Insights into the molecular basis of leukocyte tethering and rolling revealed by structures of P- and E-selectin bound to SLe(X) and PSGL-1

P-, E- and L-selectin constitute a family of cell adhesion receptors that mediate the initial tethering and rolling of leukocytes on inflamed endothelium as a prelude to their firm attachment and extravasation into tissues. The selectins bind weakly to sialyl Lewisx (SLe(X))-like glycans, but with high-affinity to specific glycoprotein counterreceptors, including PSGL-1. Here, we report crystal structures of human P- and E-selectin constructs containing the lectin and EGF (LE) domains co-complexed with SLe(X). We also present the crystal structure of P-selectin LE co-complexed with the N-terminal domain of human PSGL-1 modified by both tyrosine sulfation and SLe(X). These structures reveal differences in how E- and P-selectin bind SLe(X) and the molecular basis of the high-affinity interaction between P-selectin and PSGL-1.     

A cholesterol-independent membrane microdomain serves as a functional counter-receptor for E-selectin at the Colo201 cell surface and initiates signalling on E-selectin binding

The present study demonstrates that the functional counter-receptors for E-selectin at the cell surface of Colo201 human colon cancer cells are localized in detergent-insoluble membrane microdomains (DIM). Following isolation of counter-receptors from whole cell lysates using E-selectin-coupled magnetic beads followed by sucrose density gradient separation, both sialyl Lewis a (SLe(a))- and sialyl Lewis x (SLe(x))-carrying glycoproteins which had bound to the E-selectin-beads were distributed in detergent-soluble fractions as well as DIM. In contrast, following isolation of counter-receptors directly from the cell surface, SLe(a)-carrying glycoproteins which had bound to E-selectin-beads at the cell surface were localized only in DIM, together with a Src family kinase, Lyn, while SLe(x)-carrying glycoproteins were not detected in any fraction. The counter-receptors were distributed in a diffuse pattern on the cell surface but clustered following E-selectin binding, leading to the subsequent phosphorylation of extracellular signal-regulated kinase (ERK). Treatment of the cells with methyl-beta-cyclodextrin, a cholesterol-depleting drug, had little effect on either the association of SLe(a)-carrying glycoproteins and Lyn with the domain or ERK phosphorylation. Thus, the functional counter-receptors and Lyn are co-localized in a cholesterol-independent microdomain and create a physiological domain ('glycosynapse') at the cell surface that initiates signalling in cancer cells upon binding to E-selectin.     

Inhibition of selectin-dependent tumor cell adhesion to endothelial cells and platelets by blocking O-glycosylation of these cells.

Expression of sialosyl-Le(x) (SLe(x)) and sialosyl-Le(a) (SLe(a)) on tumor cell lines HL60, Colo205, and U937 was greatly suppressed by application of benzyl-alpha-GalNAc for inhibition of O-linked carbohydrate chain extension, which resulted in reduced adhesion of tumor cells to activated endothelial cells or platelets mediated by ELAM-1 (E-selectin) or GMP-140 (P-selectin). Inhibitors or modifiers of N-glycosylation had no effect on expression of SLe(x) or SLe(a) in these tumor cells. These findings suggest the possibility that targeting of O-glycosylation inhibitors or modifiers to tumor cells may effectively suppress metastatic potential.     

Multi-recognition capability of E-selectin in a dynamic flow system, as evidenced by differential effects of sialidases and anti-carbohydrate antibodies on selectin-mediated cell adhesion at low vs. high wall shear stress: a preliminary note.

E-selectin has a "multi-recognition" capability in terms of epitope binding specificity, depending on adhesion conditions (static vs. low- or high-shear stress dynamic systems). Specifically, (i) adhesion based on expression of alpha 2-->3 sialylated Le(x) (SLe(x)) is prominent under static or low shear stress dynamic conditions; (ii) adhesion under high shear stress dynamic conditions does not depend on the known SLe(x) species, but rather on Lex with an adjacent unidentified sialosyl substitution, which shows different susceptibility to sialidases and antibodies compared to known SLe(x).     

Design, synthesis and biological evaluation of aryl-substituted sialyl Lewis X mimetics prepared via cross-metathesis of C-fucopeptides.

Several aryl substituted C-fucopeptides have been developed as sialyl Lewis X mimetics. Although the compounds have a much simpler structure compared to SLe(x), up to 3-times higher binding affinity toward E-selectin and > 1000 times toward P-selectin was observed. Furthermore, a convenient strategy for generating a number of analogues from a SLe(x) mimetic template at a very late stage of the synthesis was introduced, using a ruthenium catalyzed cross olefin metathesis under benchtop conditions.     

PSGL-1 from the murine leukocytic cell line WEHI-3 is enriched for core 2-based O-glycans with sialyl Lewis x antigen

Leukocyte trafficking involves specific recognition between P-selectin and L-selectin and PSGL-1 containing core 2-based O-glycans expressing sialyl Lewis x (SLe(x)) antigen. However, the structural identity of the glycan component(s) displayed by murine neutrophil PSGL-1 that contributes to its P-selectin counter-receptor activity has been uncertain, since these cells express little if any SLe(x) antigen, and because there have been no direct studies to examine murine PSGL-1 glycosylation. To address this uncertainty, we studied PSGL-1 glycosylation in the murine cell line WEHI-3 using metabolic-radiolabeling with (3)H-monosaccharide precursors to detect low-abundance O-glycan structures. We report that PSGL-1 from WEHI-3 cells expresses a di-sialylated core 2 O-glycan containing the SLe(x) antigen. This fucosylated O-glycan is scarce on PSGL-1 and essentially undetectable in total leukocyte glycoproteins from WEHI-3 cells. These results demonstrate that WEHI-3 cells selectively fucosylate PSGL-1 to generate functionally important core 2-based O-glycans containing the SLe(x) antigen.     

Recombinant E-selectin-protein mediates tumor cell adhesion via sialyl-Le(a) and sialyl-Le(x).

E-selectin (previously known as ELAM-1) is one of the adhesion molecules expressed on activated endothelium. Here we show that HL-60 cells express sialyl-Le(x), but not Sialyl-Le(a) on their surface, a colon carcinoma cell line COLO 205 express both these epitopes and another colon carcinoma COLO 320 does not express either one of them. HL-60 and COLO 205 cell adhere strongly to E-selectin coated microwells, whereas COLO 320 does not adhere at all to E-selectin. Finally we provide evidence that monoclonal anti-sialyl-Le(x) can abolish part of the adherence of HL-60 cells to recombinant E-selectin. The adherence of COLO 205 cells can be decreased by either monoclonal anti-sialyl-Le(a) or anti-sialyl-Le(x) antibodies. These results indicate that cell-associated sialylated carbohydrate moieties can act as ligands for recombinant E-selectin.     

Identification of a missense mutation (G329A;Arg(110)--> GLN) in the human FUT7 gene

The human FUT7 gene codes for the alpha1,3-fucosyltransferase VII (Fuc-TVII), which is involved in the biosynthesis of the sialyl Lewis x (SLe(x)) epitope on human leukocytes. The FUT7 gene has so far been considered to be monomorphic. Neutrophils isolated from patients with ulcerative colitis were examined for apparent alterations in protein glycosylation patterns by Western blot analysis using monoclonal antibodies directed against SLe(x) and SLe(x)-related epitopes. One individual showed lower levels of SLe(x) expression and an elevated expression of CD65s compared to controls. The coding regions of the FUT7 gene from this individual were cloned, and a G329A point mutation (Arg(110) --> Gln) was found in one allele, whereas the other FUT7 allele was wild type. No Fuc-TVII enzyme activity was detected in COS-7 cells transiently transfected with the mutated FUT7 construct. The FUT7 Arg(110) is conserved in all previously cloned vertebrate alpha 1,3-fucosyltransferases. Polymerase chain reaction followed by restriction enzyme cleavage was used to screen 364 unselected Caucasians for the G329A mutation, and a frequency of < or =1% for this mutation was found (3 heterozygotes). Genetic characterization of the family members of one of the additional heterozygotes identified one individual carrying the G329A mutation in both FUT7 alleles. Peripheral blood neutrophils of this homozygously mutated individual showed a lowered expression of SLe(x) and an elevated expression of CD65s when analyzed by Western blot and flow cytometry. The homozygous individual was diagnosed with ulcer disease, non-insulin-dependent diabetes, osteoporosis, spondyloarthrosis, and Sj?gren's syndrome but had no history of recurrent bacterial infections or leukocytosis.     

Insulin/protein kinase B signalling pathway upregulates metastasis-related phenotypes and molecules in H7721 human hepatocarcinoma cell line.

The effect of insulin on cancer metastatic potential was studied in a human hepatocarcinoma cell line, H7721. Cell adhesion to human umbilical vein endothelial cells (HUVECs) and laminin as well as chemotactic cell migration and invasion were selected as the indices of metastasis-related phenotypes for assessment of metastatic potential ex vivo. The results indicated that insulin enhanced all of these metastasis-related phenotypes. After the cells were treated with specific inhibitor of PI3K (LY294002) or transfected with antisense cDNA of PKB (AS-PKB), all of the above phenotypes were attenuated, and they could not be significantly stimulated by insulin, indicating that the insulin effect on metastatic potential was mediated by PI3K and PKB. Only the monoclonal antibody to the sialyl Lewis X (SLe(x)), but not antibodies to other Lewis antigens, significantly blocked the cell adhesion to HUVECs, cell migration and invasion, suggesting that SLe(x) played a crucial role in the metastatic potential of H7721 cells. The upregulation of cell surface SLe(x) and alpha-1,3-fucosyltransferase-VII (alpha-1,3 Fuc T-VII, enzyme for SLe(x) synthesis) was also mediated by PI3K and PKB, since LY294002 and AS-PKB also reduced the expressions of SLe(x) and alpha-1,3 FucT-VII, and attenuated the response to insulin. Furthermore, the alterations in the expressions of PKB protein and activity were correlated to the changes of metastatic phenotypes and SLe(x) expression. Taken together, the insulin/PKB signalling pathway participated in the enhancement of metastatic potential of H7721 cells, which was mediated by the upregulation of the expression of SLe(x) and alpha-1,3 FucT-VII.     

Acquisition of P-selectin binding activity by en bloc transfer of sulfo Le(x) trisaccharide to the cell surface: comparison to a sialyl Le(x) tetrasaccharide transferred on the cell surface

Sialyl Le(x), NeuNAcalpha2 --> 3Galbeta1 --> 4(Fucalpha1 --> 3)GlcNAcbeta --> R, is known to be a ligand for E-selectin in various assays. The sulfated counterpart of sialyl Le(x), sulfo Le(x), (Sulfo --> 3) Galbeta1 --> 4 (Fucalpha1 --> 3) GlcNAcbeta --> R, was also shown to be a ligand for E-selectin in solid-phase assays employing immobilized oligosaccharides. In order to determine whether sulfo Le(x) structure on the cell surface also works as E-selectin or P-selectin ligand, a novel approach for in vitro transfer of oligosaccharides (S. Tsuboi, Y. Isogai, N. Hada, J. K. King, O. Hindsgaul, and M. Fukuda (1996) J. Biol. Chem. 271, 27213-27216) was utilized. A synthetic GDP-fucose harboring sialyl Le(x) or sulfo Le(x) oligosaccharide was enzymatically transferred to Chinese hamster ovary (CHO) cells with a milk fucosyltransferase. The resultant cells, CHO-sialyl Le(x) and CHO-sulfo Le(x) were tested for adhesion to E-selectin. IgG or P-selectin. IgG chimeric protein coated on plates. The results indicate that CHO-sialyl Le(x) adhered efficiently to E-selectin, while adhesion of CHO-sulfo Le(x) was very poor despite the fact that near equal number of the ligands had been attached to the cell surface. In contrast, CHO-sulfo Le(x) adhered efficiently to P-selectin, while CHO-sialyl Le(x) adhered modestly to P-selectin. These results demonstrate that sialyl Le(x) and sulfo Le(x) structures on the cell surface differ substantially in their ability to adhere to E- and P-selectin.     

Transfection of the c-erbB2/neu gene upregulates the expression of sialyl Lewis X, alpha1,3-fucosyltransferase VII, and metastatic potential in a human hepatocarcinoma cell line.

The pCMV4 plasmid containing the cancer-promoting gene, c-erbB2/neu, was cotransfected into the human hepatocarcinoma cell line 7721 with the pcDNA3 vector, which contains the 'neo' selectable marker. Several clones showing stable expression of c-erbB2/neu were established and characterized by determination of c-erbB2/neu mRNA and its encoded protein p185. Expression of Lewis antigens and alpha1,3-fucosyltransferases and the biological behavior of 7721 cells after c-erbB2/neu transfection were studied using mock cells transfected with the vectors pCMV4 and pcDNA3 as controls. SLe(x) expression on the surface of mock cells was high, whereas expression of SDLe(x), Lex and SLe(a) was absent or negligible. This is compatible with the abundant expression of alpha1,3-fucosyltransferase VII, very low expression of alphafucosyltransferase III/VI, and almost absent expression of alpha1,3-fucosyltransferase IV in the mock cells. After transfection of c-erbB2/neu, expression of SLe(x) and alpha1,3-fucosyltransferase VII were simultaneously elevated, but that of alphafucosyltransferase III/VI was not altered. The expression of both SLe(x) and alpha1,3-fucosyltransferase VII correlated positively with the expression of c-erbB2/neu in different clones, being highest in clone 13, medium in clone 6, and lowest in clone 7. In addition, the adhesion of 7721 cells to human umbilical vein endothelial cells (HUVECs) or P-selectin, as well as cell migration and invasion, were increased in c-erbB2/neu-transfected cells. These increases also correlated positively with the expression intensities of c-erbB2/neu, SLe(x) and alpha1,3-fucosyltransferase VII in the different clones, whereas cell adhesion to fibronectin correlated negatively with these variables. mAbs to SLe(x) (KM93) and SDLe(x) (FH6) significantly and slightly, respectively, abolished cell adhesion to HUVECs or P-selectin and cell migration and invasion. mAbs to SDLe(x) and SLe(a) did not suppress cell adhesion to HUVECs nor inhibit cell migration and invasion. Transfection of alpha1,3-fucosyltransferase VII cDNA into 7721 cells showed similar results to transfection of c-erbB2/neu, and the increased adhesion to HUVECs, cell migration, and invasion were also inhibited significantly by KM93 and slightly by FH6. These results indicate that expression of alpha1,3-fucosyltransferase VII and its specific product, SLe(x), and their capacity for cell adhesion, migration and invasion are closely related. Therefore, the c-erbB2/neu gene is proposed to be a metastasis-promoting gene, and its effects are at least partially mediated by the increased expression of alpha1,3-fucosyltransferase VII and SLe(x).     

Identification of new cancer biomarkers based on aberrant mucin glycoforms by in situ proximity ligation

Mucin glycoproteins are major secreted or membrane-bound molecules that, in cancer, show modifications in both the mucin proteins expression and in the O-glycosylation profile, generating some of the most relevant tumour markers in clinical use for decades. Thus far, the identification of these biomarkers has been based on the detection of either the protein or the O-glycan modifications. We therefore aimed to identify the combined mucin and O-glycan features, that is, specific glycoforms, in an attempt to increase specificity of these cancer biomarkers. Using in situ proximity ligation assays (PLA) based on existing monoclonal antibodies directed to MUC1, MUC2, MUC5AC and MUC6 mucins and to cancer-associated carbohydrate antigens Tn, Sialyl-Tn (STn), T, Sialyl-Le(a) (SLe(a)) and Sialyl-Le(x) (SLe(x)) we screened a series of 28 mucinous adenocarcinomas from different locations (stomach, ampulla of Vater, colon, lung, breast and ovary) to detect specific mucin glycoforms. We detected Tn/STn/SLe(a)/SLe(x)-MUC1 and STn/SLe(a)/SLe(x)-MUC2 glycoforms in ≥50% of the cases, with a variable distribution among organs. Some new glycoforms-T/SLe(a)-MUC2, STn/T/SLe(a) SLe(x)-MUC5AC and STn/T/SLe(a)/SLe(x)-MUC6-were identified for the first time in the present study in a variable percentage of cases from different organs. In conclusion, application of the PLA technique allowed sensitive detection of specific aberrant mucin glycoforms in cancer, increasing specificity to the use of antibodies either to the mucin protein backbone or to the O-glycan haptens alone.     

alpha(1,2)-fucosylation prevents sialyl Lewis x expression and E-selectin-mediated adhesion of fucosyltransferase VII-transfected cells.

E-selectin is a cytokine-inducible, calcium-dependent endothelial cell adhesion molecule that plays a critical role in the leucocyte-endothelium interaction during inflammation and is thought to contribute to the metastatic dissemination of tumour cells. Like the other selectins, E-selectin binds to ligands carrying the tetrasaccharide sialyl-Lewis x (NeuAcalpha2,3Galbeta1,4[Fucalpha1, 3]GlcNAc)1 or its isomer sialyl-Lewis a (NeuAcalpha2, 3Galbeta1, 3[Fucalpha1,4]GlcNAc). We examined the effect of expressing the H-type alpha(1,2)-fucosyltransferase or the alpha(2, 6)-sialyltransferase on the synthesis of sialyl-Lewis x by alpha(1, 3)fucosyltransferase. We found that H-type alpha(1, 2)-fucosyltransferase but not alpha(2,6)-sialyltransferase, strongly inhibited sialyl-Lewis x expression and E-selectin adhesion. We assume that H-type alpha(1,2)-fucosyltransferase competes with the endogenous alpha(2,3)-sialyltransferase for the N-acetyllactosamine structures assigned to further serve as acceptors for alpha(1, 3)fucosyltransferase.     

Selectin/glycoconjugate binding assays for the identification and optimization of selectin antagonists.

In this study we describe ELISA-type P- and L-selectin binding assays for the analysis of selectin antagonists. A biotinylated polyacrylamide-type glycoconjugate containing sialyl Lewis A (sLe(a)-polymer) is utilized as a synthetic ligand for both selectins analogous to the E-selectin assay we have developed recently. Following precomplexation of sLe(a)-polymer with streptavidin-peroxidase, the complex is added to microtiter plates coated with the recombinant selectins. Binding of sLe(a)-polymer to the immobilized selectins is measured by the peroxidase reaction. SLe(a)-polymer was found to bind to P- and L-selectin in a cation-dependent manner. The interaction of the polymer was blocked by neutralizing anti-P- and anti-L-selectin antibody, respectively. The reference compounds heparin and fucoidan inhibited in both assays. Sialyl Lewis X (sLe(x)) blocked binding to L-selectin by 46% at 3 mM, whereas no inhibition was observed in the P-selectin assay up to 3 mM. Control polymers containing sialic acid or beta-d-glucose instead of sLe(a) weakly bound or failed to bind to the selectins. Both assays are rapid to perform and of low variability. The P-selectin assay was successfully employed to identify and optimize novel carbohydrate-based P-selectin antagonists. The P-, L-, and E-selectin assays were used to determine the fine selectivity of several sLe(x)-related selectin antagonists. These studies together suggest that sLe(a)-polymer-based selectin assays are well suited for primary screening and the characterization of selectin antagonists.     

Conformational dynamics of sialyl Lewisx in aqueous solution and its interaction with selectinE. A study by molecular dynamics

Three dimensional structures of sialyl Lewis(x) (SLe(x)) in aqueous solution and bound to selectinE are described based on an exhaustive conformational analysis and several long molecular dynamics simulations using different glycosidic regions as starting conformations. It appears from this study that when the oligosaccharide is free in solution the NeuNAcalpha(2-3)Gal segment favors glycosidic conformation in three different regions in the (Phi,Psi) plane with propensity of populations in the ratio 1:8:1. Each one of these structures is characteristically stabilized by specific hydrogen bonding interaction between NeuNAc and Gal. On the other hand, the Gal-GlcNAc-Fuc segment can exist in four different conformational states. Based on the topology of SLe(x) we are able to predict that out of all the allowed conformations in solution only two of these structures possess a geometry that would fit without steric clashes into the binding location of selectinE. In both of these binding modes, segment Gal-GlcNAc-Fuc adopts a unique conformation. The only difference between the two SLe(x) conformers that can successfully bind to selectinE is given by two possible regions in glycosidic space in the fragment NeuNAcalpha(2-3)Gal. A large conformational departure from the crystallographic data is observed for two lysine residues at the binding site of selectinE. These two residues play an important role when SLe(x) binds selectinE in aqueous solution. These findings help reconcile the X-ray data, in which these residues appear to be 1 nm away from SLe(x), with recent liquid NMR data reporting couplings between these protein residues and the sugar.     

Adhesive dynamics simulations of sialyl-Lewis(x)/E-selectin-mediated rolling in a cell-free system

Selectin-mediated leukocyte rolling is crucial for the proper function of the immune response. Recently, selectin-mediated rolling was recreated in a cell-free system (Biophysical Journal 71:2902-2907 (1996)); it was shown that sialyl Lewis(x) (sLe(x))-coated microspheres roll over E-selectin-coated surfaces under hydrodynamic flow. The cell-free system removes many confounding cellular features, such as cell deformability and signaling, allowing us to focus on the role of carbohydrate/selectin physical chemistry in mediating rolling. In this paper, we use adhesive dynamics, a computational method that allows us to simulate adhesion, to analyze the experimental data produced in the cell-free system. We simulate the effects of shear rate, ligand density, and number of receptors per particle on rolling velocity and compare them with experimental results obtained with the cell-free system. If we assume the population of particles is homogeneous in receptor density, we predict that particle rolling velocity calculated in simulations is more sensitive to shear rate than found in experiments. Also, the calculated rolling velocity is more sensitive to the number of receptors on the microspheres than to the ligand density on the surface, again in contrast to experiment. We argue that heterogeneity in the distribution of receptors throughout the particle population causes these discrepancies. We improve the agreement between experiment and simulation by calculating the average rolling velocity of a population whose receptors follow a normal distribution, suggesting heterogeneity among particles significantly affects the experimental results. Further comparison between theory and experiment yields an estimate of the reactive compliance of sLe(x)/E-selectin interactions of 0.25 A, close to that reported in the literature for E-selectin and its natural ligand (0.3 A). We also provide an estimate of the value of the intrinsic association rate (between 10(4) and 10(5) s(-1)) for the formation of sLe(x)/E-selectin bonds.     

Cloning and characterization of a viral α2-3-sialyltransferase (vST3Gal-I) for the synthesis of sialyl Lewisx

Sialyl Lewis(x) (SLe(x), Siaα2-3Galβ1-4(Fucα1-3)GlcNAcβOR) is an important sialic acid-containing carbohydrate epitope involved in many biological processes such as inflammation and cancer metastasis. In the biosynthetic process of SLe(x), α2-3-sialyltransferase-catalyzed sialylation generally proceeds prior to α1-3-fucosyltransferase-catalyzed fucosylation. For the chemoenzymatic synthesis of SLe(x) containing different sialic acid forms, however, it would be more efficient if diverse sialic acid forms are transferred in the last step to the fucosylated substrate Lewis(x) (Le(x)). An α2-3-sialyltransferase obtained from myxoma virus-infected European rabbit kidney RK13 cells (viral α2-3-sialyltransferase (vST3Gal-I)) was reported to be able to tolerate fucosylated substrate Le(x). Nevertheless, the substrate specificity of the enzyme was only determined using partially purified protein from extracts of cells infected with myxoma virus. Herein we demonstrate that a previously reported multifunctional bacterial enzyme Pasteurella multocida sialyltransferase 1 (PmST1) can also use Le(x) as an acceptor substrate, although at a much lower efficiency compared to nonfucosylated acceptor. In addition, N-terminal 30-amino-acid truncated vST3Gal-I has been successfully cloned and expressed in Escherichia coli Origami? B(DE3) cells as a fusion protein with an N-terminal maltose binding protein (MBP) and a C-terminal His(6)-tag (MBP-Δ30vST3Gal-I-His(6)). The viral protein has been purified to homogeneity and characterized biochemically. The enzyme is active in a broad pH range varying from 5.0 to 9.0. It does not require a divalent metal for its α2-3-sialyltransferase activity. It has been used in one-pot multienzyme sialylation of Le(x) for the synthesis of SLe(x) containing different sialic acid forms with good yields.