Biology of animal lectins

Increasing diversity of animal lectin structures

The number of animal proteins known to recognize carbohydrates and the number of their biological roles continue to increase. Comparisons of primary structures show that some of the newly described lectins are akin to previously investigated lectins, whereas others represent new structural groups. Progress has been made in understanding structure-function relationships for several lectins in both the old and the new categories.

NKG2-C is a receptor on human natural killer cells that recognizes structures on K562 target cells

NKG2-C is a member of the recently discovered NKG2 family of genes and proteins, which are preferentially expressed on human natural killer (NK) cells. These potential NK cell receptors belong to a larger class of type II transmembrane proteins with a C-type lectin domain. We show here that NKG2-C is expressed as a 36-kDa glycoprotein by translation in vitro, recombinant expression and immunoprecipitation from a human NK cell clone. Further, a recombinant soluble NKG2-C-receptor binds specifically to K562 cells, which are target cells for NK cell killing, and to RPMI 8866 cells, which are feeder cells for NK cells; several other hematopoietic cell lines tested do not show any binding. The binding structures on the surface of K562 cells disappear, concomitant with a loss in susceptibility to killing when the cells are induced to differentiate with phorbol ester and Ca2+ ionophore. Our data suggest the presence of specific target molecules for NKG2-C on K562 cells, since overall glycosylation, Lewis X and Lewis Y structures, as well as the mucin-like CD43 molecule, do not change following induction of the cells. We propose that NKG2-C mediates a specific interaction of NK cells and their target cells with functional importance for NK cell killing.

X-ray crystal structures of animal lectins

Several important advances in the structure determination of animal lectins were made in the past year. The X-ray crystal structures of trimeric fragments of the human and rat mannose-binding proteins have defined for the first time the three-dimensional subunit organization of a multimeric C-type lectin. In addition, the structure of a galectin-biantennary oligosaccharide complex has provided a model for what might be biochemically relevant cross-linking interactions. Finally, in a novel variation on lectin cross-linking, independent carbohydrate-binding sites on basic fibroblast growth factor have been found to recognize opposing faces of a synthetic heparin/heparan sulphate fragment, leading to growth-factor polymerization.

A secretion inhibitory signal transduction molecule on mast cells is another C-type lectin

Secretion of inflammatory mediators by rat mast cells (line RBL-2H3) was earlier shown to be inhibited upon clustering a membrane glycoprotein by monoclonal antibody G63. This glycoprotein, named mast cell function-associated antigen (MAFA), was also shown to interfere with the coupling cascade of the type 1 Fc epsilon receptor upstream to phospholipase C gamma 1 activation by protein-tyrosine kinases. Here we report that the MAFA is expressed as both a monomer and a homodimer. Expression cloning of its cDNA shows that it contains a single open reading frame, encoding a 188-amino acid-long type II integral membrane protein. The 114 C-terminal amino acids display sequence homology with the carbohydrate-binding domain of calcium-dependent animal lectins, many of which have immunological functions. The cytoplasmic tail of MAFA contains a YXXL (YSTL) motif, which is conserved among related C-type lectins and is an essential element in the immunoreceptor tyrosine-based activation motifs. Finally, changes in the MAFA tyrosyl- and seryl-phosphorylation levels are observed in response to monoclonal antibody G63 binding, antigenic stimulation, and a combination of both treatments.

Molecular characterization of human CD94: a type II membrane glycoprotein related to the C-type lectin superfamily

Natural killer (NK) cells preferentially express several genes of the C-type lectin superfamily which have been implicated in the regulation of NK cell function. We demonstrate that CD94 is a type II membrane protein encoded by a unique gene of the C-type lectin superfamily. While homology of CD94 with the NK cell-associated NKR-P1 and NKG2 C-type lectin genes is limited to the structural motifs conserved in the carbohydrate recognition domain, all of these genes are on human chromosome 12, the syntenic of mouse chromosome 6, where genes of the NK complex (NKR-P1 and Ly-49) are located. An unexpected feature of CD94 is the essential absence of a cytoplasmic domain, implying that association with other receptors may be necessary for the function of this molecule.

Role of calcium(II) ions in the recognition of coagulation factors IX and X by IX/X-bp, an anticoagulant from snake venom

IX/X-bp, an anticoagulant protein isolated from the venom of the habu snake Trimeresurus flavoviridis, has a structure homologous to the carbohydrate-recognition domains of C-type (Ca(2+)-dependent) animal lectins, and it binds to the gamma-carboxyglutamic acid (Gla) domains of coagulation factors IX and X in a Ca(2+)-dependent fashion. In the present study, we elucidated the role of Ca2+ ions in this binding. The binding of 125I-labeled IX/X-bp to both coagulation factors required about 1 mM Ca2+ ions in this at pH 7.5. A decrease in the pH to 6.5 had a striking negative effect on the binding, and the Ca(2+)-requirement curve was shifted rightward. We investigated the binding of Ca2+ ions to IX/X-bp directly by equilibrium dialysis and identified two independent binding sites with different affinities. At pH 7.5, the apparent Kd values for these sites were 25 and 200 microM, respectively. When the pH was decreased to 6.5, the affinity of the high-affinity binding site was reduced only slightly but that of the low-affinity site was reduced considerably. Moreover, it was evident from observations of Ca(2+)-induced changes in the intrinsic fluorescence that IX/X-bp underwent a conformational change upon binding of Ca2+ ions.(ABSTRACT TRUNCATED AT 250 WORDS)

