Shreeta Acharya, S. R. PatanjaliS, S. Umadevi Sajjanp, B. GopalakrishnanS, and Avadhesha Surolia

Transcription

1 THE JOURNAL OF BIOLOGICAL CHEMISTRY by The American Society for Biochemistry and Molecular Biology, Inc. Vol. 265, No. 20, Issue of July 15, pp , 1990 Printed in U.S.A. Thermodynamic Analysis of Ligand Binding to Winged Bean (Psophocarpus tetragonolobus) Acidic Agglutinin Reveals Its Specificity for Terminally Monofucosylated H-reactive Sugars* (Received for publication, September 25, 1989) Shreeta Acharya, S. R. PatanjaliS, S. Umadevi Sajjanp, B. GopalakrishnanS, and Avadhesha Surolia From the Molecular Biophysics Unit, Indian Institute of Science, Bangalore , Zndia The sugar-specific binding of N-dansylgalactosamine to WBA II (n = 2; K. = 5.6 X lo3 M-l; AH = -21 kj*mol- ; AS = J*mol- *K-l) was utilized in substitution titrations for evaluating the association constants for the interaction of sugars with the lectin. An axial hydroxyl at C-4 and equatorial hydroxyls at C-3 and C-6 as in D-galacto configuration are crucial for binding. Both axial and equatorial hydroxyls are tolerated at C-2. Conformationally akin disaccharides such as lactose, N-acetyllactosamine, Gala l- 3GlcNAc, and Gal@ 1-3GalNAc show similar affinities. 2 -Fucosyllactose and H-disaccharide display 146 and 13 times stronger affinity over lactose and galactose, yet fucose by itself is devoid of activity. An interesting feature, noted for the first time, in protein-sugar interactions is the positive entropy change for the binding of 2 -fucosyllactose, suggesting that nonpolar interactions play an important role in stabilization of the lectin-sugar complex. 3-Fucosyllactose, lactodifucotetraose, lacto-n-fucopentaose II and III are inactive, whereas lacto-n-fucopentaose I has 14-fold lower affinity as compared with 2 -fucosyllactose. Conformational analysis indicates that the substitution at subterminal glucose or GlcNAc by L-fucose in either a 1-3 or al-4 linkage leads to its projection so as to sterically hinder the access of 3 -fucosyllactose, lactodifucotetraose, and lacto-n-fucopentaose II and III to the binding site of winged bean agglutinin II. Similarly the projection of al-3 linked Gal/GalNAc also leads to steric hindrance and hence prevents the binding of blood group A and B reactive sugars. Considering its unique specificity winged bean agglutinin II should be useful in the isolation and characterization of terminally monofucosylated H-reactive oligosaccharides from those that are difucosylated or internally fucosylated. Lectins are a class of multivalent carbohydrate binding proteins that are generally assayed as hemagglutinins (1). Because of the high degree of specificity displayed by individual lectins in their recognition of glycoconjugates, they are being employed as highly discriminating probes in studies of * This work was supported by a Science and Engineering Research Council 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 18 U.S.C. Section 1734 solely to indicate this fact. $ Research Associate in a project funded by the Department of Science and Technology, India. Research Associate at the Centre for Advanced Study funded by the University Grants Commission, India. membranes of normal and transformed cells, in blood typing, in the purification and characterization of carbohydrate containing biopolymers, and in studies of lymphocyte mitogenesis (1, 2). The sugar binding site of lectins is therefore a subject of intense scrutiny. The winged bean, Psophocarpw tetragono.!ogu.s, contains two lectins (WBA I and II) with almost identical molecular masses and number of subunits but with differing isoelectric points and hemagglutinating activities (3, 4). Winged bean basic lectin, WBA I, has p1 > 9.5 and agglutinates human type A and B but not type 0 red blood cells (4). The carbohydrate binding specificity of WBA I has been studied in considerable detail (5,6). Winged bean acidic lectin, WBA II, has pi = 5.5 and agglutinates trypsinized and desialized erythrocytes of all three (A, B, and 0) types (4, 7). D-Galactose was reported earlier to be a better inhibitor than N-acetylgalactosamine for WBA II (8). Kortt (4) subsequently reported that N-acetylgalactosamine was a better inhibitor than galactose. Utilizing lectins of well characterized specificities as competitors for the binding of WBA II to human erythrocytes, we suggested that the lectin binds to H and T antigenic determinants (and related structures) on the cell surface (7). However, unlike the anti-h lectins from Lotus tetragonolobus and Ulex europeus, binding of WBA II to erythrocytes was not inhibited by L-fucose (7, 9, 10). Since hemagglutination inhibition of lectins by sugars at its best provides information about relative affinities of ligands, a more quantitative and precise evaluation of the strengths of such interactions is imperative for elucidation of the specificity of WBA II. Moreover, as the oligosaccharide chains present on the surface of cells exhibit a wide variety of heterogeneous structures, only limited conclusions can be drawn regarding the specificities of lectins from cell binding studies and indeed the biphasic nature of the binding of WBA II to human erythrocyte precluded a quantitative evaluation of its interaction with carbohydrate ligands (7). Thermodynamic and kinetic analyses are invaluable for understanding the specificities of lectins, the nature, and the types of forces involved in these interactions in addition to their relevance for a better evaluation of their interaction with cells. Thermodynamic analysis for the binding of WBA II to a large number of natural and synthetic carbohydrate determinants was therefore carried out to answer the following questions. WBA II is inhibited by several galactose-containing disaccharides such as lactose, N-acetyllactosamine, etc., yet the bind- 1 The abbreviations used are: WBA I, winged bean agglutinin (basic): WBA II, winged bean agglutinin (acidic); Gal, galactose; Glc, glucose; Fuc, L-fucose; MeaGaca-methyl galactopyranoside; dansyl (Dns). 5-dimethvlaminonauhthalene-1-sulfonyl; DnsGalN, dansyl- &lac&samine; GkNAc, 2-acetamido-2-deoxygaladopyranose; GlcNAc, 2-acetamido-2-deoxyglucopyranose; LacNAc, N-acetyllactosamine; LNF, lacto-n-fucopentaose; Ph, phenyl. All sugars used are D-sugars unless otherwise specified

