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

Vol. 264, No. 28, Iseue of October 5, pp, 16367-16371,1989

Printed in U.S.A.

A Thermody:namic Study of the Interactionof Tubulin with Colchicine Site Ligands'@

(Received for publication, April 4, 1989)

Margarit,aMenendezSp, Jose LaynezS, F. Javier Medranol, and Jose Manuel Andreul

From the Slnstituto de Quimica Fisica "Rocasolano," Consejo Superior de Znuestigaciones Centificas, Serrano 119, and the llCentro de l'nuestigaciones Biologicas, Consejo Superior de Znvestigaciones Cientificas, Vekizquez 144, 28006 Madrid, Spain

The first step is weak binding and the backwards rate constant of the second step is very small, which results in the overall high affinity formation of an essentially stable, fluorescent end product (3,4). The difficulties encountered for a rigorous equilibrium characterization of the tubulin-colchicine interaction led to the study of simpler ligands. Single ring analogues of the trimethoxybenzene and of tropolone rings(see Chart I) were shown to bind specifically (although weakly) to the colchicine site. Their contributions to the binding of the complete ligand were analyzed by application of a simple thermodynamic model of the binding of a bifunctional ligand to a bifocal protein binding site (14). The properties of the trimethoxyTubulin is the main cellular target of antitumor and antiphenyl ligand-tubulin interactions suggested hydrophobic mitotic drugs which perturb microtubule assembly. Among binding, whereas the binding of the tropolone ring moiety these tubulin ligands the three best known are colchicine, may involve hydrogen binding and/or ring stacking interacvinblastine and taxol, which bind to different sites on the tions (1, 15). MTC' is a simple bifunctional ligand containing protein. Colchicine bilnds to a single high affinity site of the two essential rings of colchicine and lacking the middle tubulin. Under nonassembly conditions the binding of this ligand is not linked to :protein self-association (1, 2), and the ring (16) (see Chart I). This ligand binds specifically and with conformational changes induced in the system (3, 4) involve high affinity to the colchicine site without the kinetic inconvenience of colchicine, providing additional information about * This work was supported by Spanish Comision Interministerial the role of the central ring closure of colchicine in both the de Ciencia y Tecnologie Grant PB0220. The costs of publication of kinetics and theaffinity of binding. The crystalline structure
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. The abbreviations used are: MTC, 2-methoxy-5-(2,3,4-trimethoxyphenyl)-2,4,6-cycloheptatrien-l-one; Mes, 2-(N-morpholino) ethanesulfonic acid.

The bicyclic collchicine analogue 2-methoxy-5(2',3',4'-trimethox:yphenyl)-2,4,6-cycloheptatrien-lone (MTC) has been usedto study thethermodynamics of specific ligand binding to the colchicine site of tubulin, employing isothermal reaction microcalorimetry. The binding of lHTC to purified calf brain tubulin, in 10 m sodium phosphate buffer, pH 7.0, is characM terized by A H o = -19 f 1 kJ.mol", AGO = -31.8 f 0 6 . k J . mol", and A S o = 43 2 5 J mol" K" at 298 K, with a slight variation in the temperature range from 203 to 308 K. The binding thermodynamics of colchicine and allocolchicine are similar to MTC under the conditions examined, suggesting related molecular interactions of the three ligands with the protein binding site. The standard enthalpy changes of binding of colchicine and MTC at 308 K coincide within experimental error. Therefore the more favorable free energy change of binding of colchicinemustcome from a larger binding entropy change (by about 20 J.mol". K-'). This difference could be attributed to the presence of the middle ring of colchicine, which is absent in MTC. Consistently, a similar entropy change is observed by the comparison of allocolchicine to MTC binding at several temperatures. In addition, allocolchicine binding is atbout 6 k J . mol" less exothermic than MTC binding, which could be attributed to the presence in allocolcl~icine a substituted phenyl ring of instead of the colchicine-MTCtropolone ring. The present results and analysis fully compatible with the are previously proposed bifunctional binding of colchicine and MTC (through theirtrimethoxybenzene and tropolone moieties) to a bifocal protein binding site, and also with a partial immobilization of intramolecular rotation of MTC upon binding, which in colchicine is already constrained by its middle ring (Andreu, J. M., Gorbunoff, M. J., Lee, J. C., and Timasheff, S . (1984) Biochemistry 23, 1742-1752).

