Analysis of Saccharide Binding to Artocarpus integrifolia Lectin Reveals Specific Recognition of T-antigen (&D-Gal( 1+3)~-GalNAc)*

The binding of Artocarpus integrifolia lectin to N- dansylgalactosamine (where dansyl is 5-dimethylami-nonaphthalene-1-sulfonyl) leads to a 100% increase in dansyl fluorescence with a concomitant blue shift in the emission maximum by 10 nm. This binding is car-bohydrate-specific and has an association constant of 1.74 X lo4 M" at 20 "C. The lectin has two binding sites for N-dansylgalactosamine. The values of -AH and -AS for the binding of N-dansylgalactosamine are in the range of values reported for several lectin-mon-osaccharide interactions, indicating an absence of non- polar interaction of the dansyl moiety of the sugar with the combining region of the protein. Dissociation of the bound N-dansylgalactosamine from its complex with the lectin and the consequent change in its fluorescence on addition of nonfluorescent sugars allowed evaluation of the association constant for competing ligands. The thermodynamic parameters for the binding of monosaccharides suggest that the OH groups at C-2, C- 3, C-4, and C-6 in the D-galactose configuration are important loci for interaction with the lectin. The results of stopped flow spectrometry for the binding of N-dansylgalactosamine to the Artocarpus lectin are consistent with a simple

The binding of Artocarpus integrifolia lectin to Ndansylgalactosamine (where dansyl is 5-dimethylaminonaphthalene-1-sulfonyl) leads to a 100% increase in dansyl fluorescence with a concomitant blue shift in the emission maximum by 10 nm. This binding is carbohydrate-specific and has an association constant of 1.74 X lo4 M" at 20 "C. The lectin has two binding sites for N-dansylgalactosamine. The values of -AH and -A S for the binding of N-dansylgalactosamine are in the range of values reported for several lectin-monosaccharide interactions, indicating an absence of nonpolar interaction of the dansyl moiety of the sugar with the combining region of the protein. Dissociation of the bound N-dansylgalactosamine from its complex with the lectin and the consequent change in its fluorescence on addition of nonfluorescent sugars allowed evaluation of the association constant for competing ligands.
The thermodynamic parameters for the binding of monosaccharides suggest that the OH groups at C-2, C-3, C-4, and C-6 in the D-galactose configuration are important loci for interaction with the lectin. The acetamido group at C-2 of 2-acetamido-2-deoxygalactopyranose and a methoxyl group at C-1 of methyl-a-Dgalactopyranoside are presumably also involved in binding through nonpolar and van der Waals' interactions. The T-antigenic disaccharide Gal@1+3GalNAc binds very strongly to the lectin when compared with methyl-8-D-galactopyranoside, the @( 1+3)-linked disaccharides such as Galfil-+3GlcNAc, and the 8(1+4)linked disaccharides, N-acetyllactosamine and lactose. The major stabilizing force for the avid binding of Tantigenic disaccharide appears to be a favorable enthalpic contribution. The combining site of the lectin is, therefore, extended. These data taken together suggest that the Artocarpus lectin is specific toward the Thomsen-Friedenreich (T) antigen. There are subtle differences in the overall topography of its combining site when compared with that of peanut (Arachis hypogaea) agglutinin.
*This work was 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 18 U.S.C. Section 1734 solely to indicate this fact.  The Thomsen-Friedenreich antigen is a tumor-associated antigen of non-oncofetal origin and is probably one of the few chemically well-defined antigens with a proven link to malignancy; therefore, anti-T probes have enormous potentials in cancer research (1)(2)(3)(4).
Occurrence of a lectin reactive toward Thomsen-Friedenreich antigen in the seed extract of jackfruit (Artocarpus integrifolia) has been reported based on the agar gel precipitation reaction against T-antigenic glycoproteins ( 5 ) . The crude seed extract from jackfruit has also been demonstrated to be a very potent and selective mitogen of distinct T and B cell functions (6). Purification of this lectin has recently been reported (7). Among the ligands tested by these authors, MeaGal' was found to be the most effective inhibitor of the hemagglutination reaction caused by the Artocarpus lectin. This information regarding the sugar specificity is based on inhibition of hemagglutination with a limited number of monosaccharides. Moreover, these experiments, at best, provide only relative affinities.
