THE JOURNAL OF BIOLOGICAL CHEMISTRY 0 1989 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 264, No. 35, Issue of December 15, pp. 21296-21301, 1989 Printed in U. S. A.

Interaction of Tubulin and Cellular Microtubules with the New Antitumor Drug MDL 27048 A POWERFUL AND REVERSIBLE MICROTUBULE INHIBITOR*

(Received for publication, June 26, 1989)

Vincent Peyrot, Daniel Leynadier, Marcel Sarrazin, and Claudette BriandS From the Faculte de Pharmacie, Laboratoire de Physique Pharmaceutique, 27 Bd Jean Moulin, 13385, Marseille, Cedex 5, France

Ana Rodriquez, Juan Miguel Nieto, and Jose Manuel Andreu From the Centra de ZnvestiEaciones Biologicas. Conseio Superior de Znvestigaciones Cientificas, Velazquez, 144, 28006 - Madrid, Spain

~.

We have characterized the binding of trans-1-(2,5- dimethoxyphenyl)-3-[4-(dimethylamino)pheny1]-2- methyl-2-propen-1-one (MDL 27048) to purified por- cine brain tubulin, and the inhibition of microtubule assembly by this compound in vitro and using cultured cells. Binding measurements were performed by dif- ference absorption and fluorescence spectroscopy. MDL 27048 binds to one siteltubulin heterodimer with an apparent equilibrium constant Kt, = (2.8 2 0.8) X lo6 M" (50 mM 2-(N-morpholino)ethanesulfonic acid, 1 mM [ethylenebis(oxyethylenenitrilo)]tetraacetic acid, 0.5 mM MgClZ, 0.1 mM GTP buffer, pH 6.7, at 25 "C). Podophyllotoxin displaced the binding of MDL 27048, suggesting an overlap with the colchicine-bind- ing site. Assembly of purified tubulin into microtubules was inhibited by substoichiometric concentrations of MDL 27048, which also induced a slow depolymeri- zation of preassembled microtubules.

The cytoplasmic microtubules of PtK2 cells were disrupted in a concentration and time-dependent man- ner by MDL 27048, as observed by indirect immuno- fluorescence microscopy. Maximal depolymerization took place with 2 X lo-' M MDL 27048 in 3 h. When the inhibitor was washed off from the cells, fast micro- tubule assembly (-8 min) and complete reorganization of the cytoplasmic microtubule network (15-30 min) were observed. MDL 27048 also induced mitotic arrest in SV40-3T3 cell cultures. Due to all these properties, this anti-tumor drug constitutes a new and potent mi- crotubule inhibitor, characterized by its specificity and reversibility.

High cell proliferation, which appears in malignant tumors, can be treated by administration of drugs which interfere with cell division. Some of the most useful cancer therapeutic agents are metaphase inhibitors, e.g. Vinblastine or Vindesine. In order to improve therapy, efforts are being made to develop new compounds with such an effect.

It has been reported that the synthetic compound; trans-l-

* This work was supported in part by joint French-Spanish Grant 144, the Spanish DGICyT Grant PB022, and a grant from the French board of National Education. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

$ To whom correspondence should be addressed.

(2,5-dimethoxyphenyl)-3-[4-(dimethylamino)phenyl]-2- methyl-2-propen-1-one (MDL 27048) (compound 1 in Chart I) possess an antitumor effect and the ability to inhibit in vitro the assembly of microtubule proteins (1).

A part of the structure of MDL 27048, mainly the dime- thoxyphenyl ring (see compound 1, Chart I) is closely related to colchicine and to other tubulin ligands such as podophyl- lotoxin, and the bicyclic compounds 2-methoxy-5-(2,3,4-tri- methoxyphenyl)-2,4,6-cycloheptatrien-l-one ( M T C ) (com- pound 2, Chart I; (2, 3) and combretastatin (compound 3, Chart I; (4)). The rest of the MDL 27048 molecule bears no obvious resemblance to colchicine site ligands.