Rat sperm galactosyl receptor: purification and identification by polyclonal antibodies raised against multiple antigen peptides

We have previously reported the purification of rats testis galactosyl receptor, an equivalent to the Ca(2+)-dependent (C-type) minor variant of rat hepatic lectin-2/3 (RHL-2/3). We now report the purification of galactosyl receptor from rat sperm and its immunolocalization in the intact rat testis and sperm by polyclonal antibodies prepared using multiple antigen peptides (MAP) as immunogens. Two MAP antigens (designated 27-mer and 28-mer), corresponding to amino acid sequences of the carbohydrate-recognition domain (galactose) and adjacent Ca(2+)-binding sites of RHL-2/3, were used for immunization. Anti-RHL-2/3, anti-p27, and anti-p28 sera crossreacted with rat hepatocyte RHL-2/3 and its rat testis and sperm equivalent, galactosyl receptor, purified by chromatofocusing followed by galactose-Hydropore-EP affinity chromatography. Neither anti-p27 nor anti-p28 sera cross-reacted with the major hepatocyte variant, RHL-1. A RHL-1-equivalent was not detected in rat testis and sperm. Immunofluorescence studies demonstrated that anti-p27 and anti-p28 sera recognize galactosyl receptor sites at the Sertoli cell-spermatogenic cell interface and on the dorsal surface of the sperm head, overlying the acrosome. The characteristic crescent-shaped immunoreactive pattern in sperm was lost after induction of the acrosome reaction. Further studies should determine whether antisera to MAP antigens 27-mer and 28-mer, corresponding to specific protein motifs, can serve as immunological probes for examining cell-cell interaction events during spermatogenesis and at fertilization.

Carbohydrate-binding property of peptide: N-glycanase from mouse fibroblast L-929 cells as evaluated by inhibition and binding experiments using various oligosaccharides

Carbohydrate binding to peptide: N-glycanase from mouse fibroblast L-929 cells (L-929 PNGase) and inhibition by oligosaccharides of its catalytic activity were studied. L-929 PNGase was found to bind strongly with oligosaccharides having triomannosido-N,N'-diacetyl-chitobiosyl (Man3GlcNAc2) structure (Kd = approximately 10 microM). This binding was inhibited by mannotriose (Man3; Man alpha 1-->3[Man alpha 1-->6]Man) but not by N,N'-diacetylchitobiose (GlcNAc2; GlcNAc beta 1-->4GlcNAc). Scatchard analysis indicated that there exist two binding sites for Man3 on a homodimeric form of a 105-kDa subunit. Oligosaccharides having Man3GlcNAc2 structure were also shown to be strong inhibitors for the PNGase-catalyzed reaction (Ki = approximately 10 microM). The minimum structural requirements for inhibition of the PNGase activity were Man3 and GlcNAc2. Enzyme kinetic studies showed that the mechanism of inhibition by the oligosaccharides and Man3 fits well with a model wherein two inhibitor binding sites reside on L-929 PNGase. The conformity of Kd with IC50 values may be taken as an evidence for inhibition of the catalytic activity by the oligosaccharides and Man3 through the occupation of the binding sites with these molecules. On the other hand, inhibition by GlcNAc2 followed the simple competitive mode. Since the minimum substrate for the L-929 PNGase was shown to be Man beta 1-->4GlcNAc beta 1-->4GlcNAc beta 1-->peptide, GlcNAc2 may be directly accessible to the catalytic site in competition with substrate. Interestingly, alkylation of -SH group in L-929 PNGase caused complete loss of the catalytic activity, but the carbohydrate binding activity was completely retained, indicating that the catalytic site(s) is discriminated from the carbohydrate-binding sites in the active site of this enzyme. The carbohydrate-binding property seems to be unique to soluble PNGases from mammals and may be associated not only with regulation of the enzyme activity, but also with receptor and carrier functions for glycoconjugates in certain intracellular processes.

The differences in structural specificity for recognition and binding between asialoglycoprotein receptors of liver and macrophages

The Gal/GalNAc-specific lectin on the surface of rat peritoneal macrophages (macrophage asialoglycoprotein binding protein, M-ASGP-BP), which consists of a single polypeptide chain of 42 kDa, can form a homo-oligomeric receptor exhibiting high affinity for asialoorosomucoid (ASOR) [Ozaki K., Ii M., Itoh N., Kawasaki T. (1992) J Biol Chem 267: 9229-35]. In this study, the binding affinity of M-ASGP-BP was studied by using a series of synthetic or natural glycosides as inhibitors of 125I-ASOR binding to recombinant M-ASGP-BP expressed on COS-1 cells (rM-ASGP-BP), and the results were compared with those of human hepatic lectin (HHL) on Hep G2 cells. Clustering of multiple Gal (or GalNAc) residues increased the binding affinity to M-ASGP-BP as well as to HHL. In contrast to HHL and other mammalian hepatic lectins, rM-ASGP-BP bound Gal residues tighter than GalNAc residues. A galactose-terminated triantennary N-glycoside, having one N-acetyl-lactosamine unit on the 6 branch and two N-acetyl-lactosamine units on the 3 branch of the trimannosyl core structure, showed affinity enhancement of approximately 10(5) over a monovalent ligand for HHL, while the same glycopeptide showed enhancement of about 2000-fold for rM-ASGP-BP. These results suggest that spatial arrangements of sugar combining sites and subunit organization of macrophage and hepatic lectins are different.

