Thermodynamic, Kinetic, and Electron Microscopy Studies of Concanavalin A and Dioclea grandiflora Lectin Cross-linked with Synthetic Divalent Carbohydrates*

The jack bean lectin concanavalin A (ConA) and the Dioclea grandiflora lectin (DGL) are highly homologous Man/Glc-specific members of the Diocleinae subtribe. Both lectins bind, cross-link, and precipitate with carbohydrates possessing multiple terminal nonreducing Man residues. The present study investigates the binding and cross-linking interactions of ConA and DGL with a series of synthetic divalent carbohydrates that possess spacer groups with increasing flexibility and length between terminal (cid:1) -mannopyranoside residues. Isothermal titration microcalorimetry was used to determine the thermodynamics of binding of the two lectins to the divalent analogs, and kinetic light scattering and electron microscopy studies were used to characterize the cross-linking interactions of the lectins with the carbohydrates. The results demonstrated that divalent analogs with flexible spacer groups between the two terminal Man residues possess higher affinities for the two lectins as compared with those with inflexible spacer groups. Furthermore, despite their high degree of 2 , pH 7.2). All experiments were done at room temperature. Absorbances were monitored continuously until they remained constant for 30 min. After each experiment, a portion of the precipitate was treated with 100 m M (cid:2) MDM to check whether or not the precipitation was due to the binding of the saccharides. Electron Microscopy— Negative stain electron microscopy of the precipitates was performed by placing the samples on 300-mesh carbon- coated Parlodion grids that had been freshly glow-discharged for 2 min, touched to filter paper and floated on a drop of 1% phosphotungstic acid, pH 7.0, and blotted immediately. The samples were observed at 80 kV in a JEOL 1200EX electron microscope.

Lectins are carbohydrate-binding proteins that are widely conserved in nature, such as those in animals, plants, and microorganisms (1). The biological activities of many animal lectins have been determined, including receptor-mediated endocytosis of glycoproteins, cellular recognition and adhesion (cf. Ref. 2), regulation of inflammation (3), and metastasis and control of cell growth (4,5). A common feature of lectins is their multivalent binding properties (6,7). As a consequence, lectin binding to cells leads to cross-linking and aggregation of glycoprotein and glycolipid receptors. In many cases, these interactions are associated with signal transduction effects, including the arrest of bulk transport in ganglion cell axons (8), molecular sorting of glycoproteins in the secretory pathways of cells (9), apoptosis of human T cells (10,11), regulation of the T cell receptor (12,13), and growth regulation of neuroblastoma cells (14). Thus, the carbohydrate cross-linking properties of lectins are a key feature of their biological activities.
The cross-linking properties of a variety of plant and animal lectins with multivalent carbohydrates and glycoproteins have recently been reviewed (15). Studies show that a number of lectins form homogeneous cross-linked complexes with branched chain oligosaccharides and glycoproteins. For example, quantitative precipitation experiments with the Man/Glcspecific lectin concanavalin A (ConA) 1 in the presence of binary mixtures of a series of bivalent N-linked oligomannose glycopeptides indicate that each glycopeptide forms its own unique cross-linked complex with the lectin (16). Subsequent x-ray crystallographic studies have demonstrated different lattice structures of crystalline cross-linked complexes of the soybean agglutinin with four different divalent carbohydrates (17). The different lattice structures are due to differences in the structures of the cross-linking carbohydrates (17). The ability to form unique cross-linked complexes with glycoconjugates and to separate different counter-receptors into homogeneous cross-linked aggregates has recently been implicated in the apoptotic activity of galectin-1, a member of the ␤-galactosidase-specific animal lectin family (18).
Recently, galectin-3, another member of the galectin family, has been shown to form disorganized, heterogeneous crosslinked complexes with multivalent carbohydrates (19). The biological properties of galectin-3, including its anti-apoptotic activities (20) and ability to antagonize the growth inhibitory activity of galectin-1 in neuroblastoma cells (21), may relate to its ability to randomly cross-link glycoconjugates and prevent separation of different receptors. Hence, the ability of lectins to form organized or disorganized cross-linked complexes with multivalent glycoconjugate receptors, such as galectin-1 and -3, respectively, may relate to their biological activities.
In the present study, we investigated the effects of varying the flexibility and distance between terminal Man residues in divalent carbohydrate analogs on their binding thermodynamics and cross-linking interactions with ConA and the lectin from Dioclea grandiflora (DGL). Isothermal titration calorimetry (ITC), kinetic light scattering, and electron microscopy (EM) studies were used to characterize these interactions. The results demonstrated that the thermodynamics of binding and cross-linking properties of the two lectins are sensitive to the flexibility and spacing between the carbohydrate epitopes of the analogs.

