THEJOURNALOF BIOLOGICAL

CHEMISTRY Vol. 268, No. 3, Issue of January 25, pp. 2217-2222,1993

0 1993 by The American Society for Biochemistry and Molecular Biology, Inc Printed in U.S.A.

Regulatory Role ofG M 3 Ganglioside in a581 Integrin Receptor for

Fibronectin-mediated Adhesion ofFUA169 Cells*
(Received for publication, June 15,1992)

Mingzhe Zheng$§, Hang Fang$§, Tsutomu Tsuruokall, TsutomuTsujiII , Tomikazu Sasaki**, and
Sen-itiroh Hakomori$§
From $The Biomembrane Institute, Seattle, Washington 98119, the Departments of §Pathobiology and **Chemistry, University
of Washington, Seattle, Washington 98195,liMelj.i Seika Co., Morooka-cho, Kohoku-ku, Yokohama 222, Japan, and the IIFaculty
of Pharmaceutical Sciences, University of Tokyo, Bunkyo-ku, Tokyo 113, Japan

Mouse mammary carcinoma

mutant
cell
line
FUA169, characterized by high G Mganglioside
~ con-
tent, was established from parent cell line FM3A/F28-
7,which has high lactosyl ceramide (LacCer) content
but no GM3.FUA169 displays no changes in protein
glycosylation, andis a typical glycolipid mutant differ-
ing from its parent in that it contains high quantities
of GM3and GlcCer, but no LacCer (see accompanying
paper; Tsuruoka, T., Tsuji, T., Nojiri, H., Holmes, E.
H., Hakomori, S. (1993) J. Biol. Chern. 268, 2211-
2216). In contrast to parent F28-7 cells, FUA169 cells
showedclearadhesiontofibronectin(FN).Several
lines of evidence indicate that adhesion of FUA169
cells to FN requires the presence of G M ~which , sup-
portsthe function of integrinreceptor. (i) Both
FUA169 andF28-7 cells express the same quantity of
FN integrin receptor, which consists of a581 (sensitive
to RGDS peptide) anda481 (sensitive toCSl peptide).
However, adhesion to FN-coated plates, regardless of
type of FN, was much higher for FUA169 than for
F28-7 cells. (ii) F28-7 cells, which normally lack GM3
and adhere only weakly to FN, acquired G Mduring ~ ganglioside), present at the adhesion site and detected as
incubationinGM3-containingmedium, and subse-
quently adhered strongly to FN. (iii) Cholesterol-leci-
thin liposomes (cholesterol was 14C-labeled) incorpo-
rating a501 receptorisolatedfromhumanplacenta
showed clear adhesion to FN-coated plates, and this
adhesion was completely inhibited by RGDS peptide
and by anti-P1 mAb ZH1. When liposomes included a
moderatequantity of G M (0.22-0.44
~ fig (0.2-0.4
nmo1)/55 pg of phosphatidylcholine, 33 pg of choles-
terol, 5 pg of a581 in liposome), adhesionwas enhanced
significantly.Incontrast,adhesionwasgreatly re-
duced below control level forcy501 liposomes contain-
ing a higher quantity(2.2 pg; >2 nM) of G M ~Adhesion
. paper (25). Our results indicate that GM3 supports the function
to FN was also inhibited, but never enhanced, cy581 for
liposomes with similarcomposition but containing 0.4
nmol (or other quantities)of LacCer or GlcCer instead
of G M ~These
. findings suggest that the greater adhe-
sion to FN by FUA169 cells, relative to parent F28-7
cells, is due to functional supportby GMa of a501 inte-
grin receptor.
During the past decade, studies on molecular mechanisms
of cell adhesion haveclarified one “phase”of adhesion involv-
ing integrin receptors and pericellular adhesive proteins (re-
viewed in Refs. 1-4). However, there are other phases of cell
adhesion in which gangliosides play important roles.
Several linesof evidence support theidea that gangliosides
are involved in cell adhesion. Recently, we have shown that
cells expressing GMs1 ganglioside adhere to cells expressing
Gg3 or LacCer, based on direct involvement of G M in ~ cell
adhesion, i.e. interaction of G Mwith
~ these GSLs (5,6). Other
observations, as follows, indicate that gangliosides also help
control cell adhesion through modulationof integrin receptor
function. (i) In 1979, before integrin receptors were discov-
ered, Kleinman et al. (7) reported that adhesion of cells to FN
was inhibited by various gangliosides, particularly polysialo-
gangliosides. Subsequentstudies in our laboratory showed
that the inhibitory effect of gangliosides applied to lectin- and
glycosidase-mediated, as well as FN-mediated, cell adhesion
(8, 9). (ii) Most fibroblastic cells which adhere to FN lack
polysialogangliosides, but they all contain G M (the ~ simplest
detergent-insoluble matrix. Cell adhesion and spreadingwere
suppressed by exogenous G M(10). ~ (iii) In human melanoma,
G Dand~ GD3gangliosides are closely associated with vitronec-
tin receptor,based on immunohistochemical andbiochemical
analyses (11).
The purpose of the present study was to clarify the role of
G Min~ supporting integrinreceptor function in FN-mediated
cell adhesion. We employed mouse mammary carcinoma mu-
tant cell line FUA169 (characterized by high GM3 and GlcCer
content) and parent cell line FM3A/F28-7 (which has high
LacCer content but no GM3), as described inthe accompanying
of FN receptor and perhaps other types of integrin receptor.
EXPERIMENTALPROCEDURES
Materials-L-a-PC (from egg yolk), G M ~LacCer,
, GlcCer, N-ace-
tylneuraminic acid, RGDS peptide, molecular weight protein stand-
ards, gelatin-agarose, heparin-agarose, and wheat germ agglutinin-
agarose were from Sigma. Fluorescein-conjugated goat IgG fraction