both colchicine (5) and tubulin (1). Under assembly conditions, a morphologically abnormal polymerization of the tubulin-colchicine complex takes place (6, 7) with qualitative, and apparent thermodynamic, characteristics very similar to those of microtubule assembly (2). It has been proposed that the abnormal polymerization of the tubulin-colchicine complex and the substoichiometric inhibition of the microtubule growth by this ligand (8, 9) are caused by a distortion of the normal interaction geometry between the tubulin protomers (2). The primary cause of these effects is the binding of the drug to tubulin, which consists of processes not fully understood. The binding of colchicine to soluble tubulin is slow and the unoccupied binding sitedenatures rapidly, hampering equilibrium studies (10). The colchicine-tubulin interaction is known to be strongly temperature-dependent. In fact, van't Hoff enthalpy changes of 50-70 kJ .mol" have been obtained (11-13). The kinetics of binding of colchicine (col) to tubulin (TB) consists of two parallel phases (corresponding to two types of binding sites) of which only the fast phase has been characterized (3,4). This phase conforms to a two-stepmechanism: a fast reversible binding followed by a slow conformational change.
TB

k, + C O ~Kl TB .col + (TB.~ 0 1 ) ' +

k-2

(1)

16367

16368

Study of Tubulin-Colchicine Analoguesby Reaction Microcalorimetry

bulin concentrations were corrected to active protein concentrations (concentration of tubulin molecules that bind the ligand) for each individual preparation and are so expressed in the MTC calorimetric results which are consistent with this correction. Tubulin preparations with n < 0.57 were discarded. Kinetic rate constants for ligand binding site decay were determined from fluorometric measurements with saturating MTC under MTC COL ALLO the same conditions as the calorimetric measurements. The values CHART I. COL, colchicine; ALLO, allocolchicine. obtained were 4.3 X min" (298 K), 8.8 X min" (303 K) and 1.35 X min" (308 K). The later value is comparable with the of MTC is similar to colchicine (17) and the rapid kinetics of decay of colchicine binding determined with a different protein preparation and buffer conditions (3). Theserates wereemployed to binding and dissociation have been fully analyzed both in correct the tubulin concentration for the small fraction which denatheir fast and slow phases (18, 19), which conform to the tured during thermal equilibration of the calorimeter before mixing mechanism described for colchicine. The interaction of MTC with the ligand; this correction was smaller than 17% at 308 K and with tubulin has been found to be moderately dependent on not significant below 298 K. A 2-h incubation of tubulin in the temperature and van't Hoff enthalpies ranging from -7.7 to absence of ligand at 298 or 308 K did not significantly change the -28.5 kJ . mol-' have been reported from equilibrium binding equilibrium constant of binding of MTC to tubulin. The time course of 1.5 X M colchicine binding to 6 X M and kinetic measurements (19-21). This water-soluble tubulin tubulin in PG buffer at different temperatures was determined fluoligand is a specific substoichiometric inhibitor of the mitotic rometrically as described (20). After 20 min, the fractions of the final spindle and interphase microtubules of mammalian tissue- fluorescence reached were 0.97 (308 K), 0.81 (298 K), and 0.52 (283 M tubulin cultured cells. At variance with colchicine, the microtubule K). In a different type of experiment, samples of 5-7 X inhibition by MTC is morphologically and quantitatively were chromatographed on Sephadex G-25 columns equilibrated with 1.5 X 10" M colchicine, the adjusted for a reversible within few minutes a after washing out the analogue protein retention time of in which 0.98 flow had been colchicine was 20 min; & 0.06 mol of (22),' as predicted from the reversibility of the binding. bound per mol of tubulin at 298 and 308 K, 0.64 & 0.06 at 288 K, and In order to get a better understanding of the thermody- 0.52 & 0.08 at 283 K. Preincubation of tubulin for 2 h without ligand namics of ligandbinding to the colchicine site of soluble at 308 K had no significant effect on the colchicine binding time tubulin, we have employed isothermalreaction microcalorim- course, except for the fluorescence amplitude whichwas reduced etry to determine the enthalpy changes of MTC binding to according to thedecay of the site measured above with MTC. Calorimetric Measurements-Calorimetric experiments were pertubulin at several temperatures. These parameters have been formed with a LKB batch microcalorimeter equipped with