Galectin-3 is expressed in the notochord, developing bones, and skin of the postimplantation mouse embryo

The galectins are a family of low molecular weight, calcium-independent mammalian carbohydrate binding proteins that exhibit specificity for beta-galactoside derivatives. We have examined the expression pattern of galectin-3 in the developing mouse embryo by in situ hybridisation and immunohistochemistry. In the embryo proper, galectin-3 message and protein are first detected in notochord, starting from 8.5 days post coitum (dpc), and persist until this structure disappears. Galectin-3 is later found in cartilage primordia and in developing skin from 13.5 dpc. This very restricted and dynamic pattern suggests that galectin-3 may participate in the establishment and/or maintenance of notochord as well as the formation of cartilage and differentiation of skin. Finally, we find that galectin-3, which is identical to the macrophage marker Mac-2, is also expressed in embryonic macrophages.

Pulmonary surfactant protein A (SP-A) is expressed by epithelial cells of small and large intestine

Surfactant protein A (SP-A) is the most abundant protein associated with phospholipids in pulmonary surfactant. There are several lines of evidence that pulmonary and gastrointestinal epithelium produce closely related surface-active materials, although the presence of SP-A in gastrointestinal tract has so far not been reported. Indirect immunofluorescence experiments using different antibodies raised against rat pulmonary SP-A showed that some jejunal and colonic but not gastric epithelial cells positively stained for SP-A. Analysis of the proteins in cell lysates from rat small intestine and colon studied by Western blot revealed several immuno-reactive bands, including the characteristic triplet of 26-, 32-, and 38-kDa monomeric proteins, less strongly labeled than in lung cells, and higher molecular mass forms of 66 and 120 kDa also present in lung cells. The 66- and 120-kDa bands displayed the expected isoelectric pH of SP-A after two-dimensional electrophoresis. Alkylation induced conversion of the 120-kDa form (almost completely) and the 66-kDa form (partly) into the 26-38-kDa monomeric species. The presence of SP-A mRNA in rat stomach, small intestine, and colon was then searched for by conventional cDNA/reverse transcriptase-polymerase chain reaction. Products of appropriate size (372 base pairs) identical to that of pulmonary tissue were amplified in small intestine and colon but not in stomach or in other tissues used as controls. Cloning and sequencing of rat colon SP-A cDNA revealed the same sequence as the one reported for rat lung SP-A. Furthermore, analysis of the transcriptional initiation site of SP-A gene in colon by anchored-polymerase chain reaction showed that transcription was initiated at the same site in both colon and lung. These data, which demonstrate that small intestine and colon express SP-A constitutively and that this protein is present in some epithelial cells, extend the concept of intestinal surfactant and underline its close relationships to pulmonary surfactant.

The receptor DEC-205 expressed by dendritic cells and thymic epithelial cells is involved in antigen processing

Dendritic cells and thymic epithelial cells perform important immunoregulatory functions by presenting antigens in the form of peptides bound to cell-surface major histocompatibility complex (MHC) molecules to T cells. Whereas B cells are known to present specific antigens efficiently through their surface immunoglobins, a comparable mechanism for the capture and efficient presentation of diverse antigens by dendritic cells and thymic epithelial cells has not previously been described. We show here that their antigen-presentation function is associated with the high-level expression of DEC-205, an integral membrane protein homologous to the macrophage mannose receptor and related receptors which are able to bind carbohydrates and mediate endocytosis. DEC-205 is rapidly taken up by means of coated pits and vesicles, and is delivered to a multivesicular endosomal compartment that resembles the MHC class II-containing vesicles implicated in antigen presentation. Rabbit antibodies that bind DEC-205 are presented to reactive T-cell hybridomas 100-fold more efficiently than rabbit antibodies that do not bind DEC-205. Thus DEC-205 is a novel endocytic receptor that can be used by dendritic cells and thymic epithelial cells to direct captured antigens from the extracellular space to a specialized antigen-processing compartment.

The homing of hematopoietic stem cells to the bone marrow

In this article, the author discusses some of the most notable aspects of the work of Mehdi Tavassoli and others on the homing of intravenously transplanted hematopoietic stem cells to the marrow. It is well-recognized that homing of stem cells is a highly selective process, perhaps similar to the homing of lymphocytes to lymphoid tissues. The nature of the selectivity of stem cell homing is unclear, however, and may be mediated through a specific homing receptor or through a method of selective capture, retainment, or survival advantage afforded by the marrow. In this article, the focus is on current research in the identification of a specific homing receptor, the potential regulation of such a receptor by cytokines, the homing phenomenon as a multi-step process, and secondary adhesive interactions mediated by known adhesive molecules. These interactions may serve to strengthen the initial recognition and engraftment of stem cells within the hematopoietic compartment of the marrow.

