Eur. J. Biochem. 176, (1988) 0 FEBS 1988

Transcription

1 Eur. J. Biochem. 176, (1988) 0 FEBS 1988 Lectin-carbohydrate interactions Studies of the nature of hydrogen bonding between D-galactose and certain D-galactose-specific lectins, and between D-mannose and concanavalin A Lokesh BHATTACHARYYA and C. Fred BREWER Departments of Molecular Pharmacology, Atran Foundation Laboratories, and Microbiology and Immunology, Albert Einstein College of Medicine, Bronx, New York (Received March /May 6,1988) - EJB The binding of galactose-specific lectins from Erythrina indica (EIL), Erythrina arborescens (EAL), Ricinus communis (agglutinin ; RCA-I), Abrus precatorius (agglutinin; APA), and Bandeiraea simplicifolia (lectin I ; BSL- I) to fluoro-, deoxy-, and thiogalactoses were studied in order to determine the strength of hydrogen bonds between the hydroxyl groups of galactose and the binding sites of the proteins. The results have allowed insight into the nature of the donor/acceptor groups in the lectins that are involved in hydrogen bonding with the sugar. The data indicate that the C-2 hydroxyl group of galactose is involved in weak interactions as a hydrogen-bond acceptor with uncharged groups of EIL and EAL. With RCA-I, the C-2 hydroxyl group forms two weak hydrogen bonds in the capacity of a hydrogen-bond acceptor and a donor. On the other hand, there is a strong hydrogen bond between the C-2 hydroxyl group of galactose, which acts as a donor, and a charged group on BSL-I. The C-2 hydroxyl group of the sugar is also a hydrogen-bond donor to APA. The lectins are involved in strong hydrogen bonds through charged groups with the C-3 and C-4 hydroxyl groups of galactose, with the latter serving as hydrogen-bond donors. The C-6 hydroxyl group of the sugar is weakly hydrogen bonded with neutral groups of EIL, EAL, and APA. With BSL-I, however, a strong hydrogen bond is formed at this position with a charged group of the lectin. The C-6 hydroxyl groups is a hydrogen-bond acceptor for EIL and EAL, a hydrogenbond donor for APA and BSL-I, and appears not to be involved in binding to RCA-I. The data with the thiosugars indicate the involvement of the C-1 hydroxyl group of galactose in binding to EIL, EAL, and BSL-I, but not to RCA-I and APA. We have also performed a similar analysis of the binding data of fluoro- and deoxysugars to concanavalin A [Poretz, R. D. and Goldstein, 1. J. (1970) Biochemistry 9, , This has allowed comparison of the donor/ acceptor properties and free energies of hydrogen bonding of the hydroxyl groups of methyl a-d-mannopyranoside to concanavalin A with the results in the present study. On the basis of this analysis, new assignments are suggested for amino acid residues of the concanavalin A that may be involved in hydrogen bonding to the sugar. Lectins are cell-agglutinating proteins of nonimmune origin which bind carbohydrates specifically without modifying them [l]. They have been isolated and characterized from a wide variety of bacteria, viruses, invertebrates and vertebrates [2], and have been shown to be implicated in a number of biological processes including cellular recognition, adhesion, Correspondence fo C. F. Brewer, Department of Molecular Pharmacology, Albert Einstein College of Medicine of Yeshiva University, 1300 Morris Park Avenue, Bronx, New York, USA-I0461 Abbreviations. EIL, Erythrina indica lectin; EAL, Erythrina arborescens lectin; RCA-I, agglutinin I from Ricinus communis; APA, agglutinin from Ahrus precatorius; BSL-I, lectin I from Bandeiraea simplicijoliu (syn. Griffoniu simplicijolia); ConA, concanavalin A, the lectin from jack-bean seeds; MeaGal, methyl a-d-galactopyranoside; MepGal, methyl 8-D-galactopyranoside; MeaMan, methyl a-dmannopyranoside. Note. According to the modern electronic thsory, the donoracceptor relationship is determined in terms of the electrons. The group (or atom) which donates the electron pair of a bond is the donor, and the group (or atom) which accepts it is the acceptor (cf. [12]). However, in the context of the present study, we have used the definition that the group that donates the hydrogen atom of a hydrogen bond is the hydrogen-bond donor, and the group which accepts the hydrogen atom is the hydrogen-bond acceptor. and motility [2, 31. However, the best characterized lectins are from plants [4]. Plant lectins have been extensively used in the purification and characterization of glycoconjugates, in the study of cellular and subcellular