Thermodynamic and Kinetic Studies on the Mechanism of Binding of ...

Vol. 267, No. 13, Issue of May 5, pp. 89094918,1992 Printed in U.S A .

THEJOURNAL OF BLOLOGICAL CHEMISTRY 01992 by The American Society for Biochemistry and Molecular Biology, Inc.

Thermodynamic and Kinetic Studies on the Mechanism of Binding of Methylumbelliferyl Glycosides to Jacalin* (Received for publication, July 10, 1991)

Dipti GuptaS, N. V. S. A. V. Prasad RaoS, Kamal Deep Pur& Khushi L. Mattan, and Avadhesha Surolia From the Molecular Biophysics Unit,Indian Institute of Science, Bangalore 560 012, India and the YRosewell Park Memorial Institute, New York, New York 10018

The binding of Artocarpus integrifolia lectin (jacalin) to 4-methylumbelliferyl (Meumb)-glycosides, GalaMeumb, GaVMeumb, GalNAcaMeumb,GalNAcBMeumb, and Ga433GalNAcBMeumb was examined by extrinsic fluorescence quenching titration andstopped flow spectrofluorimetry. The binding was characterized by 100% quenching of fluorescence of Meumbglycosides. Their association constants range from 2.0 X lo4to 1.58 X lo6M-' at 15 "C. Entropic contribution is the major stabilizingforceforavidbinding of Meumb-glycosides indicating the existence of a hydrophobic site that is complementary to their methylumbelliferyl group. The second order association rate constants for interaction of these sugars with lectin at 16 "C vary from 8.8 X lo6to 3.24 X lo*M" s-', at pH 7.2. The first order dissociation rate constants range from 2.30 to 43.0 s-l at 15 "C. Despite the differences in their association rate constants, the overall values of association constants for these saccharides are determined by their dissociation rate constants. The second order rateconstant for theassociation of Meumbglycosides follows a pattern consistent with the magnitude of the activation energiesinvolved therein. Activation parameters for association of all ligands illustrate that the originof the barrierbetween binding of jacalin to Meumb-glycosides is entropic, and the enthalpic contribution is small. A correlation between these parameters and the structure of the ligands on the association rates underscores the importance of steric factors in determining protein saccharide recognitions.

Artocurpus integrifolia agglutinin Cjacalin), isolated from the seeds of jack fruit has attracted considerable attention owing to its interesting biological properties, which include its potentand selective stimulation of distinct T and B lymphocytes of human origin (1)and its specific recognition of IgA, from human serum (2). This property of IgAl-specific binding is clinically important and hasbeen used for increasing phagocytosis of type II/c group B streptococci by macrophages, where jacalin apparently bridged the bacteria and *This workwas supported by a grant from the Department of Science and Technology, Government of India (to A. S.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18U.S.C. Section 1734 solelyto indicate this fact. $ Research associates in a project funded by the Department of Science and Technology (to A. S.). § Senior research fellow of the Council of Scientific and Industrial Research, India.

macrophages by binding on one hand to IgA, nonspecifically adsorbed on bacteria and on the other to theexposed galactose residues on macrophages (4). The lectin is reactive toward Thomsen-Friedenreich (T)-antigen,'a chemically well-defined tumor-associated antigen linked with malignancy in exquisite man (5). Our ligand binding studies demonstrated its specificity toward T-antigenic disaccharide in contrast to conformationally related disaccharide such as N-acetyllactosamine and GalpBGlcNAc (5). Moreover, the lectin distinguishes remarkably between GalBBGalNAcaR, the tumor associated T-antigenicdeterminants,and Gal/33GalNAcflR, which occurs as a partof developmentally regulated antigen, namely the asialo GM1 ganglioside. Its mechanism of ligand binding has been elucidated using stopped flow spectrofluorimetry with N-dansylgalactosamine and also by NMR spectroscopy using I3C- and '*F-labeled sugars (5, 6). Circular dichroism spectroscopy allowed spatial assignment of saccharides in the combining region of the lectin (7). All these studies have highlighted the ability of jacalin to discriminate between GalaMe over GalpMe. The strong binding of GalaMe was attributed to a favorable nonpolar interaction between its methyl group in a-configuration with the corresponding locus in the combining site of the lectin. The poor affinity of GalbMe was surmised due to a lack of this favorable interaction coupled with an unfavorable steric interaction between its methyl group in the /3 configuration and the combining site of the lectin. In order to further probe the role of bulkier substituents at the anomeric position, we have studied the interaction between 4-methylumbelliferyl glycosides of galactose, N-acetylgalactosamine and the T-antigenic dissacharide and jacalin. These studies reveal that the binding of Meumb glycosides and of Meumbbglycosides are driven by strong entropic factors, implying that hydrophobicity plays an important role in their binding process. An interesting feature noted for the first time in protein-saccharide interaction isthat theforward rate constants ( k l ) for the association of Meumb-glycosides are dependent on ligand structures. MATERIAL AND METHODS

Ligands GalaMe, GalpMe, GalNAcaMe, GalNAcpMe, Gal@Meumb, GalaMeumb, and GalNAcpMeumb used in our experiments were purchased from Sigma. GalPBGalNAc~Meumbwas synthesized as described by Hilde De Boek et al. (8).GalNAcaMeumb was a kind gift from Prof. I. J . Goldstein, University of Michigan, Ann Arbor. The abbreviations used are: T-antigen, Thomsen-Friedenreich antigen; Meumb, methylumbelliferyl 7-hydroxy-4-methyl-coumarin; dansyl, 5-dimethylaminonaphthalene-1-sulfonyl; GalaMe, methyl-aD-galactopyranose;GalpMe, methyl-8-D-galactopyranoside;GlcDMe, methyl-P-D-glucopyranoside; GM1, monosialoganglioside. All sugars used are D-sugars unless otherwise specified.