EXPERIMENTAL PROCEDURES
ConA was purchased from Sigma and/or prepared from jack bean (Canavalia ensiformis) seeds (Sigma) according to the method of Agrawal and Goldstein (22). The concentration of ConA was determined spectrophotometrically at 280 nm using A 1%,1 cm ϭ 13.7 and 12.4 at pH 7.2 and 5.2, respectively (23), and expressed in terms of monomer (M r ϭ 25,600). DGL was isolated from D. grandiflora seeds obtained from northeastern Brazil (Albano Ferreira Martin Ltd., Sã o Paulo, Brazil) as described previously (24). The concentration of DGL was determined spectrophotometrically at 280 nm using A 1%,1 cm ϭ 12.0 at pH 7.2 and expressed in terms of monomer (M r ϭ 25,000) (24). ␣MDM was purchased from Sigma. The synthesis of carbohydrate analogs 1, 2, 3, and 4 has been reported previously (25), as have analogs 9 -13 (26). Synthesis of 5-8 will be reported elsewhere. The concentrations of carbohydrates were determined by modification of the Dubois phenolsulfuric acid method (27,28) using appropriate monosaccharides as standards. Structures of these analogs are shown in Figs. 1 and 2.
Isothermal Titration Microcalorimetry-ITC experiments were performed using an MCS ITC instrument from Microcal, Inc. (Northampton, MA). Injections of 4 ml of carbohydrate solution were added from a computer-controlled 250-or 100-l microsyringe at an interval of 4 min into the sample solution of lectin (cell volume ϭ 1.34 ml) with stirring at 350 revolutions/min. An example of an ITC experiment is shown in Fig. 3 for bivalent analog 11 with ConA at 27°C. Control experiments performed by making identical injections of saccharide into a cell containing buffer without protein showed insignificant heats of dilution. The concentrations of lectins were 0.1-0.19 mM, and the sugars were 1.0-6.0 mM, respectively. Titrations were done at pH 5.0-5.2 and at NaCl concentrations from 0.05 to 0.15 M. The experimental data were fitted to a theoretical titration curve using software supplied by Microcal, with ⌬H (enthalpy change in kcal/mol), K a (association constant in M Ϫ1 ), and n (number of binding sites/monomer) as adjustable parameters. The quantity c ϭ K a Mt(0), where Mt(0) is the initial macromolecule concentration, is of importance in titration microcalorimetry (29). All experiments were performed with c values 1 Ͻ c Ͻ 200. The instrument was calibrated using the calibration kit containing ribonuclease A (RNase A) and cytidine 2Ј-monophosphate (2Ј-CMP) supplied by the manufacturer microcal. Thermodynamic parameters were calculated from the equation ⌬G ϭ ⌬H Ϫ T⌬S ϭ ϪRT ln K a , where ⌬G, ⌬H, and ⌬S are the changes in free energy, enthalpy, and entropy of binding, respectively, T is the absolute temperature, and R ϭ 1.98 cal mol Ϫ1 K Ϫ1 .
Kinetics of Precipitation-Measured volumes of lectin and saccharide solution at stoichiometric concentration (2:1) were mixed in a 1-ml quartz cuvette, and the time-dependent development of turbidity was measured at 420 nm (30). The buffer was Hepes (0.1 M Hepes, 0.15 M NaCl, 1 mM CaCl 2 , and 1 mM MnCl 2 , pH 7.2). All experiments were done at room temperature. Absorbances were monitored continuously until they remained constant for 30 min. After each experiment, a portion of the precipitate was treated with 100 mM ␣MDM to check whether or not the precipitation was due to the binding of the saccharides.
Electron Microscopy-Negative stain electron microscopy of the precipitates was performed by placing the samples on 300-mesh carboncoated Parlodion grids that had been freshly glow-discharged for 2 min, touched to filter paper and floated on a drop of 1% phosphotungstic acid, pH 7.0, and blotted immediately. The samples were observed at 80 kV in a JEOL 1200EX electron microscope. Hence, ITC studies were performed under these conditions. ITC data for ConA binding to 1-13 at 300 K are shown in Table I. K a values for 1-4 are very similar and nearly twice as great as that of the monosaccharide ␣MDM. K a values for 5-13 are nearly 4 -6 times greater than that of ␣MDM. The Ϫ⌬H values for ConA binding to 1-13 are greater than that of ␣MDM, with values for 1-4 generally lower than those of 6 -13.

