J Biol Chem, Vol. 274, Issue 42, 29694-29698, October 15, 1999

Thermodynamic Studies of Saccharide Binding to Artocarpin, a B-Cell Mitogen, Reveals the Extended Nature of Its Interaction with Mannotriose [3,6-Di-O-(-D-mannopyranosyl)-D-mannose]^*

P. Geetha Rani, Kiran Bachhawat, Sandra Misquith^§, and Avadhesha Surolia^¶

From the  Molecular Biophysics Unit, Indian Institute of Science, and the ^¶ Jawaharlal Nehru Centre for Advanced Scientific Research, Bangalore 560 012, India

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The thermodynamics of binding of various saccharides to artocarpin, from Artocarpus integrifolia seeds, a homotetrameric lectin (M_r 65,000) with one binding site per subunit, was determined by isothermal titration calorimetry measurements at 280 and 293 K. The binding enthalpies, H_b, are the same at both temperatures, and the values range from 10.94 to 47.11 kJ mol¹. The affinities of artocarpin as obtained from isothermal titration calorimetry are in reasonable agreement with the results obtained by enzyme-linked lectin absorbent essay, which is based on the minimum amount of ligand required to inhibit horseradish peroxidase binding to artocarpin in enzyme-linked lectin absorbent essay (Misquith, S., Rani, P. G., and Surolia, A. (1994) J. Biol. Chem. 269, 30393-30401). The interactions are mainly enthalpically driven and exhibit enthalpy-entropy compensation. The order of binding affinity of artocarpin is as follows: mannotriose>Man3Man>GlcNAc₂Man₃>MeMan>Man>Man6Man>Man2Man>MeGlc>Glc, i.e. 7>4>2>1.4>1>0.4>0.3>0.24>0.11. The H for the interaction of Man3Man, Man6Man, and MeMan are similar and 20 kJ mol¹ lower than that of mannotriose. This indicates that, while Man3Man and Man6Man interact with the lectin exclusively through their nonreducing end monosaccharide with the subsites specific for the 1,3 and 1,6 arms, the mannotriose interacts with the lectin simultaneously through all three of its mannopyranosyl residues. This study thus underscores the distinction in the recognition of this common oligosaccharide motif in comparison with that displayed by other lectins with related specificity.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Carbohydrates conjugated to proteins and lipids play key structural and functional roles in essentially all living organisms. Recognition of glycoconjugates is an important event in biological systems and is frequently in the form of carbohydrate-protein interactions. The study of how biological molecules interact with one another is fundamental to understanding the chemistry of life. Among the carbohydrate binding proteins, lectins are a group of proteins or glycoproteins which stereospecifically bind carbohydrates (1). N-Linked oligomannose-type carbohydrates constitute one class of oligosaccharide chains associated with cellular glycoproteins. The oligosaccharide chains of many of the glycoproteins appear to function as receptors for lectins in a variety of biological recognition processes, such as fertilization, embryogenesis, cell migration, organ formation, immune defense, protein folding, signal transduction, and apoptosis (2-4). Detailed insights into the specificity of carbohydrate-protein interactions, however, require not only analytical data such as inhibition assays but also thermodynamic data on the complexes. Titration microcalorimetry provides a powerful tool for investigating the binding thermodynamics of macromolecule-ligand interactions and provide important insights on the nature and magnitude of forces involved therein (5-11).