a 2107compared to those of tubulin interaction with an active ring 350 LKB titration unit, in PG buffer, a t different temperatures. The C isomer of colchicine, allocolchicine, (see Chart I) (23) and calorimeter was calibrated both electrically and chemically (27). The also to the binding enthalpy change of the colchicine itself, ligand injection system was calibrated by weighing the water delivered under the restricted conditions in which attainment of equi- by the syringes. The heat of dilution of the ligand was cancelled by librium could be assured. The bindings of the three ligands, the reference cell. Whenever the calorimetric experiments were carhere studied, are moderately exothermic, with small differ- ried out using the titration unit, the small differential compression and mixing heat effects were determined separately and subtracted ences in their positive binding entropy changes. from the heat quantities evolved in the reaction experiments (28). In those experiments performed without the titration system the heat of interaction was determined by using saturating concentrations MATERIALS ANDMETHODS of the ligand for MTC and colchicine binding and saturating protein Ligands and Other Chemicals-MTCwas provided by Dr. T. J. concentrations for allocolchicine binding measurements. The heat of Fitzgerald (16). Allocolchicine was prepared by alkaline re- dilution of the protein was determined in a separate run and subarrangement of colchicine (23) by Dr. M. J. Gorbunoff at Brandeis tracted from the total signal. The equilibration time of the instrument University. Colchicine was from Aldrich; and [3H]colchicinewas from never exceeded 2 h. Du Pont-New England Nuclear. The concentrations of the three At least two experiments were performed at each temperature. The compounds in PG buffer (10 m sodium phosphate, 0.1 m GTP, reported errors are the M M root mean squares of the individual uncertainpH 7.0) were measured spectrophotometrically employing the extinc- ties. tion coefficients t343= 17,600 M" cm" (20), tZm= 11,860 M" cm" (34), and c353 = 15,950 M" cm" ( l ) ,respectively. MTC and allocolRESULTS chicine were initially dissolved in dimethyl sulfoxide and theresidual concentration of this solvent in the experiments was less than 0.1% Interaction of MTC with Tubulin-Fig. 1 shows the en(v/v). GTP, dilithium salt, was from Boehringer Mannheim, Sepha- thalpy of MTC binding to tubulin as a function of the total dex G-25 from Pharmacia LKB Biotechnology Inc., Mes and sodium ligandconcentration at 298 K. Given the affinity of the dodecyl sulfate from Sigma, and other chemicals were of analytical interaction and the high concentration of reactants, essengrade from Merck. Protein and Binding Measurements-Calf brain tubulin was puri- tially all the ligandis bound in the first titration points; fied and stored in liquid nitrogen as described (1,24,25).Sucrose was therefore the initial slope and the saturation enthalpy value removed and the protein equilibrated in PG buffer, pH 7.0, by a 5-ml indicate the stoichiometry of the interaction, whichis 1.05 k dry Sephadex G-25 centrifuge column followed by an 8-ml wet Seph- 0.07 mol of MTC/mol active tubulin from the intersection of adex G-25 column (26). Tubulin concentration was measured specthe two dashed lines in Fig 1. The results were analyzed in trophotometrically employing the extinction coefficients f275 = 1.09 terms of Equations 2 and 3 (29). liters g cm" (6 M guanidine hydrochloride) and cZT5 = 1.07 liters g " " cm" (0.4% SDS in neutral aqueous buffer) (20). l/AQ = l/AHo(l + 1/Kb [L]) (2) Binding of MTC to tubulin was measured fluorometrically (20). The average binding stoichiometry obtained with 10 tubulin prepa[Ll = [Llo - AQ [ M J d A f f " (3) rations employed in this study was n = 0.68 & 0.11, which is significantly smaller than 1 and is in agreement with previous determina- where AQ is the heat measured, [Ll0and [L] are the total and tions (20). The colchicine binding stoichiometry of these same prep- free ligand concentrations and [% I is the total protein conarations was determined by incubation of tubulin with M centration. The values of t h e standard enthalpy change,A W , colchicine followed by chromatographic separation of the stable tuKb, were iteratively bulin-colchicine complex; it was close to one (0.93 & 0.06), also in and the equilibrium association constant, agreement with previous determinations (1). The experimental tu- computed to give the best fit to the experimental results, which is indicatedby the solid line i n Fig. 1. The values obtained are AW = -18.3 0.4 kJ.mol-' and Kb (5 k 1) x ' J. M. Nieto and J. M. Andreu, unpublished results.