8909

8910

Thermodynamic and Kinetic Studies on Jacalin

Galp3GalNAcpMe was synthesized by the method of Flowers and Shapiro (91, except that GalNAcPMe was used instead of GalNAc. GalNAcpMe required for the synthesis of GalflIGalNAcPMe was prepared by the method of Neuberger and Wilson (10). The concentration of fluorescent sugars was determined using the molar extinction coefficient (e) value of 1.36 X lo4 M" cm" a t 318 nm (11). All solutions of Meumb-glycosides werefree of 7-hydroxy-4-methyl-coumarin as assessed by the absence of its fluorescence peak at 440 nm. Jacalin-Jacalin was prepared by affinity chromatography using cross-linked guargum (12), and its purity was checked by polyacrylamide gel electrophoresis. Protein concentrationwas estimated at pH 7.2 using e280 = 1.14 cm2 mg" and was expressed on the basis of molecular mass equal to 66,000 daltons, corresponding to four carbohydrate-binding sites. FluorescenceMeasurements-The fluorescence emission spectra of 4-methylumbelliferyl-glycosideswith and without the lectin were recorded on a Perkin-MPF 44A ratio recording spectrofluorimeter. Fluorescence titration were conducted on a Union Giken FS 501A fluorescence polarizer equipped with photon counting photomultipliers. The samples in 1 X 1 X 4.5-cm cuvettes were excited a t 318 nm with a 3.5-nm slit, and emission was monitored by means of a metal interference band pass filter (XI,+ = 10 nm) centered at 375 nm along with a 335-nm cut off filter. All measurements were made under continuous stirring with an inbuilt stirrer. Readings were taken after an interval of 2 min. The fluorimeter is controlled by a microprocessor which allows averaging of over 10 or more values. Constant temperature was maintained using a Lauda circulating water bath. To estimate the number of binding sites, a fixed concentration of jacalin (5 gM) was incubated with increasing concentration of GalPMeumb (1-50 p ~ 20, "C), GalNAcaMeumb (1-100 p ~ 27, "C) and GalaMe (1-100 pM, 28.2 "C) overnight prior to the fluorescence measurements. These parameters were then plotted according to Scatchard (13). Kinetic Studies-The kinetics of Meumb-glycosides interacting with jacalin was monitored on a Union Giken RA401 stopped flow spectrofluorimeter. Samples were excited at 318 nm and the emission monitored at 375 nm by using a narrow band pass filter (X, = 7 nm) at right angles to the excitation beam. Under these experimental conditions, the dead time of the instrument was 0.5 ms. The sample reservoirs and all compartments were maintained at the desired temperature (kO.1"C) using a Lauda circulating water bath. For determination of the association rate constants, 4 pM of Meumbglycosides were mixed with (40-500 gM) of jacalin. The dissociation rate constants of sugar-jacalin complexes were evaluated by dissociating the complex with a 10 mM solution of an inhibitory sugar, GalaMe. RESULTS

protein, respectively. For all the five sugars, the plot gave straight lines with intercept on the y axis equal to one, illustrating that the fluorescence quantum yield of Meumbglycosides bound to jacalin is zero, i.e., the quenching is 100%. The dataobtained by titration of a fixed concentration of the proteinwithvaryingconcentrations of GalSMeumb, GalNAcaMeumb, and GalaMeumb, analyzed according to Scatchard (13), gave a straight line (Fig. 3) with association constants (K,) of 1.74 0.118 X lo4 (S.E.;n = 4) at 20 "C, 4.38 f 0.28 x 10%(S.E.; n = 4)at 27 "C, and 1.65 k 0.07 x lo5 M" (S.E.; n = 4) at 28.2 'C and the values of the intercepts at theabscissa, n,equal to 3.6 k 0.18 (S.E.; n = 4), 4.0 f 0.25 (S.E.; n = 4), and 3.8 f 0.22 (S.E.;n = 4), respectively, for jacalin with M,of 66,000, indicating that the lectin is tetrameric in nature. The Scatchard analysis for GalNAcSMeumb and Galp3GalNAcj3Meumbwere not done because of the paucity of sugars. From the protein concentration-dependent quenching of Meumb-glycosides fluorescence, the KOvalues for their binding were calculated by the method of Chipman et al. (14) by plotting

*

- FJ/(F, - Fd] versus log[P]f log[(F, - Fc)/(Fc- F41 = log + {log[Pl,- (rn/aF")[LltJ

log[(F,

K O

where F, and F, have the same meaning as before and AF = (F, - F,) and A F , is the maximum change observed, viz. when all the ligand molecules are complexed with the protein. [PIt, [L], are total protein andligand concentrations, respectively, and [P]f is the free protein concentration which is the term in brackets on the right hand side, uiz. log[P]/ = (log[P], - (AF/AF,)[LIt). A representative plot for the binding of Galp3GalNAcpMeumb at 25"C is shown in Fig. l b (inset). The value of K. (intercept = pK.) determined by this method is 1.74 X lo5 M-' at 25 "C,assuming a stoichiometry of 1:l with the jacalin protomer of M,16,500. The values for other Meumb-glycoside range from 2.0 x IO4 (GalpMeumb) to 1.58 X lo6M-' (GalNAcaMeumb) at 15 "C. The value of KO for each of the ligands decreases with increasing temperature, as illustrated byVan't Hoff plots used for the evaluation of changes in enthalpy (AH'). From the changes in enthalpies and Gibbs free energies (AGO = -RT In K J , the changes in entropies(AS') for the association of these ligands were determined from

All the methylumbelliferyl-glycosides displayed similar fluorescence spectra. The fluorescence intensities of GalaMeumb, GalBMeumb, GalNAcaMeumb, GalNAcpMeumb, and Gal@3GalNAc@Meumb were totally quenched (Amax at 378 AGO = mo- TASO nm) on complete saturation with jacalin (193.4 pM). Quenching of the fluorescencewas observed to be completely reversed The KOvalues for the interaction of the nonfluorescent refupon addition of sufficient quantities of nonfluorescent sugars erence ligands GalaMe, GalpMe, GalNAcaMe, GalNAcpMe, such as GalatMe (0.1 M), GalNAcaMe (0.1 M), and and GalpBGalNAcSMe were also determined by substitution Gal/33GalNAcaMe (0.05 M). This shows that the binding of titration using GalNAcaMeumb and GalS3GalNAcj3Meumb Meumb-glycosides to jacalin is saccharide-specific. Failure of as theindicator ligands according to themethod of Bessler et jacalin (200 p ~ )to quench the fluorescence of al. (15) GlcNAc@Meumband GlcBMeumb, and GlcaMe to reverse the I[PlJ[PMl - 11[M], = (KL/KM)[L]f -t fluorescence quenching of GalNAcaMeumb, substantiates further that thejacalin-Meumbgalactosideinteraction is sac- where [PI,, [PM], [MIf, and [L], refer to the total protein, charide-specific. A representative plot of the changes in the protein umbelliferyl-glycoside complex, free Meumb-glycofluorescence intensity of Gal/33GalNAcpMeumbupon its ti- side, and free nonfluorescent ligand concentrations, respectration with increasing concentration of jacalin at 25 "C is tively. A plot of ([P]J[PM] - l)[M]f versus [Llf gives a shown in Fig. 1, a and b. The effect of the addition of defined straight line, the intercept and the slope of which yield the aliquots of 0.1 M GalaMe solution on the fluorescence inten- association constants for the indicating and the inhibitory sity is displayed in Fig. 2, a and b. The quantum yields of nonfluorescent ligand, respectively. Such a plot for the comMeumb-glycosides bound to jacalin were obtained by extrap- petitive binding of GalaMe to jacalin in the presence of olating a plot of F,/(F, - F,) versus inverse of protein con- Gal/33GalNAc/3Meumb at 25 "C is shown in Fig. 2. The ascentration (1/[PIt) where F, and F, arethe fluorescence sociation constantsthus obtained for GalaMe, GalSMe, intensities of free sugar and at a particular concentration of GalNAcaMe, GalNAcpMe, and Galp3GalNAcpMe are listed