In order to use this lectin as a sensitive probe, it is not only necessary to elucidate its carbohydrate specificity in detail but also to delineate the forces involved in its interaction with ligands (8,9). Thermodynamic and kinetic analyses are invaluable in understanding the specificities of these interactions in addition to their relevance for a better evaluation of the binding of lectins to cells. These apart, a question of major interest is whether lectins have extended binding sites, complementary to 2 or more pyranosyl residues. With these objectives in mind, we report here the thermodynamic parameters for the binding of A. integrifolia lectin with several synthetic ligands, including Galpl+3GalNAc, N-acetyllactosamine, Galpl+3GlcNAc, and N-dansylgalactosamine, as well as commonly available sugars, like melibiose, MeaGal, MepGal, and GalNAc, and the kinetics of its binding to N -The abbreviations used are: MecuGal, methyl-a-D-galactopyranoside; Mej3Ga1, methyl-j3-D-galactopyranoside; dansyl, 5-dimethylaminonaphthalene-1-sulfonyl; DnsGalN, N-dansylgalactosamine; Gal, D-galactopyranose; Glc, D-glUCOpyranOSe; GalNAc, 2-acetamido-2-deoxygalactopyranose; N-acetylglucosamine, 2-acetamido-2-deoxyglucopyranose. All sugars used are D-Sugars unless otherwise specified. dansylgalactosamine as studied by stopped flow spectrofluorometry.

DISCUSSION
Binding of Monosaccharides-In this paper, we have used the fluorescence of DnsGalN to probe the interaction of saccharides with Artocarpus lectin. The tetrameric lectin is bivalent, and there is no interaction between the two combining sites as the Scatchard plots in the low fractional and high saturation ranges are linear. The AH and A S values of -31.6 kJ . mol" and -26.8 J .mol". K-I, respectively, for the binding of DnsGalN to Artocarpus lectin are in the range of those obtained for several lectin-sugar interactions (20,22,30 The thermodynamic parameters for the binding of saccharides to Artocarpus lectin are highly informative regarding the specificity of this lectin. Artocarpus lectin binds MeaGal200fold stronger than MePGal; on the other hand, Griffonia simplicifolia lectin I and peanut agglutinin prefer MeaGal only by 5.5-and 1.5-fold, respectively (31-33). A methyl group in the a configuration appears to contribute positively for binding. The major stabilizing force is enthalpic. The change in enthalpy ( -A H ) for the binding of MeaGal is 13.0 kJ. mol" greater than for MepGal. A relatively large value of -AS coupled with a smaller -AH value is presumably responsible for the poor binding of MepGal. Substitutions at C-2 of Gal have marginal influence on its binding to Artocarpus lectin; Gal is 2 and 0.8 times more potent a ligand than 2deoxygalactose and galactosamine, respectively. GalNAc is about 2, 5, and 2.5 times stronger binding a ligand when compared with Gal, 2-deoxygalactose, and galactosamine, respectively. The better affinity of GalNAc over Gal is presumably due to a favorable enthalpic contribution of the acetamido group for the binding, which amounts to -17.0 kJ.mo1" and probably reflects additional van der Waals' interactions or hydrogen bonding between the acetamido group of the sugar and the protein or potentiation of the effectiveness of the existing ones. Introduction of bulky substituents as in DnsGalN does not increase affinity very markedly over GalNAc. This is in marked contrast to soybean agglutinin where introduction of bulky N-substituents increases affinity by a factor of 30, suggesting the absence of hydrophobic interaction between the dansyl group and the corresponding binding loci in the combining region of the Artocarpus lectin. This is also consistent with a less dramatic enhancement in fluorescence intensity upon binding of DnsGalN to Artocarpus lectin as compared to soybean agglutinin (34) and Pso-phocarpus tetragonolobus lectin3 and poorer entropic contribution for the association of DnsGalN when compared with soybean agglutinin. A lower enthalpic contribution of Dns-GalN binding to Artocarpus lectin suggests that the replacement of the C-2 hydroxyl group, or an acetamido group, with a dansyl group presumably abrogates hydrogen bonding at this locus.
No monosaccharide derivative with a modified substituent at C-3 of galactose was available. At C-4, inversion of the hydroxyl group as in glucose is not allowed. The C-6 hydroxyl group provides an important binding locus as shown by the observation that fucose (6-deoxygalactose) and L-arabinose do not bind to the Artocarpus lectin.