Study of Tubulin-Colchicine Analogues by Reaction Microcalorimetry

25

16369

I
a

20
I

E
2 Y
v

15

10

20

40

60

80

100

120

CMTCl

106M

FIG. 1. Calorimetric titration curveof tubulin with MTC at

298 K in PG buffer. Four independent sets of experiments using 4.0 X M tubulin. Tlne solid line corresponds to the least squares fit of the experimental (data, which is Kb = 5 X lo5 M" and AW = -18.3 kJ.mol-'.

/+"-~T

t-

300

320
Temperature (K)

300

320

FIG. 2 Thermodynamic parameters for tubulin interaction . with MTC (a)and allctcolchicine (b) as a functionof temperature. Thermodynamic data for colchicine binding at 308 K are also
AW shown (AG (0); (*I;
and TAS (0).

IO5 M-'. The enthalpy change determined by the batch procedure ("Materials and Methods") with a final total concentration of MTC three! times larger than the tubulin concentration was 20.1 f 1.5 kJ.mol-'. Likewise, substitution of 10 m Mes for phosphate in the buffer had no significant effect M on the thermodynamic parameters of binding (Kb= (4.0 f 0.5) X lo5 "' and A W = -17.6 f 0.4 kJ.mol-'). The values estimated for the equilibrium constant are the same, within experimental error, those previously determined by equilibas rium gel chromatography, fluorescence, and differential absorption titrations. Under these conditions the equilibrium constant does not depend on protein concentration and MTC does not induce any significant tubulin association (20). The enthalpy change of MTC binding to tubulin was determined at several temperatures from titration and/or batch calorimetric measurements, and theresults are shown in Fig. 2a. A heat capacity clhange, ACp" = -20 f 80 J .mol-'. K" was estimated from the least squares linear fit, taking into account the errors in the enthalpy values. Therefore, at the present level of accuracy the heat capacity change is negligible

