J. D. and HORWITZ,S. B. (1976) Effects of podophyllotoxin and VP-16-213 on microtubule assembly in vitro and nucleoside transport in HeLa cells. Biochemistry 15: 5435-5443.

122

J . J . MANFREDI and S. B. HORWITZ

LOIKE, J. D., KOZLER, V. F. and SILVERSTEIN, S. C. (1979) Increased ATP and creatine phosphate turnover in phagocytosing mouse peritoneal macrophages. J. biol. Chem. 245: 9558-9564. LOWRY, O. H. ROSEBROUGH,N. J., FARR, A. L. and RANDALL, R. J. (1951) Protein measurement with the Folin phenol reagent. J. biol. Chem. 193: 265-275. LUDUENA, R. F. (1979) Biochemistry of tubulin. In: Microtubules, pp. 65-116, ROBERTS,J. and HYAMS, J. S. (eds) Academic Press, London. LUDUENA, R. e., FELLOUS, A., FRANCON, J., NUNEZ, J. and MCMANUS, L. (1981) Effect of tau on the vinblastine-induccd aggregation of tubulin. J. Cell Biol. 89: 680-683. MANFREDI, J. J. and HORWITZ, S. B. (1984) Vinblastine paracrystals from cultured cells are calcium stable. Exp. Cell Res. 150: 205-217. MANFRED1, J. J., PARNESS, J. and HORWITZ, S. B. (1982) Taxol binds to cellular microtubules. J. Cell Biol. 94: 688-696. MARGOLIS, R. L. (1981) Role of GTP hydrolysis in microtubule treadmilling and assembly. Proc. hath. ,4cad. Sci. U.S.A. 78: 1586-1590. MARGOLlS, R. L. and WILSON, L. (1977) Addition of colchicine-tubulin complex to microtubule ends: the mechanism of substoichiometric colchicine poisoning. Proc. natn. Acad. Sci. U.S.A. 74: 3466-3470. MARGOLIS, R. L. and WILSON, L. (1978) Opposite end assembly and disassembly of microtubules at steady state in vitro. Cell 13: 1-8. MARGOUS, R. L. and WILSON, L. (1979) Regulation of the microtubule steady state in vitro by ATP. Cell 18: 673-679. MARGOLIS, R. L. and W1LSON, L. (1981) Microtubule treadmills--possible molecular machinery. Nature 239: 705-711. MARGOLIS, R. L., WILSON, L. and KIEFER, B. I. (1978) Mitotic mechanism based on intrinsic microtubule behavior. Nature 272: 450-452. MASUROVSKYE. B., PETERSON, E. R., CRAIN, S. M. and HORWITZ, S. B. (1981) Microtubule arrays in taxol-treated mouse dorsal root ganglion-spinal cord cultures. Brain Res. 217: 392-398. MASUROVSKY,E. B., PETERSON, E. R., CRAIN, S. M. and HORWlTZ, S. B. (1983) Morphologic alterations in dorsal root ganglion neurons and supporting cells of organotypic mouse spinal cord-ganglion cultures exposed to taxol. Neuroscience 10: 491-509. MATSUDAIRA, P. T. and BURGESS,D. R. (1979) Identification and organization of the components in the isolated microvillus cytoskeleton. J. Cell Biol. 83: 667-673. MELMED, R. N., KARAN1AN, P. J. and BERLIN, R. D. (1981) Control of cell volume in the J774 macrophage by microtubule disassembly and cyclic AMP. J. Cell Biol. 90: 761-768. MILLER, R. W., POWELL, R. G. and SMITH, C. R., JR. (1981) Antileukemic alkaloids from Taxus wallichiana Zucc. J. Org. Chem. 46: 1469-1474. MIZEL, S. B. and WILSON, L. (1972) Nucleoside transport in mammalian cells. Inhibition by colchicine. Biochemistry 11: 2573-2578. MOLE-BAJER, J. and BAJER, A. S. (1983) Action