2 Specificity of WBA II for Monofucosylated H-reactive Sugars ing of the lectin to erythrocytes is inhibitable by blood group H-specific lectins. Is inhibition a consequence of the overlapping determinants between WBA II and Ulex and Lotus lectins, or do they compete for binding to the same structure? L-Fucose does not inhibit WBA II-erythrocyte interaction, yet, does it form part of the determinant recognized by the lectin? If L-fucose forms part of the determinant, what kind of fucosylated structures does the lectin bind to? What is the topography of the combining site of WBA II? Whether WBA II has an extended combining region complementary to two or more hexapyranosyl residues? In addition, the elementary steps involved in the association of WBA II with DnsGalN as elucidated by stopped-flow spectrofluorimetry are also reported. MATERIALS AND METHODS Sugurs-GalNAc, Me&Gal, MefiGal, 2-deoxygalactose, L-fucose, galactosamine, galactose, L-arabinose, glucose, GlcNAc, and &dimethylaminonaphthalene-l-sulfonic acid were products of Sigma. Lactose was from British Drug House Chemicals, United Kingdom. Methyl-a-N-acetylgalactosamine was prepared as reported by Sarkar and Kabat (11). Galul-3GalaMe, GalBl-BGalpMe, and Galal-4Gal were purchased from Carbohydrate International, Sweden. Galal- 3GlcNAc and Galfil-3GalNAc (T-antigen) were prepared according to the procedures described in Refs. 12 and 13, respectively. H- disaccharide, 2 -fucosyllactose, A-trisaccharide, and B-trisaccharide were obtained from BioCarb Chemicals, Sweden. 2 -Fucosyllactose was also a kind gift from Drs. V. Ginsburg and D. D. Roberts, National Institutes of Health, Bethesda, MD. 3-Fucosyllactose and difucosyllactose were the kind gifts of Prof. E. A. Kabat, Columbia University, New York. LNF I, LNF II, and a mixture of LNF II and III were gifted by Prof. C. A. Bush, Illinois Institute of Technology, and Prof. C. Gahmberg, University of Helsinki, Finland. Prof. Bush also provided monofucosyllacto-n-hexaose. Subsequently LNF I, LNF II, LNF II, 3-fucosyllactose, and difucosyllactose were also obtained from BioCarb, Sweden. DnsGalN was available from a previous study (5, 13) and was prepared according to the method of Lartey and Derechin (14). Fuccul-PGal/31-3GalNAcaPh and Fucorl-3GalaMe were generous gifts from Dr. K. L. Matta, Rosewell Park Memorial Institute, New York. Purification of WBA II-Isolation of WBA II was carried out according to Patanjali et al. (7). The affinity purified lectin was found to be electrophoretically homogeneous. The isoelectric focussing pattern of the lectin was similar to that reported by Kortt (4). Fluorescence Measurements-Fluorescence emission spectra were recorded on a Perkin-Elmer MPF 44A spectrophotometer. Titration experiments were performed on a Union Giken FS 501 fluorescence polarizer equipped with photon counting photomultipliers essentially as described previously (5). The fluorescence intensity of DnsGalN was enhanced 2.5-fold at infinite concentration of WBA II. The association constants, K., at various temperatures were determined graphically according to the relationship (15), F, - Fo log - = F, - F, log K. + log [PI, where F,, F,, and F, are the fluorescence intensities at zero protein concentration, at a particular concentration of protein and at infinite concentration of protein. [PI, is the free protein concentration. The binding parameters were also evaluated according to the method of Scatchard (16) by titrating a fixed concentration of lectin (9.78 pm) with increasing concentration of DansGalN (2-450 FM). The change in fluorescence was recorded after equilibrating the samples for 8 h. Binding of nonfluorescent inhibitory sugars was studied by monitoring the decrease in dansyl fluorescence as it is released from its complex with the protein upon addition of the competing sugar (Fig. 3). Association constants for the ligands were determined from plots obtained from the following expression (5, 17). [ pt I pd - 1 loif = K* 5 [S]f + $ Thermodynamic parameters were calculated using the van t Hoff plots. Conformational Energy Calculations-For conformational energy calculations the Hard Sphere Exoanomeric Effect method was used (18). Pyranoside rings of D-SugarS were taken to be in the %I (D) chair, whereas that of L-fucose in icq (L) chair and the atomic coordinates for constituent monosaccharide units were taken from the crystallographic data of Longchambon et al. (19) and MO and Jenson (54). A value of 117 was set for the glycosidic bond angle. Kinetic Studies-The kinetics of interaction of WBA II and DnsGalN was monitored by a Union Giken RA401 stopped-flow spectrofluorimeter. Samples were excited at 300 nm, and the emission was monitored beyond 460 nm with the use of a cutoff filter. The dead time of the instrument was determined to be 0.5 ms. The reacting solutions were maintained at a constant temperature (+O.l C) by a Lauda circulating water bath. Experiments were carried out using pm DnsGalN and PM lectin (resultant concentrations after mixing). The dissociation rate constants of WBA II- DnsGalN interaction were determined by dissociating the complex with a 10 mm solution of an inhibitory sugar, MecuGal. Determination of Concentration-Protein concentrations were determined by Lowry s method (20). The concentration of DnsGalN was determined from its molar absorptivity of 4800 cm- at 330 nm (21). Concentrations of other sugars used for titration were determined by weight measurements. RESULTS Equilibrium Studies-The fluorescence intensity of Dns- GalN was enhanced by 2.5fold upon complete saturation of the fluorescent sugar by WBA II (Fig. 1, a and b). Addition of an inhibitory saccharide-like GalNAc totally reversed this effect. Addition of the lectin preincubated with an inhibitory sugar (N-acetylgalactosamine, 2 mm) to DnsGalN did not result in any change in the fluorescence intensity (22,23). The Scatchard plot for the binding of DnsGalN to WBA II shown in Fig. 2, gives a value of K, equal to 4.1 x lo3 M- and n equal to 1.92 for this interaction. The stoichiometry and the values of K, obtained by the method of Scatchard and the % 60.r VI (a) E E a E 40 z 0) k 0 2 f 3 m 2 20 id (ii) h (nm) Protein aliquot added@) Lo9IP1, FIG. 1. a, Fluorescence emission spectra of DnsGalN in the presence and absence of WBA II. The samples in 1 X 1 x 4.5cm cuvettes at 20 C were excited at 330 nm, and the emission spectra were collected above 450 nm. The fluorescence of 12.5 pm DnsGalN (ii) enhanced considerably in the presence of 100 pm WBA II (i). b, titration of DnsGalN with WBA II at 20 C. A 12.5 FM solution of DnsGalN (2.0 ml) was titrated with increasing amounts of WBA II. c, a graphical representation for the determination of association constants.