Thermodynamics of Binding of ConA and DGL to Analogs
The n values for ConA binding to 1-13 are between 0.68 and 0.52, as compared with 1.0 for ␣MDM.
The ITC data for DGL binding to 1-13 are shown in Table II. DGL binding to 1-13 are generally greater than that of ␣MDM (Table II) Electron Micrographs of the Precipitates of ConA and DGL with 1-13-Negative stain electron micrographs of the precipitates of ConA and DGL with 1, 2, and 3 are shown in Fig. 6. Patterns are observed for all three precipitates of ConA, whereas patterns are observed for the precipitates of DGL with 1 and 3 but not for 2. ConA and DGL precipitates with 4 failed to show patterns. Fig. 7 shows negative stain electron micrographs of the precipitates of ConA with 5, 6, and 7 and DGL with 5 and 6. The precipitates of ConA with 5-7 all show observable lattice patterns, whereas the precipitates of DGL with 5 and 6 also show lattice patterns. The precipitates of DGL with 7 failed to show a pattern. The precipitates of ConA and DGL with 8 also failed to show a pattern. Fig. 8 shows negative stain electron micrographs of the precipitates of ConA with 9, 10, and 11, and the precipitates of DGL with 9 -13. Although the precipitates of ConA with 9, 10, and 11 showed lattice patterns, the precipitates of ConA with 12 and 13 failed to show patterns. All of the precipitates of DGL with 9 -13 showed lattice patterns. The tetrameric forms of ConA and DGL are known to bind and precipitate with multivalent carbohydrates and glycoproteins (35,36). ConA has also been shown to form homogeneous cross-linked lattices with individual glycopeptides (16) and glycoproteins (37). ITC studies have shown that ConA and DGL bind with higher affinities to multivalent carbohydrates containing 2-4 mannopyranoside residues/molecule, relative to the monosaccharide ␣MDM (32). The effects of varying the flexibility and spacing between binding epitopes in multivalent carbohydrates on their thermodynamics of binding, kinetics of precipitation, and structures of their cross-linked complexes with lectins has not been investigated. The present study investigated these interactions of ConA and DGL with bivalent Man analogs 1-13 in Figs. 1 and 2. Analogs 1-13 show enhanced affinities relative to ␣MDM for ConA and DGL and different ⌬H and n values. Previous ITC studies have shown that the value of n is inversely proportional to the functional valency of carbohydrate ligands for lectins (32). Values of n of 0.5 have been observed for higher affinity carbohydrate ligands binding to ConA and DGL, with values between 0.5 and 0.8 for lower affinity bivalent ligands due to incomplete binding of the second epitope (32). The n values for ConA and DGL in Tables I and II, respectively, are consistent with lower affinity divalent carbohydrates binding to both lectins. The greater Ϫ⌬H values for 1-13 binding to both lectins,  relative to ␣MDM, are also consistent with divalent binding of the analogs (32). Interestingly, the enhanced K a values of both lectins for 1-13 appear to cluster into two groups for each lectin. Analogs 1-4 show approximate 2-fold enhanced affinities for ConA, and 3-4-fold enhanced affinities for DGL, relative to ␣MDM. On the other hand, analogs 5-13 show enhanced affinities of 4 -6fold for ConA, and 5-20-fold for DGL, relative to ␣MDM (the exception is 9, which possesses ϳ3-fold higher affinity). Thus, 1-4 possess smaller enhanced affinities for both lectins as compared with 5-13. The most obvious structural differences between these two groups of bivalent analogs is the flexibility of the linker regions between the outer two Man residues in each molecule. Analogs 1-4 possess relatively ridged linkers consisting of one or two acetylenic groups with or without a phenyl group, whereas 5-8 and 10 -13 possess relatively flexible methylene groups. 9 is absent a methylene group between the two aryl glycoside moieties, which may be part of the reason for its relatively modest enhanced affinity for DGL in that group. Thus, the degree of flexibility of the spacer groups of the analogs appears to modulate their enhanced affinities for ConA and DGL.