membranes, and in cell separation [5]. Lectin-resistant mutants have been used to investigate the biosynthesis and functions of glycoconjugates in cells [6]. The biological properties of lectins are related to their carbohydrate-binding activities. Studies with deoxy, 0- methyl, and fluorinated sugars have yielded information on the topography of the binding sites of lectins and the nature of their interactions with sugars [7-93. Hydrogen bonds have been implicated as one of the primary factors that are responsible for the stability and specificity of carbohydrate-lectin interactions [4, 10, 111. In this regard, the use of fluorinated sugars is particularly important in light of the similarity in both electronegativity and size of the fluorine atom and hydroxyl group [12], which precludes the possibility of unfavorable steric interactions (as with 0-methyl derivatives) or binding of a water molecule in the space created by substitution with a smaller substituent (as with deoxy sugars). In addition, the fluorine atom cannot act as a hydrogen-bond

2 208 Table 1. Inhibition of'hen~agglutination by galactose derivatives Sugars Minimum concentration required for complete inhibition of 4 hemagglutinating doses EIL EAL RCA-I APA BSL-I mm Galactose MeaCal MeBGal 1 -Thiogalactose Me 1 -thio-d-galactoside 2-Deoxygalactose 2-Deoxy-2-fluorogalactose 2-Deoxy-2-acetamidogalactose Mer-2-deoxy-2-acetamidogalactoside Me/l-2-deoxy-2-acetamidogalactoside 3-Deoxy-3-fluorogalactose 4-Deoxy-4-fluorogalactose 6-Deoxy-6-fluorogalactose Fucose Galacturonic acid Galactitol so n.d." n.d." n.d." a n.d. stands for not determined. The ph of the solution was adjusted to 7.2 with 5 M NaOH before use. donor (see Note on first page) but it can be a hydrogen bond acceptor, albeit weakly, due to the low polarizability of its electrons [ Indeed, hydrogen bonds involving carbonbonded fluorine have been demonstrated in many compounds [ Fluorinated derivatives have been used in studies of the specificity of binding and nature of the forces involved in enzyme-substrate [18-201, antigen-antibody [21], sugar-lectin [7-91, and sugar-transport protein interactions [22]. Recently, the free energies associated with the elimination of specific hydrogen bonds in ligand-protein complexes have been determined. Using site-specific mutagenesis of tyrosyltrna synthetase, Fersht and coworkers [23-1 have shown that the free energy associated with elimination of a hydrogen bond between an uncharged donor/acceptor pair is kj mol-'. However, deletion of a hydrogen bond between a neutral-charged pair reduces the affinity of small substrate analogs by kj mol-' [23], and of the substrate tyrosyl adenylate by about 11.3 kj mol-' []. Similar results were found in enzyme-inhibition studies of trypsin with parasubstituted benzamidines [26] and of glycogen phosphorylase using deoxy- and fluorosugars [13]. These findings show that determination of the magnitude of the free-energy loss associated with elimination of a hydrogen bond between a ligand and protein indicates the nature of the group on the protein which is involved in hydrogen bonding if the group on the jigand is known. Indeed, the conclusions reached using measurements of this type are supported by X-ray crystallographic data [14]. (In a study of ligand-binding to thermolysin using phosphonate esters and phosphonamidate analogs, about 17.1 kj mo1-l of free energy was argued to exist for a hydrogen bond between an uncharged pair of residues [27]. However, this conclusion appears to be unwarranted since the possibility of a charged residue in the hydrogen bond could not be ruled out.) In the present investigation, we have used deoxy-, thio-, and fluorogalactoses (all in the D configuration) to study the donor/acceptor relationships and free energy contributions of hydrogen bonds between the hydroxyl groups of galactose and the galactose-specific lectins from the seeds of Erythrina indica (EIL), Erythrina arborescens (EAL), Ricinus conzrnunis (agglutinin I; RCA-I), Abrus precatorius (agglutinin; APA) and Bundeiraea (Griffonia) sirnpliclfolia (lectin I; BSL-I). The free energy data were analyzed to evaluate the nature of the groups of the lectins involved in hydrogen bonding. The results were compared with data for the binding of fluoro- and deoxysugars to concanavalin A (ConA) [7], which, in turn, are used to derive information on the sugar-binding residues of the latter protein. MATERIALS AND METHODS Materials EIL and EAL were