3 11588 Specificity of WBA II for Monofucosylated H-reactive Sugars FIG. 2. Scatchard plot for the binding of DnsGalN to WBA II by fluorimetric titration at 25 C. A fixed concentration of WBA II (9.78 pm in dimer) was titrated with pm DnsGalN. On comparison of the fluorescence of equal concentration of pure ligand with that of the mixture, the amount of ligand bound and the degree of saturation (r) was determined. The slope, K. = 4104 M- and the x intercept n = 1.92 gave the number of binding sites/dimer of M, 54,000. TABLE Association constants and free energy changes observed for the binding of mono- and disaccharides to WBA II at 20 C Galactose MecuGal MePGal GalNAc MeolGalNAc DnsGalN Galactosamine L-Fucose D-Fucose L-Arabinose Lactose Lactulose Thiodigalactoside GalBl-4GlcNAc Gal/31-3GlcNAc Gal/31-3GalNAc Galpl-3GalpMe Galoll-3GalolMe Galotl-4Gal Melihiose n NB, no binding. Ligand K AG (LacNAc) M I r NB 31.2 NB kj. mole Relative affinity method of Chipman (Table I) are in good agreement, indicating the bivalent nature of the lectin (Mr 54,000). The linearity of the Scatchard plot shows that there is no interaction between the binding sites. Association constants for the interaction of various ligands with WBA II as calculated by competitive binding experiments (Fig. 3) are listed in Table I. Galactose was used as the reference sugar for computation of the relative affinities and the AG (Tables I and II). The following sugars failed to bind to the lectin when tested up to the concentrations indicated; L-fucose (200 mm), Man (200 mm), and Glc (200 mm). Among the unsubstituted monosaccharides N-acetylgalactosamine is the most complementary ligand. MecvGal is marginally better a ligand over MepGal and MeLvGalNAc is the most comple- mentary monosaccharide. Despite the preference of WBA II towards MetvGal over MepGal, disaccharides containing galactose in P-linkage were more active than those with galactose in a-linkage. In the p-linked disaccharides the lectin strongly prefers Galal-3GlcNAc over other structurally and conformationally related disaccharides such as lactose, Gal/? l- 3GalNAc, and N-acetyllactosamine. It was interesting to note that despite the failure of L-fucose to bind even at very high concentrations the lectin binds strongly when this sugar is in al-2 linkage with galactose and lactose as in H-disaccharide and 2 -fucosyllactose, respectively (Table III). Thus 2-fucosylgalactose and 2 -fucosyllactose are 13 and 146 times better inhibitors than galactose and lactose, respectively. Moreover, both of them were infinitely more active over L-fucose. Analogues of 2 -fucosyllactose, such as 3-fucosyllactose and lactodifucotetraose, were completely inactive. Steric hindrance by L-fucose linked to 3-hydroxyl group of penultimate glucose as in 3-fucosyllactose and lactodifucotetraose is presumably responsible for the failure of these ligands to bind to the lectin. Although the disaccharide Gal/31-3GlcNAc which constitutes the core of type I chains is more active as compared with lactose, LNF I (L-Fuccu1-2Ga1~1-3G1cNAc/31-3Ga1~1-4Glc) is 14 times weaker a ligand as compared with 2 - fucosyllactose, which constitutes the determinant of Type II chain. Hence, it can be concluded that WBA II prefers type II chains than type I chains of H-antigenic determinant. Likewise, Fuccul-2Gal/31-3GalNAccuPh, type III H-reactive trisaccharide, is 18 times weaker as compared with 2 -fucosyllactose. Oligosaccharides related to LNF I containing a fucosyl residue linked to GlcNAc of type I and type II chains, viz. LNF II and LNF III, respectively, failed to bind to the lectin altogether. If an N-acetylgalactosaminyl residue is attached in cy-linkage to the C-3 of the terminal nonreducing p-linked galactose of fucosyllactose, as in blood group A-tetrasaccharide, or to C-3 of galactose of H-disaccharide (L-Fuccul-2Gal) as in blood group A-trisaccharide, the binding to WBA II is abolished. Likewise blood group B-trisaccharide and A-pentasaccharide are also noninhibitory. The low affinity of monofucosyllacto-n-hexaose as compared with Galpl-3GlcNAc is probably due to conformational restrictions imposed by monofucosylated structures of the /3 l- 6 branch. Kinetic Studies-The relationship between kobs for the change in fluorescence of DnsGalN and the concentration of the reactants depends on the type of elementary step to which the relaxation corresponds. For reactions where the ligand is not transformed covalently the three most likely possibilities are considered below. The first case where the fluorescence change is related to the equilibration of D and P to form the following PD complex. P+D&PD (1) k-1 if [PI0 % [D],, where [PI,, and [D], are the initial concentrations of P and D. The observed first order rate constant for the reaction is given by kobs = k-1 + kl [Pill (2) The second example is that of a rapid pre-equilibration between P and D forming the initial complex PDi, followed by an isomerization of PDi to give the final complex PD* for which P+D &PDi+P~* (3) k-1 2