8911

Thermodynamic and Kinetic Studies on Jacalin FIG. 1. A quenching of fluorescence spectra of Galj33GalNAcBMeumb by jacalin. a, experiments were carried out in 0.02 M phosphate (pH 7.2) containing 0.1 M NaCl at 25 'C. The fluorescence spectra were recorded in the absence (curue 1 ) and after addition of several aliquots of 193.4 pM jacalin to 2.1 ml of 4.41 p~ solution of Galj33GalNAc@Meumb:(curve 2, 3 pl; curue 3, 5 pl; curue 4, 10 pl; curue 5, 15 pl; curue 6, 20 pl; curue 7, 25 pl). The lower spectra (curve 8) is for standard buffer. b, titration of GalS3GalNAcPMeumb with jacalin a t 25 "C.The fluorescence intensity is plotted as a function of added protein concentration. A 3.68 p~ solution of GalP3GalNAcj3Meumb was titrated with 501 pM jacalin. Inset gives a graphical representation for the determination of association constant K , = 1.74 X lo6 M-'. The straight line is drawn according to theregression equation ( n= 8, r = 0.9987).

I

330

L

I

I

I

0 5 10 Protein aliquot added ( P I )

52

378 A (nm)

FIG. 2. Competitive binding of MeaGal to jacalin in the presence of Galj33GalNAcj3Meumb at 26 "C. a,

1

A (nm)

Volume of

in Tables 1-111. Van't Hoff plots of K,, for these sugars are shown in Fig. 4.It may benoted that thevalues of association constants determined for these indicating ligands by substitution titrations are in the range of values obtained from direct titrationsusing the method of Chipman et al. (14).The spatial features of the combining regions of jacalin as elucidated from the thermodynamic parameters aredepicted schematically in Fig. 5. Kinetic Studies-Since the fluorescence of Meumb-glycosides decreases with time upon their rapid mixing with jacalin [PI, theelementary steps involved in these interactions were evaluated by stopped flow spectrofluorimetry. The observed rate constant, kobs, for the change in fluorescence of an indicator ligand is related in a concentration-dependent manner on thetype of elementary step to which the relaxation belongs. Three most likely possibilities for protein-ligand interactions, devoid of covalent transformation of either of the components, are considered below.

the fluorescence intensity of GalP3GalNAcPMeumb (2.1 ml, 4.41 p ~ curue , 1 ) on addition of jacalin (193.4 pM, curue 2). The fluorescence intensity after the addition of successive aliquots of MeaGal(100 mM): curue 3 , 5 pl; curue 4, 5 pb curue 5, 10 pl; curue 6, 15 pl; curue 7, 20 pl. Curve 8 is buffer control. b, substitution titration with MeaGal. To a 3.68 PM solution of Galp3GalNAcPMeumb (2.1 ml), 501 F M solution of jacalin was added (100 ~ 1 )The . mixture was then titrated with MeaGal(100 mM). The fluorescence intensity was plotted as a function of MeaGal added. Inset gives a graphical representation for the determination of the association constant for the indicator ligand (KO= 1.80 X 10' M-') and the competing ligand ( K , 15 = 2.21 X lo4 M-'). The straight line is drawn according to the regression equaMe-a-GaI added (PI) tion ( n= 9, r = 0.9959).

1

For a simple, single bimolecular association between the ligand [L] and theprotein [PI (Equation l),the observed first order change in fluorescence is related to theassociation and dissociation rate constants and the excess component [PI by Equation 2. k1

P + L " p L k-1

hobs

= k-1

+ kdplo

(2)

The second possibility involves the rapid formation of an intermediate, PLiwhich isomerizes to yield the final complex PL*

Thermodynamic andon Studies Kinetic

8912 In such a situation and [PI0 by

kobs

is related to various rate constants

where K-, =

kk+1

The last case is that of a protein undergoing slow transformation between two states of which only P* is able to associate with the ligand.

for which

Jacalin

jacalin than the concentration of the Meumb-glycoside to maintain the psuedo-first order kinetic conditions. A representative stopped flow fluorescence trace of time-dependent change in fluorescence of Galp3GalNAcPMeumb upon its rapid mixing with jacalin is shown in Fig. 6. The observed rate constant, kobs, was evaluated from the slopes of h e a r plots of ln(F, - Ft) uersm time ( t ) where F, and Ft are fluorescence at infinite time and time t, respectively. A representative plot of ka,, uersm [PI used for determination of kl and k l for Galp3GalNAcBMeumb at 25 "C is shown in Figs. 7 and 8. The values of k 1 were also determined directly by displacing indicator ligandsfrom their complexes with jacalin by GalaMe. Kinetic experiments for these ligands were carried out at several temperatures ranging from 10 to 30 "C in order to calculate the activation parameters. The activation parameters for the association and dissociation reaction for these ligands were obtained from the Arrhenius plot (Figs. 9 and 10). Activation enthalpy energies andentropies were calculated using the following equations

AH' = where

EA

ln(k/T) = -AH#/RT K-, =

kka

1

I

I

I

2.0 r

+ AS#/R + ln(k'/h)