Binding of Disaccharides-It is interesting to compare the values of association constants for monosaccharides and several of the disaccharides with those obtained for the disaccharide Galpl-3GalNAc. This disaccharide has 3-, 36-, loo-, and 610-fold higher affinity over that observed for MeaGal, GalNAc, Gal, and MepGal, respectively. It is observed that GalPl-+SGalNAc has a significantly higher K, than that for lactose, N-acetyllactosamine, and Galpl-3GlcNAc. All of these disaccharides are at least 3000 times poorer ligands as compared to Galpl-3GalNAc. It is appropriate to discuss here the carbohydrate recognition properties of Artocarpus agglutinin uis a uis peanut agglutinin. In contrast to the Artocarpus lectin, peanut agglutinin does not discriminate very markedly between fucose and galactose and between MeaGal and MePGal. It also binds reasonably well to lactose, N-acetyllactosamine, and Gal/31+3GlcNAc and poorly to GalNAc (33, 36). For the Artocarpus lectin, the increase in -AH for the disaccharide Galfil-+3GalNAc over that for MePGal amounts to about 63.0 k J . mol". This indicates that the reducing pyranosyl residue of Galpl-3GalNAc is bound in a subsite adjacent to the galactose-binding subsite. Other lectins specific for Gal show quite different thermodynamic parameters when compared to Artocarpus lectin. For example, Ricinus communis lectin (30, 35) and Momordica charantia lectin (22) show similar values of -AH for binding to monosaccharides and disaccharides. Peanut agglutinin, on the other hand, shows an increase in -AH values for binding to this disaccharide over any of the monosaccharides, although the increase in K, amounts to a factor of only 27 over that of MePGal due to a less marked change in the -AH value over the -TAS value as observed here for the Artocarpus lectin. This lectin, in contrast to peanut agglutinin, does not bind to the disaccharides lactose, N-acetyllactosamine, or Galp1-3GlcNAc (36).
The situation found here for the binding of the disaccharide by Artocarpus lectin is reminiscent of the interaction of lysozyme to chitooligosaccharides (37). Binding of lysozyme to chitooligosaccharides gave -AH values of 26.0, 48.0, and 60.0 k J mol" for N-acetylglucosamine, N,N'-diacetylchitobiose, and N,N',N"-triacetylchitotriose, respectively. The -AH values outweigh the increasingly unfavorable -AS values, so that there is a 1000-fold increase in the K , values for N,N'diacetylchitobiose over that of N-acetylglucosamine. Similarly, the binding of Galpl-3GalNAc to Artocarpus lectin is also accompanied by such a large increase in enthalpy, which adequately compensates for the increase in the A S value, resulting in a 610-fold increase in K, for the binding of the disaccharide over that for MePGal.
In order to understand the binding of disaccharides to Artocarpus agglutinin, we have used the thermodynamic parameters obtained by us for the binding of four disaccharides, uiz. GalPl+3GalNAc, Galpl+3GlcNAc, lactose, and N-acetyllactosamine, and their minimum energy conformations ob-tained by theoretical studies as well as the structures obtained by x-ray crystallography (36, 38-41). Fig. 10 shows ball and stick models of these four disaccharides. It is apparent that the conformation of the nonreducing residue in all the four disaccharides is identical.
However, the orientation of the reducing sugar moiety with respect to that of the nonreducing residue for p ( l 4 ) -l i n k e d disaccharides is different from that observed for p( 1+3)-linked disaccharides. In P(14)-linked disaccharides, the C-2 substituent of the reducing sugar is on the same side as that of the hydroxymethyl group of the nonreducing sugar moiety, whereas in P(1-+3)-linked sugars, it lies on the side opposite to the hydroxymethyl group of the nonreducing sugar (for example, compare the structures of Galpl+3GalNAc and N-acetyllactosamine in Fig. 10). Despite these differences, all these disaccharides appear similar in the topography of their ring skeleton. One would, therefore, expect that these disaccharides bind to the lectin with comparable affinities. However, our observations show that the binding of Galpl-3GalNAc is approximately 3000-fold stronger than the other disaccharides (Table I). This differa C 2-ence in affinity can be explained on the basis of the difference in configuration of the functional groups in the reducing sugar moiety in these disaccharides.