4 Specificity of WBA II for Monofucosylated H-reactive Sugars TABLE Association constants and free energy changes for the interaction. of fucosylated sugars with WBA II at 20 C Binding studies were carried out as reported in Ref. 7. Ligand Fuccul-2Gal (H-disaccharide) Galpl-4Glc I (l-2) Fuccv (2 -fucosyllactose) Galpl-4Glc l(l-31 Fuca (3-fucosyllactosel GalpI-4Glc 1(1-Z) l(l-3) Fuca Fucrv (difucosyllactose) Gal~1-3GlcNac~l-3Gal~l-4Glc l(l-2) Fucn (LNF I) Gal~1-3GlcNAc(31-3Gal~l-4Glc l(l-4) Fucn (LNF II) Gal41-4GlcNAcfll-3GalBl-4Glc I (I-3) Fuca (LNF III) Fuca l(l-3) Galpl-4GlcNAcfl I (l-6) GalpI-4Glc l(l-31 Galpl-3GlcNAcp (monofucosyllacto-n-hexaose) Fucal-3GalpMe Fucocl-2Gal41-3GalNAcolPh Relative affinity with galactose as in Ref. 7. b NB, no binding. Sugar aliquot added (~1) FIG. 3. Competitive binding of GalNAc to WBA II in the presence of DnsGalN. WBA II (88 FM) was equilibrated with DnsGalN (12.51~) and the mixture titrated with GalNAc (50 mm). In the inset, the y axis intercept gave the value for DnsGalN K. = 6200 M-l. The slope yielded value of K, = 1294 M- for GalNAc. where kobs = k-2 + kz K-, =?. [PI K-1 + [PI, In the third example the protein exists in two states undergoing a slow transformation form P to P* prior to binding of D, P$P*+D&P*D k-1 k-2 (5) II K. AG Relative affinity M kj. mole 3,200.O NB (at 25 mm) NB (at 25 mm) 2, NB (at 25 mm) NB (at 25 mm) NB (at 2 mm) for which kobs = k, + k-1 K-z K-2 + [PO] Km,=+. In each of these three mechanisms, kobs exhibits a characteristic dependence on [PI0 which can be used to distinguish between them experimentally. Equation 2 predicts that kobs will increase linearly with [PlO. Equation 4 predicts that kobs will increase with increase in [PI,, but tends to saturate when [PI0 >> k+. According to Equation 6 k,,bs will decrease as [PI0 increases. A representative plot of the time-dependent change in fluorescence of DnsGalN upon its rapid mixing with WBA II is shown in Fig. 4. Pseudo first order kinetic conditions were maintained throughout the course of the experiment by keeping the concentration of WBA II at least lo-fold higher than DnsGalN. In order to determine the K-i value more precisely and directly, the fluorescence intensity changes of the DnsGalN-WBA II complex upon binding to MeolGa were monitored kinetically (Fig. 5). The following schemes apply for such reactions. PD P+s&Ps gp+d *1 In such cases if &[S] is much greater than kl[d] the observed rate constant refers to the K-i that is kobs = k-,. The rate constant kabs was evaluated from the slopes of linear plots (6)

5 11590 Specificity of WBA II for Monofucosylated H-reactive Sugars I 1 P 4.0," G; 3.5 m Time (msrc) Time (msec) FIG. 4. A stopped-flow fluorescence trace of DnsGalN-WBA II association reaction at 25 C. DnsGalN (10 FM) and WBA II (250 pm) mixed in equal volumes in the instrument. The samples were excited at 330 nm and the emission recorded above 460 nm with a cutoff fiiter. The inset represents a typical plot used for calculating kbs- ; ILK.E b SO 120 TimeCmsec) I Time (msec) FIG. 5. Estimation of dissociation rate constant for Dns- GalN-WBA II interaction at 15 C! using MeaGal as the competing sugar. 500 pm WBA II and 20 pm DnsGalN were mixed with an equal volume of 10 mm MeolGal. The slope of the line in the inset gave the value of k-, = s-l. of ln( AF, - AF, ) versus time (t), where Ft and F, are the fluorescence at time t and at the end of the reaction respectively. kl and k-1 are obtained from the slope and the ordinate interce$ of the linear plots of lzobs versus [PI. A linearity of kob. on protein concentration indicates that Equation 2 can be used for obtaining kl and k-1. The kl and kel values determined by these methods at 20 C are 1.81 X 10 M-l s-l and s-, respectively. Slow second order rate constants for the binding of other lectins have also been reported in the past (5,13, and 24-27). The kinetic experiments were carried out at 15, 20, and 25 C in order to obtain the activation parameters. From the Arrhenius plots for such interactions the energy of activation (&) and subsequent thermodynamic parameters for each of the kinetic steps are derived. The activation parameters were calculated using the following equations, AH*=EA-RT ln(k/t) = -AH*/RT + AS*)/T + ln(k /h) AG = AH* - TAS* a where k is the appropriate rate constant, k is the Boltzman constant, and h is Planck s constant. These values are listed in Table IV. DISCUSSION In this report we have studied the binding of a variety of sugars to WBA II using DnsGalN as an indicator ligand. Analyses of carbohydrate binding by WBA II provides considerable insight about the spatial features of its combining site and reveal the exquisite specificity of the lectin for recognizing monofucosylated blood group H-related antigenic determinants. Binding of DnsGalN-The fluorescence of DnsGalN upon binding to WBA II is enhanced by 2.5-fold with a concomitant 20-nm blue shift in its emission maxima. Earlier studies on the binding of WBA I (5), soybean agglutinin (21, 27), Erythrina crystagalli lectin (28), and Artocarpus lectin (13) showed 15-, ll-, 5-, and 2-fold increase in the fluorescence intensity of DnsGalN, respectively, suggesting that the binding sites of soybean agglutinin and WBA I are considerably apolar when compared with Artocarpus lectin (13). A marginal increase in DnsGalN fluorescence intensity on binding to WBA II as compared with that observed for WBA I indicates that the dansyl moiety on the second carbon of galactose experiences a relatively polar environment upon binding to WBA II. Complete reversal of this binding with GalNAc indicates that DnsGalN-WBA II interaction is saccharide- specific. The -AH value (21.3 kj. mol- ) for this interaction is in the range of values observed for lectin-monosaccharide interactions. The dimeric lectin (M, = 54,000) is bivalent. Binding of Monosaccharides-WBA II binds three times better to MecuGal than MePGal (Tables I and III) indicating that a methyl group in (Y configuration contributes positively towards the binding. This preference is comparable with that observed for Griffonia simplicifolia isolectin Bq (32). Substitution of NH2 group at C-2 as in GalNHz weakens the binding