We asked whether tubulin would be a target for this drug and if this compound could be an inhibitor of microtubules in living cells. This report shows that MDL 27048 binds to purified tubulin, inhibits microtubules assembly in vitro as in cells, and arrests mitosis.

MATERIALS AND METHODS

Protein and Chemicals-Microtubule proteins and tubulin were purified from pig brain, stored at -80 "C, prepared for use and their concentrations measured as described (5). Taxol (NSC 125973) was a gift from Dr. Suffness (National Cancer Institute); GTP (disodium salt) was from Fluka; podophyllotoxin was from Aldrich Chemical Co., and MDL 27048 was a gift from Merrel Dow Laboratory. The concentration of MDL 27048 was measured spectrophotometrically. The extinction coefficient, t , was determined by dissolving dry crys- tals of the compound in Me2S0 (dimethyl sulfoxide) and diluting in 50 mM MES,' 1 mM EGTA, 0.5 mM MgCL (MEM) buffer, pH 6.7, then taking the UV visible spectrum. Three independent determina- tions gave t398 = 21,000 _C 1,100 M" cm".

MDL 27048 was proved pure by 'H NMR spectroscopy, which confirmed that the compound is the trans-isomer. NMR spectra were recorded at ambient temperature on a Bruker AM 200 Spectrometer in the Fourier transform mode at 200.133 MHz (Service Interuniv- ersitaire de R6sonance Magnktique Nuclkaire, Facultk de Pharmacie, Marseille).

Concentrated MDL 27048 solutions were made in M e 8 0 and kept at -20 "C.

Binding measurements were performed in MEMG buffer pH 6.7, i.e. MEM buffer containing 0.1 mM GTP.

Microtubule Assembly-Tubulin was equilibrated in assembly buffer MEM-glycerol 3.4 M, GTP 1 mM, MgCI, 10 mM, pH 6.7 (6). The reaction was started by a temperature shift from 4 to 37 "C in a thermostated 1-cm light path cell and was monitored turbidimetri-

' The abbreviations used are: MES, 2-(N-morpholino)ethane sul- fonic acid; EGTA, [ethylenebis(oxyethylenenitrilo)]tetraacetic acid; PIPES, piperazine-N,N'-bis(2-ethanesulfonic acid); PBS, phos- phate-buffered saline; BSA, bovine serum albumin; HEPES, N-2- hydroxyethylpiperazine-N'-2 ethanesulfonic acid.

21296

Inhibition of Microtubule Assembly by M D L 27048 21297

cH'o,;,u" COMERETASTATIN ( 3 ) CHaO

OCH, c; oH OCHl

, C h

CH,O "\"'

MDL-27048 ( I )

CHART I

cally (7). MDL 27048 containing samples and their controls had 1% residual Me2S0, except where otherwise indicated.

Microtubule protein assembly was carried out as described for tubulin except that the buffer was MEM buffer containing 1 mM GTP.

For electron microscopy, assembled structures were adsorbed to carbon-coated grids, negatively stained with 1.5% uranyl acetate and observed with a Philips EM 400 T microscope (Service Commun de Microscopie Electronique, Faculte des Sciences, Marseille, France).

Spectroscopic Measurements-Light absorption spectra were ob- tained with a Beckman DU 70 spectrophotometer. Absorption spectra were done employing 0.438 + 0.438 cm tandem cells (Hellma) ther- mostated at 25 & 0.5 "C with a Colora water bath. The base line was defined by placing protein and ligand in different compartments. Interaction spectra were generated by mixing protein and ligand in the same compartment and subtracting the base-line value from the interaction spectra value.

Relative fluorescence measurements were made with a Kontron SFM 25 spectrofluorimeter employing 1 X 0.2-cm cells a t 25 * 0.5 "C. The excitation and emission bandwidths were 400 and 480 nm, respectively, with slit widths of 5/5 nm.

Fluorescence spectra were obtained with a Fica MKII double-beam spectrofluorometer that gives corrected excitation and emission spec- tra.