within experimental error, which precludes any further insight into thisparameter. The thermodynamic parameters of binding of MTC to purified calf brain tubulin are summarized in Table I; about one-half of the free energy change of the of reaction is provided by its weak exothermicity and the rest it by a positive entropy change in the temperature range examined. Interaction of Allocolchicine and Colchicine with TubulinThe enthalpy change of binding of allocolchicine to tubulin was measured at different temperatures ranging from 293 to 308 K. Due to the low solubility of this ligand in water, the experiments were carried out by the batch calorimetric procedure, using an excess of protein molecules. The standard enthalpy changes per mol of protein-ligand complex formed are shown in Fig. 2b. From the linear lease squares fit, taking AW into account the errors in the values, a ACpo = -9 f 100 J .mol". K" is obtained. Allocolchicine binds specifically to the colchicine site of tubulin with high affinity and a 1 to 1 stoichiometry (34). Since the formation of this complex is relatively slow, the binding time course was followed fluorometrically under the same conditionsand reactantconcentrations as the calorimetric experiments. These parallel experiments (not shown) ensured that binding equilibrium had been attained before thermal equilibration was reached in the calorimetric experiments, even atthe lower temperature; therefore all the heat of the binding reaction has been measured. Colchicine itself binds to tubulin even more slowly which hampers rigorous equilibrium measurements underusual conditions (1).In the present case the attainment of binding equilibrium could only be ensured at 308 K and high reactant concentrations. Fig. 3a shows the fluorimetric time course, which indicates completion of more than 95% of the reaction in 20 min. In a parallel experiment, a Sephadex G-25 column was equilibrated with the same colchicine concentration at 308 K, and tubulin at the same initial concentration was passed through it, adjusting the flow for an elution timeof 20 min. Measurement of tubulin and theincrement of colchicine .6 concentration in the effluent yields 0.98 f 0 0 mol of colchicine bound per mol tubulin. Fig. 3b shows a batchcalorimetric experiment under these conditions, from which a value of A W = -23.1 f 2 kJ.mol-' was obtained after correction for the heat of dilution of the protein. Fig. 3c is a calorimetric titration under the same conditions, which indicates a high affinity, 1 to 1, binding of colchicine to tubulin with AW = -19 f 1 kJ .mol". When similar experiments were performed at lower temperatures the apparent enthalpy change measured was smaller. However, the fluorometric time courses and column binding measurements showed that binding equilibrium had not been attained when the calorimeter tracing had
TABLE I Thermodynumic parameters of binding of MTC to tubulin T "0" -AGO ASo
K

283 17.0 288 17.4 293 20.9 298 19.1 303 16.0 308 17.5
This work.

kJ.mOl" f 1 20 6 .3 . ' f 08 .' f l.lb k 13 .' f LOd 32.4 f 0.6" 47 32.1 f 02 .' k 3 o2 3 .3 . d f 03 ." f 1.5d 33.1 f 0.3" 32.8 f 0.4"

J.ml".K-' 46 f 4
f5 44 & 5 54 f 11 51 f 6 50 f 7

'Determined by the batch method.

e

From Ref. 20. Determined byboth batch and titration methods.