purified as described previously [28]. APA, purified according to Roy et al. [29], was kindly provided by Dr N. K. Sinha (Bose Institute, Calcutta, India). RCA-I and BSL-I were purchased from Sigma Chemical Co. Fluorogalactoses were prepared following the methods described previously [ The purity of the fluorogalactoses were checked by melting point, thin-layer chromatography, and high-resolution I9F-NMR at 188 MHz. Methyl a- and /3- N-acetylgalactosaminides were gifts from Dr I. J. Goldstein (University of Michigan, Ann Arbor, Michigan). Other monosaccharides were products of Sigma Chemical Co. Hernagglut ina t ion- inh ib it ion assays These were done at 22 C in 10mM sodium phosphate buffer, ph 7.2, containing 0.15 M NaCl using 3% (by vol.) suspensions of rabbit (for EIL, EAL, and RCA-I), human blood group 0 (for APA) or B (for BSL-I) erythrocytes according to Osawa and Matsumoto [34]. RESULTS AND DISCUSSION Limited data are available on the carbohydrate-binding specificities of EIL [28], EAL [28], RCA-I [35, 361, and APA

3 = Table 2. Estimated free energies of hydrogen bonding of the hydroxyl groups of galactose binding to EIL, EAL, RCA-I, APA, and BSL-I, and mannose binding to ConA The experimental error in Table 1 (less than &%) corresponds to an uncertainty of less than 1.7 kj mol-' in the values tabulated here. This uncertainty is sufficiently small to distinguish between a neutral hydrogen bond (x4.2 kj mol-') and a charged hydrogen bond ( N kj mol- '). A or D indicate that the hydroxyl group is a hydrogen-bond acceptor or donor, respectively. Data for ConA are from Porclz and Goldstein [7] 209 Hydroxyl groups Free energy of hydrogen bonding with EI L EAL RCA-I APA BSL-I ConA kj mol-' C-1 (B) -3.3 A -1.7 A - - < A A (4 c A -5.0 A -5.0 A < -2.9 D < D A < -3.3 D (axial) c-3 < -8.4 D < -8.4 D < -8.4 D < -2.9 D < D A c-4 < -8.4 D < -8.4 D < -8.4 D < -2.9 D D D C A -3.3 A b - < -2.9 D < D D a The C-1 hydroxyl group of galactose does not bind to RCA-I and APA. The C-6 hydroxyl group of galactose is not involved in RCA-I binding. [35, 371. They are galactose-specific lectins with little anomeric preference (Table 1). Studies with BSL-I show that it is a galactose-specific lectin, but that it has a strong preference for the a-anomer [9,38]. BSL-I consists of five isolectins [4] which have similar sugar-binding properties except to 2-deoxy-2- acetamidogalactose and its glycosides 19, 381. Inhibition by galactose derivatives Table 1 shows the results of inhibition of hemagglutination of the lectins by derivatives of galactose in which the hydroxyl groups of the sugar are selectively substituted. A comparison of their inhibitory potencies (relative affinities) with that of galactose or the appropriate galactosides indicates the involvement of the individual hydroxyl groups of the sugar in binding. However, the anorneric ratios of the derivatives may differ from that of galactose which contains 35% of the a- anomer in solution [39]. I9F-NMR data in the literature [32, 331 and 'H- and I9F-NMR data in the present study (not shown) indicate that the fluorogalactoses contain 33-40% a-anomer. Furthermore, other than BSL-I, the lectins show little anomeric preference. Thus, differences in the inhibitory powers of the fluorosugars are not due to variations in their anomeric ratios. Furthermore, the conformations of the C-2 and C-4 fluorogalactose derivatives have been shown to be the same [Cl(D)] as the parent sugar [32, 331. NMR data for the C-3 and C-6 derivatives are also consistent with this conformation (not shown). The free energy of hydrogen bonding with a hydroxyl group of galactose is given by the loss of binding free energy of a derivative in which the hydroxyl group is selectively substituted [7, 401. This loss can be calculated using Eqn (1): d(dg") = RTln ~ RTln - (1) KS CG in which d(dg") stands for the loss of binding free energy, KG and Ks represent the affinity constants for galactose and the substituted galactose, respectively, and cg and cs represent the respective minimum concentrations required for complete inhibition of four hemagglutinating doses of a lectin. The validity of Eqn (1) is illustrated by the finding that inhibitory concentrations of simple sugars bear a simple linear relation- KG CS ship with the affinity constant values for ConA [41]. The free energy data for the galactose-specific lectins are shown in Table 2. Table 2 also shows the free energy data for ConA which is included from