of taxol on mitosis: modification of microtubule arrangements and function of the mitotic spindle in Haemanthus endosperm. J. Cell Biol. 96: 527-540. MOLL, E., MANZ, B., MOCIKAT, S. and ZIMMERMAN,H. (1982) Fluorescent deacetylcolchicine. New aspects of its activity and localization in PtK 1 cells. Exp. Cell Res. 141:211-220. MOREJOHN, L. C. and FOSKET, D. E. (1982) Higher plant tubulin identified by self-assembly into microtubule in vitro. Nature. 297: 426-428. MOSKALEWSKI, S., THYBERG, J. and FRIBERG, U. (1980) Cold and metabolic inhibitor effects on cytoplasmic microtubules and the Golgi complex in cultured rat epiphyseal chondrocytes. Cell Tissue Res. 210: 403-415. NA, G. C. and TIMASHEFF, S. N. (1980) Thermodynamic linkage between tubulin self-association and the binding of vinblastine. Biochemistry 19: 1355-1365. NA, G. C. and TIMASHEFF, S. N. (1982) In vitro vinblastine-induced tubulin paracrystals. J. biol. Chem. 257: 10387-10391. MATH, J., FLAVIN, M. and SCHIFFMANN, E. (1981) Stimulation of tubulin tyrosinolation in rabbit leukocytes evoked by the chemoattractant formyl-methionyl-leucyl-phenylalanine. J. Cell Biol. 91: 232-239. NISHIDA, E. (1978) Effects of solution variables on the calcium sensitivity of the microtubule assembly system. J. Biochem. 84: 507-512. NISHIDA, E. and SAKAI,H. (1977) Calcium-sensitivity of the microtubule reassembly system, d. Biochem. 82: 303-306. ODA, K. and IKEHARA, Y. (1982) Taxol, a potent promoter of microtubule assembly, inhibits secretion of plasma proteins in cultured rat hepatocytes. Biochem. biophys. Res. Commun. 107: 561-567. O'FARRELL, P. Z , GOODMAN, H. M. and O'FARRELL, P. H. (1977) High resolution two-dimensional electrophoresis of basic as well as acidic proteins. Cell 12: 1133-1142. OLMSTED, J. B. and BORISY, G. G. (1973) Characterization of microtubule assembly in porcine brain extracts by viscometry. Biochemistry 12: 4282-4289. OSBORN, M. and WEBER, K. (1976) Cytoplasmic microtubules in tissue culture cells appear to grow from an organizing structure towards the plasma membrane. Proc. natn. Acad. Sci. U.S.A. 73: 867-871. OSBORN, M. and WEBER, K. (1977) The display of microtubules in transformed cells. Cell 12: 561-571. PAATERO, G. I. L. and BROWN, n . L. (1982) Effects of taxol on microtubule organization and on capping of surface immunoglobulin in mouse splenic lymphocytes. Cell Biol. int. Rep. 6: 1033-1040. PALAN1VELU,P. and LUNDUENA, R. F. (1982) Interactions of the fl-tubulin-vinblastine complex with colchicine, podophyllotoxin, and N,N'-ethylenebis(iodoacetamide). J. biol. Chem. 257:6311-6315. PALLAS,n . and SOLOMON,F. (1982) Cytoplasmic microtuhule-associated proteins: phosphorylation at novel sites is correlated with their incorporation into assembled microtubules. Cell. 30: 407-414. PARNESS, J. and HORWITZ, S. B. (1981) Taxol binds to polymerized tubulin in vitro. J. Cell Biol. 91: 479-487. PARNESS, J., KINGSTON, n . G. I., POWELL, R. G., HARRACKSINGH, C. and HORWITZ, S. B. (1982)