Artocarpin, a mannose-specific lectin isolated from jack fruit seeds is a homotetrameric protein devoid of covalently attached carbohydrates and consists of four isolectins with pI in the range of 5-6.5. Artocarpin is of considerable interest because of its potent and selective mitogenic effect on distinct T and B-cell functions, more so because of its B-cell maturation mitogenic activity (12, 13). Earlier investigations of its carbohydrate binding specificity revealed that among monosaccharides, mannose is preferred over glucose. Among mannooligosaccharides, mannotriose (Man1-3[Man1-6]Man), and mannopentaose were noted as the strongest ligands followed by Man1-3Man. Substitution of both the 1-3 or 1-6 linked mannosyl residues of mannotriose by GlcNAc in 1-2 linkage diminishes their inhibitory potencies (9). In this investigation, isothermal titration calorimetry was employed to determine the thermodynamics of the carbohydrate-artocarpin binding reaction in terms of the binding constant (K_b) and change in the free energies, enthalpies, and entropies, i.e. G_b⁰, H_b⁰, and S_b⁰.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Materials and Sample Preparation-- Glucose (Glc), Mannose (Man), methyl--glucose (MeGlc), methyl--mannopyranoside (MeMan), N-acetylglucosamine (GlcNAc), and N-acetylmannosamine (ManNAc) were obtained from Sigma. Methyl--mannopyranoside (MeMan), Man1-2Man, Man1-3Man, Man1-6Man, Man1-3[Man1-6]Man, GlcNAc2Man3, and mannopentaose were the products of Dextra Laboratories, London. All other reagents and chemicals were of the highest purity available. Artocarpin, purified as previously described (9), was dialyzed overnight against 20 mM phosphate buffer at pH 7.2 containing 150 mM sodium chloride (phosphate-buffered saline) and centrifuged to remove any insoluble material. The concentrations of the protein were determined spectrophotometrically (A₂₈₀^1% = 10.8). Solutions of the carbohydrate were prepared by weight in the dialysate to minimize differences between the protein buffer solution and ligand buffer solution in the isothermal titration calorimetry measurements.

Titration Calorimetry-- Isothermal calorimetric titration measurements were performed using an OMEGA titration calorimeter from Microcal Inc. as described previously (5, 10). A circulating water bath was used to help temperature stabilization. The instrument was allowed to equilibrate overnight. Aliquots (5-10 µl) of the ligand solution (9-36 × protein binding sites) were added from the computer-controlled 250-µl rotating syringe stirring at 395 rpm at an interval of 3 min into the lectin solution (1.34 ml) containing 0.5-8 mM binding sites. The heat changes accompanying the ligand solutions to the lectin solution were recorded. The heat of dilution was determined to be negligible in separate titrations of the ligand solution into just the buffer solution.

The total heat, Q_t was then fitted via a nonlinear least squares minimization method to the total ligand concentration (X_t) using Equation 1 (14),

(Eq. 1)

where n is the number of binding sites per monomer and V is the cell volume. The expression for the heat released per ith injection, dQ(i), is then given by Equation 2 (15),

(Eq. 2)

where dV_i is the volume of titrant added to the solution. The parameters G_b⁰ and S_b are calculated from the basic equations of thermodynamics according to Equations 3 and 4,

(Eq. 3)

(Eq. 4)

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The structures of ligands used in titration calorimetry experiments are depicted in Fig. 1. The results of a typical titration calorimetry measurement, which consisted of addition of 5-µl aliquots of Man3Man (5 mM) to artocarpin (0.5 mM) in phosphate-buffered saline, pH 7.2, at 281 K are shown in Fig. 2. The results exhibit a monotonic decrease in the exothermic heat of binding till saturation is achieved. A least squares fit of the total heat released as a function of ligand concentration to the identical site model described by Equation 1 is also shown in Fig. 2. The close fit of the data to the identical site model shows that a molecule of ligand binds to each of the four sites of artocarpin independently and with the same binding constant and stoichiometry. The thermodynamic binding parameters of mannopyranosides and mannooligosaccharides to artocarpin are listed in Table I. All experiments were performed in the C value (C = K_b × M_t, where M_t is the macromolecular concentration) range of 1<C>20, except for Man4Man where a C value of approximately 0.75 could only be achieved (binding site ~ 8 mM). In most cases the standard deviations in the values of K_a and H were within 5%, except for ligands such as glucose, mannose, MeGlc, Man2Man, etc. which have low affinity for artocarpin.