The abbreviations used are: G M ~NeuAccu2+3GaIB1+4Glcbl+

,
Cer; Gg3, GalNAc/3l-t4Gal~l+4Glc(J’l+Cer; GlcCer, glucosylcer-
amide(Glc(J’l+Cer); G D ~ GalNAcB1-*4(NeuAca2+8NeuActu2-*
.
*This study was supported by fundsfromTheBiomembrane B)GaIBl-i4Glcpl+Cer; GD3, NeuActu2+8NeuAcn2-+3Galp1+
Institute, in part undera research contract with Otsuka Pharmaceu- 4Glc(J’l+Cer; GSL, glycosphingolipid; LacCer,lactosylceramide
tical Co., and by National Cancer Institute Outstanding Investigator (Gal~l+4Glcpl-+Cer); BSA, bovine serum albumin; FN,fibronectin;
Grant CA42505 (to S. H.). This is Paper I1 in a series, “Studies with mAb, monoclonal antibody; onfFN,oncofetal fibronectin; PBS, phos-
Functional Glycosphingolipid Mutant Cells.” The costs of publication phate-buffered saline; PBSSA, PBS containing 0.1% BSA and 0.2%
of this article were defrayed in part by the payment of page charges. sodium azide; PC, phosphatidylcholine;PMSF, phenylmethylsulfonyl
This articlemusttherefore be hereby marked“advertisement”in fluoride; PAGE, polyacrylamide gel electrophoresis; TBS, Tris-buff-
accordance with 18 U.S.C. Section 1734 solely to indicate this fact. ered saline; BHK,baby hamster kidney.