Poretz and Goldstein [7] for the sake of comparison (see later). Interactions at the C-I hydroxyl group The results presented in Table 1 indicate that EIL, EAL, RCA-I, and APA do not have significant anomeric specificity. EIL seems to be equally potent for the a- and B-anomers of methyl galactopyranosides. EAL binds the a-anomer better than the p-anomer by a factor of 2, whereas RCA-I and APA have twofold higher affinity for the p-anomer. Similar results have been reported previously [28, 36, 371. For these lectins galactose is essentially as potent as its 1-0-methyl derivatives. Since a methoxy group cannot be a hydrogen-bond donor, these results indicate that either the C-1 hydroxyl group is not involved in lectin binding or the oxygen atom functions as a hydrogen-bond acceptor. Substitution of the anomeric oxygen of galactose by a sulfur atom reduces the inhibitory potency by a factor of 2-4 for EAL and EIL (compare I-thiogalactose and methyl 1-thio-P-galactoside with MejGal, Table 1 ; the former two are directly comparable since 'H-NMR at 0 MHz shows that 1-thiogalactose exists exclusively as p-anomer in solution). Since sulfur is a very weak, if at all, acceptor of hydrogen bonds [12], these results indicate that the C-1 hydroxyl group is involved in binding to EIL and EAL, and that the anomeric oxygen functions as a hydrogen-bond acceptor. Substitution of the anomeric oxygen by sulfur does not change the inhibitory potencies of the sugars for RCA-I and APA. Thus, for these two lectins the C-1 hydroxyl group of galactose does not appear to be involved in binding to the lectins. BSL-I binds MeaGal more strongly than MepGal and has an intermediate affinity for galactose, as previously reported [9]. The fourfold decrease in inhibitory potencies of 1-thiogalactose and methyl 1 -thio-p-galactoside compared to MePGal (Table 1) indicates that, like EIL and EAL, the anomeric oxygen acts as a hydrogen-bond acceptor in binding to BSL-I. The estimated free energies of hydrogen bonding of the C-1 hydroxyl group of galactose with the lectins are listed in

4 210 Table 2. The values are obtained from the loss of affinity of 1-thiogalactose and methyl 1 -thio-p-galactoside compared to MePGal for each lectin. The values of kj mol-' indicate hydrogen bonds of the hydroxyl group with neutral donor groups in EIL, EAL, and BSL-I. These interactions do not occur with RCA-I and APA. Interactions at the C-2 hydroxyf group The results in Table 1 indicate that 2-deoxygalactose binds eightfold weaker than galactose to EIL and EAL, but that 2-deoxy-2-fluorogalactose is as potent as galactose. Thus, the loss of affinity due to substitution of the hydroxyl group by hydrogen at C-2 is fully regained by substitution with a fluorine atom. Since fluorine is not a hydrogen-bond donor, the fluorine atom and, therefore, oxygen of the C-2 hydroxyl group function as hydrogen-bond acceptors. Table 2 shows that the free energy of hydrogen bonding of the C-2 hydroxyl group with EIL and EAL is 5.0 kj mol-', which indicates that the hydrogen bonds are donated by uncharged polar groups of the lectins. 2-Deoxygalactose fails to inhibit hemagglutination by RCA-I at 200 mm, but the affinity is partially restored with 2-deoxy-2-fluorogalactose (Table 1). The results suggest the presence of two hydrogen bonds between the C-2 hydroxyl group and RCA-I: one is donated and the other is accepted by the hydroxyl group. The hydrogen bond involving the oxygen atom of the hydroxyl group as an acceptor is restored by the fluorine atom in 2-deoxy-2-fluorogalactose. The free energy contribution due to the hydrogen bond accepted by the C-2 hydroxyl group is 5.0 kj mol-' (Table 2), and is, therefore, donated by an uncharged residue on the lectin. The free energy of the other hydrogen bond is more than 3.3 kj mol-'; thus, from the present data, the nature of the group of the protein associated with the latter hydrogen bond can not be deduced. In binding APA, both 2-deoxygalactose and 2-deoxy-2- fluorogalactose are more than threefold weaker than galactose, which indicates the involvement of the C-2 hydroxyl group. It is difficult to deduce the nature of this involvement because of the high concentration of galactose required for inhibition of hemagglutination by APA. However, since 2- deoxygalactose and 2-deoxy-2-fluorogalactose have at least three times less affinity than galactose, there is at least one hydrogen bond (more than 2.9 kj mol-') for which the C-2 hydroxyl group serves as a hydrogen-bond donor. 