to WBA II considerably, whereas GalNAc with an N-acetyl group at the same position is five times more potent, presumably due to additional van der Waals interactions or hydrogen bonding between the combining site of WBA II and the acetyl group or potentiation of the existing ones. A dansyl group at this locus, as in DnsGalN, further increases the potency of the ligand. A methyl substitution in (Y configuration of GalNAc converts it into the most potent monosaccharide inhibitor studied. Binding of D-talose indicates that the lectin accommodates an axial hydroxyl at C-2. Inactivity of gulose suggests that at C-3 an equatorial hydroxyl is critical. Failure of glucose and GlcNAc to bind shows the axial hydroxyl at C- 4 to be indispensable. Nonbinding of D-fucose and L-arabinose highlights the importance of C-5 hydroxymethyl group for binding to WBA II. Binding of Disaccharides-WBA II, like peanut lectin, binds better to MeaGal than MePGal but among disaccharides prefers P-linked disaccharides over the a-linked ones. LacNAc, Galpl-3GlcNAc, and T-antigen are better ligands than a-linked sugars like melibiose (29-31, 33). Artocarpus lectin likewise prefers T-antigen over melibiose (13). Among the P-linked disaccharides, Galpl-3GlcNAc has the strongest affinity and lactose the least, whereas GalPl- 3GalNAc, LacNAc, and thiodigalactoside have intermediate affinities. The change in enthalpy (AH) for lactose, Galal- 3GlcNAc, and LacNAc is greater by 24, 10, and 12 kj. mol-, respectively, over that for MepGal, indicating that the second pyranosyl residues of these disaccharides is bound in a site adjacent to the galactose binding subsite, i.e. the lectin has an extended binding site.

6 Specificity of WBA II for Monofucosylated H-reactive Sugars TABLE III Association constants and thermodynamic parameters for the binding of sugars to WBA II 10-Z x K. Ligand AC AH AS 15 C 20 C 25 C M-1 kj. mol- J.mot.K- Galactose MeaGal MepGal GalNAc MePGalNAc DnsGalN GalBl-4Glc Thiodigalactoside Galpl-4GlcNAc H-disaccharide Fucosyllactose C FIG. 6. Ball and stick plots (ORTEP) of disaccharides: a, GalBl-4Glcg; b, Galfil-4GlcNAq9; c, Galal-3GalNAqS; d, Gal@l-3GlcNAcB. The torsional angles are defined according to the IUPAC-IUB convention = O(ring)-Cl-01-C(i), $ = Cl- 01-C(i)-C(i - 1). Dihedral angles sugars are 4 = -5O, # = 110 and for the pl-a-linked sugars = -7O, # = WBA II appears to recognize these disaccharides in a similar fashion which can be understood better by examining their ball and stick plots (13,33-35) (Fig. 6). The pl-4-linked disaccharides, in contrast to /?l-3-linked compounds, have their reducing sugar residue rotated with respect to the nonreducing one in such a way that its substituents at C-2 atom are on opposite sides (13, 33-35). Despite this significant conformational difference between fil-4- and p l-3-linked sugars, these compounds possess striking conformational similarity to be sterically accommodated in the binding site of the lectin. However, the hydroxyl group at C-4 of the reducing sugar of Gal@l-3GalNAc is axially oriented as opposed to an equatorial orientation of the hydroxyl group attached to the spatially related carbon atom uiz. C-3 of the reducing sugar of lactose and LacNAc and C-4 of the reducing moiety in the case of GalPI-3GlcNAc. Since WBA II, unlike Artocarpus lectin (13), does not distinguish strongly between these disaccharides, an axial orientation of the hydroxyl group attached to C-4/C-3, as the case may be, of their reducing sugar moieties is not crucial for binding. Relatively stronger binding of Galpl-SGlcNAc, Gal@l-3GalNAc, and LacNAc over lactose is presumably due to the favorable entropic contribution made by the C-2 acetamido group of their reducing sugar moiety. Thiodigalactoside shows lesser -AH value than lactose yet interacts better than the former, presumably due to an increased statistical probability of binding through any of its galactopyranosyl residues. Binding of Fucosylated Oligosaccharides-Among the var- d TABLE IV Rate constants and activation parameters for the interaction of WBA II with DnsGalN Dissociation rate constants were measured directly by the inhibition of WBA II-DnsGalN interaction with MeotGal. Temperature lo- x 1O-3 x K. IO-~ x K. k,, k-, (kinetics) (equilibrium) C M.s- s-1 M M AH,, = kj.mol- AH-, = kj. mol- AS,, = J. mol-. K- AS-, = J. mol-. K- AG,, = 42.9 kj.mol- AG-, = kj.mol- E,, = kj. mol- E-, = kj-mol- ti = AH+l - AH-, = kj.mol- AS = AS,, - AS-] = J.mol-l.K- AGO = AG,, - AC-1 = kj.mol- EA = E+, - E-1 = kj.mol- ious fucosylated saccharides tested, 2 -fucosyllactose was found to be the best ligand for WBA II. Both the monosaccharides attached to the P-linked galactose, L-fucose, and glucose, participate in binding as removal of either of these sugars from 2 -fucosyllactose substantially reduces the binding ability of the resulting disaccharide. Thus, removal of reducing terminal glucose and terminal L-fucose from 2 - fucosyllactose lowers the affinities of the resulting disaccharides by 13- and 146-fold, respectively, when compared with the parent compound. This shows that the contribution of L- fucose to the binding energy is two times higher than that of reducing end glucose. The H-disaccharide and 2 -fucosyllactose show an increase in entropy and enthalpy over galactose and lactose (Table V). A relatively positive change in AH and AS has been reported earlier for the interaction of lima bean lectin with A-trisaccharide (36). But the positive value of AS for the binding of 2 -fucosyllactose has been noted for first time in lectin-sugar interactions. It is therefore appropriate to draw conclusions based on other studies (37). A positive increase in the values of AS and AH upon fucosylation of galactose and lactose indicates that fucosylation enhances nonpolar interactions between the lectin and these compounds. In other words the fucose residue of H-disaccharide and fucosyllactose probably encounters a hydrophobic locus in the combining region of the lectin. From the overall values of thermodynamic parameters for the association of 2 -fucosyllactose with the lectin, it is obvious that the major contributing factor in the stabilization