16370

Study of Tubulin-Colchicine Analoguesby Reaction Microcalorimetry

a

ers (which have sufficiently different protonation enthalpies (30, 31)) are the same indicates that MTC does not induce any changes in the protonation state of tubulin. The value Z obtained at 298 K, A F = -19 & 1 kJ.mol-' (Table I), is comprised within the previous van't Hoff estimates from equilibrium measurements by Andreu et al. (20) (-6.7 f 2.9 kJ.mol-'1 and Bane et al. (21) (-28.4 & 4.2 kJ-mol-') and agrees fairly well with the value which can be calculated from the kinetic analysis and data of Engelborghs and Fitzgerald (18, 19) (-19 kJ.mol-', for 60% fastphase and 40% slow phase). In general, calorimetry provides a better measure of enthalpy than do measurements of the temperature variation of equilibrium constants. Therefore, previous van't Hoff enthalpy changes (18,21) should be regarded simply as approximations. In particular, the deviations of the value previously estimated by one of us (20) with respect tothe present calorimetric measurements, underthe same conditions, could be attributed to the weak temperature dependence of the fluorimetrically measured equilibrium constants and also to the uncertainty of these values in both extremes of the temperature interval (see Fig. 10 in Ref. 20). On the other hand, o / the calorimetric and thevan't Hoff enthalpies do not have to q / be necessarily identical, since species may be present which I I I I 1 2 3 are detected by calorimetry and not by fluorescence, even if [colchicine]/[tubulin] their contributions to the free energy change are insignificant. FIG. 3. Colchicine binding tubulin at 308 K. a, fluorescence to The measurements of the enthalpy change of the MTCtime course of the binding of 60 p~ tubulin and 150 PM colchicine. b, tubulininteraction at several temperatures have been exdirectcalorimetric recording of theheat evolved during tubulintended to allocolchicine and, partially, to colchicine. The ; colchicine complex formation([tubulin] = 32.7 p ~ [colchicine] = results, which are compared in Fig. 2 , clearly indicatea 112 p ~ ) the calorimetric recording corresponding totheMTC; tubulininteractionis also included as example of a fast reaction moderately exothermic binding with also a significant en. to the free energy change of tubulin-ligand (dashed line, [tubulin] = 45 PM; [MTC] = 17.7 p ~ ) Note that the tropic contribution heat measurements in b lag after the binding measurements in a due complex formation in the three cases. These positive entropy to theslow response time of the calorimeter. This technical limitation changes would be compatible with a contribution of hydrois very clearly exemplified by the calorimetric recording of the tubulin-MTC interaction which is practically instantaneous in the time phobic interactions (29) that should be established through scale of this experiment. c, calorimetric titration curve of tubulin with the trimethoxyphenyl moiety, as suggested previously (1, 14), since the tropolone methyl ether binding to tubulin is assocolchicine in PG buffer ([tubulin] = 63 "). ciatedto negative entropy and enthalpy changes (1, 14). Furthermore, the negative enthalpy changes associated to the TABLE I1 formation of the three complexes, constitute an additional Thermodynamic parameters of allocolchicine support to theinvolvement of van der Waals and/or hydrogen and colchicine binding to tubulin bond interactionsin the binding of colchicine to tubulin, Ligand T -AH"" -AGO A ' S suggested to be established through the C ring (1,14,15). The K kJ.mol" J.rnol".K" small ACp values estimated could be accounted for by the Allocolchicine 273 31.8 f 0.2' effect of accompanying polar groups, as exemplified by transf 1.5 293 10.7 fer heat capacity changes in model systems (29, 35). 34.7 f 0.5' 80 f 7 298 10.8 f 2.0 82 f 7 308 10.6 f 1.5 36.0 f 0.3' Effects of Ligand Modification-The fact that the thermoColchicine 308 21 f 2 43' 71 f 6 dynamic parameters of binding of MTC, allocolchicine, and colchicine to tubulin are comparable suggests that the basic This work. * From Ref. 34. molecular interactions leading to the observed binding are From Ref. 3. Allocolchicine was measured by the batch method; similar in the three cases. However, small but significant colchicine was measured by both batch and titration methods (see differences are apparent (Fig. 2). Let us now analyze the the text). differences between the thermodynamic parameters of MTC binding and those of colchicine and allocolchicine, AAG", returned to its base line, indicating that the heatflow should AAH", and AAS", (defined as AAX(1igand) = AX(1igand) fall below the instrumental sensitivity due to the slowness of AX(MTC) at 308 K). The values obtained result not only the binding process. Table I1 summarizes the thermodynamic from the differences in the contacts made between protein parameters for allocolchicine and colchicine binding to tubu- and ligand, but also from any other variation in the solvation lin. and mobility changes in the system when the protein and ligand bind. Since these characteristics are presently unavailDISCUSSION able, the following analysis is necessarily limited, but useful. The difference between the structures of colchicine and Thermodynamics of Specific Ligand Binding to the Colchicine Site of Tubulin-We have measured directly the enthalpy MTC is the presence of ring B (Chart I) which holds the change of binding of the bicyclic colchicine analogue MTC to acetamido group and restricts to two the possible rotational tubulin. This compound is at present the best characterized states of the biaryl bond between rings A and C ( 5 ) . This MTC. The difference colchicine site ligand (see the Introduction). The fact that the biaryl bond is however, free to rotate in enthalpy changes measured in either phosphate or Mes buff- between the bindings of colchicine and MTC, AAG", is roughly