2217
2218 Receptor
Function
Integrin
Regulated by G M ~
anti-rabbit IgG F(ab')* was from Cappel Laboratories, Organon Tek- Hialeah, FL),lo4 cells were analyzed for forward and right angle light
nika, Westchester, PA. Rabbit anti-mouse FN receptor polyclonal scatter. Fluorescence intensity was expressed on a log scale.
antiserum 3675 (12) was provided by S. Akiyama (National Institutes Cell Labeling and Preclpitation of FN Receptor-Cells (IOs) were
of Health). Mouse mAb ZH1 uersus human FN receptor subunit labeled with 2 mCi of carrier-free sodium lz5iodideby lactoperoxidase-
was established in this laboratory.* [methyl-3H]Thymidine (6.7 Ci/ glucose oxidase method (22). Labeled cells were washed with TBS
mmol), [4-'4C]cholesterol (51mCi/mmol), [3H]sodiumborohydride and lysed in cold TBS containing 1% Triton X-100,0.5% Nonidet P-
(50 Ci/mmol), and '251-radionuclide (17 Ci/mg, reductant-free) were 40,0.1% BSA, 2 mM PMSF, 50 pg/ml aprotinin, 1 mM CaCl,, 1 mM
from Du Pont-New England Nuclear. ['H]GM3 (1 pg 2 0.7 pCi) was MgCl,, and 1 mM MnCL a t 4 "C for 3 h. After centrifugation, cell
prepared by reductive tritiation of the olefinic bond of sphingosine lysate supernatant was pre-cleared by incubation with Sepharose 4B
(13). andre-centrifugation.Lysates were absorbedonto 1 ml of FN-
Puri/ication o/ FN and FN Receptor-Plasma fibronectin was Sepharose 4B (7.5 mg of protein/ml of gel) a t 4 "C overnight with
purified from citrated normal human plasma by affinity chromatog- shaking. FN receptors were washed and eluted with 10 mM EDTA in
raphy ongelatin-agarose (14) and further purified on heparin-agarose TBS containing 0.5% Nonidet P-40. The receptor preparation was
(15). Human hepatoma cell line HUH7 (16) was cultured in serum- further purified by affinity chromatography on wheat germ agglu-
free, chemically defined medium (17), and onfFN was purified from tinin-agarose and eluted with 0.2 M N-acetylglucosamine. The result-
this culturemedium. DeglycosylatedonfFN was prepared from HUH7 ing preparations were analyzed on SDS-PAGE (23). Proteins were
culture medium inthepresence of tunicamycin(2 pg/ml, 24 h). visualized by exposing Kodax X-Omat AR film to the dried gel a t
Human FN receptor was purified from humanplacenta by -80 "C.
modification* of the method of Pytela etal. (18). Synthesis of CSl Peptide-Polymeric peptide KKTDELPQLV-
Cell Adhesion Assay"FUA169 and F28-7 cells were labeled with TLPHPNLHGPEILDVPSTVQKY, spanning the CS1 region of the
["Hlthymidine (1 pCi/ml, 24 h), washed 3 times, and suspended in IIICS domain of FN, was synthesized using an Applied Biosystems
medium supplemented with 0.1% heat-denatured BSA. 2 X lo5 cells (Foster City, CA) 430A peptide synthesizer. t-Butoxycarbonyl-pro-
were seeded into each well of 96-well assay plates (Pro-bind,Falcon, tected amino acids were used for dicyclohexylcarbodiimide-mediated
Becton-Dickinson, Oxnard, CA) previously coatedwith FNand coupling to the peptide chain. p-Methylbenzhydrylamine resin was
blocked with 1% heat-denatured BSA inPBS. Cells were leftto used as solid support. Peptides were cleaved from resin by treatment
adhere a t 37 "C for 2 h, and non-adherent cells were removed by (0 "C, 2 h) with a cleavage mixture, composed of 1.94 ml of trimeth-
gentle washing with PBS. Adherent cells were dissolved in 1% SDS ylsilyl trifluoromethanesulfonate, 2 ml of Scavenger mixture (thioan-
with 0.1 N NaOH and quantitated by scintillation counter. In cell iso1e:ethylenediamine tartrate:m-cresol, 1.2:0.6:0.2 ml), and 6.89 ml
adhesion inhibition assay, cells were preincubated with inhibitors at of trifluoroacetic acid. After precipitation with ether and treatment
room temperature for 30 min before transfer to coated plates. with 1 M ammonium fluoride, peptides were purified by chromatog-
Liposome Preparation and Liposome Binding Assay-Liposomes raphy on a Sephadex G25 column, followed by reverse-phase high-
incorporating integrinwere prepared by a modification of the method performance liquid chromatography on a CRI C18 column (Column
originally described by Mimms et al. (19) and adaptedby others (20, Resolution, Inc., SanJose, CA).
21). In these earlier studies,only PC was used as a liposome constit-
uent. We modified the method for inclusion of cholesterol and G M ~ RESULTS
(or LacCer or GlcCer), asfollows.
PC (55 pg) and 33 pg ['4C]cholesterol alone or with G M 3 (three Mouse mammarycarcinomamutant cellline FUA169
different concentrations for experiments shown in Fig. 7; eight con- (characterized by high GM3 and GlcCer content) and parent
centrations for Fig. 8) or GlcCer or LacCer (three concentrations
equivalent to those for G M 3 in Fig. 7) were dried onto a glass tube
cell line FM3A/F28-7 (which has high LacCer content but no
under N, stream and dissolved in 0.5 ml of TBS (50 mM Tris-HC1, GM3)showed a striking difference in integrin-mediated adhe-
p H 7.4, 0.14 M NaC1) containing50 mM octylglucoside, 0.1 mM CaC12, sion to various types of FN. This GSL mutant cell line has
and 5 pg of n5p1 receptor, and 0.1 pg of 1251-labeleda581 receptor as proved useful in demonstrating therole of G Min ~ supporting
indicator. [%]GM~was included for some liposome preparations for the function of murine integrin receptorsfor FN.
the purpose of monitoring co-incorporationof G Mand ~ cholesterol in Differential Adhesion of FUA169 and F28-7 Cells to Plates
the same liposome. Detergent was removed by dialysis against TBS Coated withVariousTypes of FN"FUA169 cells adhered
containing 0.1 mM CaC12 and 0.1 mM PMSF at 4 "C for 24 h. The
suspension was adjusted to60% in sucrose, overlaid with 3 ml of 30% significantly to plates coated with plasma fibronectin (Fig.
sucrose and 1 ml of 10% sucrose (made with TBS containing 0.1 mM l A ) , onfFN isolated from HUH7 cells (Fig. l B ) , or deglyco-
CaC12), and centrifuged (275,000 X g) a t 4 "C for 18 h. Liposomes sylatedonfFN (Fig. IC). Adhesion toonfFN was notably
were recovered at the topof the 10% sucrose layer. greater than that to plasma fibronectin and slightly greater
Liposomes were prepared withPC:cholesterol:FN receptorin a than that to deglycosylated onfFN. Under the same condi-
ratio of 55:33:5 (pg). In such liposomes containing 0.08, 0.4, and 2.0 tions, F28-7 cells showed much lower adhesion to FN-coated
nM G M S , Gwd incorporation was 81, 83, and 86% respectively; FN
receptorincorporation was 42.9, 43.3, and 43.7%; and cholesterol plates, particularly a t low FN concentration (Fig. 1, A-C).
incorporation was 86-88%. Comparable ranges for incorporation of Adhesion to FN-coated plates was strongly inhibitedby prein-
GSL and other components were observed when similar liposomes cubation of FUA169 cells with RGDS or CS1 peptide for 30
were prepared with LacCer or GlcCer instead of G M ~Purified
. lipo- min a t room temp, andcompletely inhibited by preincubation
somes were dialyzed against TBS containing 1 mMCaC12, 1 mM with a mixture of 0.5 mM RGDS plus 50 pg/ml CS1 (Fig. 2).
MgCl,, 1 mM MnCl,, and 1 mM PMSF at 4 "C for 12 h. Liposome These findings indicate thatFUA169 adhesion to FN depends
binding assay was performed in Pro-bind assay plates coated with
FN. Liposome suspensions were adjusted to contain equal concentra- on the known mechanism in which integrin receptor recog-
tions of FN receptor and 1% BSA, then added to wells and allowed nizes the RGDS (for a5Dl) or CS1 (for ~ ~ 4 6sequence
1) present
to bind with gentle shaking a t 4 "C for 24 h. Nonspecifically bound on FN.
liposomes were removed by washing 3 times with TBS. Bound lipo- The difference in adhesion of FUA169 uersus F28-7 cells
somes were dissolved in 1% SDS and quantitated by scintillation to FN-coated plates is not due to adifference in integrin
counter. Liposomes containing no FN receptor were used as negative receptor expression at the cell surface. The two cell lines
controls. Data were adjusted with total /3 counts of added sample.
Flow Cytometry"FUA169 and F28-7 cells were suspended at lo6 showed nearlyidenticalintegrin receptorexpression, as
cells/ml and incubatedwith rabbit anti-mouse FN receptor polyclonal probed with antiserum against murine FN receptors by flow
antibody 3675 (1:lOO dilution of antiserum) at4 "C for 1 h. Cells were cytometry (Fig. 3). F N receptors were cell surface-labeled by
washed with PBSSA and incubated withfluorescein isothiocyanate- lZ5I-iodination, extracted, and affinity-purified on FN-Seph-
conjugated goat IgG fraction anti-rabbit IgG F(ab')* in PBSSA a t arose column. FN receptor fractions thus preparedwere sep-
4 "C for 30 min. Stained cells were washed and fixed with cold 2% arated by SDS-PAGE. FUA169 and F28-7 cells expressedFN
paraformaldehyde in PBS. Non-immune rabbit serum was used as
negative control.Using flow cytometry (EPICSProfile, Coulter Corp., receptor to the samedegree (Fig. 4).
Role of GM3 Surface Expressioni n Adhesion of GM3-enriched
* M. Zheng, H. Fang, and S. Hakomori, unpublished observations. F28-7 Cells and FUA169 Cells to FN-coated Plates-The fact
 Receptor
Function
Regulated
Integrin by G M ~ 2219
1oor Adhesion of Integrin Liposomes to FN, and Effect of G Y ~ ,
LacCer, and GlcCer on Adhesive Function of a5PI in Lipo-
somes-a5~1 integrin receptor wasprepared andpurified from
human placenta and incorporated into cholesterol-lecithin
liposomes(cholesterolwas14C-labeled), as described under
“Experimental Procedures” and elsewhere.’ Adhesion of these
a5pI-containing liposomes to FN-coated plates is shown in
Plasma FN Onf-FN Deqlycosylated onf-FN Fig. 7. When GM3 was included in liposomes (see “Experimen-
(Wmll (pq/ml) (pq/ml)
tal Procedures”), their adhesion was greatly altered. When a
FIG. 1. Comparative binding of FUA169 versus F28-7 cells small quantity of G Mwas ~ included (0.05-0.08 nmol of G M ~ /
to immobilized FN. Plates were coated with 100 pl/well plasma 55 fig of PC, 33 pg of cholesterol, 5 pg of a5Pl in liposome),
fibronectin (panel A), onfFN (panel E?), or deglycosylated onfFN adhesion to FN-coated plateswas essentially the same asfor
(panel C) a t various concentrations (abscissa).FUA169 (0)and F28-
7 (0) cells were labeled with [3H]thymidine and left to adhere to liposomes lacking G M (Fig.
~ 7A, line 2). When G M content
~
coated platesa t 37 “C for 2 h. FN-mediated adhesionwas quantitated was increased t o 0.4 nmol in a5Pl liposomes (composition
by scintillation counting. Results are expressed as percent adherent otherwiseas described above),adhesionincreased signifi-
cells. Points represent meansof triplicate experiments.Bars represent cantly (Fig. 7 A , line I). However,when G M content ~ was
standard deviations. further increased to2.2 kg or higher (>2 nmol), adhesionwas
reduced, particularly for plates coated by incubation with 7-
Concentratim of CS-1 (ug/ml) 20 pg/ml FN. Optimal concentration of GM3 (for enhancing
0 1.6 6.25 25 100
adhesion to FN-coated plates) was in the range of 0.2-0.5
nmol for a5Pl liposomeswith the above composition. In
contrast, similar a5Pl liposomes with equivalent nanomolar
quantities of LacCer or GlcCer instead of G”3 showed weak
but consistent inhibition of adhesion (Fig. 7, B and C ) . Re-
gardless of concentration, a5Pl liposomes with inclusion of
LacCer or GlcCer showed adhesion lower than thatof control
liposomes (Fig.7, B and C ) .Adhesion of liposomes containing
various quantities (0.08-10.0 nmol) of G M to ~ plastic plates
coatedwith a constantconcentration (12.5pg/ml) of FN
peaked around 0.2-0.4 nmol of G M and ~ fell below control
level for quantities above 1 nmol(Fig. 8). Theenhanced
Concentration of RGDS (uM) adhesion of a5Pl liposomes containing G Mwas ~ weakly sup-
FIG. 2. Inhibition of FUA169 cell adhesion to FN by RGDS pressed by preincubation of liposomes with a n t i - G ~mAb ~
and CS1 peptides. Labeled cells were incubated with RGDS ( O ) , DH2, and strongly suppressedby preincubation with anti-P1
CS1 (O),or RGDS plus CS1 (B) at room temperature for 30 min, mAb ZH1 or RGDS peptide (Fig. 9).
transferred to Pro-bind platespreviously coated with FN (50 pg/ml,
100 pl/well), and left to adhere. The
degree ofinhibition was expressed DISCUSSION
as a percentage, with 0% defined as thelevel for control wells (lacking
inhibitor) and 100% defined as total absence of cells in a well after An effective approach for elucidating the functional roles
washing. of GSLs is to establish and study mutant cell lines, which
differ from parentcells only in expressionof a particular GSL
t h a t integrin receptor expression on FUA169 and F28-7 cells but show identicalprotein glycosylation pattern. FUA169
is the same suggests that their differential adhesion to FN- cells, isolated as described in the accompanying paper (25),
coated plates is explained by the presence of G Mor ~ GlcCer express ahigh level of GM3 and moderate GlcCer (but no
o n FUA169, but not onF28-7; i.e. G Mor~ GlcCer may enhance LacCer), whereas parent F28-7 cells express high LacCer (but
or support the function of integrin receptor in some way. no G M ~
or GlcCer). We have shown previously that GMs in
When F28-7 cells were incubated inmedia containing various baby hamster kidney (BHK) cells is closely associated with
quantities of Gh.13, they acquiredGh13 as indicatedby cytofluo- an adhesion matrix (or plaque) which may have an essential
rometricanalysiswith a n t i - G ~ 3mAb DH2 (Fig. 5, inset). role in an unidentified adhesion mechanism (possibly FN-
These G~~-enriched cells, mediated BHK cell adhesion, since BHK cells excrete FN)
F28-7 cells, in contrast to untreated
showed enhancedadhesiontoFN-coatedplates (Fig.5). (10). Before development of theconcept of FNreceptors
GlcCer (100 pg/ml) weakly inhibited F28-7 adhesion to FN- (integrins), it was shown that FN-dependent cell adhesion
coated plates (Fig. 5), suggesting t h a t G M ~ , GlcCer, sup- was inhibited by gangliosides, particularly polysialoganglio-
not
ports integrin receptor function. On the other hand, adhesion sides (7). Such inhibitionwas observed for adhesion not only
of FUA169 cells to FN-coated plates was strongly inhibited to FN-coated plates but also to plates coated with various
by exogenous addition of G M (i.e. ~ incubation of cells with lectins andglycosidases. The phenomenonwas not considered
mediumcontaining 6-50 pg/ml G M ~ )weakly , inhibited by tobe specific, norto imply a role of gangliosides as FN
LacCer, and essentially not inhibited by sialic acid (Fig. 6). receptors (8, 9, 24). Later,there was renewed interest in
Weak inhibition was observed when FUA169 cells were in- possible involvement of gangliosides (particularly GDz and
cubated with GlcCer (data not shown). A n t i - G ~mAb ~ DH2 G D ~ in ) adhesion of humanmelanoma cells tovitronectin
produced only weak inhibition of FUA169 adhesion to FN- receptor, based on two findings. (i) Bindingof anti-GD2mAbs
coated plates (20 and 38-40% inhibition by 1.6-3.1 and 50- and anti-vitronectin receptor mAbs was co-localized at focal
100 pg of DH2 IgG/ml, respectively). Thus, exogenous addi- adhesion plaques, as demonstrated by double-labeled trans-
tion of G M may ~ affect optimal G M concentration
~ in mem- mission immunoelectron microscopy and by immunofluores-
brane, asnecessary for fullfunction of integrin receptor. This cence. (ii) G D ganglioside
~ and vitronectin receptor were co-
hypothesis was supported by additional experiments as de- purified from RGD-containing peptide in presence
the ofCa2+.
scribed below. It was thereforepostulatedthat G D ganglioside
~ plays an
2220 Receptor
Function
Integrin
Regulated by GM,
A. F28-7 1. FUA169 100 -

50 -

0 I In 100 IO00 0 I In 100 1000

LFL w L

FIG. 3. FN receptor expression on surfaces of FUA169 and

F28-7 cells. F28-7 (panel A ) and FUA169 (panel B ) cells were 0 1.6 625 25 50
stained with anti-mouse FN receptor polyclonal antibody 3675 as f% Concentration for Coating (uwrni)
described under “Experimental Procedures.” Fluorescence intensity
was expressed on a log scale (LFL); means of LFL measurements are FIG. 5. Effect of preincubation with G M on ~ adhesion of
indicated. F28-7 cells to FN-coated plates. F28-7 cells were cultured in
media containing 50 or 100 pg/ml G Mfor ~ 24 h at 37 “C, and binding
of anti-& mAb DH2 was determined by cytofluorometry (enhance-
ment of fluorescence is shown in inset). The,cells were metabolically
labeled with [3H]thymidine, and seeded onto Pro-bind plates coated
205 - with various concentrations of FN (abscissa). 0, 100 pg/ml G Mused ~
for incubation (curve 3); A, 50 pg/ml G M (curve
~ 2); 0, control (no
G M ~(curve
) 1 ); A, 100 pg/ml GlcCer (curve 4 ) . Inset, cytofluorometric
-FNR curves for F28-7 cell adhesion underfour conditions. Numbers shown
116 - are mean fluorescence intensity for each condition. Curve 0, negative
control (nonspecific mouse IgG). Curve 1 , no G Mincubation.
~ Curve
h
97.4 - 2, incubated with 50 pg/ml G M ~Curve
. 3,100 pg/ml G M ~ .

-
n
u
66 -
s
E

45 -

29 -
1 2
Inhibitor Concentrution (ug/rni)
FIG. 4. Precipitation of FN receptors from surface-labeled
FUA169 and F28-7 cells. F28-7 (lane I ) and FUA169 (lane 2) FIG.6. Inhibition of FUA169 cell adhesion to FN by GMS.
cells were labeled with andFN receptors were precipitated as GM3 (0)or LacCer (0)were dried in glass tubesunder NP and
described under “Experimental Procedures.” Molecular weight pro- dissolved in medium with sonication. Sialic acid (A) was dissolved
tein standards: myosin, 205 kDa; p-galactosidase, 116 kDa; phospho- directly in medium. Labeled cells were incubated with inhibitors at
rylase b,97.4 kDa; BSA, 66 kDa; ovalbumin, 45 kDa; carbonic various concentrations (abscissa) at room temp for 30 min and then
anhydrase, 29 kDa. SDS-PAGE was performed using 7.5% gel under left to adhere to plates previously coated with FN (50 pg/ml, 100 PI/
reducing conditions. well). Degree of inhibition was expressed as in Fig. 2. Increases in
anti-GM3 mAb binding to FUA169 cells after incubation with various
concentrations of G M (0, ~ 1.6,3.1, 6.2, 12.5, 25, 50, and 100 rg/ml)
essential role in vitronectin receptor function(11).However, were, respectively, 14.8,19.5, 20.9,21.4,22.9,26.9,29.6, and 38.5
expression of G Dand~ G Dis~ limited to human melanoma cells (mean LFL).
(theyarenot found on most normal cells), so the above
observations are not relevant to thegeneral function of inte- was significantly enhanced, whereas for G M concentrations
~
grin receptors. above the optimal range,adhesion was actually decreased
The present study showed a clear differenceof FN/integrin- below control (no G M ~levels.
) When various concentrations
mediated adhesion of G~3-containing mutantFUA169 cells, of LacCer or GlcCer were included in a5pl liposomes, adhe-
in contrast to thatof LacCer-containing parent F28-7 cells. sion to FNwas consistently lower than for controls. Thus, it
FUA169 cells showed much stronger adhesion to FN-coated appears likely that thedifference between FUA169 and F28-
plates (regardless of type of FN) thanF28-7 cells, despite the 7 cells in regards to FN/integrin-mediatedadhesion is due to
fact that bothcell lines express identical quality and quantity (i) the effect of G M ~which
, enhances integrin FN receptor
of integrin receptor. Definitive evidence that Gm3 plays a role function withina certain optimalrange of concentration, and
in integrin-dependent adhesion was provided bythe observa- (ii) the effect of LacCer or GlcCer, which consistently sup-
tion thatweakly adherent F28-7 cells became strongly adher- pressintegrin function. Integrin function(particularly of
ent to FN after G Mwas ~ inserted in the membrane through a5pl)may be enhanced at anoptimal G Mconcentration
~ and
incubation of cells in G~3-COntainingmedium. impaired by excessive GMS, which could explain the suppres-
Another crucial experiment in the present study involved sion of FUA169 cells to FN-coated plates under the latter
cu5pl integrin-containing liposomes with varying content of condition.
G M ~Laccer,
, or GlcCer. Within a certain optimal range of The present study, utilizing aGSL mutantcell line express-
G Mconcentration,
~ adhesion of liposomes to FN-coated plates ing high & and a parent cell line expressing high LacCer
 Integrin Receptor Function Regulatedby G M S 2221

p-:;r7
B

"""""""

1
-0-
0 0.8 2.3 7 20 60 0.8 2.3 7 20 60 0.8 2.3 7 20 60

FN Concentration (pg/rnl) for Coating

I D Control FNR-Liposome

FN Concentration (pg/rnl) for Coaling

FIG. 7. FN binding by a581 integrin liposomes containing various quantities of GSLs. Purified 01501 integrin receptorwas
incorporated in liposomes as described under "Experimental Procedures." Liposomes of constant composition in termsof PC, ['4C]cholesterol,
and (~501,with various quantities of G"3, LacCer, or GlcCer were prepared, and aliquots of -1.25 pg of cholesterol/well were added to Pro-
bind plates previously coated by 1:3 sequence dilution of FN (abscissa). Specific binding was quantitated by scintillation counting, and
expressed as nanograms of cholesterol bound. Panel A , a501 liposomes (55 pg of PC, 33 pg of ['4C]cholesterol, 5 pgof a5Pl receptor)
containing G M2.2 ~ pg (2 nM; A, curue 31, 0.44 pg (0.4 nM; 0, curue 1 ), and 0.088 pg (0.08 nM; 0, curue 2) were compared with control FN
receptor liposomes containing no G M (dashed
~ line C) in terms of binding activity. Panel B, all experimental conditions and symbols as in
panel A except using LacCer instead of G M ~Panel
. C, conditions and symbols as in panel A except using GlcCer instead of GM3. Panel D,
adhesion of control FN receptor liposomes without GSL.

DH (pg/ml) ZH-l ( p g / m l ) RGDS (JAM)

Concentrotion of Inhibitor
in n m o l Gontained in MR-Liposome FIG. 9. Effect of anti-GM3mAb, anti-a581 mAb,and syn-
thetic RGDS peptide on FN binding of a581 liposomes. a501
FIG. 8. Adhesion of FN receptor liposomes containing var- liposomes (55 pg of PC, 33 pg of [14C]cholesterol,0.44 pg of (0.4 nmol)
ious concentrations of GMSto plastic plates coated with a GM3,5 pg of a561 receptor) were prepared as described under "Ex-
constant concentration of FN. Each well of Pro-bind assay plates perimental Procedures." Liposomes were preincubated with anti-GM3
was coated with 12.5 pg/ml FN in PBS, and saturated with 1%BSA mAb DH2 IgG fraction, anti-& mAb ZH1 IgG, fraction, or synthetic
i n PBS. Binding assays were performed using PC/[14C]cholesterol/ peptide RGDS for 1 h at 4 "C, and thenleft to adhere to plastic plates
F N receptor liposomes containing various concentrations of G M(see ~ precoated with FN (50 pglml, 100 pl/well). Degree of inhibition was
"Experimental Procedures").Abscissa, GM3 quantity(nmol)con- expressed as in Fig. 2.
tained in liposomes. Ordinate, difference of adhesion (expressed as
nanograms of cholesterol bound/well) of GMs-containing liposome
versus control liposome. Value for control (i.e. FN receptor liposome and no G M ~provides
, a good opportunity to study therole of
without G M ~was) 26 ng of cholesterol bound/well.
GM3in integrin receptor function. Such mutants may help in
elucidating the functionsof other specific GSLs.
2222 Receptor
Function
Integrin
Acknowledgments-We thank Drs. Steve Akiyama and Ken Ya-
mada (National Institutes of Health, Bethesda, MD) for kind dona-
tion of polyclonal antiserum 3675, and Dr. Steve Anderson for sci-
entific editing and preparation of the manuscript.
REFERENCES
1. Hynes, R. 0.(1987) Cell 48,549-554
2. Yamada, K. M. (1989) in Fibronectin (Mosher, D.F., ed) pp. 47-121,
Academic Press, San Diego
3. Hemler, M. E. (1990) Annu. Reu. Imrnunol. 8, 365-400
4. Ruoslahti, E. (1991) J. Clin. Inuest. 87, 1-5
5. Kojima, N., and Hakomori, S. (1989) J. Biol. Chem. 2 6 4 , 20159-20162
6. Kojima, N., and Hakomori, S. (1991) J . Biol. Chern. 266, 17552-17558
7. Kleinman, H. K., Martin, G. R., and Fishman, P. H. (1979) Proc. Natl.
A d . Sci. U. S. A. 76,3367-3371
8. Rauvala, H., Carter, W. G., and Hakomori, S. (1981) J. Cell Biol. 88, 127-
137
9. Carter, W. G., Rauvala, H., and Hakomori, S. (1981) J. Cell Biol. 88, 138-
1AX
10. OLaxa, Y., Mugnai, G., Bremer, E. G., and Hakomori, S. (1984) Exp. Cell
Res. 155,448-456
11. Cheresh, D. A,, Pytela, R., Pierschbacher, M. D., Klier, F. G., Ruoslahti,
Regulated by G M ~
E., and Reisfeld, R. A. (1987) J. Cell Biol. 105, 1163-1173
12. Akiyama, S. K., and Yamada, K. M. (1987) J . B i d . Chem. 262, 17536-
17542
13. SchwaGmann, G . (1978) Biochim. Biophys. Acta 529, 106-114
14. Engvall, E., and Ruoslahti, E. (1977) Int. J. Cancer 20.1-5
15. Sekieuchi. K.. Siri. A,. Zardi.. L... and Hakomori. S. (1985) J. Biol. Chem.
26o0,5io5-5114’
16. Matsuura, H., Takio, K., Titani, K., Greene, T., Levery, S. B.,Salyan, M.
E. K., and Hakomori, S. (1988) J. Bwl. Chem. 263,3314-3322
17. Nakabayashi, H., Taketa, K., Miyano, K., Yamane, T., and Sato, J. (1982)
Cancer Res. 42,3858-3863
18. Pytela, R., Pierschbacher, M. D., Argraves, W. S., Suzuki, S., and Ruoslahti,
E. (1987) Methods Enzymol. 144,475-489
19. Mimms, L. T., Zampighi, G., Nozaki, Y., Tanford, C., and Reynolds, J. A.
(1981) Biochemistry 20,833-840
20. Pytela, R., Pierschbacher, M. D., and Ruoslahti, E. (1985) Cell 40, 191-
19R
21. St&&, W. D., Raj ara,S. M., Wayner, E. A., Carter, W. G., and Santoro,
S. A. (1989) J. All Blol. 108, 1917-1924
22. Hvnes. R. 0.(1973) Proc. Natl. Acd. Sci. U. S.A. 70.3170-3174
2 3 . Liemmh~U. k ~ ( 1 9 7 0Nature
) 227,680-685
24. Rauvala,’H., and Hakomori, S. (1981) J. Cell Biol. 88, 149-159
25. Tsuruoka. T., Tsuii, T., Noiiri, H., Holmes, E. H., and Hakomori, S. (1993)
J. Biol. Chem. 268,221i-2216