AG' = AH' - TAS#

Since kobs depends in a characteristic manner on [PIo it could be exploited to discriminateexperimentally between the three mechanisms outlined above. From Equation2, it follows that k& would increase linearly with[PIo; Equation4 predicts that kobs would increase linearly with [PIo butwould tend to saturate as [PIoincreases from a value much lower than l/Kl to P >> l/Kl; while according toEquation6, kobs would decrease as [PIoincreases. The forward rate constants were measured by performing the experiment using at least 10-fold higher concentration of

0

- RT

4.0

FIG. 3. Scatchard plot for binding of MeumbaGalNActo jacalin by fluorimetric titration at 27 OC. A fixed concentration of jacalin (5 FM) was titrated with varying concentrations of MeumbaGalNAc (1-50 PM), and the datawere analyzed according to Scatchard. The straight line is drawn according to the regression equation ( n = 7, r = 0.9538).K. is calculated as 4.38 & 0.28 X lo5M" (S.E.; n = 4) with 4.00 +. 0.25 (S.E.; n = 4) binding sites.

where k istheappropriaterateconstant, k' is Boltzman constant, and h is the Planck's constant.Activation parameters thus obtained arelisted in Tables IV-VIII. The association rate constants for bindingof Meumb-glycosides to jacalin determined in this study vary from 0.88 X lo6 to 3.24 X IO6 M" s" at 15 "C.Values of obtained from the y intercept of this plot are consistent with those estimated from the first order rate constants for the displacement of Meumb-glycosides complexed with jacalin by a large excess of GalaMe. The values thus calculatedfor various Meumb-glycosides range from 2.30 to 43.0 s" a t 20 "C. DISCUSSION

A systematic thermodynamic and kinetic analysis of the interaction of jacalin with GalaMeumb, GalpMeumb, GalNAcaMeumb, GalNAcpMeumb, and GalB3GalNAcpMeumb reported here in conjunction with the knowledge of its carbohydratespecificity reveals severalinteresting features pertaining to itsligand binding properties.The stoichiometry of saccharide binding to jacalinwas redetermined, due to the fact that ourperception of the molecular weight of the protein has changed considerably (uiz.from M, 40,000-66,000) during recent years(16). The Scatchard analysesperformedwith GalpMeumb, GalNAcaMeumb, and GalaMeumb gave values of 3.6, 4.0, and 3.8, respectively, for the number of binding sites indicating thatjacalin with M , 66,000 has four binding sites, which is in agreement with the value of2.2 reported earlier (5) with M , 40,000. The total quenching of fluorescence intensities of all the Meumb-glycosides upon interaction withjacalinindicates that the Meumb group experiences a similar environment a t or near the combining region in all the sugars and implies that it occupies a similar position in the protein-bindingsite. Among the possible factors that can cause the quenching of the Meumb fluorophore, anchored a t or near the combining region as a result of carbohydrate specific binding, is the polarity of the environment. This evident is fromthe decrease in the relative fluorescence of all the ligands, as a function of the solvent polarity, determined in solution with increasing concentration of p-dioxane; the decrease is identical for all five ligands withfluorescence quenched to 90% of initial value in pure dioxane (data not shown). Thus, the environment of

8913

Thermodynamic and Kinetic Studies onJacalin TABLEI Association constants and thermodynamic parameters for binding of a-sugars to jacalin Values in parentheses indicate S.E. values ( n = 4). 10-~ X K. Sugar

25 "C

30 "C

"' GalaMe

47.80 (f1.25)(k2.58) GalaMeumb292.00 394.00 (27.7) GalNAcaMe 92.60 (f6.37) GalNAcaMeumb 1580.00 (k71.76) (k104.7)

-ASu*

-AGu*n

-AH0

20 'C

15 "C

J.mol-1.K"

k J . mol"

22.00 (51.15) 209.00 (k17.02) 35.35 (f2.35) 710.00 (k47.88)

30.30 (k14.56) 55.59 (+.4.01) 1120.00

25.14 15.90 (k0.85) 30.66 148.00 (33.89) 23.21 (k1.61) 330.00 (k24.46)

56.47 (k2.99) 48.48 (k3.01) 66.88 (k2.06) 66.71 (k3.52)

106.93 (k8.25) 60.78 (k5.20) 137.41 (k7.99) 111.50 (k11.58)

(kO.19) (k0.12) 26.62 (k0.20) 34.04 (k0.15)

'*, Values calculated at 20 "C. TABLEI1 Association constants and thermodynamic parameters for binding of /%sugars to jacalin Values in Darentheses indicate S.E. values ( n= 4). lo-? x K , Sugar

20 "C

25 "C ~~

~

30 "C .~ ~

GalPMeumb GalNAcBMe GalNAcpMeumb Galp3GalNAcpMe Galp3GalNAcpMeumb

0.22 (k0.007) 20.00 (k1.00) 0.44 (k0.03) 72.50 (k4.89) 0.41 38.49 (f0.015) 275.00 (k16.38)

J . mol". K"

k J . mol"

"1

GalPMe

-ASu*

-AGO*"

-AH0

15 "C

0.20 0.12 (k0.007) (k0.004) 16.00 23.58 10.70 41.57 (50.74) (k2.33) (f0.57) (k0.64) 51.30 35.50 (f3.16)(k3.02) (f1.67) (k2.45) 0.30 0.25 (kO.O1l) (k0.008) 214.00110.00 174.00 (k13.88) (k0.16) (k3.57) (f6.39) (klO.90)

0.08 12.87 (f0.003) 8.90

44.85 (2.02)

109.15 (k4.90) 61.40 (k3.12)

(kO.09) (kO.ll)

25.70

52.76 (k1.56) 35.04

14.5gffb (k0.14) 32.03 (k0.15) 13.92 (kO.09) 29.90

70.75 (k5.19) 83.86 (k2.19) 17.54 (k1.56)

(I*, Values calculated at 20 "C. b#,

Value calculated at 15 "C.

A

0

H

GalQlMc GalQlMoumb GalPMe GalBMournb GalNAcQlMo

5.0L GalNAcQlMournb

GalNAcPMo

GalNAcBMcurnb

Gal~93GalNAc~Me

Galp3GalNAc~Meurnb

4.0 -0.3

3.45 3.40 3.35

FIG. 5. A schematic representation of the combining region of jacalin. Subsite H is the hydrophobic site which accommodates the Meumb and methyl grow of Meumb-glycosides and Me sugars

subsite B alone. On the other hand aas consequence of the dominating

the indicating group Of all these saccharides in the combiningfavorable influence ofthe Meumb group, the GalNAc and Gal residues site of jacalin appearsto be essentiallynonpolar. of Galp3GalNAcPMeumb are able tointeract with subsites A and B, Jacalin is able to distinguish remarkably between GalaMe respectively.