Study of Tubulin-Colchicine Analogues by Reaction Microcalorimetry

16371

-10 kJ .mol-' favorabbe to colchicine, which arises predomi- would seem more consistent with a strongly unfavorable kinetic process. nantly fromapositivedifference inthebindingentropy changes (AAS" = 20 f 12 J . mol-'. K-'). This is fully consistAcknowledgments-We are indebted to Drs. T. J. Fitzgerald and ent with a partial immobilization of the intramolecular rotation of MTC upon binding, as suggested previously (20). On M. Gorbunoff for gifts of MTC and allocolchicine, respectively. We wish to thank Dr. S. N. Timasheff for helpful discussions. the other hand, it has'been stated in the literature that ring B is a third determinant recognized by the colchicine site (32, REFERENCES 33). However, such interaction of ring B with tubulin would 1. Andreu, J. M., and Timasheff, S. N. (1982) Biochemistry 2 1 , be likely to produce ,a largedifferential enthalpychange 6465-6476 between colchicine and MTC binding, which accordingto our 2. Andreu, J . M., Wagenknecht, T., and Timasheff, S. N. (1983) Biochemistry 2 2 , 1556-1566 experimentalmeasurelnentis practically negligible within 3. Garland, D. L. (1978) Biochemistry 17,4266-4272 experimental error (AAH" = -3.5 f 3.5 kJ.mo1). Therefore, 4. Lambeir, A., and Engelborghs, Y. (1981) J. Biol. Chem. 2 5 6 , although compensating effects giving a negligible net enthalpy 3279-3282 difference are possible, the simplest interpretation of the 5. Detrich, H. W., 111, Williams, R. C.,Jr., Macdonald, T. L., Wilson, present thermodynamic data is that there is little basis for a L., and Puett, D. (1981) Biochemistry 20,5999-6005 6. Andreu, J. M., and Timasheff, S. N.(1982) Proc. Natl. Acad. Sei. direct tubulin-ring B interaction in theequilibrium complex. U. S A. 214,151-157 . Actually, most published results are more simply rationalized 7. Saltarelli, D., and Pantaloni, D. (1982) Biochemistry 2 1 , 2996B constitutes a marked kinetic in lightof the finding that ring 3006 impedement to binding: and dissociation, without any strong 8. Margolis, R. L., and Wilson, L. (1977) Proc. Natl. Acad. Sci. U. direct effects on the tlnermodynamics of binding of several S. A. 74,3466-3470 9. Lambeir, A., and Engelborghs, Y. (1980) Eur. J. Biochem. 1 0 9 , ligands (34). 619-624 The differencesbetween the structures of allocolchicine 10. L., and and MTC (see Chart 1:)are the presence of ring B and also 11. Wilson,J. (1972)Bryan, J. (1974)1Adu. Cell. Mol. Biol. 3 , 21-72 Bryan, Biochemistry 1 , 2611-2616 the substitution of the carbomethoxyphenyl ring C' for the 12. Bhattacharyya, B., and Wolff, J. (1974) Proc. Natl. Acad. Sci. U. tropolone methyl ether ring C. The binding of allocolchicine S. A. 7 1 , 2627-2631 is slightly favoredwith respect the binding MTC. Similar 13. Barnes, L. D., Robinson, A. K., Williams, R. F., and Horowitz, to of P. M. (1983) Biochem. Biophys. Res. Commun. 116,886-872 to colchicine, there is a differential entropy change, AAS" = 31 k 13 J.mol-' .K" which may be mainly attributed to the 14. Andreu, J. M., and Timasheff, S. N. (1982) Biochemistry 21, 534-543 immobilization effect of ring B, as discussed above. In addi- 15. Rava, R. P., Hastie, S. B., and Myslik, J. C. (1987) J. Am. Chem. tion, the binding of allocolchicine is clearly less exothermic, SOC.1 0 9 , 2002-2003 of by approximately 7 k J . mol", than the binding MTC. This 16. Fitzgerald, T. J. (1976) Biochem. Pharmacol. 