View larger version (20K): [in this window] [in a new window]   Fig. 1.   Structures of monosaccharides and mannooligosaccharides used in the study.

View larger version (27K): [in this window] [in a new window]   Fig. 2.   Calorimetric titration of artocarpin (0.5 mM) with Man1-3Man (5 mM) at 281 K. Top, data obtained for 50 automatic injections, each 5 µl, of Man1-3Man; bottom, the integrated curve shows experimental points () and the best fit (). The buffer was 20 mM phosphate buffer at pH 7.2 containing 150 mM NaCl.

The thermodynamic parameters for the binding of mannose and mannooligosaccharide show that the binding reactions are essentially enthalpically driven with little dependence of the enthalpy on temperatures from 280.1 to 293.5 K. The values for H_b range from 47.11 kJ for mannotriose to 9.84 kJ for Man1-2Man. The binding constants for different sugars range from 21,200 M¹ for mannopentaose to 150 M¹ for glucose. The binding reactions for artocarpin to monosaccharides and mannooligosaccharides exhibit enthalpy-entropy compensation as shown in Fig. 3.

View larger version (12K): [in this window] [in a new window]   Fig. 3.   Enthalpy-entropy compensation plot for the binding of artocarpin to monosaccharides and mannooligosaccharides at 293 K. The plot shows a linear relationship with a slope of 1.2 with a correlation coefficient to 0.93.

Whereas glucose and mannose bind to the lectin, albeit weakly, GlcNAc and ManNAc do not bind at all even at high concentrations (8 mM protein binding sites and 160 mM sugars). Binding of the lectin with Man4Man is barely detected at the above concentrations, so much so that the thermodynamic parameters for its binding could not be ascertained. MeMan, Man3Man, and Man6Man display similar changes in enthalpies (~26-28 kJ mol¹), whereas mannotriose and mannopentaose exhibit 45-47 kJ mol¹ of H values, viz. 20 kJ more favorable change in enthalpy over them.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

In an earlier study, artocarpin was demonstrated to differ considerably from all the other mannose/glucose binding lectins studied so far. Using an enzyme based assay, it was shown that among mannooligosaccharides artocarpin binds best to mannotriose and mannopentaose, highlighting the possibility of an extended combining region for these saccharides in the lectin (9). To examine in greater detail the nature of these interactions, the binding of several monosaccharides and mannooligosaccharides to artocarpin was studied by isothermal titration calorimetry.

In Table I, the values of the binding constants and resultant relative affinities determined by titration calorimetry are listed together with the relative potencies determined by inhibition assays using enzyme-linked lectin absorbent essay (ELLA) (9). The relative affinities for binding of these saccharides are in good agreement with the inhibition studies, except for GlcNAc₂Man₃ which gives higher value of K_b from that expected from inhibition data (9).

Thermodynamics of the interaction of artocarpin with mannose and mannooligosaccharides are essentially enthalpically driven and exhibit compensatory changes in H_b and TS as shown in Fig. 3. This compensatory behavior has been attributed to the accompanying changes in solvent reorganization. Reduction in soft vibrational modes restricting mobility of water molecules involved in the binding reaction at the interface between the lectin and the sugar molecules could account for enthalpy-entropy compensation. Displacement of water molecules from the interacting complementary surfaces account for the favorable entropic contribution. However, this gain in entropy from previously imposed motional restriction is offset by loss of a certain amount of enthalpic interactions, i.e. water-mediated hydrogen bonds and van der Waals interactions in the combining site. If a group of similar ligands interact by the same mechanism, then a linear relationship between enthalpy and entropy can be expected, with a slope of unity reflecting a complete compensation. Our H versus TS plots show linearity with slope of 1.2 (correlation coefficient = 0.93) indicating an underlying common mechanism in artocarpin-sugar reaction, unlike ConA¹ where profound deviations have been noted, for example, between the other sugars used in these studies and GlcNAc₂Man₃ (16, 17). A value of >1 for the slope ( = 1.2) of enthalpy-entropy compensation plot indicates the primacy of enthalpic forces in determining the overall free energies of artocarpin-ligand interactions. This indicates that the major driving force for artocarpin-sugar interactions is hydrogen bonding and Van der Waals interactions. Additionally, temperature independence of enthalpies highlights that these reactions for the most part occur with little changes in the conformation of either the ligand or the protein.