Kinetics of Precipitation of ConA and DGL with Analogs 1-13-At pH 7.0 and high salt concentration, ConA and DGL are tetramers and precipitate with analogs 1-13. The kinetics of precipitation of both lectins with the analogs is shown in the time-dependent light-scattering profiles in Figs. 4 and 5. Because the concentrations of the two lectins and the concentrations of the analogs are the same, a comparison of the precipitation rate profiles of different analogs with the two lectins can be made. Differences in the precipitation rates are due to sev-eral factors, including the affinities of the analogs, the rates of formation of soluble cross-linked complexes, and the solubility constants of their cross-linked lattices. Fig. 4a shows the kinetics of cross-linking and precipitation of ConA with analogs 1-4. Analog 1 shows the slowest rate of precipitation with ConA followed by 4, 3, and 2, respectively. Analog 3, however, shows a greater degree of precipitation with ConA than 2. Thus, the effects of different spacer groups of the analogs are observed in their kinetics and extent of precipitation with ConA. Fig. 4b shows the time-dependent light-scattering profiles of DGL with 1-4. Although analog 1 is the slowest to precipitate with DGL, similar to ConA, the order of kinetics and extent of precipitation of DGL with 2-4 is different from that with ConA. Analogs 3 and 4 show the fastest and largest degree of precip-itation of DGL, whereas 2 and 3 are fastest with ConA. Thus, the different spacer groups of analogs 1-4 exhibit different kinetics and extents of precipitation with ConA and DGL, even though the structures of the two proteins are very similar (31). Fig. 5 shows the time-dependent light-scattering profiles of ConA and DGL with analogs 9 -12, respectively. Similar to 1-4, analogs 9 -12 show differential kinetics and the extent of precipitation with ConA and DGL. Furthermore, the relative kinetics and extent of precipitation of the analogs differs for the two lectins. These results are similar to those of analogs 1-4 with the two lectins. These results demonstrate that differences in the structures of the two lectins affect their kinetics of cross-linking interactions with 1-4 and 9 -12 (Figs. 4 and 5).
Electron Microscopy of the Cross-linked Lattices of ConA and DGL with 1-13-We have previously used negative stain EM to observe the presence of organized lattices in cross-linked complexes of lectins with multivalent carbohydrates (cf. Ref. 15). In the present study, differences in the spacer groups in 1-13 are observed to affect the structures of their cross-linked complexes with ConA and DGL (15). For example, ConA shows EM patterns for precipitates with 1-3, 5-7, and 9 -11. No patterns are observed for the precipitates of 4, 8, 12, and 13. The lack of patterns observed for these precipitates correlates with the increased distance and flexibility separating the Man residues in the analogs, which, in turn, prevents the formation of organized cross-linked complexes.
DGL also shows a pattern of structures for its precipitates with 1-13. EM patterns for precipitates are observed with 1, 3, 5, 6, and 9 -13. No patterns are observed for 2, 4, 7, and 8. With the exception of 2, the lack of patterns observed with 4, 7, and 8 correlate with the increased distance and flexibility between the Man residues in the molecules, a finding similar to that for ConA.
Although the structures and binding specificities of ConA and DGL are very similar, both lectins show differences in their patterns of lattice structures with 1-13. DGL shows patterns with 12 and 13, unlike ConA, which shows no patterns with these two analogs. ConA shows a pattern for precipitates with 2, whereas DGL shows no pattern. ConA also shows a pattern for precipitates with 7, whereas DGL shows no pattern. Thus, two highly homologous lectins show differences in the observed patterns of their precipitates with 1-13. The detailed lattice structures of the cross-linked complexes of the two lectins with the analogs in Figs. 6 and 7 will await x-ray fiber diffraction and image reconstruction of the EMs (cf. Ref. 38) or x-crystallographic analysis of crystals of the respective cross-linked complexes (cf. Ref. 17) CONCLUSIONS The results demonstrated that bivalent Man analogs with flexible spacer groups exhibit higher affinities for ConA and DGL than analogs with rigid spacer groups. ConA and DGL also showed differences in their kinetics of precipitation with the bivalent analogs and differences in the EM patterns of their precipitates with 1-13. The present findings indicated that the spacing and flexibility of carbohydrate epitopes in divalent carbohydrates affects their thermodynamics of binding, kinetics of precipitation, and structures of their cross-linked complexes with different lectins. These results have important implications for the interaction of lectins with multivalent carbohydrate receptors in biological systems (39). FIG. 8. Negative stain electron micrographs of ConA and DGL cross-linked with 9, 10, and 11. The ratio of the molar concentrations of the lectins to the carbohydrates was 2:1. Cross-linking of ConA with 12 and 13 did not give rise to observable patterns; however, crosslinking DGL with the same sugars showed patterns.