Binding Measurements-First, equilibrium difference spectropho- tometric measurements of binding were obtained at constant total MDL 27048 concentration and increasing tubulin concentrations. The change in absorbance at 404 nm was plotted uersus [protein]" and extrapolated to infinite protein concentration; this yielded the increment of extinction coefficient, Atqo4, of MDL 27048 upon binding to tubulin. Second, protein-ligand interaction spectra were obtained at different total MDL 27048 concentrations. The concentration of bound MDL 27048 was measured as .L4A104/At404 and the concentration of free ligand calculated as the difference from the total concentration.

Binding of MDL 27048 to tubulin was also measured fluorimetri- cally. MDL 27048, bound to tubulin, was excited at 400 nm and its fluorescence measured at 480 mm. Inner filter corrections were made by the procedure of Mertens and Kagi (8). Titration of 2 p~ MDL 27048 with excess tubulin gave the fluorescence/pmol of MDL 27048 bound. Then, tubulin was titrated with various total MDL 27048 concentrations. The concentration of MDL 27048 bound was meas- ured by the fluorescence of the solution after mixing and the free ligand concentration was considered to be the difference from the known total concentration.

Cell Culture and Immunofluorescence Procedures"PtK2 cells (9) and SV40-3T3 cells (10) were grown in DMEM (Sigma) with 44 mM NaHCOs, 10% (v/v) calf serum (Flow), 2 mM-glutamine and 0.004%- gentamycin sulfate, a t 37 "C in an humidified atmosphere containing 5% co,.

PtK2 cells were plated onto 9 X 9-mm glass coverslips (Bellco) a t a density of 200,000 cells/ml, cultured for 1 day, washed twice with culture medium without serum, and treated with the inhibitor in DMEM with 8.3 mM NaHCOa and 18 mM HEPES, pH 7.2, a t 37 "C. The coverslips were washed with PEM (100 mM PIPES, 1 mM EGTA, and 1 mM MgCI2, pH 6.8) containing 4% polyethylene glycol 8,000 (Sigma), permeabilized for 90 s with PEM containing 0.5% Triton X-100, washed again with PEM-polyethylene glycol, and the cover-

slip-attached cytoskeletons were fixed with 3.7% formaldehyde in PEM-l% dimethyl sulfoxide for 30 min and kept in PBS at 4 "c (11). The coverslips were placed into a humidified container a t 37 "C, overlaid with 2 pg/ml affinity purified anti-tubulin monospecific rabbit antibodies (to sequence positions a415-443; (12) in PBS con- taining 10 mg/ml bovine serum albumin and 0.02% NaN3 and incu- bated for 1 h. After washing twice with PBS and gentle agitation (10 min), the coverslips were overlaid with fluoresceinated goat anti- rabbit antibodies (Behring Diagnostics, diluted 1:20 in PBS/bovine serum albumin/NaNJ, together with phalloidine-rhodamine (Molec- ular Probes, diluted 1:20) incubated for 45 min, washed with PBS in the dark and mounted with anti-fading solution (13). The cytoskele- tons were photographed on Kodak Tri-X film, employing a Zeiss photomicroscope 111, equipped with epiilumination, a 40 X plana- pochromatic objective and the appropriate filters for fluoresceine and rhodamine.

Exponentially growing SV40-3T3 cells were placed in culture me- dium without serum and immediately treated with MDL 27048. Giemsa-stained metaphasic plates were prepared from culture tryp- sinates (14) and the percent of cells with condensed chromosomes (mitotic index) was counted. To measure the cellular DNA content, 1 ml of resuspended cells were fixed by addition of 0.05 ml formal- dehyde, collected by centrifugation, and resuspended in 0.25 ml of 0.02 M citrate-phosphate buffer, pH 3.0, containing 0.2 M sucrose, 0.1 mM EDTA, and 0.1% Triton X-100. These cells were mixed with 0.5 ml of 0.002% acridine orange in 0.01 M citrate-phosphate, 0.1 M NaCI, pH 3.8, and the fluorescence of the DNA-bound acridine orange/cell (15) was measured with a Coulter Epics CS cytofluorometer.