8914

Thermodynamic and Kinetic Studies on Jacalin (a)

; : z

I

.6I

Y

-

l

-"

41

1 .-rn

I

I

I

I

0

I

20

LO

60

80

Time ( m s e c . )

I

GaINAcBMeumb-jacalin association reaction at 25 "C. Galp3GalNAcpMeumb (4 PM) and jacalin (40PM) were allowed to mix in equal amounts in a stopped flow instrument at 25 "C. The samples were excited at 318 nm, and emission was recorded above 375 nm. The spectrum represents the average of 10 measurements. The inset represents a plot drawn according to theregression equation ( n = 13, r = 0.9993), the slope of which yields kobs value of 95.6 & 4.06 (S.E.; n = 4).

t

800

400

1200

1600

2000

Time fmsec)

FIG.8. Evaluation of the dissociation rate constantfor Ga433GalNAcBMeumb-jacalin interactionat 25 "C using GalaMe as the competing ligand. Jacalin (80 PM) and indicator ligand (4 PM) were mixed with an equal volume of 100 mM GalaMe. The slope of the line drawn according to the regression equation ( n = 9, r = 0.9998) in the inset gave the value of k-l = 8.96 s-'.

d

I

8001

L

3.30

400! 200

160 240 320 400

80

Time (rnsoc.)

0

FIG.6. A stopped flow fluorescence trace of Galj33-

'Oo0

0

100

I

I

3.355 3.41 103/1 ( K - 1

I

I

3.47

FIG.9. Arrhenius plots for the association (0)and dissociation (0) rate constants for the binding Ga433of GalNAcBMeumb to jacalin weredrawn according tothe regression equation with n = 4, r = 0.9965 and n = 4, r = 0.9925, respectively. kl and k-I plotted are the mean values of 910 independent runs. S.E. for the parameters thus determined are listed in Table VIII.

/ [PI (PM)

FIG.7. Determination of rate constants for the association of jacalin and Galfl3GalNACBMeumb. kobs values determined at 25 "C were plotted against the protein concentration [PI (after mixing). The concentration of the sugar wasfixed at 2.00 p~ (after mixing). The slope of the line yielded a kl value of 1.56 X lo6 M-' s-'. The straight lineis drawn according to the regression equation ( n = 9, r = 1.0000).

and GalPMe. The ability of jacalin to discriminate between the a- and P-anomers is further substantiated by the higher association constants of Meumbaglycosides (3.94 X lo5 and 15.8 X lo5M-' at 15 "C for GalaMeumb and GalNAcaMeumb, respectively), as compared to their p-anomers (2.0 X lo4 and 7.25 X lo4M-' for GalPMeumb and GalNAcPMeumb, respectively). This stronger affinity of a-anomers over p-anomers of Meumb-glycosidesis due to more favorable enthalpic contributions for the binding of the former. Galfl3GalNAcflMeumb is afourtimesstronger ligand than GalNAcpMeumb. GalNAcPMeumb and GalP3GalNAcPMeumb show 164- and 670-fold greateraffinities over GalNAcPMe and Gal/33GalNAc@Me,respectively, indi-

40

o 30

a

"r

2"

01

AGO

I

\I I 11 1

Reaction coordinate

FIG. 10. Thermodynamics/kinetic profile forthe binding of GalNAcaMeumb (a)and G a M e u m b to jacalin ( b ) .

cating a positive contribution of Meumb group for the association of Meumbb-glycosides. Since the enthalpic increase is insignificant the betteraffinity of Meumb-glycosidesover the corresponding Me-glycosides is probably due to favorable entropic contributions of Meumb group (Tables I and 11), reflecting the additional hydrophobic interaction between the Meumb group of ligands and the protein. This difference is

Thermodynamic and Kinetic Studieson Jacalin TABLE111 Differences between A H " , AS', and AGO values of Me-glycosides and corresponding Meumb-glycosides Sugar

-AGO

-ASn

-AH0

GalaMe GalaMeumb

106.93 25.14 30.66

56.47 48.47

60.78

GalPMe GalpMeumb

12.87 23.58

44.85 41.57

GalNAcaMe GalNAcaMeumb

26.52 34.04

GalNAcpMe GalNAcpMeumb Galp3GalNAcpMe GalS3GalNAcSMeumb

k,,

"c

M".s"

15

1.52 (kO.09) 1.65 (fO.ll) 1.70 (f0.12) 1.60 1.86 (fO.ll)

20 25 30

k- I

X

S"

3.86 (f0.26) 5.65 (f0.39) 7.84 (f0.53) 11.61 (f0.82)

AH?, =

7.90 kJ.mol-' (f0.55) AS!, = -98.70 J . mol". K" (f2.04) AG?, = 35.08 kJ.mol-' (k0.15) E!, = 7.06 kJ.mol-' (f0.59)

AH" =

AAS' J . mol" .K"

AAH'

kJ-mol"

-5.52

8.00

46.15

109.15 61.40

-10.71

3.28

47.75

66.88 66.71

137.41 111.50

-7.42

1.17

25.91

14.59 26.42

52.76

70.75

13.92 29.90

38.49 35.04

83.86 17.54

3.45

66.32

-11.83

10-5 X K.

10-5 X K. (equilibrium)

(kinetics)

T

"c

"1

"1

-15.98

TABLEV Rate constants and activation parameters for the interaction to jacalin with GalPMeumb Values in Darentheses indicate S.E. ( n = 4).