2 5 , 1383-1387 suggests a difference in the interactions of rings C' and C 17. Rossi, M., Link, J., and Lee, J. C. (1984) Arch. Blochem. Biophys. 23 1,470-476 with the protein. One possible explanation might be the loss 18. Engelborghs, Y., and Fitzgerald, T. J. (1986) Ann. N . Y. Acad. of a hypothetical hydrogen bond (1,29), although this cannot Sci. 466,709-717 be established at this moment. 19. Eneelborehs. Y .and Fitzeerald. T. J. (1987) J. Biol. Chem. 262. . . . Exothermic Binding clf Colchicine-Careful measurement of 604-5509 20. Andreu. J. M.. Gorbunoff. M. J.. Lee. J. C.. and Timasheff. S. N. the enthalpy change of colchicine binding to tubulin under ( m u j ~iochemistry 23,147211752 the restricted conditions inwhich attainment of equilibrium 21. Bane, S., Puett, D., Macdonald, T. L., and Williams, R. C., Jr. could be ascertained (high temperature and reactants concen(1984) J. Biol. Chem. 259, 7391-7398 trations; Fig. 3) hasindicated moderatelynegativevalues, 22. Diez, J. C., Avila, J., Nieto, J . M., and Andreu, J. M. (1987) Cell similar toMTC. In contrast to this, the bindingcolchicine of Motil. Cystoskel. 7 , 178-186 to tubulin commonly consideredto be strongly temperature- 23. Fernholtz, M. (1950) Justus Liebigs Ann. Chem. 568,63-72 is dependent and apparently possesses a markedly positive van't 24. Weisenberg, R. C., Borisy, G. G., and Taylor, E. (1968) Biochemistry Hoff enthalpy change (see the Introduction). However, most 25. Lee, J. 7,4466-4479R. P., and Timasheff, S. N. (1973) J. Biol. C., Frigon, experimental attempts to measure colchicine equilibrium Chem. 248,7253-7262 binding may be biased by the very slow kinetics, not allowing 26. Frigon, R. P., and Timasheff, S. N. (1975) Biochemistry 1 4 , attainment of equilibrium without substantial protein dena4559-4566 27. Wadso, I. (1968) Acta Chem. S c a d . 2 2 , 927-937 turation, as discussed previously (14). The calorimetric enthalpy change suggests that the equilibrium binding of col- 28. Chen, A., and Wadso, I. (1982) J. Biochem. Biophys. Methods 6, 307-316 chicine should be weakly dependent on temperature (as the 29. Eftink, M., and Biltonen, R. (1980) in Biological Microcalorimetry binding of MTC and allocolchicine), but it simply cannot be (Beezer, A. E., ed) pp. 349-350, Academic Press, London measured rigorously at low temperatures. 30. Christensen, J. J., Hansen, L. D., and Izzat, R. M. (1976) Handbook of Proton I o n i ~ ~ t i ~148, John Wiley & Sons, New York p. n , Finally, comparison of the calorimetric enthalpy change of colchicine binding with the kinetic analysis the fast phase 31. Amaralis Vega, C.,and Bates, R.G. (1976) Anal. Chem. 48,1293of 1296 by Lambeir andEngelblorghs (4) suggests that the backwards 32. Bhattacharyya, B., Howards, R., Maity, S. N., Brossi, A., Sharma, activation energy of the second step (Equation 1) should be P. N., and Wolff, J. (1986) Proc. Natl. Acad. Sci. U. S. A. 8 3 , in the order of magnit.ude of the activation energy of the 2052-2055 33. Banerjee, A., Barnes, L. D., and Ludueiia, R. F. (1987) Biochim. forwardsecond step (IO0 k J . mol-') and not as small as Biophys. Acta 9 1 3 , 138-144 depicted in a recent kinetic pathway (Fig. 4 in Ref. 19), 34. Medrano, F. J., Andreu, J. M., Gorbunoff, M. J., and Timasheff, possibly on the basis previous van't Hoff estimations ofthe of S. N.(1989) Biochemistry 28,5589-5599 enthalpy change of colchicine binding to tubulin. Further- 35. Franks, F., Reid, D. S., and Suggett, A. (1973) J. Solution Chem. more,ahigh activation energyfor this backward reaction 2,99-105
1

'

~~