A comparison of the thermodynamic parameters for the binding of artocarpin with carbohydrates provide interesting insights on the topography of the combining site of artocarpin. As the artocarpin-sugar interactions are enthalpically driven and enthalpy-entropy compensation is observed, the differences in binding enthalpies can be explained by considering that the initial interaction of the carbohydrate ligands with the solvent is similar for all the ligands.

The fact that glucose is about 9-fold poorer a ligand, as compared with mannose, and displays H_b, which is 9 kJ mol¹ less positive than the latter, emphasizes the importance of an axial orientation for the hydroxyl group at C2 of a hexapyranose for its interaction with the corresponding locus on the combining site of artocarpin. The preference of the lectin for mannopyranoside is also reflected when one considers the binding of MeMan and MeGlc, where the former is 6-fold better a ligand and exhibits 7.7 kJ mol¹ more favorable binding enthalpy as compared with the latter. The preference for mannopyranoside over glucopyranoside is apparently the highest for artocarpin as compared with the other mannose/glucose-specific lectins such as those from pea, lentil, ConA, or Diocleinae family of lectins (18-20). MeMan was inactive, highlighting the indispensability of mannopyranoside in  configuration for its recognition by artocarpin. Nonbinding of N-acetylmannosamine suggests that the bulky acetamido group at C-2 in manno configuration leads to a steric hindrance.

It is interesting to compare the binding potencies of various mannooligosaccharides with ConA, Dioclea lectins, and snowdrop lectin with that of artocarpin. Dioclea lectins and especially ConA recognize mannobioses in the order Man1-2ManMan1-6Man>Man1-3Man in a site that essentially accommodates a monosaccharide, a fact also borne out by their nearly similar enthalpies of binding (21, 22). The greater affinity of ConA for Man1-2Man has been related to statistically increased probability of binding of either the reducing or the nonreducing mannose residue to a site that essentially accommodates a single mannopyranosyl residue (21, 22). Dioclea lectins in most cases recognize these mannobioses by a similar mechanism (18, 21). ConA and Dioclea lectins in general bind to the trimannoside, present in the core regions of N-linked glycans, with about 30-60- and 200-fold better affinities, respectively. It has also been demonstrated that, whereas the 1-6 mannose occupies the primary site viz. the high affinity monosaccharide binding site, the 1-3 linked mannose occupies a secondary site; and both together, with the reducing end 3,6-disubstituted core mannose residue, constitute the extended binding site discovered by the pioneering work of Brewer and co-workers (16, 17, 23-25). Snow drop (Galanthus nivalis agglutinin; GNA) lectin on the other hand binds preferentially to Man1-3Man over Man1-6Man and fails to recognize Man1-2Man. Man1-3Man is nearly equivalent in its affinity to that of mannotriose, and it has been proposed by Chevernak and Toone (27) that it has specific sites for binding to 1-6 and 1-3 arms of the trisaccharide (26).