TABLEIV Rate constants and activation parametersfor the interaction to jacalin with GalaMeumb Values in parentheses indicate S.E. values ( n = 4). T

AAC'

J . mol". K"

kJ.mol"

0.79

k,, x

k-1

M".s"

10

3.93

3.94

2.92

2.92

15

2.15

2.09

20

1.48

25

(k0.06) 0.88 (k0.06) 1.oo (k0.07)

S"

28.57 (k2.00) 43.00 (f2.98) 61.63 (k4.40)

56.54 kJ. mol" (f1.93) AS?, = -12.62 J.mol".K" (f3.56) = 68.76 kJ. mol" (f0.23) Et1= 67.70 kJ .mol" (f2.50)

AH?, - AH!,= -60.45 (f2.99)

AS' = AS?,

- AS!,

= -91.34 J.mol-'.K"

AGO = AG?,

- AG?,

= -33.68 kJ mol"

(f3.67)

(f1.03) AS?,= -79.11 J.mol".K" (f4.95) A G ~= , 38.05 kJ.mo1" (k0.17) E t , = 17.31 kJ.mol-' (f1.00)

AH' kJ.mol"

.

X

K.

(kinetics)

10" X K. (equilibrium)

"'

"'

2.60 2.04

1.58

1.61

1.12 1.07

AH?,= 14.87 kJ. mol" AH?,=

10"

AH?,

=

AS?,

=

A@,

=

E?, =

54.22 kJ.mol-' (k2.26) -25.43 J . mol". K" (k2.30) 61.67 kJ.mol-' (f0.56) 56.66 kJ .mol" (k2.56)

AH?, - AH!,= -39.35

kJ.mo1" (f2.57) AS" = As!, - AS!, = -53.68 J.mol-I. K" (f2.21) AGO = AGf, - AGEl = -23.62 kJ.mol-' =

fk0.78)

(f2.61)

Meumb, Four-fold etc. higher affinity of Gal@3more pronounced for ligandscontaining theMeumb group in GalNAcPMeumb over GalNAcPMeumb suggests that the @@-linkagethan in a-linkage. galactopyranoside group of this disaccharide interacts with a Complete quenching of their fluorescence intensities to- secondary subsite designated as subsite B, which contributes gether with positive entropies involved in these interactions to the increased value of its association constant. In other is indicative of the hydrophobic contribution of the 4-meth- words, the Meumb group alsoappears to facilitate the binding ylumbelliferyl moieties of these glycosides to their affinities of the galactose residue of GalB3GalNAc in addition to that for jacalin. It is therefore apparent that the Meumb group of its GalNAc residue. Thus, the combining region of jacalin binds toa hydrophobic site (subsiteH), adjacent to the sugar- could be considered to consist of three subsites: subsite H binding subsitesA and B (Fig. 5). This indicates that althoughwhich is nonpolar in nature and which interacts with the the methyl group of GalPMe destabilizes the interaction, the Meumb group of Meumb-glycosides, and subsites A and B Meumb group at the same locus increases the affinity due to which accommodate the penultimate saccharide moiety of favorable entropiccontributions.TheMeumb group in p- monosaccharides and the terminal galactose residue, respeclinkage shows increased affinity either due to its interaction tively, of the Galp3GalNAcpMeumb disaccharide. The model with a nonpolar site that lies distant from the position occu- proposed here is in accordancewith the model proposed pied by the methyl group of GalpMe or it spans, in addition, previously (7). As pointed out earlier, subsite A is specific for the nonpolar locus that iscomplementary to the methyl group a-linked orfree galactose and galactosamine residues,whereas of GalaMe. In either of the events, the Meumb group in @- only @-linked monosaccharides, including those from @-linked linkage is located at a position different from that of the disaccharides, are accommodated in subsite B. As the binding methyl group of Gal@Me, GalNAcpMe, Galp3GalNAcp- of Meumbpglycosides is dominated by the contribution of

Thermodynamic and Kinetic Studies on Jacalin

8916

TABLEVI Rate constants and activation parameters for the interaction to jacalin with GalNAcaMeumb Values in parentheses indicate S.E. values ( n = 4). T

k+, x

k- 1

lo-'

x K,, (kinetics)

X

K.

(eauilibrium)

TABLEVI11 Rate constants and activation parameters for the interaction to jacalin with Galb3GalNAcpMeumb Values in parentheses indicate S.E. values ( n = 4).

T

"c 10 15 3.40

3.17 1.48 (k0.21) 3.24 (k0.22)

2.19 (k0.15) 2.30 (k0.15) 3.36 (k0.23) 5.38 (k0.33)

20 25

(k0.24) 0.67 3.59 (k0.26)

2.19

1.25

1.411.58

1.58 1.41

1.01

1.12 0.71 0.67

0.71

AH!,= 65.07 kJ. mol"

AH$, = 4.62 kJ. mol" (k0.70) AS?, = -103.96 J.mol".K" (k3.35) AG?, = 35.08 kJ. mol" (k0.70) E?, = 7.06 kJ. mol" (k0.79) AHo =

1.48

AH!,- AH!,

AS" = AS?,

-

AGO = AG?,

-

AS!, = -12.62 J . mol". K" (k1.26) AG!, = 68.76 kJ. mol" (k0.53) E!, = 67.70 kJ. mol" (k3.35)

(f3.23) AS!, = -91.34 J.mol".K" (k2.31) AG!, = -33.68 kJ.mol-' (f3.40)

TABLEVI1 Rate constants and activation parametersfor the interaction of jacalin with GalNAcPMeumb Values in parentheses indicate S.E. values ( n = 4).

2.26

T

k+, x

"c

"'.s-'

10

2.11 (f0.15)

"'

15 20

2.50

k-1

(k0.15) 2.38 (k0.16)

25 (k0.18)

S"

20.67 (k1.49) 7.25 28.90 (k2.05) 5.13 45.45 (k3.19) 3.55 61.96 (k4.33)

AH?,= 4.78 kJ. mol-'

AH$l

x K,

(kinetics)

10-4 x K. (eauilibrium)

"' 10.20 7.82 5.19 4.03

AH!,= 55.43 kJ .mol"

(k0.43) ASfl =-106.49 J.mol".K" (f3.10) AG?, = 35.98 kJ. mol" (k0.39) E t 1 =7.21 kJ.mol-' (k0.67) AHo =

io-'

M-'.s-l

k- 1 S"

10-~x K. (kinetics)

10-~ x K.

(equilibrium)

"'

"1

15 2.3020 25 5.3830

AH?I

(k0.08) 1.40 (kO.09) 1.56 (fO.10) 1.64 (kO.11) = 4.62

(k0.39) (k0.50) 1.12 3.36 (k0.65)

1.01

(kO.80)

kJ. mol"

(k0.59) AS$, = -103.96 J.mol".K" (f2.32) AG?, = 35.08 kJ. mol" (k0.29) E?, = 7.06 kJ. mol" (k0.85)

AH!,= 65.07 kJ. mol" (k2.16)

AS!l = -12.62 J.mol".K" (k3.99)

AG!, = 68.76 kJ.mol-' (kO.91)

E!, = 67.70 kJ. mol" (k2.56)

AH" = A H $ , - AHf1 = -60.45 kJ.mol-'

kJ.mo1"

= -60.45

k+l x

(k2.91)

As!, = -23.86 J . mol". K" (k2.09)

AGX, = 62.42 kJ. mol" (k0.86)

E!, = 57.87 kJ.mol-' (k3.13)

- AH!,= -50.65 kJ.mol" (zk3.00)

AS" = ASfl - AS!, = -82.63 J.mol".K" (k2.19) AGO = AGfl - AGE, = -26.44 kJ.mol-' (k1.09)

Meumb group (i.e., - AAG = -10.71 to -15.98 k.J.mol"), it holds the saccharide moiety of GalPMeumb and GalNAcBMeumb to subsite A, in spite of the propensity of their @-linkedsugars to reach the subsite B. The stronger binding of GalB3GalNAcBMeumb can also be explained in the same manner, i.e. the interaction of the Meumb group with the subsite H promotes the association of GalNAc and Gal residues in subsites A and B, respectively, thereby im-

(k2.56) AS" = AS+, - AS!, = -91.34 J.mol".K" (k3.80) AGO = AG$1 - AG!, = -33.68 kJ.mol-' (k1.31)

parting it a higher affinity as compared to GalNAcpMeumb. It has been suggested by Jencks (17) that Gibbs free energy changes for the binding of proteins to the individual components of ligands can be determined in terms of their respective intrinsic binding energies ( G ) and a "connection Gibbs energies (G").The latter (G")derives largely, if not exclusively, from changes in translationalandrotational energy. An analysis of this type is helpful in the interpretation of the observed AH" and T A P values of binding in aqueous solutions. Binding of jacalin to Meumb-glycosides is not amenable to such an analysis due to the failure of the binding of 4methylumbelliferone by itself. Thus, despite its strong and positive contribution to the binding process, which is largely due to nonpolar interactions, the umbelliferyl moiety by itself has no access to the binding cleft and has to be positioned appropriately in the hydrophobic site by the saccharide component of Meumb-glycoside. Another important difference between these studies and analysis of Jencks involves the influence of the configuration of the bond linking the sugar moiety with the Meumb group on both the thermodynamic and kinetic parameters'\of interaction between jacalin and Meumb-glycoside. Kinetic Studies-The kinetics follow a one-step binding mechanism. An interesting feature noted for the first time in protein-sugar interactions is the dependence of the second order rate constants for the binding of jacalin to Meumbglycosides onthe nature of the linkage of the aglycon umbelliferyl moiety as well as the sugar per se. GalNAcaMeumb binds four times faster than GalpMeumb. The association rate constant obtained here for a-anomers ( k , = 1.65 X lo6 and 3.40 X lo6M" s-l for GalaMeumb and GalNAcaMeumb, respectively) are nearly 1.5 times higher than those for the corresponding p-anomers ( k , 1.0 X lo6 and 2.38 X lo6 M" s" for GalpMeumb and GalNAcpMeumb, respectively). Association rate constants for the a- and 8-anomers of GalNAc are found to be about 2-fold higher than those for the corresponding anomers of galactose. The kl values observed here are

Thermodynamic Kinetic and faster than most other lectin-sugar interactions. It is also noted that thebinding of Meumb-glycosides and especially of their a-anomer9 are more than an order of magnitude faster than GalNAc itself (6). These second order rate constants, however, are two orders of magnitude slower than the diffusion controlled rate constants. One generally explains such reactions by invoking the formation of an intermediate (PLi) that isomerizes to form the final complex (PL*). If 12-1 >> kz then kobawould increase linearly when P > l/Kl. The failure to observe PLi could be attributed to itslow l/Kl value, so that significant quantities of it do not accumulate during the course of the reaction. Since our kObsversus [PIo plots are linear up to 250 PM of the latter, l/Kl has to be lower than 4000 M-'. Despite lower second order rateconstants observed here rather than thediffusion controlled reactions, the probability of occurrence of reaction intermediate(s) within the dead time of the instrument or otherwise is ruled out since the fluorescence change appearing in the stopped flow traces, for both the association and dissociation reactions, are similar in magnitude to those observed in the steady state titrations. The agreement between kinetically determined values of association constants (kl/k-l) and changes in enthalpies and those determined by fluorescence titrations indicates that the K,, and the enthalpy changes are related to the total binding process and that there does not exist any unobserved faster process that contributes appreciably to these parameters for the saccharide binding. Linearity of Arrhenius plots rules out the occurrence of dramatic conformational changes in the lectin molecule in the temperature range studied. Thus, kinetics of binding of Meumb-glycosides to jacalin is both quantitatively and qualitatively consistent with a single step bimolecular association reaction. Inspection of activation parameters of Meumb-glycosides reveals that theenergy of activation is the limiting factor for the differences in forward rate constantfor these saccharides, as E+1is lower for the ligands which display faster association rate constants. A substantial activation enthalpy for dissociation of Meumb-glycosides-jacalincomplex could be due to the energy required to break the bonds between these complexes duringtheir dissociation process. Much of the activation energy component involved in the association of Meumbglycosides is due to a high entropic contribution since the enthalpicbarrieris insignificant. This indicates thatthe interaction between Meumb-glycosides and jacalin is sterically restricted. It is noteworthy that theenergy of activation for the binding of Meumb-glycosidesto jacalin is significantly lower than those observed for association of N-dansylgalactosamine to jacalin f l + l = 56.4 kJ.mol-'; AS!, = 59.94 J mol-' K-'). This implies that a large amount of activation energy is required to fit the bulky dansyl substituent at the C-2 position of galactosamine in the binding site of the protein. GalNAcaMeumb and GalNAcPMeumb show greater kl values than the a- and B-anomers of GalNAc for binding to jacalin (6). It is therefore interesting to compare the values of activation energies for a and P-anomers of [13C]GalNA~ ( f l + l is 8.68 and 8.186 kJ.mol-' for a- and P-anomers, respectively) with those of GalNAcMeumb (6). In spite of the presence of a large bulky Meumb group, activation energy of a- and p-anomers of GalNAcMeumb are lower than a- and p-anomers of GalNAc. It is thereforeapparent thatthe Meumb group contributes favorably to both the binding affinity and rate constant.The activation entropy involved in the binding of Meumb-glycosides to jacalin can be utilized in overcoming some steric constraints, such as some conformational changes in the protein and protein-ligand complex.

Studies on Jacalin

8917

These appear very probable in view of the considerable ligand induced changes in the CD pattern of the protein (7).However, such conformational changes are likely to occur only at the tertiary structural level and not at the secondary S t N C tural level, as no change in the CD spectra was observed in the secondary structural region (7). Additionally, this term may also reflect perturbation, release, or shift in the state of water upon ligand binding. This is similar to the binding of dyes to cyclodextrins (18) where the role of steric factors has indeed been demonstrated. A comparison of the energy diagram for the binding of GalNAcaMeumb and GalPMeumb to jacalin, which are the fastestand slowest binding ligands, respectively, provides some insights into the mechanism of the interaction of jacalin with Meumb-glycosides. Despite the unfavorable activation entropy which is the principal barrier for jacalin-Meumbglycosides association, the noticeable differences in thesecond order rate constants for the binding of GalNAcaMeumb over GalPMeumb could be accounted for by differences in their activation enthalpies. Thus, apart from the requirement of precise configuration of reactants the overall rate constants are determined by energetic features. This energetic barrier could be utilized in the breaking of hydrogen bonds between the solvent molecules and the sugar, between the protein and the solvent molecules, and for the formation of newer ones between the sugar and the protein in the complexes. The slower binding ligands thus need more energy to form the complex. Despite the failure to overrule the predominant role of steric factors in determining protein-saccharide interaction, we as well as others have explained that smaller second order rate constants in the association of lectins with sugars result from the formation of an unobservable reaction intermediate which isomerizes to form the final complex (5, 6, 8, 19-24). These studies demonstrating the favorable influence of the Meumb group on the rates of binding of Meumb-sugars visa-vis anomers of GalNAc (6) apparently helpresolve this dilemma and indicate that steric factors could indeed play an important role in determining the overall rates of proteinsaccharide recognition processes. CONCLUSION

In addition to underlining the unique ability of jacalin to distinguish between a- and p-anomers of saccharides, it has been possible to highlight several salient features of its combining region whichexhibit notonly subsite A and B for sugar binding, but also a hydrophobic site for accommodating the Meumb moiety. The rateconstantsestimated from these studies are two to three orders smaller than those expected from diffusion-controlled processes, due to a significant entropic barrier for the formation of jacalin-sugar complexes, and suggest that the steric factors control jacalin-saccharide association to a large extent, Acknowledgments-We thank Dr. M. I. Khan for helpful discussions during the preparation of the manuscript. We thank Dr. V. K. Mishra for his help in recording the fluorescence spectra and Professor P. Balram for extending to us the use of a Perkin-Elmer-44A fluorescence spectrophotometer. The assistance of K. Radhakrishnan in typing the manuscript is also acknowledged. REFERENCES 1. Bunn-Moreno, M. M., and Campos-Neto, A. (1981) J. Immunol. 127,427-429 2. Roque-Barreira,M. C., and Campos-Neto, A. (1985) J.Immunol. 134, 1740-1743 3. Kondoh, H., Kobacgshi, K., Hagiwara, K., and Kajii, T. (1986) J. Immunol. Methods. 88, 171-173

8918

Thermodynamic and Kinetic Studies on Jacalin

4. Payne, N. R., Concepcion, N. F., and Anthony, B. F. (1990) Infect. Immun. 58,3663-3670 5. Sastry, M.V. K., Banerjee, P., Patanjali, S. R., Swamy, M. J., Swarnalatha, G. V., and Surolia, A. (1986)J. Biol. Chem. 261, 11726-11733 6. Sastry, M. V. K., Swamy, M. J., and Surolia, A. (1988)J. Biol. Chem. 263,14826-14831 7. Mahanta, S. K., Sastry, M. V. K., and Surolia, A. (1990)Biochem. J. 286,831-840 8. DeBoeck, H., Matta, K.L., Claeyssens, M., Sharon, N., and Loontiens, F. G. (1983)Eur. J. Biochem. 131,453-460 9. Flowers, H. M., and Shapiro, D. (1965)J. Org. Chem. 30, 20412043 10. Neuberger, A., and Wilson, B. M. (1971)Carbohydr. Res. 17,8195 11. Loontiens, F. G., Clegg, R. M., and Jovin, T. M. (1977)Biochemi s t 16,159-166 ~ 12. Appukuttan, P. S., Surolia, A., and Bachhawat, B.K. (1977) Indian J. Biochem. Biophys. 14,382-384 13. Scatchard, G. (1949)Ann. N. Y. Acad. Sci. 51, 660-672

14. Chipman, D.M., Grisario, V., andSharon, N. (1967)J. Biol. Chem. 242,4388-4394 15. Bessler, W., and Goldstein, I. J. (1974)Arch. Bioehem. Biophys. 166,444-445 16. Banerjee, R., Dhanaraj, V., Mahanta, S. K., Surolia, A., and Vijyan, M. (1991)J. Mol. BWL, in press 17. Jencks, W. P. (1981)Proc. Natl. Acad. Sci. U. S. A. 78, 40464050 18. Cramer, F., Saenger, W., and Spatz, H.C. (1967)J. Am. Chem. SOC.89, 14-20 19. Acharya, S., Patanjali, S. R., Sajjan, S. V., Gopalkrishnan, B., and Surolia, A. (1990)J. Biol. Chem. 266, 11586-11594 20. Farina, R.D., and Wilkins, R. G. (1980)Biochim. Biophys. Acta. 631,428-438 21. Swamy, M. J., Sastry, M.V.K., Khan, M. I., and Surolia, A. (1986)Biochem. J. 234,515-522 22. Khan, M. I., Sastry, M. V. K., and Surolia, A. (1986)J. Biol. Chem. 261,3013-3019 23. Lewis. S. D.. Shafer. J. A.. and Goldstein. I. J. (1976) . , Arch. Biochem. Biophys. 172,689-695 24. Clegg, R. M., Loontiens, F. G., and Jovin, T. M. (1977)Biochemi s t 16,167-175 ~