Carbohydrate recognition by a natural killer cell receptor, Ly-49C

Ly-49 represents a family of type II transmembrane proteins containing C-type lectin domains. At least two members of the Ly-49 family, namely Ly-49A and Ly-49C, are expressed by distinct subsets of natural killer cells and bind to class I major histocompatibility complex antigens on the surface of target cells. In this report we have established that Ly-49C mediates carbohydrate recognition. The sulfated glycans fucoidan, lambda-carrageenan, and dextran sulfate were found to be potent inhibitors of Ly-49C-mediated cell adhesion, whereas other polysaccharides of similar size, charge, or sulfate content were noninhibitory. All of the polysaccharides which inhibited Ly-49C adhesion also blocked the binding of the antibody 5E6 to Ly-49C-expressing COS cells, confirming the direct protein-carbohydrate interaction. The enzymatic removal of specific carbohydrates from the target cell surface has shown that Ly-49C-mediated adhesion is not sialic acid-dependent, but is significantly decreased following fucosidase treatment. These results suggest an important role for carbohydrate recognition by natural killer cell receptors.

X-ray crystal structure of the soybean agglutinin cross-linked with a biantennary analog of the blood group I carbohydrate antigen

Soybean agglutinin (SBA) (Glycine max), which is a tetrameric GalNAc/Gal-specific lectin, has recently been reported to form unique, highly organized cross-linked complexes with a series of naturally occurring and synthetic multiantennary carbohydrates with terminal GalNAc or Gal residues [Gupta, D., Bhattacharyya, L., Fant, J., Macaluso, F., Sabesan, S., & Brewer, C. F. (1994) Biochemistry 33, 7495-7504]. In order to elucidate the nature of these complexes, the X-ray crystallographic structure of SBA cross-linked with a biantennary analog of the blood group I carbohydrate antigen is reported. The structure reveals that lattice formation is promoted uniquely by the bridging action of the bivalent pentasaccharide (beta-LacNAc)2Gal-beta-R, where R is -O(CH2)5COOCH3 and the beta-LacNAc moieties are linked to the 2 and 6 positions of the core Gal. The structure of SBA complexed with the synthetic biantennary pentasaccharide has thus been determined by molecular replacement techniques and refined at 2.6 A resolution to an R value of 20.1%. The crystals are hexagonal with a P6(4)22 space group, which differs significantly from that of crystals of the free protein. In the structure, each monomeric asymmetric unit contains a Man9 oligomannose-type chain at Asn 75, with only the first two GlcNAc residues visible. The overall tertiary structure of the SBA subunit is similar to that of other legume lectins as well as certain animal lectins. However, the dimer interface in the SBA tetramer is unusual in that only one complete peptide chain is sterically permitted, thus requiring juxtapositioning of one C-terminal fragmented subunit together with an intact subunit. Association between SBA tetramers involves binding of the terminal Gal residues of the pentasaccharide at identical sites in each monomer, with the sugar cross-linking to a symmetry-related neighbor molecule. The cross-linking pentasaccharide is in a conformation that possesses a pseudo-2-fold axis of symmetry which lies on a crystallographic 2-fold axis of symmetry of the lattice. Hence, the symmetry properties of the bivalent oligosaccharide as well as the lectin are structural determinants of the lattice. The results are discussed in terms of multidimensional carbohydrate-lectin cross-linked complexes, as well as the signal transduction properties of multivalent lectins.

Role for a synapse-specific carbohydrate in agrin-induced clustering of acetylcholine receptors

Lectins such as VVA-B4, which bind N-acetylgalactosaminyl (GalNAc)-terminated saccharides, selectively stain the neuromuscular junction, thus defining a synapse-specific carbohydrate. In seeking roles for this carbohydrate, we asked whether VVA-B4 affected the distribution of acetylcholine receptors (AChRs) on cultured muscle cells. We found that incubation of myotubes with VVA-B4 induced formation of AChR clusters and potentiated the effect of a nerve-derived clustering factor, agrin. Additional experiments implicated GalNAc-terminated glycoconjugates as modulators of agrin-induced AChR clustering. Enzymatic removal of GalNAc residues or treatment with a multivalent protein-GalNAc conjugate blocked agrin-induced clustering, whereas enzymatic unmasking of additional GalNAc residues induced clustering in the absence of added agrin. Moreover, incubation with agrin led to redistribution of VVA-B4-binding material on myotubes. Together, these results suggest that agrin-induced clustering of AChRs involves a GalNAc-dependent step.