7 11592 Specificity of WBA II for Monofucosylated H-reactive Sugars TABLE V Differences in AH and AS values between H-disaccharide and galactose, 2 -fucosyllactose, and lactose sugar AG AH AS AAG AAH AAS kj. t7wt' J. mole. K- kj. mot J.mol-.K- Galactose Fucosylgalactose Lactose P -Fucosyllactose FIG. 7. ORTEP plots of the following fucosylated sugars along with the values of torsional angles. a, 2 -fucosyllactose: I#JFUC = -8O, $Fuc = 140 ; 4Gal= -6O, J/Gal = 110 ; b, 3- fucosyllactose: $Fuc = -69, $Fuc = -94 = -63, +Gal = 128. c, = $Fuc = -94 = -63, $Gal= -69, 128 ; $Fucoll-2 = -8O, $Fuc = 140. d, A- tetrasaccharide: $GalNAc = 56, $GalNAc = -176 = -63, $Gal = 128 = -66, $Fuc = 130. of the complex is entropic, rather than enthalpic, in origin. This large positive value of AS suggests that hydrophobic interactions play an important role in stabilizing the 2 - fucosyllactose-wba II complex (38). That polar interactions also contribute toward the stability of the complex is evident from the negative value of AH. LNF I which constitutes the type I H-reactive antigen is weaker a ligand as compared to 2 -fucosyllactose, the type II H-reactive sugar, thereby indicating that the lectin is able to distinguish between type I and type II H structures. Interestingly though the lectin fails to bind several other fucosylated sugars such as 3-fucosyllactose, difucosyllactose, blood group A-reactive tri-, tetra-, and pentasaccharides, and blood group B-reactive tri- and pentasaccharides, LNF II and LNF III. In order to determine the site of attachment of WBA II to fucosylated oligosaccharides ball and stick plots of the minimum energy conformations of these saccharides as obtained by HSEA calculations were used (6, 39-44). Since 2 -fucosyllactose is the most complementary ligand for WBA II it is in order to discuss the molecular model for this compound first (Fig. 7~). The bottom, top, and proximal regions of this compound are designated as (Y, p, and y sides, respectively (45, 46). Since all the constituent sugars of 2 -fucosyllactose make a contribution to the binding process, the combining site of WBA II encompasses almost all of this trisaccharide. A comparison of the binding propensities and conformations of 2 -fucosyllactose with other fucosylated determinants as outlined below suggests that all the three surfaces, viz. LY, /3, and y sides of this compound, interact with the lectin. One C can readily see that in 3-fucosyllactose (Fig. 7b) L-fucose in (~1-3 linkage to the reducing glucose masks substantial area on the p side of both the terminal /3-galactosyl group and the reducing end glucose residue. Thus important regions of the determinants necessary for binding to WBA II are no longer accessible. From the foregoing discussion the important role played by the p side of 2 -fucosyllactose for binding to WBA II is obvious. In difucosyllactose, (Fig. 7c), the cul-3-linked fucose covers, in addition to p-galactosyl and the reducing glucosyl moiety, parts of the cul-2-linked L-fncosyl residue, thereby sterically blocking the access of difucosyllactose to the combining site of WBA II. The terminal cyl-3-linked GalNAc and Gal in blood group A and B saccharides, respectively, are in the plane of the oligosaccharide projecting on the y side of 2 -fucosyllactose (Fig. 7d). Such a projection of Gal and GalNAc on the y side of 2 -fucosyllactose would sterically prevent the binding of the saccharides as observed experimentally with blood group A and B saccharides. These results highlight the important role played by the y side of 2 -fucosyllactose in the binding process. Projection of L- fucose in al-3 linkage to galactose on the y side likewise explains the nonbinding of L-fucal-3Gal structure. Relatively poor binding of LNF I (Fig. 8~) as compared with 2 -fucosyllactose is presumably due to the projection of the acetamido group of its penultimate GlcNAc on the (Y side which occupies a considerable portion below (~1-2 fucosyl unit. This perhaps reduces favorable contacts of the fucosyl residue within the combining site. This result indicates that the lectin binds to at least a portion of determinants on the LY side. Nonbinding

8 Specificity of WBA II for Monofucosyhted H-reactive Sugars a b FIG. 8. Conformational models of the following fucosylated pentasaccharides using the program ORTEP. a, LNF I: +Fuc = -68, #FIX = 130 = -59, #Gal = -108 ; 4GlcNAc = -64, $GlcNAc = -126 = -64, $Gal = 122. b, LNF II: &Glc = -64, $Glc = 122 ; +Galfi3-I = -64, $Gal = -126 ; I#JFUC = 66, $Fuc = -105 ; $GlcNA@3-1 = -67, y GlcNAc = 136 c, LNF III: &Glc = -64, $Glc = 122 = -64, llga1 = -127 ;,$~GlcNAc/34-1= -66, $GlcNAc = 129 ; cpfuc = -68, $Fuc = -95. of LNF II and LNF III is obviously related to the shielding of a substantial portion of the terminal P-galactosyl and subterminal fi-glcnac residues by L-fucose in al-3 and al-4 linkage, respectively, on the /l side of these compounds (Fig. 8, b and c). Thus the analysis of ligand affinities together with their conformation indicates that WBA II binds to 2 - fucosyllactose on its top and bottom surfaces, as well as to the y side. Based on these results the topography of the combining region of WBA II uis ti uis 2 -fucosyllactose is depicted in Fig. 9. Since L-fucose in al-2 linkage to the p- galactosyl residue makes a major contribution to the binding process as compared with the reducing glucosyl residue, the continuous line is meant to indicate that FUCLV~-2Gal portion is buried in a deep cleft in the combining region of WBA II. The broken line indicates that the reducing glucose unit interacts with a shallow region in the combining site. A comparison of the specificities of WBA II with the well characterized fucose binding lectins from U. europeus I (9), L. tetragonolobus (lo), G. simplicifolia isolectin IV (47), E. europeus (47, 48), Aleuria aurantia (49, 50), and Anguilla anguilla (51) sheds some light on the unique carbohydrate recognition activity of the lectin. Lotus, Ulex, Aleuria, and Anguilla lectins bind to L-fucose itself reasonably well in contrast to WBA II, Aleuria lectin does not distinguish appreciably between 2 - and 3-fucosyllactose as well as between FIG. 9. Stereodiagram of the conformational model of 2 - fucosyllactose in the combining region of WBA II. As compared with Fig. 7a, this model has been given a rotation of +45 along the x axis. The continuous line indicates the deep cleft in which the Fucol-2Galp unit resides, whereas the dashed line outlines the shallow region encompassing the Glc residue in the binding site. LNF I and LNF II and binds reasonably well to LNF III (48, 49). WBA II, unlike Aleuria lectin, fails to bind 3-fucosyllactose and LNF II and III. Thus WBA II should be useful in purifying and characterizing glycoconjugates with terminal fucosylation from those fucosylated in the core region. Lotus and Ulex lectins bind to blood group H type II chains strongly and fail to recognize type I chains (9, 10). Difucosyllactose, which is one of the best inhibitors for Lotus and Ulex lectins, does not bind to WBA II, which can thus distinguish between terminally fucosylated H-antigenic determinants from their difucosylated counterparts. G. simplicifolia isolectin IV reacted more strongly with Leb and difucosylated H-antigenic oligosaccharides than to Le oligosaccharides and 3-fucosyllactose (46). It did not bind to 2 -fucosyllactose and several other cyl-2fucosylated H-reactive substances as well as to L-fucose. Excepting similarity regarding nonbinding to L-fucose, the specificity of WBA II is exactly opposite to that of G. simplicifolia lectin IV. Euonymus lectin binds to type I and II blood group B tetrasaccharides, and type I and II H-trisaccharides were equally potent (47-49). L-Fucose did not bind to the lectin. In contrast to Euonymus lectin the blood group B-reactive sugars are not recognized by WBA II. Conformational analysis of ligands that are reactive with Ulex and Evonymus lectins have been reported (48, 52). Ulex lectin binds principally from the (Y side of H-reactive oligosaccharide. It fails to bind to type I H chain as a consequence of the shielding of the terminal Lul-2-linked fucose determinant by the acetamido group of the subterminal GlcNAc residue on the cy side of the molecule. Euonymus lectin binds principally from the p side and hence it interacts equally well to type I and type II H determinants. WBA II is able to interact with both 2 -fucosyllactose and LNF I, albeit preferentially with the former one. Its failure to recognize blood group B saccharides in contrast to Euonymus lectin indicates that it binds to blood group H-reactive trisaccharide from the 1y, p, and y side. Kinetic Studies-Kinetics of sugar binding to the lectin is qualitatively consistent with a single step bimolecular reaction mechanism (Equation 1). The possibility of formation of a reaction intermediate within the dead time of the instrument is ruled out as the kinetically determined value of association constant at each temperature is in agreement with values determined from fluorescence titrations, and the fluorescence change observed for both the association and dissociation upon the rapid mixing of DnsGalN with the lectin is equivalent to that observed in equilibrium titrations. Moreover, the value of -AH determined from kinetic measurements is in agreement with the value from equilibrium measurements, indicating that the enthalphy change is related to the total

9 11594 Specificity of WBA II for Monofucosylated H-reactive Sugars binding event and that there does not exist a faster step which contributes significantly to the reaction enthalpy. The linearity of the Arrhenius plot rules out any significant conformational transitions in the lectin molecule within the temperature range studied. The large activation energy is also consistent with the slow reaction rates observed for the association of the sugar with the lectin. The activation parameters (Table IV) for the association and dissociation processes indicate that they are enthalpically controlled, although there is some entropic contribution. The k+, values for WBA II-DnsGalN association are certainly not in the range of values expected for a diffusion-controlled reaction. The low values for second order rate constant suggest that the reaction proceeds through the formation of a rapid but observable reaction intermediate PDi followed by a slower observable isomerization of this complex to form the final complex PD as depicted in Equation 3. Alternatively, the rates of interaction between DnsGalN and WBA II are limited by steric factors. Such steric factors have been demonstrated by Cramer et al. (53) to control