In the four disaccharides mentioned above, the C-4 hydroxyl group of the reducing sugar moiety in the p(1-3)linked disaccharides occupies a position similar to that of the C-3 hydroxyl group of the reducing pyranosyl residue of the p(1-A)-linked disaccharides (Fig. 10). However, the configuration of these hydroxyl groups in these disaccharides is not the same. The C-3 hydroxyl groups of the reducing residue in p(l-A)-linked disaccharides and the C-4 hydroxyl group in Galpl-3GlcNAc are in the equatorial configuration, whereas the C-4 hydroxyl group of the reducing residue in GalB1-3GalNAc is in the axial configuration. This difference alone seems to be responsible for the very weak binding of Galp1+ 3GlcNAc, N-acetyllactosamine, and lactose when compared with Galpl+3GalNAc. The structure and conformation of Galpl+3GalNAc and Galpl-3GlcNAc are identical except for the configuration of the C-4 hydroxyl group of the reducing residue. It, therefore, appears probable that the change in    configuration of this hydroxyl group not only leads to a loss of a favorable contact of the disaccharide with the lectin combining site, but also leads to an unfavorable interaction. It is possible that the hydrophilic hydroxyl group in the equatorial configuration is coming in contact with a hydrophobic region in the ligand-binding site of the lectin. The major stabilizing force for the binding of Galpl+3GalNAc to the lectin is apparently provided by the C-4 hydroxyl group as well as the C-2 acetamido group of the reducing sugar moiety of this disaccharide. The involvement of the C-6 hydroxymethyl group of the reducing sugar moiety can be ascertained only on the analyses of the interaction and structure of Galpl+3~-arabinose.
Kinetic Studies-The forward rate constants listed in Table  I1 are several orders of magnitude slower than the diffusioncontrolled reactions, and the kinetics of the sugar binding to the lectin is qualitatively consistent with a single-step binding mechanism (Equation 1). Bimolecular association rate constants which are slower than the diffusion-controlled process have also been reported for the binding of chromogenic/ fluorogenic ligands to concanavalin A (23, 25-27), R. communis agglutinin (24), and soybean agglutinin (34). The flu-orescence change appearing in stopped flow traces on completion of the reaction is equivalent to that observed in equilibrium titrations for association and dissociation rates. The possibility of formation of reaction intermediate within the dead time of the instrument is, therefore, ruled out. The values of overall association constants determined from kinetic measurements between 12 and 27 "C are in agreement with values determined from the fluorescence titrations.
The value of -AH determined from kinetic measurements also agrees well with the value from equilibrium measurements. This indicates that the enthalpy change is related to the total binding event, and there does not exist any faster process which we are not observing, but which contributes significantly to the reaction enthalpy for saccharide binding. Linearity of ln(k,/T) versus (1/T) plot indicates that dramatic conformational transitions in the lectin molecule are absent in the temperature range studied.
The large activation energy is also consistent with the slow reaction rates observed here. The apparent activation entropy is quite large when compared with facile bimolecular reactions. This large activation entropy for the binding reaction suggests that a specific configuration of reactants is required for sugar binding to the lectin. However, the principal barrier for the sugar binding appears to be energetic.
Generally, when the second order rate constants for the ligand binding to a protein differ by several orders of magnitude from that expected for a diffusion-controlled reaction, the binding is presumed to involve a putative intermediate complex as depicted in the equation below (D represents Ndansylgalactosamine and P represents lectin), Conclusions-Considering the thermodynamic parameters in the light of suggestions by Ross and Subrahmanian (43), the major forces for the binding of sugars to Artocarpus lectin are hydrogen bonding and van der Waals' interactions coupled with some contributions from nonpolar interactions (44). Thus, several points of hydrogen bonding and nonpolar interactions between the Artocarpus lectin and saccharides may be inferred from these studies. The C-2, C-4, and C-6 hydroxyl groups of Gal are probably involved in hydrogen bonding. Nonpolar contacts play an important role in the association of MeaGal through its methyl group (since MeaGal binds about 30 times stronger than galactose). The T-antigenic disaccharide binds to the lectin very strongly due to a favorable enthalpic contribution. In contrast to the peanut agglutinin, the only other anti-T lectin reported, Artocarpus agglutinin binds extremely poorly to other disaccharides such as lactose, N-acetyllactosamine, and Gal@l+3GlcNAc, which are topographically similar to Galplj3GalNAc. The differences in the binding property of Artocarpus lectin when compared with peanut agglutinin reflect subtle differences in the topographies of their combining region; and this lectin should, therefore, prove to be a valuable probe for studying the expression of the Thomsen-Friedenreich antigen on cell surfaces.  From the temperature dependent association constants, p l o t s a s shorn i n Fig. 4 and a r e l i s t e d i n Table 1. I t C a n be Seen f r m t h e change than that observed for the association of "Gal. GalNAc. Gal and MePGal.