2-Deoxygalactose binds BSL-I about 63-fold weaker than galactose (Table l), indicating strong hydrogen bonding between the lectin and the C-2 hydroxyl group of the sugar. 2-Deoxy-2-fluorogalactose loses affinity by another factor of more than 4 (more than 3.3 kj mol-'). The higher affinity of 2-deoxygalactose with respect to 2-deoxy-2-fluorogalactose may be due to the free energy contribution to sugar binding from a hydrogen bond formed by the lectin with a water molecule in the space previously occupied by the C-2 hydroxyl group. Hydrogen-bond formation with solvent water molecule(s) in the space between protein and ligand has been demonstrated and shown to affect protein-ligand binding affinity [lo, 141. Therefore, the free energy of hydrogen bonding with the C-2 hydroxyl group is calculated on the basis of inhibition data obtained with 2-deoxy-2-fluorogalactose (Table 2). Thus, the C-2 hydroxyl group is involved in one hydrogen bond in which it functions as a donor. The free energy data indicate that the hydrogen bond is accepted by a charged group of the lectin. Interestingly, 2-deoxy-2-acetamidogalactose and its methyl glycosides bind to EIL and EAL slightly better than the respective galactose analogs, indicating the presence in these proteins of a locus for the acetamido group of the sugar. Such a locus is absent in RCA-I and APA, since these sugars fail to inhibit hemagglutination by the lectins. Similar results were obtained by Van Wauwe et al. [36] with RCA-I using p-nitrophenyl derivatives of 2-deoxy-2-acetamidogalactose for the inhibition of precipitation of alfalfa galactan by the lectin. However, the results differ from the findings of Absar et al. [37] who found by the ultraviolet difference spectroscopic studies that 2-deoxy-2-acetamidogalactose binds APA nearly as strong as galactose. The isolectins of BSL-I also appear to differ in their locus for binding to the acetamido group at the C-2 position of galactose [9, 381. Interactions at the C-3 hydroxyl group Table 1 shows that 3-deoxy-3-fluorogalactose does not inhibit hemagglutination by the lectins at the highest concentrations tested. p-nitrophenyl3-deoxy-3-fluoro-fl-galactoside failed to inhibit precipitation of BSL-I by guaran [9]. The results indicate that the C-3 hydroxyl group is essential for binding of galactose to these lectins. In the absence of data with 3-deoxygalactose, it is difficult to predict with certainty the number of hydrogen bonds between the C-3 hydroxyl group and each lectin. However, the results clearly show that the hydroxyl group is involved in at least one hydrogen bond with each lectin and serves as a donor. Table 2 shows that the hydrogen bond involves a large free energy change of more than 8.4 kj mol-' for EIL, EAL, RCA-I, and BSL-I. These results are consistent with a hydrogen bond accepted by a charged residue of each lectin. The free energy contribution for APA of greater than 2.9 kj mol-', however, does not indicate the nature of the acceptor group on the protein. Interactions at the C-4 hydroxyl group Table 1 shows that 4-deoxy-4-fluorogalactose fails to inhibit hemagglutination by EIL, EAL, RCA-I, and APA. Thus, at least one hydrogen bond is donated by the C-4 hydroxyl group in lectin binding. The free energy data (Table 2) suggest that, like the C-3 hydroxyl group, the hydrogen bond with the C-4 hydroxyl group is accepted by charged groups of EIL, EAL, and RCA-I. The data allow no conclusion about the nature of the acceptor group of APA. 4-Deoxy-4-fluorogalactose inhibits hemagglutination by BSL-I, albeit 63-fold weaker than galactose. This result is different from that obtained by Murphy and Goldstein [9] who reported that methyl 4-deoxy-4-fluoro-a-galactoside did not inhibit precipitation of BSL-I by guaran and concluded that the sugar did not bind BSL-I. This difference may be due to the lower concentration of sugar used (20 mm) and/or different method employed to monitor the binding of 4-deoxy- 4-fluorogalactose in that study [9]. In any case, the result indicates that the C-4 hydroxyl group donates a hydrogen bond to the lectin. The free energy value of 10.0 kj mol-' suggests participation of a charged acceptor group at this position in the proteins. Interactions at the C-6 hydroxyl group The results in Table 1 show that fucose (6-deoxygalactose) is fourfold weaker than galactose in binding to EIL and EAL, but that 6-deoxy-6-fluorogalactose is as potent as galactose.

5 21 1 The results indicate that the C-6 hydroxyl group serves as an acceptor of a weak hydrogen bond which contributes a free energy of about 3.3 kj mol-' for EIL and EAL (Table 2). Therefore, the C-6 hydroxyl group is hydrogen-bonded to an uncharged residue in both lectins. The data for RCA-I (Table 1) suggest the lack of involvement of the C-6 hydroxyl group in lectin binding. The results with APA show the presence of a hydrogen bond donated by the C-6 hydroxyl group to a neutral acceptor group of the lectin. The data with BSL-I lead to a similar conclusion; however, the free energy data (Table 2) indicate that the hydrogen bond is accepted by a charged residue of the protein. The higher affinity of fucose for APA and BSL-I compared to 6-deoxy-6-fluorogalactose (Table 1) may be due to binding of a solvent water molecule in the space provided by the absence of the C-6 hydroxyl group in fucose. These results also provide insight into the specificity of RCA-I binding to asparagine-linked carbohydrates. RCA-I has been found to bind complex-type oligosaccharides more strongly if the penultimate galactose residue is linked a(2-6) to a nonreducing terminal sialic acid residue [42, 431. On the other hand, the presence of an a(2-3)-linked sialic acid greatly reduces binding. This has been attributed to the higher rotational freedom of the 42-66) linkage [43]. The present results indicate that a free hydroxyl group at the C-3 position is necessary for binding to RCA-I, and that the C-6-hydroxyl group does not participate. Thus, substitution at the C-3 position, but not at the C-6 position, greatly reduces galactose binding to RCA-I, which explains the pattern of binding of the above complex-type oligosaccharides to the lectin [42,43]. Interestingly, galactitol fails to inhibit the hemagglutination by EIL, EAL, RCA-I, and APA, indicating that the ring structure of galactose is necessary for binding. However, BSL-I is inhibited by galactitol, albeit very weakly. Galacturonic acid failed to inhibit hemagglutination by all of the lectins, which may, in part, be due to the presence of a bulky group or anionic charge at the 6-position of the sugar. Comparisons with Con A ConA is a lectin isolated from the seeds of the jack bean which specifically binds glucose and mannose residues. Studies of the binding of deoxy- and fluorosugars to ConA have provided free energy data and donor/acceptor relationships for hydrogen bonding of the hydroxyl groups of MeaMan with the protein [7, 441. Thus, it is of interest to compare the pattern of hydrogen bonding of MeaMan to ConA with that of galactose binding to the galactose-specific lectins. Furthermore, assignments have been made of amino acid groups in ConA that are postulated to be hydrogen-bonded to the monosaccharide, based on computer-docking studies [45] that use the X-ray crystal coordinates of the protein [46]. Therefore, analysis of the binding data for ConA can be compared with the conclusions of the docking study. The free energies of binding of the hydroxyl groups of MeaMan to ConA and their donor/acceptor.properties are shown in Table 2, as determined from binding data with deoxy, fluoro, and 1-0-methyl derivatives of glucose and mannose [7]. Analysis of the data suggest that the C-1 (a-configuration) and C-2 (axial) hydroxyl groups accept hydrogen bonds from uncharged groups, the C-3 hydroxyl accepts a hydrogen bond from a charged group, and that both the C-4 and C-6 hydroxyl groups donate a hydrogen bond to charged groups of the protein (Table 2). These findings show similarit- ies and differences compared to the pattern of hydrogen bonding of galactose to the galactose-specific lectins. The C-1 hydroxyl groups of galactose and glucose/mannose form weak hydrogen bonds with EIL, EAL, and BSL-I, and ConA but are not hydrogen bonded to RCA-I and APA. The C-2 hydroxyl groups of the sugars form weak hydrogen bonds with EIL, EAL, and ConA, but a strong hydrogen bond with BSL-I. The hydrogen-bonding pattern is unique with RCA-I, which forms two bonds with the hydrogen and oxygen atoms of the C-2 hydroxyl group of galactose. The C-3 and C-4 hydroxyl groups of MeaMan form strong hydrogen bonds with ConA, which also occurs with galactose and the galactose-specific lectins, although the donor/acceptor properties at these positions are different for ConA and the galactose-specific lectins. The C-6 hydroxyl group of galactose is not hydrogen-bonded to RCA-I, forms weak hydrogen bonds with EIL and EAL, and binds strongly to BSL-I. The C-6 hydroxyl group of MeaMan also binds strongly to ConA. ConA possesses a donor group at this position, similar to BSL-I and APA. Interestingly, both ConA and the galactose-specific lectins (except APA for which the data are not conclusive) demonstrate strong hydrogen bonds with the C-3 and C-4 hydroxyl groups of the sugars via charged groups in the proteins, even though the C-4 hydroxyl groups in galactose and mannose/ glucose are configurationally different. For comparison, neutral - charged-pair hydrogen bonds between the C-3 and C-6 hydroxyl groups of glucose and glycogen phosphorylase have recently been reported [14]. Assignment of residues in ConA that are involved hydrogen bonding to MeaMan The hydrogen-bonding pattern of MeaMan binding to ConA and the donor/acceptor properties of the protein can be used, together with X-ray crystallographic data [46], to assign residues in the protein which may be involved in binding to the monosaccharide, and to check the conclusions of recent computer-docking studies [45]. Carver et al. [45] using such techniques concluded that the C-3 hydroxyl group of MeaMan forms a hydrogen bond to the carboxyl group of Asp-208 of ConA. However, the data in Table 2 shows that the C-3 hydroxyl group is a hydrogen-bond acceptor, and since the carboxyl group of Asp-208 is not likely to be a hydrogen-bond donor at neutral ph, this residue is unlikely to be hydrogen-bonded to the sugar at this position. The authors [45] also reported that the C-4 hydroxyl group of the sugar is hydrogen-bonded to the backbone -NH group of Arg-228, which is an uncharged polar group, and that the C-6 hydroxyl group is not hydrogen-bonded at all. The hydrogenbonding free energies for C-4 and C-6 hydroxyl groups (Table 2) are clearly incompatable with these assignments. Thus, alternative assignments of groups in the protein that are consistent with the data in Table 2 need to be considered. Hardman et al. [46] concluded from X-ray crystal data of ConA that Gly-98, Leu-99, Ser-168, Asp-208, Thr-226, and Arg-228 were involved in binding to MeaMan. Thus, two residues with charged side chains, viz. Asp-208 and Arg-22X, were initially assigned to be part of the monosaccharide-binding site of the protein. However, Carver et al. [45] noted that the carboxyl group of Asp-16 is at salt-bridging distance from the guanido group of Arg-228. Therefore, this carboxyl group could also be involved in hydrogen-bonding with the sugar. ConA thus possesses three charged groups which could form neutral -charged-pair hydrogen bonds with the C-3, C-4, and

6 212 C-6 hydroxyl groups of MeaMan. The protonated guanido group of Arg-228 is hydrogen-bond donor which could be bonded to the C-3 hydroxyl group of the sugar. The ionized carboxyl group of Asp-16, which is in close proximity to the side-chain group of Arg-228 [45], could accept a hydrogen bond from the C-4 hydroxyl group of the sugar. Finally, the ionized carboxyl group of Asp-208 which would function as a hydrogen-bond acceptor could be bonded with the C-6 hydroxyl group of the monosaccharide. Ser-168 and Thr-226 are the amino acid residues at the sugar-binding site [46] which could provide uncharged side chains for hydrogen-bonding to the C-1 and C-2 hydroxyl groups. Alternatively, these hydroxyl groups may be hydrogen-bonded to the - NH groups of the polypeptide backbone. The above assignments, though speculative, are consistent with the energetics and polarities of the hydrogen bonds between MeaMan and ConA listed in Table 2. In this regard, it is important to note that evidence has been presented from 13C-NMR studies that the a- and p- anomers of glucose bind with different orientations in the binding site of ConA [47]. These findings would affect the interactions of the hydroxyl groups of the sugar with the protein, which, in part, could account for some of the discrepancies between the results of the docking experiments [45] and those found in the present study. At present, similar assignments of the binding residues in the galactose-specific lectins are not possible due to the lack of crystallographic data. The authors wish to thank Dr I. J. Goldstein (University of Michigan, Ann Arbor, Michigan) and Dr N. K. Sinha (Bose Institute, Calcutta, India) for kind gifts of methyl glycosides of N-acetylgalactosamine, and APA, respectively. This work was supported by grant CA-I6054 from the National Cancer Institute, Department of Health, Education and Welfare, and Core Grant P30 CA from the same agency. The NMR facility was supported by Instrumentation grant I-S10-RR02309 from the National Institute of Health and DMB from the National Science Foundation. REFERENCES 1. Goldstein, I. J., Hughes, R. C., Monsigny, M., Osawa, T. & Sharon, N. (1980) Nature (Lond.) 285, Liener, I. E., Sharon, N. & Goldstein, I. J. (eds) (1986) The lecrins, Academic Press, New York. 3. Brandley, B. K. & Schnaar, R. L. (1986) J. Leukocyte Biol. 40, Goldstein, I. J. & Poretz, R. D. (1986) in The lectins (Liener, I. E., Sharon, N. & Goldstein, I. J., eds) pp , Academic Press, New York. 5. Lis, H. & Sharon, N. (1984) in Biofugy of carbohydrates (Ginsburg, V. & Robbins, P. W., eds) pp. 1-85, John Wiley, New York. 6. Stanley, P. (1980) in The biochemistry of glycoproteins and proteoglycans (Lennarz, w. J., ed.) pp , Plenum Press, New York. 7. Poretz, R. D. & Goldstein, I. J. (1970) Biochemistry 9, Alter, G. M. & Magnuson, J. A. (1974) Biochemistry 13, Murphy, L. A. & Goldstein, I. J. (1979) Biochemistry 18, Ochoa, J. L. (1981) J. Chromatogr. 215, Quiocho, F. A. & Vyas, N. K. (1984) Nature (Lond.) 310, Pauling, L. (1980) The nature of the chemical bond, 3rd edn, Cornell University Press, Ithaca N. Y. 13. Penglis, A. A. E. (1981) Adv. Carbohydr. Chem. Biochern. 38, Street, I. P., Armstrong, C. R. &Withers, S. G. (1986) Biochemistry, Murray-Rust, P., Stallings, W. C., Monti, C. T., Preston, R. K. & Glusker, J. P. (1983) J. Am. Chem. Soc. 105, Karipides, A. & Miller, C. (1984) J. Am. Chem. Soc. 106, Withers, S. G., Street, I. P. & Rettig, S. J. (1986) Can. J. C hem. 64, Thomas, P., Bessel, E. M. & Westwood, J. H. (1974) Biuchem. J. 139, Machado de Domenech, E. E. & Sols, A. (1980) FEBS Lett 119, Kasumi, T., Tsumuraya, Y., Brewer, C. F., Kersters-Hilderson, H., Claeyssens, M. & Hehre, E. J. (1987) Biochemistry 26, Ittah, Y. & Glaudemans, C. P. J. (1981) Carhohydr. Res. 95, Rees, W. D. & Holman, G. D. (1981) Biochim. Biophys. Actu Fersht, A. R., Shi, J.-P., Knill-Jones, J., Lowe, D. M., Wilkinson, A. J., Blow, D. M., Brick, P., Carter, P., Waye, M. M. Y. & Winter, G. (1985) Nature (Lond.) 314, Bedouelle, H., Carter, P., Waye, M. M. Y., Winter, G., Lowe. D. M., Wilkinson, A. J. & Fersht, A. R. (1985) Biochimie (Paris) 67, Wells, T. N. C. & Fersht, A. R. (1986) Biochemistry, Mares-Guia, M., Nelson, D. L. & Rogana, E. (1977) J. Am. Chem. Soc. 99, Bartlett, P. A. & Marlowe, C. K. (1987) Science ( Wash. DC) 235, Bhattacharyya, L., Das, P. K. & Sen, A. (1981) Arch. Biochrm. Biophys. 211, Roy, J., Som, S. & Sen, A. (1976) Arch. Biochem. Biophys. 174, Taylor, N. F. &Kent, P. W. (1985)J. Chem. Soc., Brimacombe, J. S., Foster, A. B., Hems, R. & Hall, L. D. (1968) Carbohydr. Res. 8, Adamson, J. & Marcus, D. M. (1970) Carbohydr. Rex 13, Marcus, D. M. & Westwood, J. H. (1971) Carbohydr. Res. 17, Osawa, T. & Matsumoto, I. (1972) Methods Enzymol. 28, Olsnes, S., Saltvedt, E. & Pihl, A. (1974) J. Bid. Chem. 249, Van Wauwe, J.-P., Loontiens, F. G. & De Bruync, C. K. (1973) Biochim. Biophys. Acta 313, Absar, N., Yamasaki, N. & Funatsu, G. (1986) J. Biochem. ( Tokyo) 99, Goldstein, I. J., Blake, D. A., Ebisu, S., Williams, T. J. & Murphy, L. A. (1981) J. Biol. Chem. 6, Finar, 1. L. (1985) Organic chemistry, vol. 1, Longman, Essex, England. 40. Jencks, W. P. (1981) Proc. Nut1 Acad. Sci. USA 768, Loontiens, F. G., Van Wauwe, J. P. & De Bruyne, C. K. (1975) Carbohydr. Res. 44, Baenziger, J. U. & Fiete, D. (1979) J. Bid. Chem. 4, Debray, H., Decout, D., Strecker, G., Spik, G. & Montreuil, J. (1981) Eur. J. Biochem. 117, Goldstein, I. J. (1975) Adv. Exp. Med. Biol. 55, Carver, J. P., Mackenzie, A. E. & Hardman, K. D. (1985) Biopolymew 24, Hardman, K. D., Agrawal, R. C. & Freiser, M. J. (1982) J. Mol. Biol. 157, Brewer, C. F., Sternlich, H., Marcus, D. M. & Grollman, A. P. (1973) Biochemistry 12,