Poor affinity of artocarpin for Man1-2Man (K_b = 490 M¹) as compared with mannose (1378 M¹) itself together with its low enthalpy (18.6 kJ mol¹ versus 25.1 kJ mol¹ of mannose) suggests that it is able to recognize only a part of mannose epitope in this disaccharide. It is therefore apparent that the substitution of the reducing end mannose at C-2, an essential locus for the recognition of artocarpin, precludes its binding, whereas the orientation of the reducing mannose in the disaccharide is such that some of its determinants are inaccessible to the combining site of the lectin (Fig. 4). Man1-3Man is ~1.5-2-fold poorer a ligand, whereas Man1-6Man is 17-fold weaker a ligand as compared with mannotriose. This indicates that 1-3-linked mannose occupies the primary binding site, and the 1-6Man occupies the secondary subsite with both sites being specific to the 1,3 and 1,6 arms of the two oligosaccharides. They, together with 3,6-disubstituted mannose residue, constitute the extended combining site of artocarpin. The equivalence of binding constants (~9000-10,800 M¹) and H (~46-47 kJ mol¹) values for mannotriose and mannopentaose suggests that artocarpin binds to the trimannosyl moiety located at the 1,6 arm of mannopentaose. Poor binding of GlcNAc₂Man₃ (K_b = 2902 M¹), a complex type glycan, and especially its low values of H (10.94 kJ mol¹) underscore the unfavorable consequences of substituting the 1,3- and 1,6-linked mannosyl residues at their C-2 positions for interactions with artocarpin. In this respect, artocarpin differs strikingly from ConA where this structure binds better than mannotriose (17). More so, its binding mechanism also differs significantly from those of mannooligosaccharides, whereas for artocarpin its location on the enthalpy-entropy compensation plot attests to its mechanism of interaction akin to the other ligands used (16).

View larger version (19K): [in this window] [in a new window]   Fig. 4.   Schematic representation of artocarpin extended binding site for mannotriose. The 1-3-linked mannose occupies the primary binding site, and the 1-6 Man occupies the secondary subsite with both sites being specific to the 1-3 and 1-6 arms of the two oligosaccharides. They together with core mannose constitute the extended combining site of artocarpin.

At a superficial level, the combining site of artocarpin would appear similar to that of GNA. Nevertheless a detailed examination of thermodynamic parameters reveals that for GNA the binding of Man1-3Man and mannotriose are accompanied by values of H that differ very little from each other, i.e. 12.97 and 16.32 kJ mol¹, respectively (27). For artocarpin the situation is different. The order of binding affinity is as follows: mannotriose>Man3Man>GlcNAc₂Man₃>MeMan>Man> Man6Man>Man2Man>MeGlc>Glc, i.e. 7>4>2>1.4>1> 0.4>0.3>0.24>0.11. Moreover, the H_b values for the binding of MeMan, Man3Man, and Man6Man are close to each other (27-29 kJ mol¹), which are about 20 kJ mol¹ lower than that of mannotriose. This indicates that, while the two disaccharides may be binding to artocarpin mostly through their nonreducing end monosaccharides to the respective subsites specific for 1-3 and 1-6 arms of the combining site, the addition of core mannose makes approximately 20 kJ mol¹ contribution to the binding enthalpy. In other words, in contrast to GNA where the core mannose makes little contribution, 3.2 kJ mol¹, to the overall binding process, in artocarpin it seems to make significant contribution to the interaction. Taken together, these data suggest that the extended binding site of artocarpin has features that diverge from both ConA family of lectins and GNA (16, 17, 20, 27).

In summary, the extended combining site of artocarpin exhibits interesting differences from the mannose/glucose lectins studied so far and provides an additional paradigm for investigation of the recognition of the trimannoside motif found commonly in the N-linked glycoproteins.

	FOOTNOTES

^* This work was supported by the Department of Science & Technology, and the OMEGA titration calorimeter was provided by a grant from the Department of Biotechnology, Government of India (to A. S.).The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

^§ Permanent address: Dept. of Chemistry, St. Joseph's College, Bangalore 560 025, India.

To whom correspondence should be addressed: Molecular Biophysics Unit, Indian Institute of Science, Bangalore 560012, India. Tel.: 91-80-3092389; Fax: 91-80-3348535/3341683; E-mail: surolia@mbu.iisc.ernet.in.

	ABBREVIATIONS

The abbreviations used are: ConA, concanavalin A; GNA, Galanthus nivalis agglutinin.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES