- Email: [email protected]
Sulfated glycolipids and cell adhesion
ARCHIVES
OF BIOCHEMISTRY
AND BIOPHYSICS
Vol. 267, No. 2, December, pp. 405-415,1988
INVITED
PAPER
Sulfated Glycolipids and Cell Adhesion DAVID D. ROBERTS Laboratory
of Structural
VICTOR GINSBURG’
AND
Biology, National Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of Health, Bethesda, Maryland 208% Received June 6,1988, and in revised form July 27,1988
The adhesive glycoproteins laminin, thrombospondin, and von Willebrand factor bind specifically and with high affinity to sulfatides, and it is this binding that probably accounts for their ability to agglutinate glutaraldehyde-fixed erythrocytes. The three proteins differ, however, in the inhibition of their binding to sulfatides by sulfated polysaccharides. Fucoidan strongly inhibits binding of both laminin and thrombospondin, but not of von Willebrand factor, suggesting the involvement of laminin or thrombospondin, or other unknown sulfatide-binding proteins in specific cell interactions that are also inhibited by fucoidan. Thrombospondin adsorbed on plastic promotes the attachment and spreading of some melanoma cells. Interestingly, fucoidan and an antibody against the sulfatide-binding domain of thrombospondin selectively inhibit spreading but not attachment to thrombospondin-coated surfaces. Sulfatides, but not neutral glycolipids or gangliosides, when adsorbed on plastic also promote attachment and spreading of some cultured cell lines. Direct adhesion of melanoma cells requires high densities of adsorbed sulfatide. In the presence of laminin, however, specific adhesion of some cell types to sulfatide is strongly stimulated and requires only low densities of adsorbed lipid, suggesting that laminin is mediating adhesion by crosslinking receptors on the cell surface to sulfatide adsorbed on the plastic. Although thrombospondin also binds to sulfatides and to melanoma cells, it does not enhance but rather inhibits direct and laminin-dependent melanoma cell adhesion to sulfatide, presumably because it is unable to bind simultaneously to ligands on opposing surfaces. Thus, sulfated glycolipids can participate in both laminin- and thrombospondin-mediated cell adhesion, but their mechanisms of interaction are different. o 1988 Academic Press. k.
Glycolipids containing sulfate esters are common constituents of cell membranes (1, 2). The sulfate is usually on galactose but also occurs on N-acetylglucosamine, N-acetylgalactosamine, glucuronic acid, sialic acid, and glucose. Structures of reported sulfated glycolipids in animal tissues are given in Table I (3). Whereas they are most abundant in the white matter of brain, sulfated glycolipids are also found in other tissues including kidney, spleen, 1 To whom correspondence should be addressed at: National Institutes of Health, Building 8, Room 110, Bethesda, MD 20892.
granulocytes, erythrocytes, platelets, testis, stomach, and intestine as well as in meconium. Many functions have been suggested for sulfated glycolipids (l), including a possible role in sodium transport, in the binding of opiates to their receptors, in the activation of oxygen radical generating systems, and in the activity of blood coagulation factor XII. Another possible function of sulfated glycolipids is in cell adhesion. Recently we reported that the adhesive proteins laminin (27), thrombospondin (28), and von Willebrand factor (29) bind specifically 405
0003-9861/88 $3.00 Copyright All rights
0 1988 by Academic Press, Inc. of reproduction in any form reserved.
Mouse small intestine
Gangliotetraosylcer-W-SO4
Trig1ucosyla1kylacylglycerol-IIIs-SO4
Galactosylfll-3alkylacylGro-IaSOd
Glycerolipids
gastric mncosa
Human saliva and
Testis
Sea urchin
G%N
(24
Human peripheral nerve
VIa(3’S0,GlcA)1actoneohexaosylCer
Sea urchin
(19) (20,21) (2%21) ce w (2%23)
Bovine gastric mucosa Human peripheral nerve
(‘5)
(10)
c&4-9)
Reference
IV8(3’S0,GlcA)lactoneotetraosy1Cer
Bovine gastric mncosa
Rat kidney Hog gastric mucosa
LactoneotetraosylCer-HP-SOI
Hog gastric mncosa
TrihexosylCer-HP-SO4
Gangliotetraosylcer-II”,IV8-bisSO1
Rat kidney Hog gastric mncosa
GangliotriaosylCer-IIs,IIIs-bisS0,
LactotriaosylCer-IIIs-SO4
,.
Rat kidney
Gangliotriaosylcer-II*-SOa
(seminolipid)
Human kidney
Many tissues
LactosylCer-II*-SO,
Gal(3SO,)61-1Cer
Source
Rat kidney
(sulfatide)
structure
GlucosylCer-P-SO,
GalactosylCer-P-SO,
Glyeosphingolipids
Sulfated glycolipid
I
STRUCTURES OF SULFATED GLYCOLIPIDS
TABLE
:
SULFATED
GLYCOLIPIDS
SULFATIDE- AND TYPE IV COLLAGEN-BINDING DOMAINS
ii’ :
CELL MEMBRANE RECEPTOR-BINDING
HEPARlN-BINDING
FIG. 1. Binding domains of laminin
DOMAIN
DOMAIN
(27,33).
and with high affinity to sulfatide (galactosylceramide 13-sulfate), and that this binding probably accounts for their ability to agglutinate aldehyde-fixed sheep erythrocytes. The criteria for specific binding to sulfatide include the demonstration that binding is not due to low affinity ionic interactions indicated by similar binding to anionic phospholipids and gangliosides as to sulfatides; that binding is labile to acid solvolysis but stable to base or neuraminidase treatment of lipids; that acidic but not neutral lipid fractions are active; that the protein binds with high affinity to purified sulfatide but not to cholesterol 3-sulfate; and that binding to sulfatide is saturable. Some proteins bind nonspecifically to sulfatide. For example, fibronectin binds nonspecifically and with low affinity to all anionic lipids tested, including phospholipids and gangliosides as well as sulfatides (27, 30,31), and some IgM antibodies also bind to sulfatides (32). In contrast, thrombospondin binds to at least lOOO-fold lower levels of sulfatide in solid phase assay than to unrelated anionic lipids and to loo-fold lower levels than to cholesterol 3-sulfate (28). Von Willebrand factor has a similar specificity (29). Laminin is less specific: cholesterol 3-sulfate is approximately 10% as active as sulfatide and phosphatidylserine is 1% as active on a weight basis (27). However, the concentration dependence of laminin binding indicates that binding to cholesterol 3-sulfate is very low affinity, whereas binding to sulfatide is high afinity and saturated at 20 pg/ml laminin (27).
AND
CELL
ADHESION
407
The proposed sulfatide-binding domains of laminin and thrombospondin are shown diagrammatically in Figs. 1 and 2. The location of the sulfatide-binding domain of von Willebrand factor is not known. Laminin is an 850-kDa cross-shaped glycoprotein with three short arms and a long arm (33) (Fig. 1). Sulfatide- and type IV collagen-binding domains are probably at the ends of the short arms, a heparin-binding domain is at the end of the long arm, and a cell membrane receptor-binding domain, which may involve a sequence YIGSR, is near the center of the cross (33). Thrombospondin is a 450-kDa trimeric glycoprotein with globular carboxy-terminal and amino-terminal domains (34-36) (Fig. 2). The carboxy-terminal domains bind to platelets and melanoma cells (37) and are required for haptotaxis of melanoma cells on an immobilized gradient of thrombospondin (38). The amino-terminal domains bind to heparin and sulfatide and are required for chemotaxis of melanoma cells in a soluble gradient of thrombospondin (38). The relative affinities of thrombospondin, laminin, and von Willebrand factor for various sulfated glycolipids are shown in Table II (39). Surprisingly, galaetosylceramide 13-sulfate is the best ligand for all three proteins. Its specificity relative to the synthetic isomer galactosylceramide I’-sulfate is moderate, ranging from 2-fold for thrombospondin to almost 6-fold for von Willebrand factor. In contrast, the presence of one additional sugar between galactose 3-sulfate and the ceramide as in lactosyl sulfatide decreases binding from 5- to lo-fold. The longer chain lipids examined, even those with two sulfate esters,
AMINO-TERMINA AND HEPARIN-BI DOMAINS
CARBOXY-TERMINAL MELANOMA CELLAND ING DOMAINS ATIDE- PLATELET-B
FIG. 2. Binding domains of thrombospondin 36).
(28,34-
ROBERTS
AND TABLE
GINSBURG II
RELATIVE AFFINITIES OFTHROMBOSPONDIN, LAMININ, AND VONWILLEBRANDFACTOR FORSULFATEDGLYCOLIPIDS(39) Protein Sulfated glycolipid
Thrombospondin
Laminin
von Willehrand factor
Relative affinity” GalCer-13-SO4 GalCer-I’-SO, LacCer-113-SO, GgOse&er-113-SO, GgOse&er-113,1113-bisS0, GgOseJer-I13,1V3-bisS0,
1.00 (at 30 pmol) 0.52 0.10 0.009 0.043
1.00 (at 100 pmol) 0.29 0.12 0.06
1.00 (at 100 pmol) 0.17 0.20 0.03 0.13
“Binding of each protein to galactosyl ceramide I%ulfate measured in solid phase assay was assigned a value of 1.00 at the indicated concentrations of lipid. Relative affinities for other lipids were calculated from the number of moles of test lipid required to give binding identical to that of the indicated quantity of galactosyl ceramide 13-sulfate.
are even weaker. Furthermore, addition of a terminal N-acetylgalactosamine to lactosyl sulfatide reduced binding from 2- to 11-fold. Thus, the three proteins prefer sulfated glycolipids with sulfate esters on terminal nonreducing residues over lipids with internal sulfate esters. Thrombospondin also does not bind to the sulfated triglucosylglycerolipid, which has a terminal 6-sulfate (25). Proximity of the sulfate ester to the ceramide is also strongly preferred, even though the accessibility of the sulfate esters to the proteins probably increases with increasing distance from the lipid. Preference for galactosyl sulfatide could be due to specificity for a conformation of the sugar which occurs only in the simple sulfatide or could indicate specific interactions between the proteins and the ceramide head groups as has been proposed for antibodies to seminolipid (40). Alternatively, interactions between galactose and the head group of the ceramide may stabilize the active conformation of the sugar (41). The two possibilities will be difficult to distinguish. The inhibition by anionic polysaccharides of the binding of thrombospondin, laminin, and von Willebrand factor to immobilized sulfatide is shown in Table III (39). Laminin and thrombospondin are similar in that the sulfated fucan, fucoi-
dan, is the best inhibitor and the order of activities of most of the other inhibitors is the same. Heparin, however, is a potent inhibitor of thrombospondin but not of laminin binding, as might be predicted from the location of their respective binding domains (Figs. 1 and 2). Von Willebrand factor behaves differently from the other proteins as fucoidan is relatively weak and heparin is inactive as an inhibitor of its binding to sulfatides (29). Instead, high-molecular-weight dextran sulfate is the only potent inhibitor for von Willebrand factor binding of those tested. Binding of laminin and thrombospondin to immobilized sulfatide is inhibited by some simple anions and anionic monosaccharides (28,39). As expected for predominantly ionic binding, increasing the ionic strength by addition of chloride inhibits binding of laminin to sulfatide (Fig. 3). The increasing ratio of specific binding at 2 and 10 pg/ml laminin indicates that the affinity of binding increases with decreasing ionic strength. Relative to isotonic conditions, binding of subsaturating concentrations of laminin is inhibited 50% at 120 mM added chloride. Sulfate is 4.4-fold more potent than chloride (39). Whereas galactose and glucose are not inhibitory, some anionic sugars are more active than expected
SULFATED
GLYCOLIPIDS TABLE
INHIBITION
AND
CELL
409
ADHESION
III
POLYSACCHARIDES OF THE BINDING OF THROMBOSPONDIN LAMININ, AND VON WILLEBRAND FACTOR TO SULFATIDES (39)
BY ANIONIC
Protein Inhibitor
Thrombospondin
Fucoidan Dextran sulfate (M, 500,000) Dextran sulfate (Mr 5000) Heparin Keratan sulfate C4 Keratan sulfate 2A Chondroitin sulfate Hyaluronate Colominic acid
0.3 2.2 28
10 >lOOO ND* >lOOO 140 ND
Laminin
von Willebrand
4 30 120 600 300
160
>lOOO >lOOO >lOOO 11000
a Concentration giving 50% inhibition of protein binding to 75,200, or 250 ng of sulfatide/well thrombospondin, and von Willebrand factor, respectively, in solid phase microtiter plate assays. *Not determined.
simply from their contribution to the ionic strength. The most potent inhibitor of those compounds tested is methyl-a-DGlcNAc 3-sulfate which is 15-fold more potent than chloride and 3.7-fold more potent than sulfate. The structures of the most ac-
r/2 FIG. 3. Effect of ionic strength on laminin binding to sulfatides. Binding of laminin at 2 fig/ml (open symbols) or 10 pg/ml (closed symbols) to wells coated with 200 ng sulfatide (circles) or uncoated wells (squares) is presented as a function of ionic strength adjusted by addition of NaCl to 50 mM Tris buffer, pH 7.8, containing 5 mM CaCl, and 1% bovine serum albumin.
factor
10 >lOOO >lOOO 700 >lOOO >lOOO >lOOO ND for laminin,
tive inhibitors, however, seem unrelated, and galactose 3-sulfate is not one of the most active. Galactose 6-sulfate is weaker than galactose 3-sulfate by only a factor of 1.1, whereas the respective glycolipid isomers differ 3.4-fold in binding activity. Again, the conformation of the ceramidelinked sugars may be different and could account for the greater binding specificity obtained with sulfated glycolipids. Some sugar phosphates also inhibit laminin binding to sulfatides. Mannose 6-phosphate and glucose l-phosphate are more potent than any of the hexose sulfates examined, which may be relevant to reports that both mannose 6-phosphate and fucoidan inhibit binding of lymphocytes to high endothelial venules (42). Although it is unlikely that laminin is involved in this interaction, these results demonstrate that hapten inhibition can be misleading when used to characterize receptors to which proteins bind primarily by ionic interactions. Most sulfated sugars, however, were no more potent inhibitors of thrombospondin binding than expected based on their contribution to the ionic strength (28). The lack of specificity among galactose sulfates is consistent with the finding that throm-
410
ROBERTS AND GINSBURG
SULFATED
GLYCOCONJUGATES
CELL RECEPTOR
PLASTIC A-ITACHMENT
.
SPREADING
FIG. 4. Attachment and spreading of melanoma cells on thrombospondin-coated plastic (37).
bospondin does not strongly prefer galactosylceramide 13-sulfate over the 6-sulfate isomer. This contrasts with the specificity and relatively high inhibitory activity of certain sulfated polysaccharides (Table III), suggesting that tight binding of the adhesive proteins requires more than one sulfate residue and that these sulfate residues must be in a specific orientation. A requirement for more than one sulfate residue for tight binding could explain the finding that gangliosides are remarkably potent inhibitors of the binding of laminin to sulfatides when both glycolipids are adsorbed on plastic (43). For example, the binding of laminin is inhibited 50% at a ratio of gangliosides GM1 to sulfatide of 1:lO and inhibited 100% at a ratio of 1:l. Thus, it is likely that a laminin molecule must interact simultaneously with several sulfatides in a specific orientation for tight binding, and a single ganglioside molecule can prevent this interaction. Gangliosides absorbed on erythrocyte membranes render the erythrocytes no longer agglutinable by laminin (43). This inhibition results from a masking of the erythrocyte sulfatides by the absorbed gangliosides. Based on these findings, results obtained using “soluble” gangliosides or other glycolipids as inhibitors of agglutination (44,45) or other biological activities (46-54) must be interpreted with caution. Thus, the inhibition by gangliosides of laminin-mediated agglutination is indirect as is the inhibition by gangliosides of binding to lymphoma cell lines of antibodies against gangliotriaosylceramide or globotriaosylceramide (55-57). Gangliosides can both stimulate (58) and inhibit (59) neuronal cell adhesion and
neurite outgrowth on several substrates including laminin. However, sulfated glycolipids also influence neuronal adhesion. Sulfatide, sulfated glucuronosylparagloboside, and the sulfated oligosaccharide released from the latter glycolipid inhibit cerebellar cell adhesion to a laminin substrate and neuron-neuron and neuron-astrocyte adhesion, whereas gangliosides have no effect in this assay (60). Although laminin was reported to bind directly to gangliosides (59), binding was observed on thin-layer plates only when loo-fold higher amounts of lipids than are required to detect binding to sulfatides were used. Therefore the inhibition by gangliosides of neurite outgrowth on laminin is probably indirect. A protein receptor involved in neurite outgrowth on laminin has also been purified (61). Gangliosides may modulate this receptor as was reported for integrin family receptors (62). Other effects of gangliosides associated with adsorption onto the cell surface include inhibition of lectin and ionophore stimulation of lymphocytes (4’7,48), inhibition of growth of 3T3 cells (49), and an increase in affinity of platelet-derived growth factor binding and inhibition of growth factor-stimulated tyrosine phosphorylation in 3T3 cells (50). Whether gangliosides in these systems are actually receptors, or whether they are simply masking the “true” receptors remains to be seen. Thrombospondin is secreted by many cell lines in culture and occurs in the extracellular matrix (34). It is involved in the aggregation of thrombin- or ionophore-activated platelets (35,36, 63, 64) and in the adhesion of erythrocytes parasitized with
SULFATED
GLYCOLIPIDS TABLE
AND
CELL
411
ADHESION
IV
ADHESIONOFMELANOMACELLSINTHEPRESENCEORABSENCEOFLAMININTOSULFATIDES ADSORBEDONPLASTIC(80) Cells attached/mm’ a Lipid None Sulfatide Sulfatide Sulfatide Globoside GM3 GM1
GDla GDlb GD3
GT3 Phosphatidylethanolamine Phosphatidylserine
Concentration (&ml)
-1aminin
20 5 1.25 50 20 50 20 20 20 20 20 20
0.2 31 4 0.6 1 3 2 3 2 1 1 1 1
+laminin (5 rg/ml) 1.4 83 62 26 2 7 0 3 3 1 3 5 9
a Melanoma cells (2’70/mma) were applied to plastic disks coated with various lipids and adhesion in the presence and absence of laminin was measured after 90 min at 37°C.
falciparum malaria (65). Thrombospondin also mediates the attachment and spreading of some human melanoma cell lines on plastic (36). Attachment and spreading on thrombospondin is similar to that obtained with the well-characterized cell adhesion protein, laminin (33). The aminoterminal sulfatide- and heparin-binding domain of thrombospondin is required for cell spreading but not for attachment (37) as a proteolytic fragment of thrombospondin, which lacks this domain, promotes attachment but not spreading. Spreading of cells on thrombospondin is also inhibited by fucoidan and dextran sulfate (37). A possible mechanism of cell adhesion to thrombospondin in illustrated diagrammatically in Fig. 4. Attachment of melanoma cells is mediated by the carboxy-terminal domains of thrombospondin binding to cell receptors analogous to the platelet receptors whereas cell spreading is mediated by the amino-terminal domains binding to sulfated glycoconjugates. A third region of thrombospondin may also interact with melanoma cells since an antibody that inhibits attachment (37) binds to the core region of thrombospondin (66), which contains a binding site for type
V collagen (67). Fragments of thrombospondin containing this region but not the carboxy-terminal domain, however, do not support attachment (37). Multiple interactions between adhesive proteins and cells may be a general phenomenon. The attachment and spreading activities of laminin can be separated using proteolytic fragments (68,69), and several distinct receptors for laminin have been identified (61, 70, 71). Multiple sites may also be involved in interaction of fibronectin with cells (72, 73) in addition to the Arg-Gly-Asp sequence recognized by integrin family receptors (74,75). Malignant melanoma can metastasize locally through epidermal basement membrane or distantly through blood or lymphatic vessels. Melanoma cell interaction with adhesive molecules may play a role in metastasis (76). Because thrombospondin promotes melanoma cell adhesion and migration (37, 38) and is associated with endothelium in viva (77), interactions of melanoma cells with thrombospondin may be important for passage of melanoma cells into or out of circulation. Interestingly, thrombospondin injected into mice prior to injection of cancer cells increases me-
412
ROBERTS
MELANOMA
AND
GINSBURG
CELL
SULF RECE
CELL RECEPTOR
ELANOMA CELL
CELL RECEPTOR
ROMBOSPONDIN SULFATIDES
PLASTIC FIG. 5. Laminin-mediated crosslinking of melanoma bition of binding of melanoma cells to sulfatide-coated
tastases (78) whereas fucoidan, which inhibits cell spreading on thrombospondin, decreases metastases (79). Several melanoma cell lines have lectins on their surfaces that bind to sulfated glycolipids and mediate the attachment and spreading of the cells on sulfatide-coated plastic (80). Attachment and spreading requires high surface densities of sulfatide and does not occur with some other lipids
cells to sulfatide-coated plastic (A) and inhiplastic by thrombospondin (B) (76).
tested (Table IV). Based on antibody inhibition data, the cell surface lectin is not laminin or thrombospondin (80). In the presence of laminin, some tumor cell lines adhere to low densities of sulfatide but not to other lipids including gangliosides (Table IV), presumably because the laminin crosslinks cell surface laminin receptors to the sulfatide adsorbed on plastic, as illustrated in Fig. 5A. The abil-
8E 60 -
Q I III,1 01
1 THROMSOSPONDINI~g/mll
10
FIG. 6. Inhibition by thrombospondin of melanoma cell adhesion to sulfatide adsorbed on plastic (76). Adhesion to plastic coated with sulfatide and albumin (0) or with albumin alone (0) is presented as percentage of control adhesion to sulfatide-coated plastic in the absence of added thrombospondin.
SULFATED
GLYCOLIPIDS TABLE
AND
CELL
413
ADHESION
V
INHIBITIONBY POLYSACCHARIDES OFTHE LAMININ-DEPENDENTADHESIONAND THE DIRECT ADHESION OFMELANOMACELLSTO SULFATIDESADSORBEDONPLASTIC(76) Inhibitor
Laminin-dependent adhesion
Direct adhesion
&Oh&W Fucoidan Dextran sulfate (M, 500,000) Heparin S. plicuta sulfated glycan Fl S. plicata sulfated glycan F2 Chondroitin sulfate Hyaluronate Colominic acid H. ho&ii phosphomannan
0.2 0.6
0.05 0.05
24
0.8
8 >lOO >lOO >lOO >lOO >lOO
>lOO >lOO 1100 >lOO rlO0 ZlOO
a Concentration of inhibitor giving 50% inhibition of control cell attachment when included as a competitive inhibitor in the adhesion assay.
ity of laminin to promote cell attachment to sulfatide-coated surfaces provides a possible mechanism for laminin-mediated cell-cell adhesion. Although thrombospondin adsorbed on plastic promotes adhesion of melanoma cells (3i’), and thrombospondin binds with high affinity to sulfatides, in contrast to laminin it does not promote but rather strongly inhibits cell attachment to sulfatides adsorbed on plastic (Fig. 6) (80), presumably because it is unable to bind simultaneously to ligands on opposing surfaces, as illustrated in Fig. 5B. This possible property of thrombospondin may explain a previous report that it also inhibits attachment of endothelial cells to some substrates (81). Fucoidan and dextran sulfate, which inhibit the binding of thrombospondin and laminin to sulfatides (Table III) and the spreading of melanoma cells on thrombospondin (37), are also potent inhibitors of both the laminin-dependent and direct adhesion of melanoma cells to sulfatides adsorbed on plastic (Table V). Interestingly, these sulfated polysaccharides promote lymphocyte aggregation (82) and adhesion (83) and also inhibit several cell interactions including lymphocyte binding to high endothelial venules (84) and recirculation (85), endothelial cell spreading (86), and human immunodeficiency virus binding to
T lymphocytes (87, 88). Possibly some of these effects result from the inhibition of sulfatide binding. REFERENCES
1. FAROOQUI,A. A., AND HORROCKS,L.A. (1985) Mol. Cell. Biochem. 66,87-95.
2. RADIN, N. S. (1983) in Handbook of Neurochemistry (Lajtha, A., Ed.), Vol. 3, 2nd ed., pp. 163177, Plenum, New York. 3. ROBERTS,D. D. (1986) Chem. Phys. Lipids 42,173183; (1987) in Methods in Enzymology (Ginsburg, V., Ed.), Vol. 138, pp. 473-483, Academic Press, San Diego. 4. YAMAKAWA, T., KISO, N., HANDA, S., MAKITA, A., AND YOKOYAMA,S. (1962) J Biochem 52,226227. 5. STOFFYN,P., AND STOFFYN, A. (1963) Biochim. Biophys. Acta 70,218-220. 6. MARTENSSON, E. (1966) Biochim. Biophys. Acta 116,521-531. 7. SARLIEVE, L. L., ZALC, B., NESKOVIC,N. M., ZANET~A, J.-P., AND REBEL, G. (1984) Biochim. Biophys. Acta 795,166-168. 8. HANSSON,C. G., KARLSSON,K.-A., AND SAMUELSSON,B. E. (1978) J. Biochem. (Tokyo) 83, 813819.
9. KARLSSON,K.-A., AND LARSON,G. (1981) J.
Biol.
Chem. 256,3512-3524.
10. ISHIZUKA, I., IIDA, N., TOIDA, T., KUSHI, Y., AND HANDA, S. (1987) Proceedings, 9th International Symposium on Glycoconjugates. [Abstract 311 11. TADANO,K., AND ISHIZUKA,I. (1982)J. BioL Chem. 257,1482-1490.
414
ROBERTS
AND
12. TADANO, K., AND ISHIZUKA, I. (1982)J. Bid C&m. 257,9294-9299. 13. SLOMIANY, A., SLOMIANY, B. L., AND ANNESE, C. (1980) Eur. J. Biochem. 109,471-474. 14. SLOMIANY, B. L., SLOMIANY, A., AND HOROWITZ, M. I. (1974) B&him. Biophys. Acta 348, 388396. 15. TADANO, K., ISHIZUKA, I., MATSUO, M., AND MATSUMOTO, S. (1982) J. Biol. Chem. 257, 1341313420. 16. SLOMIANY, B. L., AND SLOMIANY, A. (1978) J. Biol. Chem 253,351’7-3520. 17. LEFFLER, H., HANSSON, G. C., AND STROMBERG, N. (1986) J. Biol. Chem 261,1440-1444. 18. SLOMIANY, B. L., KOJIMA, K., BANAS-GRUSZKA, Z., MURTY, V. L. N., GALICKI, N. I., AND SLOMIANY, A. (1981) Eur. J. B&hem 119,647-650. 19. SLOMIANY, A., KOJIMA, K., BANAS-GRUSZKA, Z., AND SLOMIANY, B. L. (1981) Biochem. Biophys. Res. Commun 100,778-784. 20. CHOU, K. H., ILYAS, A. A., EVANS, J. E., QUARLES, R. H., AND JUNGALWALA, F. B. (1985) Biochem Biophys. Res. Commun 128,383-388. 21. CHOU, D. K. H., SCHWARTING, G. A., EVANS, J. E., AND JUNGALWALA, F. B. (1985) J. Neurochem. 49,865-8’73. 22. KOCHETOV, N. K., SMIRNOVA, G. P., AND CHEDAREVA, N. V. (1976) B&him. Biophys. Acta 424, 274-283. 23. PROKAZOVA, N. V., MIKHALILOV, A. T., KoCHAROV, S. L., MALCHENKO, L. A., ZVEZDINA, N. D., BUZNIKOV, G., AND BERGELSON, L. D. (1981) Eur. J B&hem. 115.671-677. 24. ISHIZUKA, I., SUZUKI, M., AND YAMAKAWA, (1973) J. B&hem 73,77-87.
T.
25. SLOMIANY, B. L., SLOMIANY, A., AND GLASS, G. B. J. (1977) Eur. J. B&hem 78,33-39. 26. NARASIMHAN, R., BENNICK, A., PALMER, B., AND MURRAY, R. K. (1982) J. Biol. Chem. 257,1512215128. 27. ROBERTS, D. D., RAO, C. N., MAGNANI, J. L., SPITALNIK, S. L., LIOTTA, L. A., AND GINSBURG, V. (1985) Proc. Natl. Accd Sci. USA 82,13061310. 28. ROBERTS, D. D., HAVERSTICK, D. M., DIXIT, V. M., FRAZIER, W. A., SANTORO, S. A., AND GINSBURG, V. (1985) J. Biol. Chem. 260,9405-9411. 29. ROBERTS, D. D., WILLIAMS, S. B., GRALNICK, H. R., AND GINSBURG, V. (1986) J. Biol. Chem. 261. 3306-3309. 30. ROSSI, J. D., AND WALLACE, B. A. (1983) J. Biol. Chem. 258,3327-3331. 31. MATYA, G. R., EVERS, D. C., RADINSKY, R., AND MORRE, D. J. (1986) Exp. Cell Res. 162,296-318. 32. HOFSE~ER, W., HEUSSER, C. H., HERSCHKOWITZ, N., AND BLASER, K. (1983) J. Immunol. Methods 57,99-109.
GINSBURG 33. MARTIN, G. R., AND TIMPL, R. (1987) Annu. Rev. Cell Biol. 3,57-85. 34. FRAZIER, W. A. (1987) J. Cell Biol. 105,625-632. 35. LAWLER, J., DERICK, L. H., CONNOLLY, J. E., CHEN, J.-H., AND CHAO, F. C. (1985) J. Biol. Chem 260, 3762-3772. 36. DIXIT, V. M., HAVERSTICK, D. M., O’ROURKE, K. M., HENNESSY, S. W., GRANT, G. A., SANTORO, S. A., AND FRAZIER, W. A. (1985) Proc. Natl. Acad. Sci. USA 82,3472-3476. 37. ROBERTS, D. D., SHERWOOD, J. A., AND GINSBURG, V. (1987) J. Cell Biol. 104,131-139. 38. TARABOLETTI, G., ROBERTS, D. D., AND LIOTTA, L. A. (1987) J. Cell BioL 105,2409-2415. 39. ROBERTS, D. D., RAO, C. N., LIOTTA, L. A., GRALNICK, H. R., AND GINSBURG, V. (1986) J. Biol. Chem. 261,6872-6877. 40. GOUJET-ZALC, C., GUERCI, A., DUBOIS, G., AND ZALC, B. (1986) J Neurochem. 46,435-439. 41. BOGGS, J. M., RANGARAJ, G., MOSCARELLO, M. A., AND KOSHY, K. M. (1985) Biochim Biophys. Acta 816.208-220. 42. STOOLMAN, L. M., TENFORDE, T. S., AND ROSEN, S. D. (1984) J. Cell Biol. 99,1535-1540. 43. ROBERTS, D. D., LIOTTA, L. A., AND GINSBURG, V. (1986) Arch. Biochem Biophys. 250,498-504. 44. KENNEDY, D. W., ROHRBACH, D. H., MARTIN, G. R., MOMOI, T., AND YAMADA, K. M. (1983) J Cell Physiol. 114,257-262. 45. YAMADA, K. M., KENNEDY, D. W., GROTENDORST, G. R., AND MOMOI, T. (1981) J. Cell Physiol 109, 343-351. 46. KLEINMAN, H. K., MARTIN, G. R., AND FISHMAN, P. H. (1979) Proc. Natl. Acad Sci. USA 76,33673371. 47. KRISHNARAJ, R., LIN, J., AND KEMP, R. G. (1983) Cell Immunol. 78,152-160. 48. WHISLER, R. L., AND YATES, A. J. (1980) J. Immunol. 125,2106-2111. 49. KEENEN, T. W., SCHMID, E., FRANKE, W. W., AND WIEGANDT, H. (1975) Exp. Cell Res. 92,259-270. 50. BREMER, E. G., HAKOMORI, $BOWEN-POPE, D. F., RAINES, E., AND Ross, R. (1984) J. Biol. Chem. 259,6818-6825. 51. KUWATA, T., HANDA, S., FUSE, A., AND MORINAGA, N. (1978) Biochem. Biophys. Res. Cornmun 85,77-84. 52. RAUVALA, H., CARTER, W. G., AND HAKOMORI, S. (1981) J. Cell Biol. 88,127-137. 53. GRIFFITHS, S. L., PERKINS, R. M., STREULI, C. H., AND CRITCHLEY, D. R. (1986) J. Cell BioC 102, 469-476. 54. GALVIN, N. J., DIXIT, V. M., O’ROURKE, K. M., SANTORO, S. A., GRANT, G. A., AND FRAZIER, W. A. (1985) J. Cell Biol 101,1434-1441. 55. URDAL, D. L., AND HAKOMORI, S. (1983) J. Biol. Chem. 258,6869-6874.
SULFATED
GLYCOLIPIDS
56. KANNAGI, R., STROUP,R.,COCHRAN,N.A., URDAL, D. L., YOUNG, W. W., JR., AND HAKOMORI, S. (1983) Cancer Res. 43,4997-5005. 57. WIELS,J.,HOLMES,E.H.,COCHRAN,N.,TURSZ,T., AND HAKOMORI, S. (1984) J. Biol. Chem 259, 14783-14787.
58. BYRNE, M. C., LEDEEN, R. W., ROISEN, F. J., YoRKE,G., AND SELAFANI,J.R.(~~~~)J. Neurochem. 41,1214-1222.
59. LAITINEN,J.,LOPPONEN, R., MERENMIES,J., AND RAUVALA, H. (1987) FEBSLett. 217,94-100. 60. KUNEMUND,V.,JUNGALWALA,F.B.,FISCHER,G., CHOU, D. K. H., KEILHAUER, G., AND SCHACHNER,M.(~~~~)J.C~~~BioZ. 106,213-223. 61. KLEINMAN,H.K.,OGLE,R.C.,CANNON,F.B.,LITTLE, C. D., SWEENEY,T. M., AND LUCKENBILLEDDS, L. (1988) Proc. Natl. Acad. Sci. USA 85, 1282-1286. 62. CHERESH, D. A., PYTELA, R., PIERSCHBACHER, M.D.,KLEIN,F. G.,RUOSLAHTI,E., AND REISFELD,R.A.(~~~~)J. Cell Biol. 105,1163-1173. 63. GARTNER,T.K., WALZ,D. A., AIKEN,M.,STARRSPIRE&L., AND OGILVIE,M.L.(~~~~) Biochem. Biophys. Res. Commun. 124,290-295.
64. LEUNG, L. L. K. (1984) J. Clin. Invest. 74, 17641772. 65. ROBERTS,D. D., SHERWOOD,J. A., SPITALNIK, S. L., PANTON, L. J., HOWARD, R. J., DIXIT, V.M., FRAZIER, W. A., MILLER, L. H., AND GINSBURG,V. (1985) Nature (London) 318, 6466. 66. DIXIT,V.M.,GALVIN,N.J.,O'ROURKE,K.M.,AND FRAZIER,~. A.(1986)J. Biol. Chem. 261,19621968. 67. MUMBY, S. M., RAUGI, G. J., AND BORNSTEIN,P. (1984) J. Cell Biol. g&646-652. 68. TIMPL,R.,JOHANSSON,S.,VANDELDEN,V.,OBERBAUMER,~.,AND Hoo~,M.(1983)J. Biol. Chem. 258,8922-8927. 69. GOODMAN,S. L., DEUTZMANN,R., AND VON DER Ma~~,K.(1987)J. Cell Biol. 105,585-598.
AND
CELL
415
ADHESION
70. RAO,N.C.,BARSKY,S.H.,TERRANOVA,V.P.,AND LIOTTA, L. A. (1983) Biochem Biophys. Res. Commun. 111,804-808. 71. SMALHEISER,N. R., AND SCHWARTZ,N.B. (1987) Proc. Natl. Acad Sci. USA 84,6457-6461.
72. AKIYAMA, S. K., HASEGAWA, E.,HASEGAWA, T., AND YAMADA, K. M.(1985)J. BioL Chem. 260, 13256-13260.
73. HUMPHRIES,M.J.,KOMORIJA, A.,AKIYAMA,S.K., OLDEN,K.,AND YAMADA,K.M.(~~~~)J. Biol. Chem. 262,6885-6892.
74. PIERSCHBADER,M.D., AND RuosLAHTI,E.(~~~~) Nature (London) 309,30-33. 75. RUOSLAHTI,E., AND PIERSCHBADER,M.D.(~~~~) Science238,491-497. 76. LIOTTA, L. A., RAO, C. N., AND WEWER, U. M. (1986) Annu. Rev. Biochem. 55,1037-1057. 77. WIGHT, T. N., RAUGI, G. J., MUMBY, S. M., AND BORNSTEIN,P. (1985) J. Histochem. Cytochem. 33,295-302. 78. TUSZYNSKI,G. P., GASIC, T. B., ROTHMAN,V. L., KNUDSEN,K.A.,AND G~~~,G.J.(1987)Cancer Res. 47,4130-4133.
79. COOMBE,D. R., PARISH, C. R., RAMSHAW, I. A., AND SNOWDEN,J. M.(198'7) Int. J. Cancer 39, 82-86. 80. RoBERTs,D.D.,WEWER,U.M.,LIOTTA,L.A.,AND GINSBURG,V.(~~S~)C~~~~~Res.,48,3367-3373. 81. LAHAV, J., DASDIK, R., AND STEIN, 0. (1987) Semin. Throb.
Hemostasis
13,352-360.
82. THURN, A.L., ANDUNDERHILL,C.B.(~~~~)J. Cell Physiol. 126,352-358.
83. BRANDLEY,B.K.,Ross,T.S.,AND
SCHNAAR,R.L.
(1987) J. Cell Biol. 105,991-997.
84. STOOLMAN,L.M., AND RosEN,S.D.(~~~~)J.
Cell
Biol. 96,722-729.
85. SPANGRUDE,G.J.,BRAATEN,B. A.,ANDDAYNES, R. A. (1984) J. Immunol 132,354-362. 86. GLABE, C. G., YEDNOCK, T., AND ROSEN, S. D. (1983) J. Cell Sci. 61,475-490.
87. UENO,R., AND KUNO, S.(1987)Lancef 1379. 88. MITSUYA, H., LOONEY,D. J., KUNO, S., UENO, R., WONG-STAHL,F., AND BRODER,S. (1988) Science240,646-649.
OF BIOCHEMISTRY
AND BIOPHYSICS
Vol. 267, No. 2, December, pp. 405-415,1988
INVITED
PAPER
Sulfated Glycolipids and Cell Adhesion DAVID D. ROBERTS Laboratory
of Structural
VICTOR GINSBURG’
AND
Biology, National Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of Health, Bethesda, Maryland 208% Received June 6,1988, and in revised form July 27,1988
The adhesive glycoproteins laminin, thrombospondin, and von Willebrand factor bind specifically and with high affinity to sulfatides, and it is this binding that probably accounts for their ability to agglutinate glutaraldehyde-fixed erythrocytes. The three proteins differ, however, in the inhibition of their binding to sulfatides by sulfated polysaccharides. Fucoidan strongly inhibits binding of both laminin and thrombospondin, but not of von Willebrand factor, suggesting the involvement of laminin or thrombospondin, or other unknown sulfatide-binding proteins in specific cell interactions that are also inhibited by fucoidan. Thrombospondin adsorbed on plastic promotes the attachment and spreading of some melanoma cells. Interestingly, fucoidan and an antibody against the sulfatide-binding domain of thrombospondin selectively inhibit spreading but not attachment to thrombospondin-coated surfaces. Sulfatides, but not neutral glycolipids or gangliosides, when adsorbed on plastic also promote attachment and spreading of some cultured cell lines. Direct adhesion of melanoma cells requires high densities of adsorbed sulfatide. In the presence of laminin, however, specific adhesion of some cell types to sulfatide is strongly stimulated and requires only low densities of adsorbed lipid, suggesting that laminin is mediating adhesion by crosslinking receptors on the cell surface to sulfatide adsorbed on the plastic. Although thrombospondin also binds to sulfatides and to melanoma cells, it does not enhance but rather inhibits direct and laminin-dependent melanoma cell adhesion to sulfatide, presumably because it is unable to bind simultaneously to ligands on opposing surfaces. Thus, sulfated glycolipids can participate in both laminin- and thrombospondin-mediated cell adhesion, but their mechanisms of interaction are different. o 1988 Academic Press. k.
Glycolipids containing sulfate esters are common constituents of cell membranes (1, 2). The sulfate is usually on galactose but also occurs on N-acetylglucosamine, N-acetylgalactosamine, glucuronic acid, sialic acid, and glucose. Structures of reported sulfated glycolipids in animal tissues are given in Table I (3). Whereas they are most abundant in the white matter of brain, sulfated glycolipids are also found in other tissues including kidney, spleen, 1 To whom correspondence should be addressed at: National Institutes of Health, Building 8, Room 110, Bethesda, MD 20892.
granulocytes, erythrocytes, platelets, testis, stomach, and intestine as well as in meconium. Many functions have been suggested for sulfated glycolipids (l), including a possible role in sodium transport, in the binding of opiates to their receptors, in the activation of oxygen radical generating systems, and in the activity of blood coagulation factor XII. Another possible function of sulfated glycolipids is in cell adhesion. Recently we reported that the adhesive proteins laminin (27), thrombospondin (28), and von Willebrand factor (29) bind specifically 405
0003-9861/88 $3.00 Copyright All rights
0 1988 by Academic Press, Inc. of reproduction in any form reserved.
Mouse small intestine
Gangliotetraosylcer-W-SO4
Trig1ucosyla1kylacylglycerol-IIIs-SO4
Galactosylfll-3alkylacylGro-IaSOd
Glycerolipids
gastric mncosa
Human saliva and
Testis
Sea urchin
G%N
(24
Human peripheral nerve
VIa(3’S0,GlcA)1actoneohexaosylCer
Sea urchin
(19) (20,21) (2%21) ce w (2%23)
Bovine gastric mucosa Human peripheral nerve
(‘5)
(10)
c&4-9)
Reference
IV8(3’S0,GlcA)lactoneotetraosy1Cer
Bovine gastric mncosa
Rat kidney Hog gastric mucosa
LactoneotetraosylCer-HP-SOI
Hog gastric mncosa
TrihexosylCer-HP-SO4
Gangliotetraosylcer-II”,IV8-bisSO1
Rat kidney Hog gastric mncosa
GangliotriaosylCer-IIs,IIIs-bisS0,
LactotriaosylCer-IIIs-SO4
,.
Rat kidney
Gangliotriaosylcer-II*-SOa
(seminolipid)
Human kidney
Many tissues
LactosylCer-II*-SO,
Gal(3SO,)61-1Cer
Source
Rat kidney
(sulfatide)
structure
GlucosylCer-P-SO,
GalactosylCer-P-SO,
Glyeosphingolipids
Sulfated glycolipid
I
STRUCTURES OF SULFATED GLYCOLIPIDS
TABLE
:
SULFATED
GLYCOLIPIDS
SULFATIDE- AND TYPE IV COLLAGEN-BINDING DOMAINS
ii’ :
CELL MEMBRANE RECEPTOR-BINDING
HEPARlN-BINDING
FIG. 1. Binding domains of laminin
DOMAIN
DOMAIN
(27,33).
and with high affinity to sulfatide (galactosylceramide 13-sulfate), and that this binding probably accounts for their ability to agglutinate aldehyde-fixed sheep erythrocytes. The criteria for specific binding to sulfatide include the demonstration that binding is not due to low affinity ionic interactions indicated by similar binding to anionic phospholipids and gangliosides as to sulfatides; that binding is labile to acid solvolysis but stable to base or neuraminidase treatment of lipids; that acidic but not neutral lipid fractions are active; that the protein binds with high affinity to purified sulfatide but not to cholesterol 3-sulfate; and that binding to sulfatide is saturable. Some proteins bind nonspecifically to sulfatide. For example, fibronectin binds nonspecifically and with low affinity to all anionic lipids tested, including phospholipids and gangliosides as well as sulfatides (27, 30,31), and some IgM antibodies also bind to sulfatides (32). In contrast, thrombospondin binds to at least lOOO-fold lower levels of sulfatide in solid phase assay than to unrelated anionic lipids and to loo-fold lower levels than to cholesterol 3-sulfate (28). Von Willebrand factor has a similar specificity (29). Laminin is less specific: cholesterol 3-sulfate is approximately 10% as active as sulfatide and phosphatidylserine is 1% as active on a weight basis (27). However, the concentration dependence of laminin binding indicates that binding to cholesterol 3-sulfate is very low affinity, whereas binding to sulfatide is high afinity and saturated at 20 pg/ml laminin (27).
AND
CELL
ADHESION
407
The proposed sulfatide-binding domains of laminin and thrombospondin are shown diagrammatically in Figs. 1 and 2. The location of the sulfatide-binding domain of von Willebrand factor is not known. Laminin is an 850-kDa cross-shaped glycoprotein with three short arms and a long arm (33) (Fig. 1). Sulfatide- and type IV collagen-binding domains are probably at the ends of the short arms, a heparin-binding domain is at the end of the long arm, and a cell membrane receptor-binding domain, which may involve a sequence YIGSR, is near the center of the cross (33). Thrombospondin is a 450-kDa trimeric glycoprotein with globular carboxy-terminal and amino-terminal domains (34-36) (Fig. 2). The carboxy-terminal domains bind to platelets and melanoma cells (37) and are required for haptotaxis of melanoma cells on an immobilized gradient of thrombospondin (38). The amino-terminal domains bind to heparin and sulfatide and are required for chemotaxis of melanoma cells in a soluble gradient of thrombospondin (38). The relative affinities of thrombospondin, laminin, and von Willebrand factor for various sulfated glycolipids are shown in Table II (39). Surprisingly, galaetosylceramide 13-sulfate is the best ligand for all three proteins. Its specificity relative to the synthetic isomer galactosylceramide I’-sulfate is moderate, ranging from 2-fold for thrombospondin to almost 6-fold for von Willebrand factor. In contrast, the presence of one additional sugar between galactose 3-sulfate and the ceramide as in lactosyl sulfatide decreases binding from 5- to lo-fold. The longer chain lipids examined, even those with two sulfate esters,
AMINO-TERMINA AND HEPARIN-BI DOMAINS
CARBOXY-TERMINAL MELANOMA CELLAND ING DOMAINS ATIDE- PLATELET-B
FIG. 2. Binding domains of thrombospondin 36).
(28,34-
ROBERTS
AND TABLE
GINSBURG II
RELATIVE AFFINITIES OFTHROMBOSPONDIN, LAMININ, AND VONWILLEBRANDFACTOR FORSULFATEDGLYCOLIPIDS(39) Protein Sulfated glycolipid
Thrombospondin
Laminin
von Willehrand factor
Relative affinity” GalCer-13-SO4 GalCer-I’-SO, LacCer-113-SO, GgOse&er-113-SO, GgOse&er-113,1113-bisS0, GgOseJer-I13,1V3-bisS0,
1.00 (at 30 pmol) 0.52 0.10 0.009 0.043
1.00 (at 100 pmol) 0.29 0.12 0.06
1.00 (at 100 pmol) 0.17 0.20 0.03 0.13
“Binding of each protein to galactosyl ceramide I%ulfate measured in solid phase assay was assigned a value of 1.00 at the indicated concentrations of lipid. Relative affinities for other lipids were calculated from the number of moles of test lipid required to give binding identical to that of the indicated quantity of galactosyl ceramide 13-sulfate.
are even weaker. Furthermore, addition of a terminal N-acetylgalactosamine to lactosyl sulfatide reduced binding from 2- to 11-fold. Thus, the three proteins prefer sulfated glycolipids with sulfate esters on terminal nonreducing residues over lipids with internal sulfate esters. Thrombospondin also does not bind to the sulfated triglucosylglycerolipid, which has a terminal 6-sulfate (25). Proximity of the sulfate ester to the ceramide is also strongly preferred, even though the accessibility of the sulfate esters to the proteins probably increases with increasing distance from the lipid. Preference for galactosyl sulfatide could be due to specificity for a conformation of the sugar which occurs only in the simple sulfatide or could indicate specific interactions between the proteins and the ceramide head groups as has been proposed for antibodies to seminolipid (40). Alternatively, interactions between galactose and the head group of the ceramide may stabilize the active conformation of the sugar (41). The two possibilities will be difficult to distinguish. The inhibition by anionic polysaccharides of the binding of thrombospondin, laminin, and von Willebrand factor to immobilized sulfatide is shown in Table III (39). Laminin and thrombospondin are similar in that the sulfated fucan, fucoi-
dan, is the best inhibitor and the order of activities of most of the other inhibitors is the same. Heparin, however, is a potent inhibitor of thrombospondin but not of laminin binding, as might be predicted from the location of their respective binding domains (Figs. 1 and 2). Von Willebrand factor behaves differently from the other proteins as fucoidan is relatively weak and heparin is inactive as an inhibitor of its binding to sulfatides (29). Instead, high-molecular-weight dextran sulfate is the only potent inhibitor for von Willebrand factor binding of those tested. Binding of laminin and thrombospondin to immobilized sulfatide is inhibited by some simple anions and anionic monosaccharides (28,39). As expected for predominantly ionic binding, increasing the ionic strength by addition of chloride inhibits binding of laminin to sulfatide (Fig. 3). The increasing ratio of specific binding at 2 and 10 pg/ml laminin indicates that the affinity of binding increases with decreasing ionic strength. Relative to isotonic conditions, binding of subsaturating concentrations of laminin is inhibited 50% at 120 mM added chloride. Sulfate is 4.4-fold more potent than chloride (39). Whereas galactose and glucose are not inhibitory, some anionic sugars are more active than expected
SULFATED
GLYCOLIPIDS TABLE
INHIBITION
AND
CELL
409
ADHESION
III
POLYSACCHARIDES OF THE BINDING OF THROMBOSPONDIN LAMININ, AND VON WILLEBRAND FACTOR TO SULFATIDES (39)
BY ANIONIC
Protein Inhibitor
Thrombospondin
Fucoidan Dextran sulfate (M, 500,000) Dextran sulfate (Mr 5000) Heparin Keratan sulfate C4 Keratan sulfate 2A Chondroitin sulfate Hyaluronate Colominic acid
0.3 2.2 28
10 >lOOO ND* >lOOO 140 ND
Laminin
von Willebrand
4 30 120 600 300
160
>lOOO >lOOO >lOOO 11000
a Concentration giving 50% inhibition of protein binding to 75,200, or 250 ng of sulfatide/well thrombospondin, and von Willebrand factor, respectively, in solid phase microtiter plate assays. *Not determined.
simply from their contribution to the ionic strength. The most potent inhibitor of those compounds tested is methyl-a-DGlcNAc 3-sulfate which is 15-fold more potent than chloride and 3.7-fold more potent than sulfate. The structures of the most ac-
r/2 FIG. 3. Effect of ionic strength on laminin binding to sulfatides. Binding of laminin at 2 fig/ml (open symbols) or 10 pg/ml (closed symbols) to wells coated with 200 ng sulfatide (circles) or uncoated wells (squares) is presented as a function of ionic strength adjusted by addition of NaCl to 50 mM Tris buffer, pH 7.8, containing 5 mM CaCl, and 1% bovine serum albumin.
factor
10 >lOOO >lOOO 700 >lOOO >lOOO >lOOO ND for laminin,
tive inhibitors, however, seem unrelated, and galactose 3-sulfate is not one of the most active. Galactose 6-sulfate is weaker than galactose 3-sulfate by only a factor of 1.1, whereas the respective glycolipid isomers differ 3.4-fold in binding activity. Again, the conformation of the ceramidelinked sugars may be different and could account for the greater binding specificity obtained with sulfated glycolipids. Some sugar phosphates also inhibit laminin binding to sulfatides. Mannose 6-phosphate and glucose l-phosphate are more potent than any of the hexose sulfates examined, which may be relevant to reports that both mannose 6-phosphate and fucoidan inhibit binding of lymphocytes to high endothelial venules (42). Although it is unlikely that laminin is involved in this interaction, these results demonstrate that hapten inhibition can be misleading when used to characterize receptors to which proteins bind primarily by ionic interactions. Most sulfated sugars, however, were no more potent inhibitors of thrombospondin binding than expected based on their contribution to the ionic strength (28). The lack of specificity among galactose sulfates is consistent with the finding that throm-
410
ROBERTS AND GINSBURG
SULFATED
GLYCOCONJUGATES
CELL RECEPTOR
PLASTIC A-ITACHMENT
.
SPREADING
FIG. 4. Attachment and spreading of melanoma cells on thrombospondin-coated plastic (37).
bospondin does not strongly prefer galactosylceramide 13-sulfate over the 6-sulfate isomer. This contrasts with the specificity and relatively high inhibitory activity of certain sulfated polysaccharides (Table III), suggesting that tight binding of the adhesive proteins requires more than one sulfate residue and that these sulfate residues must be in a specific orientation. A requirement for more than one sulfate residue for tight binding could explain the finding that gangliosides are remarkably potent inhibitors of the binding of laminin to sulfatides when both glycolipids are adsorbed on plastic (43). For example, the binding of laminin is inhibited 50% at a ratio of gangliosides GM1 to sulfatide of 1:lO and inhibited 100% at a ratio of 1:l. Thus, it is likely that a laminin molecule must interact simultaneously with several sulfatides in a specific orientation for tight binding, and a single ganglioside molecule can prevent this interaction. Gangliosides absorbed on erythrocyte membranes render the erythrocytes no longer agglutinable by laminin (43). This inhibition results from a masking of the erythrocyte sulfatides by the absorbed gangliosides. Based on these findings, results obtained using “soluble” gangliosides or other glycolipids as inhibitors of agglutination (44,45) or other biological activities (46-54) must be interpreted with caution. Thus, the inhibition by gangliosides of laminin-mediated agglutination is indirect as is the inhibition by gangliosides of binding to lymphoma cell lines of antibodies against gangliotriaosylceramide or globotriaosylceramide (55-57). Gangliosides can both stimulate (58) and inhibit (59) neuronal cell adhesion and
neurite outgrowth on several substrates including laminin. However, sulfated glycolipids also influence neuronal adhesion. Sulfatide, sulfated glucuronosylparagloboside, and the sulfated oligosaccharide released from the latter glycolipid inhibit cerebellar cell adhesion to a laminin substrate and neuron-neuron and neuron-astrocyte adhesion, whereas gangliosides have no effect in this assay (60). Although laminin was reported to bind directly to gangliosides (59), binding was observed on thin-layer plates only when loo-fold higher amounts of lipids than are required to detect binding to sulfatides were used. Therefore the inhibition by gangliosides of neurite outgrowth on laminin is probably indirect. A protein receptor involved in neurite outgrowth on laminin has also been purified (61). Gangliosides may modulate this receptor as was reported for integrin family receptors (62). Other effects of gangliosides associated with adsorption onto the cell surface include inhibition of lectin and ionophore stimulation of lymphocytes (4’7,48), inhibition of growth of 3T3 cells (49), and an increase in affinity of platelet-derived growth factor binding and inhibition of growth factor-stimulated tyrosine phosphorylation in 3T3 cells (50). Whether gangliosides in these systems are actually receptors, or whether they are simply masking the “true” receptors remains to be seen. Thrombospondin is secreted by many cell lines in culture and occurs in the extracellular matrix (34). It is involved in the aggregation of thrombin- or ionophore-activated platelets (35,36, 63, 64) and in the adhesion of erythrocytes parasitized with
SULFATED
GLYCOLIPIDS TABLE
AND
CELL
411
ADHESION
IV
ADHESIONOFMELANOMACELLSINTHEPRESENCEORABSENCEOFLAMININTOSULFATIDES ADSORBEDONPLASTIC(80) Cells attached/mm’ a Lipid None Sulfatide Sulfatide Sulfatide Globoside GM3 GM1
GDla GDlb GD3
GT3 Phosphatidylethanolamine Phosphatidylserine
Concentration (&ml)
-1aminin
20 5 1.25 50 20 50 20 20 20 20 20 20
0.2 31 4 0.6 1 3 2 3 2 1 1 1 1
+laminin (5 rg/ml) 1.4 83 62 26 2 7 0 3 3 1 3 5 9
a Melanoma cells (2’70/mma) were applied to plastic disks coated with various lipids and adhesion in the presence and absence of laminin was measured after 90 min at 37°C.
falciparum malaria (65). Thrombospondin also mediates the attachment and spreading of some human melanoma cell lines on plastic (36). Attachment and spreading on thrombospondin is similar to that obtained with the well-characterized cell adhesion protein, laminin (33). The aminoterminal sulfatide- and heparin-binding domain of thrombospondin is required for cell spreading but not for attachment (37) as a proteolytic fragment of thrombospondin, which lacks this domain, promotes attachment but not spreading. Spreading of cells on thrombospondin is also inhibited by fucoidan and dextran sulfate (37). A possible mechanism of cell adhesion to thrombospondin in illustrated diagrammatically in Fig. 4. Attachment of melanoma cells is mediated by the carboxy-terminal domains of thrombospondin binding to cell receptors analogous to the platelet receptors whereas cell spreading is mediated by the amino-terminal domains binding to sulfated glycoconjugates. A third region of thrombospondin may also interact with melanoma cells since an antibody that inhibits attachment (37) binds to the core region of thrombospondin (66), which contains a binding site for type
V collagen (67). Fragments of thrombospondin containing this region but not the carboxy-terminal domain, however, do not support attachment (37). Multiple interactions between adhesive proteins and cells may be a general phenomenon. The attachment and spreading activities of laminin can be separated using proteolytic fragments (68,69), and several distinct receptors for laminin have been identified (61, 70, 71). Multiple sites may also be involved in interaction of fibronectin with cells (72, 73) in addition to the Arg-Gly-Asp sequence recognized by integrin family receptors (74,75). Malignant melanoma can metastasize locally through epidermal basement membrane or distantly through blood or lymphatic vessels. Melanoma cell interaction with adhesive molecules may play a role in metastasis (76). Because thrombospondin promotes melanoma cell adhesion and migration (37, 38) and is associated with endothelium in viva (77), interactions of melanoma cells with thrombospondin may be important for passage of melanoma cells into or out of circulation. Interestingly, thrombospondin injected into mice prior to injection of cancer cells increases me-
412
ROBERTS
MELANOMA
AND
GINSBURG
CELL
SULF RECE
CELL RECEPTOR
ELANOMA CELL
CELL RECEPTOR
ROMBOSPONDIN SULFATIDES
PLASTIC FIG. 5. Laminin-mediated crosslinking of melanoma bition of binding of melanoma cells to sulfatide-coated
tastases (78) whereas fucoidan, which inhibits cell spreading on thrombospondin, decreases metastases (79). Several melanoma cell lines have lectins on their surfaces that bind to sulfated glycolipids and mediate the attachment and spreading of the cells on sulfatide-coated plastic (80). Attachment and spreading requires high surface densities of sulfatide and does not occur with some other lipids
cells to sulfatide-coated plastic (A) and inhiplastic by thrombospondin (B) (76).
tested (Table IV). Based on antibody inhibition data, the cell surface lectin is not laminin or thrombospondin (80). In the presence of laminin, some tumor cell lines adhere to low densities of sulfatide but not to other lipids including gangliosides (Table IV), presumably because the laminin crosslinks cell surface laminin receptors to the sulfatide adsorbed on plastic, as illustrated in Fig. 5A. The abil-
8E 60 -
Q I III,1 01
1 THROMSOSPONDINI~g/mll
10
FIG. 6. Inhibition by thrombospondin of melanoma cell adhesion to sulfatide adsorbed on plastic (76). Adhesion to plastic coated with sulfatide and albumin (0) or with albumin alone (0) is presented as percentage of control adhesion to sulfatide-coated plastic in the absence of added thrombospondin.
SULFATED
GLYCOLIPIDS TABLE
AND
CELL
413
ADHESION
V
INHIBITIONBY POLYSACCHARIDES OFTHE LAMININ-DEPENDENTADHESIONAND THE DIRECT ADHESION OFMELANOMACELLSTO SULFATIDESADSORBEDONPLASTIC(76) Inhibitor
Laminin-dependent adhesion
Direct adhesion
&Oh&W Fucoidan Dextran sulfate (M, 500,000) Heparin S. plicuta sulfated glycan Fl S. plicata sulfated glycan F2 Chondroitin sulfate Hyaluronate Colominic acid H. ho&ii phosphomannan
0.2 0.6
0.05 0.05
24
0.8
8 >lOO >lOO >lOO >lOO >lOO
>lOO >lOO 1100 >lOO rlO0 ZlOO
a Concentration of inhibitor giving 50% inhibition of control cell attachment when included as a competitive inhibitor in the adhesion assay.
ity of laminin to promote cell attachment to sulfatide-coated surfaces provides a possible mechanism for laminin-mediated cell-cell adhesion. Although thrombospondin adsorbed on plastic promotes adhesion of melanoma cells (3i’), and thrombospondin binds with high affinity to sulfatides, in contrast to laminin it does not promote but rather strongly inhibits cell attachment to sulfatides adsorbed on plastic (Fig. 6) (80), presumably because it is unable to bind simultaneously to ligands on opposing surfaces, as illustrated in Fig. 5B. This possible property of thrombospondin may explain a previous report that it also inhibits attachment of endothelial cells to some substrates (81). Fucoidan and dextran sulfate, which inhibit the binding of thrombospondin and laminin to sulfatides (Table III) and the spreading of melanoma cells on thrombospondin (37), are also potent inhibitors of both the laminin-dependent and direct adhesion of melanoma cells to sulfatides adsorbed on plastic (Table V). Interestingly, these sulfated polysaccharides promote lymphocyte aggregation (82) and adhesion (83) and also inhibit several cell interactions including lymphocyte binding to high endothelial venules (84) and recirculation (85), endothelial cell spreading (86), and human immunodeficiency virus binding to
T lymphocytes (87, 88). Possibly some of these effects result from the inhibition of sulfatide binding. REFERENCES
1. FAROOQUI,A. A., AND HORROCKS,L.A. (1985) Mol. Cell. Biochem. 66,87-95.
2. RADIN, N. S. (1983) in Handbook of Neurochemistry (Lajtha, A., Ed.), Vol. 3, 2nd ed., pp. 163177, Plenum, New York. 3. ROBERTS,D. D. (1986) Chem. Phys. Lipids 42,173183; (1987) in Methods in Enzymology (Ginsburg, V., Ed.), Vol. 138, pp. 473-483, Academic Press, San Diego. 4. YAMAKAWA, T., KISO, N., HANDA, S., MAKITA, A., AND YOKOYAMA,S. (1962) J Biochem 52,226227. 5. STOFFYN,P., AND STOFFYN, A. (1963) Biochim. Biophys. Acta 70,218-220. 6. MARTENSSON, E. (1966) Biochim. Biophys. Acta 116,521-531. 7. SARLIEVE, L. L., ZALC, B., NESKOVIC,N. M., ZANET~A, J.-P., AND REBEL, G. (1984) Biochim. Biophys. Acta 795,166-168. 8. HANSSON,C. G., KARLSSON,K.-A., AND SAMUELSSON,B. E. (1978) J. Biochem. (Tokyo) 83, 813819.
9. KARLSSON,K.-A., AND LARSON,G. (1981) J.
Biol.
Chem. 256,3512-3524.
10. ISHIZUKA, I., IIDA, N., TOIDA, T., KUSHI, Y., AND HANDA, S. (1987) Proceedings, 9th International Symposium on Glycoconjugates. [Abstract 311 11. TADANO,K., AND ISHIZUKA,I. (1982)J. BioL Chem. 257,1482-1490.
414
ROBERTS
AND
12. TADANO, K., AND ISHIZUKA, I. (1982)J. Bid C&m. 257,9294-9299. 13. SLOMIANY, A., SLOMIANY, B. L., AND ANNESE, C. (1980) Eur. J. Biochem. 109,471-474. 14. SLOMIANY, B. L., SLOMIANY, A., AND HOROWITZ, M. I. (1974) B&him. Biophys. Acta 348, 388396. 15. TADANO, K., ISHIZUKA, I., MATSUO, M., AND MATSUMOTO, S. (1982) J. Biol. Chem. 257, 1341313420. 16. SLOMIANY, B. L., AND SLOMIANY, A. (1978) J. Biol. Chem 253,351’7-3520. 17. LEFFLER, H., HANSSON, G. C., AND STROMBERG, N. (1986) J. Biol. Chem 261,1440-1444. 18. SLOMIANY, B. L., KOJIMA, K., BANAS-GRUSZKA, Z., MURTY, V. L. N., GALICKI, N. I., AND SLOMIANY, A. (1981) Eur. J. B&hem 119,647-650. 19. SLOMIANY, A., KOJIMA, K., BANAS-GRUSZKA, Z., AND SLOMIANY, B. L. (1981) Biochem. Biophys. Res. Commun 100,778-784. 20. CHOU, K. H., ILYAS, A. A., EVANS, J. E., QUARLES, R. H., AND JUNGALWALA, F. B. (1985) Biochem Biophys. Res. Commun 128,383-388. 21. CHOU, D. K. H., SCHWARTING, G. A., EVANS, J. E., AND JUNGALWALA, F. B. (1985) J. Neurochem. 49,865-8’73. 22. KOCHETOV, N. K., SMIRNOVA, G. P., AND CHEDAREVA, N. V. (1976) B&him. Biophys. Acta 424, 274-283. 23. PROKAZOVA, N. V., MIKHALILOV, A. T., KoCHAROV, S. L., MALCHENKO, L. A., ZVEZDINA, N. D., BUZNIKOV, G., AND BERGELSON, L. D. (1981) Eur. J B&hem. 115.671-677. 24. ISHIZUKA, I., SUZUKI, M., AND YAMAKAWA, (1973) J. B&hem 73,77-87.
T.
25. SLOMIANY, B. L., SLOMIANY, A., AND GLASS, G. B. J. (1977) Eur. J. B&hem 78,33-39. 26. NARASIMHAN, R., BENNICK, A., PALMER, B., AND MURRAY, R. K. (1982) J. Biol. Chem. 257,1512215128. 27. ROBERTS, D. D., RAO, C. N., MAGNANI, J. L., SPITALNIK, S. L., LIOTTA, L. A., AND GINSBURG, V. (1985) Proc. Natl. Accd Sci. USA 82,13061310. 28. ROBERTS, D. D., HAVERSTICK, D. M., DIXIT, V. M., FRAZIER, W. A., SANTORO, S. A., AND GINSBURG, V. (1985) J. Biol. Chem. 260,9405-9411. 29. ROBERTS, D. D., WILLIAMS, S. B., GRALNICK, H. R., AND GINSBURG, V. (1986) J. Biol. Chem. 261. 3306-3309. 30. ROSSI, J. D., AND WALLACE, B. A. (1983) J. Biol. Chem. 258,3327-3331. 31. MATYA, G. R., EVERS, D. C., RADINSKY, R., AND MORRE, D. J. (1986) Exp. Cell Res. 162,296-318. 32. HOFSE~ER, W., HEUSSER, C. H., HERSCHKOWITZ, N., AND BLASER, K. (1983) J. Immunol. Methods 57,99-109.
GINSBURG 33. MARTIN, G. R., AND TIMPL, R. (1987) Annu. Rev. Cell Biol. 3,57-85. 34. FRAZIER, W. A. (1987) J. Cell Biol. 105,625-632. 35. LAWLER, J., DERICK, L. H., CONNOLLY, J. E., CHEN, J.-H., AND CHAO, F. C. (1985) J. Biol. Chem 260, 3762-3772. 36. DIXIT, V. M., HAVERSTICK, D. M., O’ROURKE, K. M., HENNESSY, S. W., GRANT, G. A., SANTORO, S. A., AND FRAZIER, W. A. (1985) Proc. Natl. Acad. Sci. USA 82,3472-3476. 37. ROBERTS, D. D., SHERWOOD, J. A., AND GINSBURG, V. (1987) J. Cell Biol. 104,131-139. 38. TARABOLETTI, G., ROBERTS, D. D., AND LIOTTA, L. A. (1987) J. Cell BioL 105,2409-2415. 39. ROBERTS, D. D., RAO, C. N., LIOTTA, L. A., GRALNICK, H. R., AND GINSBURG, V. (1986) J. Biol. Chem. 261,6872-6877. 40. GOUJET-ZALC, C., GUERCI, A., DUBOIS, G., AND ZALC, B. (1986) J Neurochem. 46,435-439. 41. BOGGS, J. M., RANGARAJ, G., MOSCARELLO, M. A., AND KOSHY, K. M. (1985) Biochim Biophys. Acta 816.208-220. 42. STOOLMAN, L. M., TENFORDE, T. S., AND ROSEN, S. D. (1984) J. Cell Biol. 99,1535-1540. 43. ROBERTS, D. D., LIOTTA, L. A., AND GINSBURG, V. (1986) Arch. Biochem Biophys. 250,498-504. 44. KENNEDY, D. W., ROHRBACH, D. H., MARTIN, G. R., MOMOI, T., AND YAMADA, K. M. (1983) J Cell Physiol. 114,257-262. 45. YAMADA, K. M., KENNEDY, D. W., GROTENDORST, G. R., AND MOMOI, T. (1981) J. Cell Physiol 109, 343-351. 46. KLEINMAN, H. K., MARTIN, G. R., AND FISHMAN, P. H. (1979) Proc. Natl. Acad Sci. USA 76,33673371. 47. KRISHNARAJ, R., LIN, J., AND KEMP, R. G. (1983) Cell Immunol. 78,152-160. 48. WHISLER, R. L., AND YATES, A. J. (1980) J. Immunol. 125,2106-2111. 49. KEENEN, T. W., SCHMID, E., FRANKE, W. W., AND WIEGANDT, H. (1975) Exp. Cell Res. 92,259-270. 50. BREMER, E. G., HAKOMORI, $BOWEN-POPE, D. F., RAINES, E., AND Ross, R. (1984) J. Biol. Chem. 259,6818-6825. 51. KUWATA, T., HANDA, S., FUSE, A., AND MORINAGA, N. (1978) Biochem. Biophys. Res. Cornmun 85,77-84. 52. RAUVALA, H., CARTER, W. G., AND HAKOMORI, S. (1981) J. Cell Biol. 88,127-137. 53. GRIFFITHS, S. L., PERKINS, R. M., STREULI, C. H., AND CRITCHLEY, D. R. (1986) J. Cell BioC 102, 469-476. 54. GALVIN, N. J., DIXIT, V. M., O’ROURKE, K. M., SANTORO, S. A., GRANT, G. A., AND FRAZIER, W. A. (1985) J. Cell Biol 101,1434-1441. 55. URDAL, D. L., AND HAKOMORI, S. (1983) J. Biol. Chem. 258,6869-6874.
SULFATED
GLYCOLIPIDS
56. KANNAGI, R., STROUP,R.,COCHRAN,N.A., URDAL, D. L., YOUNG, W. W., JR., AND HAKOMORI, S. (1983) Cancer Res. 43,4997-5005. 57. WIELS,J.,HOLMES,E.H.,COCHRAN,N.,TURSZ,T., AND HAKOMORI, S. (1984) J. Biol. Chem 259, 14783-14787.
58. BYRNE, M. C., LEDEEN, R. W., ROISEN, F. J., YoRKE,G., AND SELAFANI,J.R.(~~~~)J. Neurochem. 41,1214-1222.
59. LAITINEN,J.,LOPPONEN, R., MERENMIES,J., AND RAUVALA, H. (1987) FEBSLett. 217,94-100. 60. KUNEMUND,V.,JUNGALWALA,F.B.,FISCHER,G., CHOU, D. K. H., KEILHAUER, G., AND SCHACHNER,M.(~~~~)J.C~~~BioZ. 106,213-223. 61. KLEINMAN,H.K.,OGLE,R.C.,CANNON,F.B.,LITTLE, C. D., SWEENEY,T. M., AND LUCKENBILLEDDS, L. (1988) Proc. Natl. Acad. Sci. USA 85, 1282-1286. 62. CHERESH, D. A., PYTELA, R., PIERSCHBACHER, M.D.,KLEIN,F. G.,RUOSLAHTI,E., AND REISFELD,R.A.(~~~~)J. Cell Biol. 105,1163-1173. 63. GARTNER,T.K., WALZ,D. A., AIKEN,M.,STARRSPIRE&L., AND OGILVIE,M.L.(~~~~) Biochem. Biophys. Res. Commun. 124,290-295.
64. LEUNG, L. L. K. (1984) J. Clin. Invest. 74, 17641772. 65. ROBERTS,D. D., SHERWOOD,J. A., SPITALNIK, S. L., PANTON, L. J., HOWARD, R. J., DIXIT, V.M., FRAZIER, W. A., MILLER, L. H., AND GINSBURG,V. (1985) Nature (London) 318, 6466. 66. DIXIT,V.M.,GALVIN,N.J.,O'ROURKE,K.M.,AND FRAZIER,~. A.(1986)J. Biol. Chem. 261,19621968. 67. MUMBY, S. M., RAUGI, G. J., AND BORNSTEIN,P. (1984) J. Cell Biol. g&646-652. 68. TIMPL,R.,JOHANSSON,S.,VANDELDEN,V.,OBERBAUMER,~.,AND Hoo~,M.(1983)J. Biol. Chem. 258,8922-8927. 69. GOODMAN,S. L., DEUTZMANN,R., AND VON DER Ma~~,K.(1987)J. Cell Biol. 105,585-598.
AND
CELL
415
ADHESION
70. RAO,N.C.,BARSKY,S.H.,TERRANOVA,V.P.,AND LIOTTA, L. A. (1983) Biochem Biophys. Res. Commun. 111,804-808. 71. SMALHEISER,N. R., AND SCHWARTZ,N.B. (1987) Proc. Natl. Acad Sci. USA 84,6457-6461.
72. AKIYAMA, S. K., HASEGAWA, E.,HASEGAWA, T., AND YAMADA, K. M.(1985)J. BioL Chem. 260, 13256-13260.
73. HUMPHRIES,M.J.,KOMORIJA, A.,AKIYAMA,S.K., OLDEN,K.,AND YAMADA,K.M.(~~~~)J. Biol. Chem. 262,6885-6892.
74. PIERSCHBADER,M.D., AND RuosLAHTI,E.(~~~~) Nature (London) 309,30-33. 75. RUOSLAHTI,E., AND PIERSCHBADER,M.D.(~~~~) Science238,491-497. 76. LIOTTA, L. A., RAO, C. N., AND WEWER, U. M. (1986) Annu. Rev. Biochem. 55,1037-1057. 77. WIGHT, T. N., RAUGI, G. J., MUMBY, S. M., AND BORNSTEIN,P. (1985) J. Histochem. Cytochem. 33,295-302. 78. TUSZYNSKI,G. P., GASIC, T. B., ROTHMAN,V. L., KNUDSEN,K.A.,AND G~~~,G.J.(1987)Cancer Res. 47,4130-4133.
79. COOMBE,D. R., PARISH, C. R., RAMSHAW, I. A., AND SNOWDEN,J. M.(198'7) Int. J. Cancer 39, 82-86. 80. RoBERTs,D.D.,WEWER,U.M.,LIOTTA,L.A.,AND GINSBURG,V.(~~S~)C~~~~~Res.,48,3367-3373. 81. LAHAV, J., DASDIK, R., AND STEIN, 0. (1987) Semin. Throb.
Hemostasis
13,352-360.
82. THURN, A.L., ANDUNDERHILL,C.B.(~~~~)J. Cell Physiol. 126,352-358.
83. BRANDLEY,B.K.,Ross,T.S.,AND
SCHNAAR,R.L.
(1987) J. Cell Biol. 105,991-997.
84. STOOLMAN,L.M., AND RosEN,S.D.(~~~~)J.
Cell
Biol. 96,722-729.
85. SPANGRUDE,G.J.,BRAATEN,B. A.,ANDDAYNES, R. A. (1984) J. Immunol 132,354-362. 86. GLABE, C. G., YEDNOCK, T., AND ROSEN, S. D. (1983) J. Cell Sci. 61,475-490.
87. UENO,R., AND KUNO, S.(1987)Lancef 1379. 88. MITSUYA, H., LOONEY,D. J., KUNO, S., UENO, R., WONG-STAHL,F., AND BRODER,S. (1988) Science240,646-649.