Taxol

123

Structure-activity study of cytotoxicity and microtubule assembly in vitro by taxol and related taxanes. Biochem. biophys. Res. Commun. 105: 1082-1089.

PARNFAS,J., ASNES,C. F. and HORWITZ, S. B. (1983) Taxol binds differentially to flagellar outer doublets and their reassembled microtubules. Cell. Motility 3: 123-130. PENNINGROTH, S. i . and KIRSCHNER,i . W. (1977) Nucleotide binding and phosphorylation in microtubule assembly in vitro. J. molec. Biol. 115: 643~573. PETERSON, S. P. and BERNS, M. W. (1980) The centriolar complex. Int. Rev. CytoL 64: 81-106. PICK, E., HONIG, S. and GRIFFEL,B. (1979) The mechanism of action of soluble lymphocyte mediators. Int. Archs Allergy appL Immun. 58: 149-159. PICKETT-HEAPS,J. D. (1969) The evolution of the mitotic apparatus: an attempt at comparative ultrastructural cytology in dividing plant cells. Cytobios. 1: 257-280. RANG, D., HOBBLE, R. and KIRSCI-INER,M. W. (1982) Mitosis in a cell with multiple centrioles. J. Cell Biol. 94: 549-556. ROBBINS, E., JENTZSCH,G. and MICALI, A. (1968) The centriole cycle in synchronized HeLa cells. J. Cell Biol. 36: 329-339. ROBERTS, R. L., NATH, J., FRIEDMAN,M. M. and GALL1N,J. I. (1982) Effects of taxol on human neutrophils. J. Immun. 129: 2134--2141. SANDOVAL, I. V. and WEBER, K. (1980) Guanosine 5'-(~,/~-methylene) triphosphate enhances specifically microtubule nucleation and stops the treadmill of tubulin protomers. J. biol. Chem. 255: 6966-6974. SANDOVAL, I. V., MACDONALD, E., JAMESON, J. L. and CUATRECASAS,P. (1977) Role of nucleotides in tubulin polymerization: effect of guanylyl 5'-methylene diphosphonate. Proc. natn. Acad. Sci. U.S.A. 74: 4881-4885. SANDOVAL,I. V., JAMESON,J. L., NmDEL, J., MAcDoNALD, E. and COATRECASAS,P. (1978) Role of nucleotides in tubulin polymerization: effect of guanosine Y-methylene diphosphonate. Proc. natn. Acad. Sci. U.S.A. 75: 3178-3182. SC.AXTEN, G., SCHATTEN,H., BESXOR,T. H. and BALCZON, R. (1982) Taxol inhibits the nuclear movements during fertilization and induces asters in unfertilized sea urchin eggs. J. Cell Biol. 94: 455-465. SCH~LE, R. B. and BORISY, G. G. (1979) In vitro assembly of microtubules. In: Microtubules, pp. 175-254, ROBERTS, K. and HYAMS,J. S. (eds) Academic Press, London. ScrufF, P. B., FANT, J., AUSTrJt, L. A. and HORWlTZ, S. B. (1978) Effects of taxol on cell growth and in vitro microtubule assembly. J. Supramolec. Struct. 8: (Suppl. 2) 328. Scmrr, P. B., FANT, J. and HORWlTZ, S. B. (1979) Promotion of microtubule assembly in vitro by taxol. Nature 277: 665~i67. ScrufF, P. B. and HORWITZ,S. B. (1980) Taxol stabilizes microtubules in mouse fibroblast cells. Proc. natn. Acad. Sci. U.S.A. 77: 1561-1565. Scm~, P. B. and HORWlTZ, S. B. (1981a) Tubulin: a target for chemotherapeutic agents. In: Molecular Actions and Targets for Cancer Chemotherapeutic Agents, pp. 483-507, SARTORELLI,A. C., LAZLO,J. S. and BERT1NO, J. R. (eds) Bristol-Myers Cancer Symposium, Vol. 2, Academic Press, New York. ScrufF, P. B. and HORWlTZ, S. B. (1981b) Taxol assembles tubulin in the absence of exogenous guanosine 5'-triphosphate or microtubule-associated proteins. Biochemistry 20: 3247-3252. Scm~, P. B., KENDE, A. S. and HORWITZ, S. B. (1978) Steganacin: an inhibitor of HeLa cell growth and microtubule assembly in vitro. Biochem. biophys. Res. Commun. 85: 737-746. SCHLIWA, M. (1976) The role of divalent cations in the regulation of microtubule assembly. 3. Cell Biol. 70: 527-540. SCHUWA, M. and VAN BLERKOM,J. (1981) Structural interaction of cytoskeletal components. J. Cell Biol. 90: 222-235. SCHHWA, M., EOTENEOER, U., BOUNSRI, J. C. and IZANT, J. G. (1981a) Calcium lability of cytoplasmic microtubules and its modulation by microtubule-associated proteins. Proc. natn. Acad. Sci. U.S.A. 78: 1037-1041. SCHLIWA, M., VAN BLERKOM,J. and PORTER, K. R. (1981b) Stabilization of cytoplasmic ground substance in detergent-opened cells and a structural and biochemical analysis of its composition. Proc. natn. Acad. Sci. U.S.A. 78: 4329-4333. SCHUWA, M., PRYZWANSKY,R. B. and EUTENEOER,U. (1982) Centrosome splitting in neutrophils: an unusual phenomenon related to cell activation and motility. Cell. 31: 705-717. SECRET, C. J., HADrmLD, J. R. and BEER, C. T. (1972) Studies on the binding of [3H]vinblastine by rat blood platelets in vitro. Biochem. Pharmac. 21: 1609-1624. SEE, Y. P., CARLSEN,S. A., TmL, J. E. and LING, V. (1974) Increased drug permeability in Chinese hamster ovary cells in the presence of cyanide. Biochim. biophys. Acta 373: 242-252. SERPINSKAYA,A. S., GELEAND,V. I. and KOPNIN, B. P. (1981) Comparison of mitostatic effect, cell uptake and tubulin-binding activity of colchicine and colcemid. Biochim. biophys. Acta 673: 86-92. SHARP, G. A., OSBORN,M. and WEBER,K. (1981) Ultrastructure of multiple microtubule initiation sites in mouse neuroblastoma cells. 3. Cell Sci. 47: 1-24. SrERLlr~, P. and MASCAROO,R. (1982) Epidermal growth factor-induced centrosomal separation: mechanism and relationship to mitogenesis. J. Cell Biol. 95: 316-322. SI-ERLINE,P. and MUNDY,G. R. (1977) Role of the tubulin-microtubule system in lymphocyte activation. J. Cell Biol. 74: 371-376. SrmRLINE, P., SCmAVONE,K. and BROCATO,S. (1979) Endogenous inhibitor of colchicine-tubulin binding in rat brain. Science 205: 593-595. SLOBODA, R. D., RUDOU'H, S. A., ROSENBAUM, J. L. and GREENGARD, P. (1975) Cyclic AMP-dependent endogenous phosphorylation of a microtubule-associated protein. Proc. natn. Acad. Sci. U.S.A. 72:177-18 I. SOLOMON, F. (1977) Binding sites for calcium on tubulin. Biochemistry 16: 358-363. SOLOMON, F. (1980) Neuroblastoma cells recapitulate their detailed neurite morphologies after reversible microtubule disassembly. Cell 21: 333-338.