-0.006 I I I I I I , I A

320 346 372 398 424 450

WAVELENGTH (nm)

0 0.2 0.4 0.6 0.8 1.0 1.2

R h o 1 bound I 105g tubulin)

tion of tubulin with MDL 27048 in MEM buffer 0.1 mM GTP, pH FIG. 1. A , difference absorption spectra generated by the interac-

6.7, at 25 "C, tubulin (2.8 pM) and MDL 27048 1 pM ( I ), 3 pM ( Z ) , 5 pM ( 3 ) , 7 p M ( 4 ) , 9 pM ( 5 ) . (- - -) is the direct spectrum of MDL (1 p ~ ) in the same buffer. The inset shows the 260-500 nm spectrum generated by the interaction of tubulin (8 p ~ ) with MDL 27048 (20 p ~ ) . B, difference spectrophotometric titration presented as a Scat- chard plot. The ranges of protein and MDL 27048 concentrations employed in these experimentations were 2-3 p M and 1-9 pM, re- spectively.

21298 Inhibition of Microtubule

I 1 ln C 3

.- I

2 5 0 300 350 4 0 0 4 5 0 500 5 5 0 600

WAVELENGTH (nm)

0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6

R( mol bound I 10'g tubulin

f I I

IC I , I I L

10 20 30 40 TIME ( m i d

FIG. 2. A , fluorescence changes produced by the tubulin-MDL 27048 interaction ( a ) fluorescence emission (excitation at 364 nm) and ( b ) excitation (emission at 495 nm). Spectra of a solution con- taining 5 pM tubulin and 10 p~ MDL 27048 (the ligand alone gave negligible fluorescence in this experiment). B, Scatchard plot of the fluorimetric measurements. The inset shows the fluorimeter titration of 2.3 hM tubulin with various MDL 27048 concentrations. C, time course of fluorescence change of 2 p~ MDL 27048 binding to 3 p~ tubulin at 25 "C; 100 p~ podophyllotoxin was added at the time indicated by the arrow.

RESULTS

Binding of MDL 27048 to Purified Tubulin-The interac- tion of MDL 27048 with tubulin in a MEMG buffer, pH 6.7, was measured by difference absorption spectroscopy and flu- orescence.

Fig. 1A (inset) shows the spectrum generated by this inter- action, which is characterized by maxima a t 296 and 404 nm, and minima at 356 and 448 nm. Control experiments were performed with similar concentrations of MDL 27048 in the

Assembly by MDL 27048

0.3 I

0 5 10 15 20 25 30

TIME b i n )

- 0.41 E, A k" I -------

OV I I I

0 12 24 36 I I 84 96

TIME (mid

FIG. 3. A , turbidity time course of in vitro microtubule assembly: 10 pM tubulin ( I ) , with 0.5 p~ MDL 27048 (2) , with 1 p~ MDL 27048 ( 3 ) , 5 p~ MDL 27048 ( 4 ) . Inset, in a different experiment the plateau absorbance values of tubulin polymerization were measured in the absence (a) or in the presence of 0.75 ~ L M MDL 27048 ( b ) . B, effect of MDL 27048 on assembled microtubules (tubulin 10 p ~ ) . At the time indicated by the arrolu 20 PM MDL 27048 was added (dashed line) then samples were cooled at 4 "C during 30 min and after rewarmed at 37 "C.

absence of protein: no perturbation was detected (Fig. l A , dashed line). The change measured in the absorption coeffi- cient of MDL 27048 upon binding to tubulin was At404 nm = 4129 f 697 M" cm". Tubulin solutions were titrated with various MDL 27048 concentrations (see "Materials and Meth- ods"). A Scatchard (16) plot of the data could be linearly fitted (Fig. 1B) by 0.96 0.20 binding sites with an apparent binding equilibrium constant Kb = (2.1 f 0.4) X lo6 M-'.