the reaction rates involved in the binding of dyes to cyclodextrins. In conclusion these studies highlight the unique ability of WBA II, in contrast to Ulex I and Lotus lectins, to distinguish terminally monofucosylated blood group H oligosaccharides from difucosylated and other fucosylated substances. Conformational analysis, in addition, explains not only the binding activity of various saccharides but also the spatial arrangement of the sugars in the combining region of the lectin. Acknowledgments-We wish to express our gratitude to Dr. C. A. Bush and Dr. C. G. Gahmberg for LNF I, LNF II, and LNF III, Dr. E. A. Kabat for 3-fucosyllactose and difucosyllactose, Dr. V. Ginsburg and Dr. D. D. Roberts for 2 -fucosyllactose. Dr. Bush also provided monofucosyllacto-n-hexaose. We wish to thank Dr. K. L. Matta for Fucal-BGalal-3GalNAcaPh andfucal-3galpme. The ORTEP program used was adapted to HPlOOO/A 700 minicomputer by Dr. C. Ramakrishnan. REFERENCES 1. Lis, H., and Sharon, N. (1986) The Lectin.s: Properties, Functions and Applications in Biology and Medicine, pp , Academic Press Inc., Orlando, FL 2. Lis, H., and Sharon, N. (1986) Annu. Rev. B&hem. 66, Kortt. A. A. (1984) Eur. J. Biochem Kortt; A. A. (1985) Arch. Biochem. Biophis. 236, Kahn, M. I., Sastry, M. V. K., and Surolia, A. (1986) J. Biol. Chem. 261, Matsuda, T., Kabat, E. A., and Surolia, A. (1989) Mol. Zmmunol. 26, Patanjali, S. R., Sajjan, S. U., and Surolia, A. (1988) Biochem. J. 242, Pueppke, S. G. (1979) Biochim. Biophys. Acta. 681, Pereira, M. E. A., and Kabat, E. A. (1974) Biochemistry 13, Pereira, M. E. A., Kisailus, E. C., Gruezo, F. G., and Kabat, E. A. (1978) Arch. Biochem. Biophys. 185, Sarkar, M., and Kabat, E. A. (1979) Carbohydr. Res. 69, Flowers, H. M., and Shapiro, D. (1965) J. Org. Chem. 30, Sastry, M. V. K., Banarjee, P., Pantanjali, S. R., Swamy, M. J., Swarnalatha, G. V., and Surolia, A. (1986) J. Biol. Chem. 261, Lartey, P. A., and Derechin, M. (1979) Prep. Biochem. 9, Chinman. D. M.. Grisaro. V.. and Sharon. N. (1967) J. Biol. &em. i42, Scatchard, G. (1949) Ann. N. Y. Acad. Sci. 61, I 17. Bessler, W., Shaffer, J. A., and Goldstein, I. J. (1974) J. Biol. Chem Bock, K. (1983) Pure Appl. Chem. 55, Longchambon, P. F., Ohannessian, J., Avenel, D., and Neuman, A. (1975) Acta. Crystallogr. B31, Lowry, 0. H., Rosebrough, N. J., Farr, A. L., and Randall, R. J. (1951) J. Biol. Chem. 193, De Boeck, H.. Lis. H.. van Tilbeurzh. H.. Sharon. N.. and Loontiens, F: G. (l984) j. Biol. Chem.-2~9, d Christie, D. J., Alter, G. M., and Magnuson, J. A. (1978) Biochemi.&y17, Khan, M. I., Mazumder, T., Pain, D., Gaur, N., and Surolia, A. (1981) Eur. J. Biochem Gray, R. D., and Glew, R. H. (1973) J. Biol. Chem. 248, van Landschoot, A., Loontiens, F. G., Clegg, R. M., and Jovin, T. M. (1980) Eur. J. Biochem. 83, Lewis, S. D., Shaffer, J. A., and Goldstein, I. J. (1976) Arch. Biochem. Biophys. 172, Swamy, M. J., Krishna Sastry, M. V., Kahn, M. I., and Surolia, A. (1986) Biochem. J. 234, DeBoeck, H., Loontiens, F. G., Lis, H., and Sharon, N. (1984) Arch. Biochem. Biophys. 234, Pereira, M. E. A., Kabat, E. A., Lotan, R., and Sharon, N. (1976) Carbohydr. Res. 51, Neurohr, K. J., Young, N. M., and Mantsch, H. H. (1980) J. Biol. Chem. 255, Lotan, R., Skutelsky, E., Danon, D., and Sharon, N. (1975) J. Biol. Chem Goldstein, I. J., Blake, D. A., Ebisu, S., Williams, T. J., and Murnhv. L. A. (1981) J. Biol. Chem Neurohr, K. J., Bundle, D. R., Young, N. M., and Mantsch, H. H. (1982) Eur. J. Biochem. 123, Longchambon, P. F., Ohannessain, J., Pandraud, H. G., Duchet, D., Jacquinet, J. C., and Sinay, P. (1981) Acta. Crystallogr. Sect. B. Struct. Crystallogr. Cry&. Chem. B37, Hirotsu, K., and Shimadu, A. (1974) Bull. Chem. Sot. Jpn. 47, Roberts, D. D., and Goldstein, I. J. (1984) J. Biol. Chem. 259, Joel, K. S., Phillip, istry26, W. Mi, and Scheraga, H. A. (1987) Biochem- 38. Spohr, U., Hindsgaul, O., and Lemieux, R. U. (1985) Can. J. Chem. 63, Dua, V. K., Kazne, G., Dube, E. V., and Bush, C. A. (1985) J. Chromatogr. 328, Ramachandran, G. N., and Sasisekharan, V. (1968) Adv. Protein Chem. 23, Biswas. M.. and Rae. V. S. R. (1980) Biowlvmers Lemieux, R. U., Bock, K., Delbaere; L. T. J., Koto, S., and Rao, V. S. (1980) Can. J. Chem Thorgensen, H., Lemieux, R. U.; Bock, K., and Meyer, B. (1982) Can J. Chem. 60, Rao, B. N. N., Dua, V. K., and Bush, C. A. (1985) Biopolymers s 45. Kihlberg, J., Hultgren, S. J., Normark, S., and Magnusson, G. (1989) J. Am. Chem. Sot. 111, Petryniak, J., and Goldstein, I. J. (1986) Biochemistry 13, Shibata, S., Goldstein, I. J., and Baker, D. A. (1982) J. Biol. Chem. 257, Petryniak, J., Pereira, M. E. A., and Kabat, E. A. (1977) Arch. Biochm. Biophys. 178, Kochibe, N., and Furukawa, K. (1980) Biochemistry 19, Yamashita, K., Kochibe, N., Ohkura, T., Ueda, I., and Kobata, A. (1986) J. Biol. Chem. 260, Springer, F. G., and Desai, P. (1971) Biochemistry 10, Hindsgaul, O., Norberg, (1982) Carbohydr. Res. T., Le Pendu, J., and Lemieux, R. U. 53. Cramer, F., Saenger, W., and Spatz, H. C. (1967) J. Am. Chem. Sot. 89, MO, F., and Jensen, L. H. (1975) Acta. Crystallogr. B31,

10 Thermodynamic analysis of ligand binding to winged bean (Psophocarpus tetragonolobus) acidic agglutinin reveals its specificity for terminally monofucosylated H-reactive sugars. S Acharya, S R Patanjali, S U Sajjan, B Gopalakrishnan and A Surolia J. Biol. Chem. 1990, 265: Access the most updated version of this article at Alerts: When this article is cited When a correction for this article is posted Click here to choose from all of JBC's alerts This article cites 0 references, 0 of which can be accessed free at