124

J . J . MANFRED1 and S. B. HORWITZ

SOLOMON,F., MAGENDANTZ,M. and SALZMAN,A. (1979) Identification with cellular microtubules of one of the co-assembling microtubule-associated proteins. Cell 18: 431-438. SPIEGELMAN, B. i . , PENNINGROTH, S. M. and KRISCHNER, i . W. (1977) Turnover of tubulin and the N site GTP in Chinese hamster ovary cells. Cell 12: 587-600. SPIEGELMAN,B. M., LOPATA,M. A. and KIRSCHNER, M. W. (!979a) Multiple sites for the initiation of microtubule assembly in mammalian cells. Cell 16: 239-252. SPIEGELMAN, B. M., LOPATA, M. A. and KIRSCHNER, i . W. (1979b) Aggregation of microtubule initiation sites preceding neurite outgrowth in mouse neuroblastoma cells. Cell 16: 253-263. STERNLICHT,H. and R1NGEL, I. (1979) Colchicine inhibition of microtubule assembly via copolymer formation. J. biol. Chem. 254: 10540-10550. SUMMERS, K. and K1RSCHNER, M. W. (1979) Characteristics of the polar assembly and disassembly of microtubules observed in vitro by darkfield light microscopy. J. Cell Biol. 83: 205-217. TAYLOR, E. W. (1965) The mechanism of colchicine inhibition of mitosis. J. Cell Biol. 25: 145-160. THOMPSON, W. C., DEAN1N, G. G. and GORDON, M. W. (1979) Intact microtubules are required for rapid turnover of carboxyl-terminal tyrosine of ct-tubulin in cell cultures. Proc. hath. Acad. Sci. U.S.A. 76: 1318-1322. THOMPSON, W. C., PURICH, n . L. and WILSON, L. (1981a) Taxol slows microtubule "treadmilling" in vitro. J. Cell Biol. 91: 329a. THOMPSON, W. C., WILSON, L. and PURICH, D. L. (1981b) Taxol induces microtubule assembly at low temperature. Cell Motility 1: 445-454. TIMASHEFF, S. N. and GRISHAM, L. M. (1980) In vitro assembly of cytoplasmic microtubules. A. Rev. Biochem. 49: 565-591. TOYAMA, Y., FORRY-SCHAUDIES, S., HOFFMAN, B. and HOLTZER, H. (1982) Effects of taxol and colcemid on myofibrillogenesis. Proc. hath. Acad. Sci. U.S.A. 70: 6556-6560. TUCKER, R. W. (1980) Cytoplasmic microtubules, centriole ciliation and mitogenesis in Balb/c 3T3 cells. In: Microtubules and Microtubule Inhibitors, pp. 497-508, DEBRABANDER, M. and DEMEY, J. (eds) Elsevier, Amsterdam. VALLEE,R. B. 0982) A taxol dependent procedure for the isolation of microtubules and microtubule-associated proteins (MAPs). J. Cell Biol. 92: 435-442. WANI, M. C., TAYLOR, H. L., WALL, M. E., COGGON, P. and MCPHAIL, A. T. (1971) Plant antitumor agents. VI. The isolation and structure of taxol, a novel antileukemic and antitumor agent from Taxus brevifolia. J. Am. Chem. Soc. 93: 2325-2327. WARFIELD, R. K. N. and BOUCK, G. B. (1974) Microtubule-macrotubule transitions: intermediates after exposure to the mitotic inhibitor vinblastine. Science 186: 1219-1221. WARR, J. R., FLANAGAN, n . J. and ANDERSON, M. (1982) Mutants of Chinese hamster ovary cells with altered sensitivity to taxol and benzimidazole carbamates. Cell Biol. int. Rep. 6: 455-460. WEBER, K., POLLACK, R. and BIBRING, T. (1975) Antibody against tubulin: the specific visualization of cytoplasmic microtubules in tissue culture cells. Proc. hath. Acad. Sci. U.S.A. 72: 459-463. WEHLAND, J. and SANDOVAL,I. V. (1983) Cells injected with guanosine 5'-(ct,fl-methylene)triphosphate, an ct,fl-nonhydrolyzable analog of GTP, show anomalous patterns of tubulin polymerization affecting cell translocation, intracellulr movement, and the organization of Golgi elements. Proc. hath. Acad. Sci. U.S.A. 80: 1938-1941. WEINGARTEN, i . n., LOCKWOOD, A. H., Hwo, S. and KIRSCHNER, M. W. (1975) A protein factor essential for microtubule assembly. Proc. hath. Acad. Sci. U.S.A. 72: 1858-1862. WEISENBERG, R. C. (1972) Microtubule formation in vitro in solutions containing low calcium concentrations. Science 177:1104-1105. WEISENBERG, R. C. (1980) Role of co-operative interactions, microtubule-associated proteins and guanosine triphosphate in microtubule assembly: a model. J. molec. Biol. 139: 660-677. WEISENBERG, R. C. and DEERY, W. J. (1981) The mechanism of calcium-induced microtubule disassembly. Biochem. Biophys. Res. Commun. 102: 924-931. WEISENBERG, R. C., DEERY, W. J. and DICKINSON, P. J. (1976) Tubulin-nucleotide interactions during the polymerization and depolymerization of microtubules. Biochemistry 15: 4248-4254. WHITF1ELD, C. F. and SCHWORER, M. E. (1978) Alteration of hexose transport in avian erythrocytes by vinblastine and colchicine. Arch. Biochem. Biophys. 191: 727-733. WILSON, L. and FRIEDKIN, M. (1967) The biochemical events of mitosis. II. The in vivo and in vitro binding of colchicine in grasshopper embryos and its possible relation to inhibition of mitosis. Biochemistry 6: 3126-3135. WILSON, L., BAMBURG,J. R., MIZEL, S. B., GRISHAM, L. M. and CRESWELL, K. M. (1974) Interaction of drugs with microtubule proteins. Fedn Proc. 33: 158-166. WILSON, L., CRESWELL, K. M. and CHIN, D. (1975) The mechanism of action of vinblastine. Binding of acety1-3H-vinblastine to embryonic chick brain tubulin and tubulin from sea urchin sperm tail outer doublet microtubules. Biochemistry 14: 5586-5592. WRIGHT, M. and MOISAND, A. (1982) Spatial relationships between centrioles and the centrosphere in monoasters induced by taxol in Physarum polycephalum amoebae. Protoplasm. 113: 67-79. WRIGHT, M., MOISAND,A. and OUSTRIN,M. L. (1982) Stabilization of monoasters by taxol in the amoebae of Physarum polycephalum (Myxomycetes). Protoplasma 113: 44-56. ZIEVE,G. and SOLOMON,F. (1982) Proteins specifically asociated with the microtubules of the mammalian mitotic spindle. Cell 28: 233-242.

Taxol

~

125

ZIMlVlERMAN,H., DOENGES, K. H. and MOLL, E. (1982) Deacetylcolchicine conjugated to a heme-nonapeptide acts with both soluble tubulin and microtubules polymerized in vitro. Eur. J. Cell Biol. 26: 310-314.