The molecular chaperone calnexin binds Glc1Man9GlcNAc2 oligosaccharide as an initial step in recognizing unfolded glycoproteins

Calnexin is a molecular chaperone that resides in the membrane of the endoplasmic reticulum. Most proteins that calnexin binds are N-glycosylated, and treatment of cells with tunicamycin or inhibitors of initial glucose trimming steps interferes with calnexin binding. To test if calnexin is a lectin that binds early oligosaccharide processing intermediates, a recombinant soluble calnexin was created. Incubation of soluble calnexin with a mixture of Glc0-3Man9GlcNAc2 oligosaccharides resulted in specific binding of the Glc1Man9GlcNAc2 species. Furthermore, Glc1Man5-7GlcNAc2 oligosaccharides bound relatively poorly, suggesting that, in addition to a requirement for the single terminal glucose residue, at least one of the terminal mannose residues was important for binding. To assess the involvement of oligosaccharide-protein interactions in complexes of calnexin and newly synthesized glycoproteins, alpha 1-antitrypsin or the heavy chain of the class I histocompatibility molecule were purified as complexes with calnexin and digested with endoglycosidase H. All oligosaccharides on either glycoprotein were accessible to this probe and could be removed without disrupting the association with calnexin. Furthermore, the addition of 1 M alpha-methyl glucoside or alpha-methyl mannoside had no effect on complex stability. These findings suggest that once complexes between calnexin and glycoproteins are formed, oligosaccharide binding does not contribute significantly to the overall interaction. However, it is likely that the binding of Glc1Man9GlcNAc2 oligosaccharides is a crucial event during the initial recognition of newly synthesized glycoproteins by calnexin.

The Merck Frosst Award Lecture 1994/La conference Merck Frosst 1994. Calnexin: a molecular chaperone with a taste for carbohydrate

Calnexin is an integral membrane protein of the endoplasmic reticulum (ER) that binds transiently to a wide array of newly synthesized membrane and secretory proteins. It also exhibits prolonged binding to misfolded or incompletely folded proteins. Recent studies have demonstrated that calnexin functions as a molecular chaperone to facilitate the folding and assembly of proteins in the ER. It is also a component of the quality control system that prevents proteins from progressing along the secretory pathway until they have acquired proper tertiary or quaternary structure. Most proteins that are translocated into the ER are glycosylated at Asn residues, and calnexin's interactions are almost exclusively restricted to proteins that possess this posttranslational modification. The preference for glycoproteins resides in calnexin's ability to function as a lectin with specificity for the Glc1Man9GlcNAc2 oligosaccharide, an early intermediate in the processing of Asn-linked oligosaccharides. Calnexin also has the capacity to bind to polypeptide segments of unfolded glycoproteins. Available evidence suggests that calnexin utilizes its lectin property during initial capture of a newly synthesized glycoprotein and that subsequent association (and chaperone function) is mediated through polypeptide interactions. Unlike other molecular chaperones that are soluble proteins, calnexin is an intrinsic component of the ER membrane. Its unique ability to capture unfolded glycoproteins through their large oligosaccharide moieties may have evolved as a means to overcome accessibility problems imposed by being constrained within a lipid bilayer.

Galectin-8. A new rat lectin, related to galectin-4

A protein of 35 kDa which has the characteristic properties of galectins (S-type lectins) was cloned from rat liver cDNA expression library. Since names for galectins 1-7 were already assigned, this new protein was named galectin-8. Three lines of evidence demonstrate that galectin-8 is indeed a novel galectin: (i) its deduced amino acid sequence contains two domains with conserved motifs that are implicated in the carbohydrate binding of galectins, (ii) in vitro translation products of galectin-8 cDNA or bacterially expressed recombinant galectin-8 are biologically active and possess sugar binding and hemagglutination activity, and (iii) a protein of the expected size (34 kDa) that binds to lactosyl-Sepharose and reacts with galectin-8-specific antibodies is present in rat liver and comprises approximately 0.025% of the total Triton X-100-soluble hepatic proteins. Overall, galectin-8 is structurally related (34% identity) to galectin-4, a soluble rat galectin with two carbohydrate-binding domains in the same polypeptide chain, joined by a link peptide. Nonetheless, several important features distinguish these two galectins: (i) Northern blot analysis revealed that, unlike galectin-4 that is confined to the intestine and stomach, galectin-8 is expressed in liver, kidney, cardiac muscle, lung, and brain; (ii) unlike galectin-4, but similar to galectins-1 and -2, galectin-8 contains 4 Cys residues; (iii) the link peptide of galectin-8 is unique and bears no similarity to any known protein; (iv) the N-terminal carbohydrate-binding region of galectin-8 contains a unique WG-E-I motif instead of the consensus WG-E-R/K motif implicated as playing an essential role in sugar-binding of all galectins. Together with galectin-4, galectin-8 therefore represents a subfamily of galectins consisting of a tandem repeat of structurally different carbohydrate recognition domains within a single polypeptide chain.

Synthesis of an endogeneous lectin, galectin-1, by human endothelial cells is up-regulated by endothelial cell activation

The pattern of expression of an endogenous lectin, galectin-1, was examined in human lymphoid tissue. Galectin-1 was detected in the endothelial cells lining specialized vessels, termed high endothelial venules, in activated lymphoid tissue, but not in a resting lymph node. Cultured endothelial cells (human aortic and umbilical vein endothelial cells (HAECs and HUVECs)) expressed galectin-1. Activation of the cultured endothelial cells increased the level of galectin-1 expression, as determined by ELISA. Northern blot analysis and high throughput cDNA sequencing. These results suggest that galectin-1 expressed by endothelial cells may bind to and affect the trafficking of cells emigrating from blood into tissues.

Immunohistochemical and glycohistochemical localization of the beta-galactoside-binding S-type lectin in human placenta

The localization of the beta-galactoside binding lectin was studied immunohistochemically on acetone-fixed cryostat sections of full-term placental tissue using a biotinylated monoclonal antibody and glycohistochemically applying biotinylated asialofetuin and lactosylated bovine serum albumin. On blots of placental tissue lysates the lectin is recognized by the biotinylated lactosylated bovine serum albumin. The glycoconjugate recognition of the lectin on blots was inhibited in the presence of 0.1 M lactose showing the specificity of the interactions. The anti-lectin monoclonal antibody stained syncytiotrophoblast and trophoblastic cells. Both reagents applied for glycohistochemistry stained syncytiotrophoblast and trophoblastic cells of placental villi and the trophoblastic layer of extraplacental membranes. A strong uniform cytoplasmic staining was characteristic for syncytiotrophoblast and to a lower extent for cytotrophoblastic cells. The localization of the lectin is discussed with respect to a possible immunosuppressive function.

Trimeric structure of a C-type mannose-binding protein

BACKGROUND

Mannose-binding proteins (MBPs) are C-type (Ca(2+)-dependent) animal lectins found in serum. They recognize cell-surface oligosaccharide structures characteristic of pathogenic bacteria and fungi, and trigger the neutralization of these organisms. Like most lectins, MBPs display weak intrinsic affinity for monovalent sugar ligands, but bind avidly to multivalent ligands.

RESULTS

We report physical studies in solution and the crystal structure determined at 1.8 A Bragg spacings of a trimeric fragment of MBP-A, containing the carbohydrate-recognition domain (CRD) and the neck domain that links the carboxy-terminal CRD to the collagen-like portion of the intact molecule. The neck consists of a parallel triple-stranded coiled coil of alpha-helices linked by four residues to the CRD. The isolated neck peptide does not form stable helices in aqueous solution. The previously characterized carbohydrate-binding sites lie at the distal end of the trimer and are separated from each other by 53 A.

CONCLUSIONS

The carbohydrate-binding sites in MBP-A are too far apart for a single trimer to bind multivalently to a typical mammalian high-mannose oligosaccharide. Thus MBPs can recognize pathogens selectively by binding avidly only to the widely spaced, repetitive sugar arrays on pathogenic cell surfaces. Sequence alignments reveal that other C-type lectins are likely to have a similar oligomeric structure, but differences in their detailed organization will have an important role in determining their interactions with oligosaccharides.

Expression of different members of the Ly-49 gene family defines distinct natural killer cell subsets and cell adhesion properties

The murine Ly-49 antigen belongs to a family of type II transmembrane molecules containing lectin-like domains. The original member of this family, Ly-49A, has been demonstrated to be expressed by a subpopulation of natural killer (NK) cells, bind certain class I major histocompatibility complexes (MHC), and act as a negative regulator of lytic activity. The expression patterns and functional activities of the other Ly-49s, however, is unknown. We extended the study of this family by isolating cDNAs encoding two new Ly-49 molecules. The reactivity of these and previously identified Ly-49 molecules with NK antibodies was tested in a COS cell expression system. YE1/32 and YE1/48 bound Ly-49A specifically, and 5E6 reacted only with Ly-49C. Three-color flow cytometric analysis demonstrated Ly-49A and Ly-49C expression defines complex, but distinct subsets within NK1.1+ cells. Some NK1.1-CD3+ as well as NK1.1-CD3- cells expressing Ly-49A or C were also detected. Analysis of MHC congenic strains of mice demonstrated that YE1/32+ and YE1/48+ NK cells are not deleted, as has been shown with the Ly-49A mAb A1. Furthermore, COS cells transfected with Ly-49A bound H-2d and H-2k cell lines, whereas Ly-49C transfectants bound H-2d, H-2k, H-2b, and H-2s. The antibodies 5E6 and 34-1-2S (anti-class I MHC) inhibited the binding of Ly-49C to an H-2s cell line. These results imply that the NK cell antigens Ly-49A and C bind to different repertoires of class I MHC molecules.

Identification and purification of a stress associated nuclear carbohydrate binding protein (M(r) 33,000) from rat liver by application of a new photoreactive carbohydrate probe

A photoreactive alpha-D-glucose probe has been designed for the specific detection of carbohydrate binding proteins (CBPs). The probe consists of four parts: (i) an alpha-D-glucose moiety; (ii) the digoxigenin tag; (iii) the photoreactive cross-linker; and (iv) the lysyl-lysine backbone. After incubation with lectins in the dark, the probe is activated and cross-linked to the CBPs after being treated by several flashes. Using this method we have identified a new alpha-D-glucose CBP of M(r) = 33,000, termed CBP33, in the nuclei of rats exposed to transient immobilization stress. Monoclonal antibodies were raised against the partially purified protein and subsequently used to enrich CBP33. It was purified (> 2400-fold) to apparent homogeneity from a 0.6 M nuclear salt extract by two subsequent affinity chromatography steps (antibody-affinity as well as alpha-D-glucose affinity column).

Human mannose-binding protein carbohydrate recognition domain trimerizes through a triple alpha-helical coiled-coil

Human mannose-binding protein is a hexamer of trimers with each subunit consisting of an amino-terminal region rich in cysteine, 19 collagen repeats, a 'neck', and a carbohydrate recognition domain that requires calcium to bind ligand. A 148-residue peptide, consisting of the 'neck' and carbohydrate recognition domains forms trimers in solution and in crystals. The structure of this trimeric peptide has been determined in two different crystal forms. The 'neck' forms a triple alpha-helical coiled-coil. Each alpha-helix interacts with a neighbouring carbohydrate recognition domain. The spatial arrangement of the carbohydrate recognition domains suggest how MBP trimers form the basic recognition unit for branched oligosaccharides on microorganisms.

Fusion between large phagocytic vesicles: targeting of yeast and other particulates to phagolysosomes that shelter the bacterium Coxiella burnetii or the protozoan Leishmania amazonensis in Chinese hamster ovary cells

This report examines the fusion of phagocytic vesicles with the large phagolysosome-like vacuoles induced in Chinese hamster ovary cells by the bacterium Coxiella burnetti or the Protozoan flagellate Leishmania amazonensis. Infection by these organisms is compatible with cell survival and multiplication. Fusion was inferred from the transfer of microscopically identifiable particles from donor to target vesicles. Donor vesicles contained heat-killed yeast, zymosan, beta-glucan or latex beads taken up by the host cells. Yeast and zymosan were also coated with Concanavalin A to increase their uptake by the cells (Goldman, R., Exp. Cell Res. 104, 325–334, 1977). Particle localization, routinely ascertained by phase-contrast microscopy, was confirmed by confocal laser fluorescence and by transmission electron microscopy. Coxiella vacuoles admitted all the particles tested and transfer took place whether the particles were given to the cells prior to or after infection. Transfer of uncoated or Concanavalin-A-coated yeast or zymosan was dependent on the number of particles ingested and on the incubation period (between 2 and 24 hours). Furthermore, the transfer step was quite efficient, since over 85% of the particles ingested entered Coxiella vacuoles at all particle to cell ratios examined. The fraction of uncoated or Concanavalin-A-coated yeast or zymosan transferred to Leishmania vacuoles was consistently lower and diminished at higher particle loads. In addition, only rarely did latex beads enter these vacuoles. The models proposed may be useful for the delineation of biochemical and molecular mechanisms involved in the fusion of large phagocytic vesicles and the modulation of the latter by cellular and pathogen-derived signals.

Does an animal peptide: N-glycanase have the dual role as an enzyme and a carbohydrate-binding protein?

Recently, we have reported purification and characterization of a de-N-glycosylating enzyme, peptide: N-glycanase (PNGase) found in C3H mouse fibroblast L-929 cells, and designated L-929 PNGase [Suzuki T, Seko A, Kitajima K, Inoue Y, Inoue S (1994) J Biol Chem 269, 17611-18]. The unique properties of L-929 PNGase are that the enzyme had a high affinity to the substrate glycopeptide (e.g. Km = 114 microM for fetuin derived glycopentapeptide) and that the PNGase-catalysed reaction is strongly inhibited by the released free oligosaccharides but not by the free peptides formed, suggesting that L-929 PNGase is able to bind to a certain type of carbohydrate chain. In this study, we report the new findings of the mannan-binding property of L-929 PNGase: the de-N-glycosylating enzyme activity of L-929 PNGase was inhibited by yeast mannan and triomannose, Man alpha 1-->3(Man alpha 1-->6)Man, but not by mannose and alpha-methyl-D-mannoside. Furthermore, L-929 PNGase was revealed to bind to the glycan moiety of yeast mannan by using mannan-conjugated Sepharose 4B gel as a ligand, suggesting that L-929 PNGase could serve not only as an enzyme but also as a carbohydrate recognition protein in vivo. Such 'dual' properties found for animal-derived L-929 PNGase are unique and are not shared with other previously characterized plant- and bacterial-origin PNGases--PNGase A and PNGase F, respectively.

The CD69 receptor: a multipurpose cell-surface trigger for hematopoietic cells

CD69 was initially described as being restricted to recently activated lymphoid cells, but is now known to be expressed on the surface of all hematopoietically derived leukocytes. Crosslinking of CD69 generates intracellular signals in all cell lineages studied, both mouse and human, and results in a variety of cellular end responses. Since a specific ligand has not yet been identified, a definite functional identity for CD69 is still missing. However, as discussed here by Roberto Testi and colleagues, the broad expression of CD69 and its conserved ability to generate intracellular signals suggests a general role for the CD69 receptor in the biology of hematopoietic cells.

Cell type-dependent alterations of binding of synthetic blood group antigen-related oligosaccharides in lung cancer

Blood group antigen-related oligosaccharides have been implicated in growth regulation, cell mobility control and adhesion; we are therefore interested in the localization of receptors for these oligosaccharides in tumour cells. Labelled neoglycoconjugates that carry synthetic sugar structures are suitable tools to determine: whether such binding sites are present in human lung cancer; whether structural alterations of the glycoligand part will affect extent of binding; and whether cell type-associated alterations can be detected. Sections from 121 cases of lung cancer, representing small cell and non-small cell lung carcinoma, mesothelioma and metastases from extrapulmonary primary carcinomas were used to study the binding of nine synthetic AH- and Le-related oligosaccharides. Probes with fucose-alpha 1-3/4-N-acetylglucosamine-beta 1-R, an A-like disaccharide and 3'-sulfated galactose as ligand appear to bind less well to small cell than to non-small cell lung cancer cases, whereas Lec-disaccharide distinguishes mesothelioma from metastatic carcinoma. The latter ligand, A-like disaccharide and H (type III)-like trisaccharide exhibit evident cell type-associated differences in extent of binding. Thus, tailor-made neoglycoconjugates constitute a promising class of histopathological tools that warrants further study.

Synthetic glycosylation of proteins using N-(beta-saccharide) iodoacetamides: applications in site-specific glycosylation and solid-phase enzymic oligosaccharide synthesis

A simple and efficient synthetic glycosylation method suitable for use in solid-phase enzymic oligosaccharide synthesis and site-specific glycosylation of recombinant proteins to produce defined glycoforms is described. This strategy utilizes N-(beta-saccharide) haloacetamides for attaching oligosaccharides specifically to cysteine residues of proteins in solution to form neoglycoproteins. The alkylation reaction was tested using N-(beta-chitotriose) bromoacetamide and an unprotected synthetic hexapeptide containing a single cysteine residue. The glycosylated product was confirmed by amino acid and hexosamine analyses as well as laser desorption mass spectrometry. Similarly N-(beta-chitotriose) iodoacetamide was covalently linked to non-reduced BSA to produce a defined glycoform of this protein. The specific attachment of chitotriose at the single cysteine residue in non-reduced serum albumin was suggested by Ellman's assay for free thiols. This was verified by amino acid sequencing of tryptic glycopeptide derived from this neoglycoprotein. Multiple sugar attachment was accomplished using fully reduced serum albumin as demonstrated by the formation of two neoglycoproteins using iodoacetamide derivatives of galactose beta 1-3-N-acetylgalactosamine (Gal beta 1-3GalNAc) and the major xylose/fucose-class plant-type oligosaccharide of horseradish peroxidase. These two neoglycoproteins with an average of 18-21 sugar residues attached were assayed positively for binding to peanut agglutinin and a sugar-specific anti-(horseradish peroxidase) monoclonal antibody YZ1/2.23 respectively. Sialylation of the neoglycoprotein containing Gal beta 1-3GalNAc was accomplished using alpha-2,3-sialyltransferase and radiolabelled CMP-N-acetylneuraminic acid. Significantly, glycan attachment using this conjugation method is reversible as demonstrated by the release of oligosaccharides from these two neoglycoproteins using hydrazinolysis. Therefore this method could provide invaluable reagents for many glycobiological studies.

Analysis of carbohydrate-protein interactions with synthetic N-linked neoglycoconjugate probes

Recently we have describe a simple efficient chemical method of generating an asparagine side-chain linker with beta-stereochemistry at the anomeric position of neutral oligosaccharides. We now report the 1-N-glycyl beta-derivatization of sialylated saccharides. Several neoglycoconjugates formed using these N-linked inter-mediates were investigated for their usefulness in probing carbohydrate-protein interactions. First, biotinyl derivatives of two xylose/fucose class plant-type oligosaccharides purified from horseradish peroxidase were effective in demonstrating the carbohydrate specificity of polyclonal anti-(horseradish peroxidase) antibodies. Secondly, a fluorescein-labelled asialo- and digalactosylated biantennary complex sugar was synthesized and shown to bind to a Ricinus communis agglutinin column. This galactose-specific recognition was abolished by treating this fluorescein-labelled oligosaccharide with jack bean beta-galactosidase. Finally, two 1-N-glycyl beta-saccharide derivatives were modified with thiophosgene to form their corresponding isothiocyanate derivatives. Coupling of these isothiocyanate derivatives of sugars to BSA, amino-derivatized polystyrene plates and glass-fibre discs resulted in multiple sugar presentation. The binding of an anti-N-acetylglucosamine monoclonal antibody to N,N'-diacetylchitobiose residues presented on BSA and solid supports was shown by e.l.i.s.a. Similarly the binding of concanavalin A to asialo-, agalactosylated biantennary complex oligosaccharide residues attached to BSA was demonstrated by a competitive e.l.i.s.a. Our results demonstrate that N-linked neoglycoconjugates could be made readily available and they are valuable tools for the detailed analyses of carbohydrates and carbohydrate-binding proteins.