MDL 27048 from 0.4 to 20 pM in the absence of tubulin was found to have a very weak fluorescence in the MEMG buffer, pH 6.7. At higher concentrations, from 40 to 200 pM, a nonlinear increase in fluorescence occurred (data not shown). Addition of tubulin 3.2 HM to the former system resulted in an large increase in relative fluorescence as shown in Fig. 2 A . Bound MDL 27048 had an emission maximum at 494 nm. Excitation maxima were at 380 and 284 nm. The latter value suggests a possible energy transfer from aromatic residues of tubulin to the bound ligand. Quenching of the intrinsic protein fluorescence was also observed (not shown). Fig. 2B (inset) shows a fluorimetric titration of 2.3 p~ tubulin with MDL 27048 a t a concentration range of 0.4-16 p ~ . The results were analyzed using a linear Scatchard (16) plot. The equilibrium parameters were respectively estimated as (3.50 k 0.8) X lo6 M-' and 1.30 & 0.10 for the apparent binding constant and number of binding sites (Fig. 2B). In other

Inhibition of Microtubule Assembly by MDL 27048 21299

FIG. 4. Depolymerization of cel- lular microtubules by MDL 27048. PtK2 cells were incubated during 3 h at 37 "C in culture medium without serum and processed for immunofluorescence microscopy. a, microtubules of control cells without inhibitor; b-e, microtubules of cells treated with 0.2, 0.5, 0.8, and 2 pM MDL 27048, respectively; f , actin microfilaments of cells treated with 5 p~ MDL 27048 (these were indistinguisha- ble from actin filaments in the control, which are not shown). The bar corre- sponds to 20 pm.

experiments aimed at following the displacement of MDL 27048 from tubulin by using a trimethoxy phenyl compound, Podophyllotoxin, which binds to tubulin through its trime- thoxy benzene ring, the MDL 27048 was added to tubulin and the relative fluorescence was recorded. Then podophyllotoxin was added, and the time course of fluorescence change was followed. As shown in Fig. 2C the effect of podophyllotoxin indicated that the dissociation of MDL-tubulin complex was essentially complete in less than 20 min.

Microtubule Assembly Studies-Fig. 3A shows the effects of MDL 27048 on the turbidimetric time course of microtubule assembly. A clear inhibition was noticed, and the rate of assembly as well as the final amount of microtubules were lower in the presence of substoichiometric MDL 27048 con- centrations. With a tubulin or microtubule proteins concen-

tration of 1 mg/ml, the half-inhibitory concentration of MDL 27048 was 1 pM. The process was examined when varying the protein concentration a t a fixed MDL 27048 concentration (0.75 p ~ ) and the results, which are depicted by Fig. 3A (inset), indicated an increase on the critical protein concen- tration required for assembly (abscissa intercept of the plot) and a change in the slope of the plot.

In Fig. 3B 20 p~ MDL 27048 was added to a microtubule preparation a t equilibrium, a slow decrease in turbidity was recorded, and a new equilibrium was reached. An aliquot was saved and negatively stained; electron microscopic examina- tion showed microtubule structures (data not shown). The obtained microtubules could be also depolymerized by cooling to 4 "C for 30 min; the residual turbidity observed was due to amorphous aggregates of protein. Rewarming to 37 "C induced

21300 Inhibition of Microtubule Assembly by MDL 27048

FIG. 5 . I , upper panel: time course of microtubule depolymerization by MDL 27048. PtK2 cells were treated with 10 p~ MDL 27048 during different periods and processed for immunofluorescence of microtubule. a, control (0 rnin); b, 10 min; c, 20 min; d, 80 min; e, 180 min. I I , lower panel: time course of microtubule repolymerization following removal of the inhibitor. PtK2 cells were treated with 2 p~ MDL 27048 for 4 h, giving a maximal depolymerization, as in Fig. 4e and Fig. 5Ie. The inhibitor was washed off and the cells were fixed after several time periods a t 37 "C. a, 2 min; b, 4 min; c, 8 min; d, 15 min; e, 30 rnin. The bar represents 20 pm.

A 1

3 h

6 h

9 h

27 h

B 3 h

6 h

A 9 h

27 h

Fluorescence intensity

FIG. 6. Effects of MDL 27048 on cell division. Control (col- umn A ) and 1 p~ inhibitor treated (column R ) SV40-3T3 cell cultures were fixed after the times indicated in the Fig. (3, 6, 9, and 27 h). The relative DNA content/cell was measured by fluorocytometry ("Materials and Methods") of 1000 cells/sample.

assembly of the control, whereas in the presence of MDL 27048 no reassembly occurred. In another experiment (not shown) designed to follow the reversibility of the inhibition induced by MDL 27048, taxol(20 p ~ ) was added to a mixture of tubulin (10 p ~ ) and MDL 27048 (5 p ~ ) . An immediate increase in turbidity was recorded and negative stained sam- ples showed microtubule structures.

Inhibition of Interphase Microtubules and Mitotic Arrest of

Cultured Mammalian Cells by MDL 27048"Since this ligand bound reversibly to purified tubulin and inhibited in vitro microtubule assembly, it was of clear interest to test whether it could inhibit the microtubules of living cells. Fig. 4 shows the effects of various concentrations of MDL 27048 on the cytoplasmic microtubule network of PtK2 cells, visualized by immunofluorescence microscopy. At 0.2 p~ inhibitor, undu- lating microtubules were observed in the cell periphery (Fig. 4b). Between 0.5 and 1 p~ inhibitor several degrees of depo- lymerization of the microtubule network were observed (Fig. 4, c-e). The depolymerization was maximal a t 2-5 PM drug, leaving a few drug-resistant microtubules/cell (Fig. 4e), but this did not affect the actin filaments (Fig. 4f ) . This dose- dependent microtubule depolymerization effect was very sim- ilar to the effects of colchicine, the bicyclic colchicine analogue MTC, and nocodazole in the same assay (not shown). How- ever l and 2 p~ MDL 27048 in the presence of 10% fetal calf serum had a weaker effect (similar to Fig. 4b), which suggests the interaction of this inhibitor with one or several compo- nents, e.g. serum albumin. Next, the time course and reversi- bility of the inhibition of cellular microtubules by MDL 27048 were tested. Fig. 51 shows representative time frames of a depolymerization experiment with 10 p~ inhibitor, which show a full effect within 3 h. A comparative experiment indicated qualitatively similar inhibition time courses with 10 p~ MDL 27048, colchicine, and MTC (not shown). When MDL 27048 was washed off from cells with their microtubules depolymerized, a fast microtubule reassembly was observed. Representative results are shown in Fig. 511, where microtu- bule growth can be observed as early as 2 min after drug removal (Fig. 511, frame a ) and newly polymerized microtu- bule networks, which are undistinguishable from controls, after 15-30 min (frames d and e). This rapid reversibility is similar to that of MTC and does not take place after colchicine inhibition (24).

To functionally characterize the effects of MDL 27048 on mitotic spindle microtubules, mitoses, and DNA content/cell were examined in drug-treated cultures of the transformed line SV40-3T3. No significant effect was observed at 0.05 p~

Inhibition of Microtubule Assembly by MDL 27048 21301

inhibitor and a significant mitotic arrest was observed at a concentration of 0.2 p ~ . At 1 p~ drug the mitotic index increased from 1% (control values) to 2.7, 10.2, and 12.7% in cells treated for 3, 6, and 9 h, respectively. Fig. 6 shows the cytofluorometer profiles of the control ( A ) and 1 pM drug- treated cell cultures ( B ) . There is a marked shift to larger DNA content/cell in the drug-treated population (i.e. the average fluorescence nearly doubles at 6-9 h, indicating an accumulation of cells that have completed their S phase but which have not proceeded through mitosis. The deviation to the left in the profiles of the control cells at long time periods was probably due to the serum deprivation during the exper- iments.

DISCUSSION

The Tubulin-MDL 27048 Interaction-MDL 27048 binding to tubulin as well as the localization of its high affinity binding site have been investigated using different absorption and fluorimetric measurements. We observed no important differ- ence in the results between the two methods. The apparent binding constant to the tubulin dimer was (2.1-3.5) X lo6 "I, which is higher than that of the Vincaalkaloids (about 2 X lo4 M - I ) (18, 191, and of the colchicine analogue, MTC (4.9 X IO5 M - I ) (3), but close to the best estimates of the binding of colchicine (ca lo7 M", (20)). No binding parameters are described in the literature for combretastatin analogues. MDL 27048 binding measurements indicated that there is one bind- ing site/tubulin dimer, as for colchicine and its analogues. Although herein we did not notice any trend of the apparent MDL 27048 affinity with the tubulin concentration, possible tubulin self-association reactions induced by the binding of this ligand remain to be investigated. It is well known that podophyllotoxin competes with colchicine binding to tubulin, probably through the trimethoxylbenzene ring which is com- mon to the two alkaloids (21,22). The binding of MDL 27048 to tubulin was rapid and reversible. Our results (Fig. 2C) indicate that podophyllotoxin displaces the MDL 27048 from its binding site: we can preliminarily assume that the trime- thoxylbenzene ring of podophyllotoxin binds at or near the dimethoxylbenzene ring of MDL 27048. There is no indication at present of where the rest of the MDL 27048 molecule binds.

Structurally, MDL 27048 consists of two substituted ben- zene rings linked by a methylpropenone. The two phenyl rings are oriented in trans (see Chart I) and one is substituted by two methoxy groups. Our results showed that the effects of MDL 27048 on microtubules are similar to those of colchicine analogues or combretastatin analogues. Moreover, it should be noted that the combretastatin analogues present two phenyl rings with methoxy groups linked by a saturated two carbon bridge (4). Lin et al. (4) reported that "trans-isomers were less active than the corresponding cis-isomers, that reduction of double bound also yielded agents with less activ- ity and that the enhancement of the bridge length decreased the activity." Thus, all of these structure-activity relation- ships cannot be applied to MDL 27048, but they must, be considered.

Inhibition of Polymerization-The inhibition of microtu- bule assembly by MDL 27048 is apparently related to an increase in the critical concentration required for polymeri- zation and a decrease in the fraction of assembly-active pro- tein (complete inhibition occurred at 5 p~ MDL 27048). The results clearly indicated that MDL 27048 inhibited substo- ichiometrically the assembly of microtubules, such as colchi- cine, podophyllotoxin, and vinblastine.

I n vitro, at low or high concentration, MDL 27048, like colchicine analogues, does not disrupt preformed microtu-

bules. In the same experimental conditions (glycerol and high Mg concentration) vincaalkaloids were reported to induce a complete disruption of preformed microtubules and formation of a large number of spirals (23, 24).

MDL 27048 Effects on Cellular Microtubules and Cell Diui- sion-MDL 27048 presents a marked depolymerization effect on the cytoplasmic microtubule network of PtK2 cells, which is maximal at 2-5 p~ drug concentration. Moreover, the time course and the high degree of reversibility of the inhibition of cellular microtubules by MDL 27048 are similar to those of MTC (17) and of Nocodazole (25). This reversibility may be linked with the reversibility of binding to purified tubulin and also indicates a good plasma membrane permeability of the drug. The mitotic index and DNA content/cell examined in drug treated cultures of the SV40-3T3 line indicate a strong inhibitory effect on cell growth.

Conclusions-The results show that MDL 27048 can be considered as a specific antitubulin drug which binds rapidly and reversibly to this protein. In spite of other possible interaction with other subcellular components, the antitumor activity of MDL 27048 (1) may well be due to its inhibition of microtubule formation.

Because of its particular chemical structure, MDL 27048 constitutes a new type of antitubulin agent, which may over- lap the colchicine-binding site, and which could prove to be of value as an experimental inhibitor in the study of micro- tubules and microtubule-mediated functions.

Acknowledgments-We wish to thank Prof. Consuelo de la Torre for extended use of her fluorescence microscope, Dr. M. A. Arevalo for the affinity purified antibodies, Javier Medrano for help with fluorescence measurements, Dr. R. Faure for NMR measurements, and Servicio de Citometria de Flujo, Centro de Investigaciones Bio- logicas.

Addendum-Preliminary results show that MDL 27048 ( 5 FM) added in bovine serum albumin solution (1 mg/ml) became fluores- cent this indicates an interaction between bovine serum albumin and MDL 27048.

REFERENCES 1. Sunkara, P. S., Stemerick, D. M., and Edwards, M. L. (1988) Proc. Am.

2. Fitzgerald, T. J. (1976) Biochem. Pharnacol. 25 , 1381-1387 3. Andreu, J. ,M., Gorbunoff, M. J., Lee, J. C., and Timasheff, S. N. (1984)

4. Lin, C. M., Singh, S. B., Chu, P. S., Dempcy, R. O., Schmidt, J. M., Pettit,

5. Chaudoreille, M. M., Peyrot, V., Braguer, D., and Crevat, A. (1987) Mol.

6. Peyrot, V., Briand, C., Codaccioni, F., and Sari, J. C. (1987) Chem. Biol.

7. Gaskin, F., Cantor, C. R., and Shelanski, M. L. (1974) J. Mol. Biol. 8 9 ,

Assoc. Cancer Res. 29, 328

Biochemwtry 23,1742-1752

G. R., and Hamel, E. (1988) Mol. Pharmacol. 34 , 200-208

Pharmacol. 32, 731-736

Interactions 61, 151-158

717-75R

9. 8.

10.

12. 11.

13.

14.

15.

16. 17.

18. 19.

20. 21.

22. 23.

24.

25.

Mertens, L. M., and Kagi, J. H. R. (1979) Anal. Biochem. 96,448-455 Shaw, M. V., and Krooth, R. S. (1964) Cytogenetics 3 , 199-203 Todaro, G. J., and Aaronson, S. A. (1969) Virology 3 8 , 174-202 Osborn, M., and Weber, K. (1982) Methods Cell Biol. 24,98-132 De La Vina, S., Andreu, D., Medrano, F. J., Nieto, J. M., and Andreu, J.

Johnson, G. A., and Nogueira Araujo, G. M. (1981) J. Immunol. Methods

Moorhead, P. S., Nowell, P. C., Mellmann, W. J., Battips, D. M., and

Traganos, F., Darzynkiewicz, Z., Sharpless, T., and Melaured, M. R. (1977)

Scatchard, G. (1949) Ann. N. Y . Acad. Sci. 51,660-672 Diez, J. C., Avila, J., Nieto, J. M., and Andreu, J. M. (1987) Cell Motil.

Na, G. C., and Tirnasheff, S. N. (1986) Biochemistry 25,6214-6222 Singer, W. D., Hersh, R. T., and Himes, R. H. (1988) Biochem. Pharmacol.

Garland, D. L. (1978) Biochemistry 17,4266-4272 Cortese, F., Bhattacharyya, B., and Wolff, J. (1977) J. Biol. Chem. 252,

Andreu, J. M., and Timasheff, 5. N. (1982) Biochemistry 21,6465-6476 Himes, R. H., Kersey, R. N., Heller-Bettinger, I., and Samson, F. (1976)

Jordan, M. A., Margolis, R. L., Himes, R. H., and Wilson, L. (1980) J . Mol.

De Brabander, J. J., Van der Viern, R., Aerts, F. E. M., Borgers, M., and

." . ."

M. (1988) Biochemistry 27,5352-5365

43,349-350

Hurgesford, D. A. (1960) Exp. Cell Res. 20,613-616

J. Histoehem. Cytochem. 25,46-56

Cytoskeleton 7, 178-186

37,2691-2696

1134-1140

Cancer Res. 36 , 3798-3802

Biol. 187,61-73

Janssen, P. A. J. (1976) Cancer Res. 36, 905-916