INHIBITION OF FIBRONECTIN-MEDIATED ADHESION OF HAMSTER … · 2005-09-06 · ADHESION OF HAMSTER FIBROBLASTS TO SUBSTRATUM: EFFECTS OF TUNICAMYCIN AND SOME CELL SURFACE MODIFYING

J. Cell Sci. 44, 33-58 (1980)Printed in Great Britain © Company of Biologists Limited 1080

INHIBITION OF FIBRONECTIN-MEDIATED

ADHESION OF HAMSTER FIBROBLASTS

TO SUBSTRATUM: EFFECTS OF TUNICAMYCIN

AND SOME CELL SURFACE MODIFYING

REAGENTS

TERRY D. BUTTERS, VINOD DEVALIA», JOHN D. APLINAND R.COLIN HUGHESNational Institute for Medical Research, Mill Hill, London NWy 1AA, U.K.

SUMMARYUsing baby hamster kidney (BHK) fibroblasts we have studied the effect of tunicamycin,

a specific inhibitor of protein glycosylation, on the ability of trypsinized cells to attach andspread onto fibronectin. Tunicamycin inhibited mannose incorporation into total acid-precipitable glycoproteins by at least 95 % while glucosamine and leucine incorporation wereless or hardly inhibited. Hydrolysis and analysis of PHJglucosamine-labelled glycoproteinsshowed that radioactivity incorporated into cells exposed to tunicamycin was present pre-dominantly as galactosamine, presumably present in O-glycosidically linked glycan chainswhose assembly is insensitive to the drug.

Treated cells exhibit reduced amounts of surface-associated fibronectin and adhere relativelypoorly to plastic or collagen surfaces pre-coated with plasma or BHK cell-derived fibronectinsat the minimum concentrations required to induce nearly quantitative attachment and spreadingof untreated cells. Drug-treated cells do adhere and spread into a bipolar configuration onsurfaces saturated with fibronectin. Cells treated with tunicamycin and then grown in theabsence of the drug revert to a more normal behaviour, indicating that under certain conditionsthe effects of the drug are reversible.

Fibronectin-mediated spreading of trypsinized BHK cells is also inhibited by pre-treatmentof cells with several non-penetrating reagents reactive with cell surface amino groups, namelypyridoxal phosphate, trinitrobenzene sulphonate and fluorescein 5-isothiocyanate. Analysis ofsurface substitution indicates a strong correlation between the extent of amino group substi-tution and inability of treated cells to interact with a fibronectin lattice.

While the extent of attachment under these conditions is normal, cells pretreated with aspecific non-penetrating thiol reagent, p-chloromercuribenzenesulphonate fail to attach tofibronectin-coated culture dishes in a dose-dependent fashion, indicating that a biochemicaldistinction can be made between the processes of attachment and spreading.

We conclude that both iV-glycosidically linked carbohydrate moieties of BHK cell surfaceglycoproteins and primary amine groups present in surface proteins or lipid head groups playa role in interactions of cells with fibronectin, leading to the formation and maintenance of astable well-spread morphology. Both AMinked glycans and surface sulphydryl groups appearto be required for an attachment process which precedes spreading.

• Present address: University of Glasgow, Scotland.Address for correspondence: Dr R. C. Hughes, National Institute for Medical Research, Mill

Hill, London NW7 IAA, U.K.

34 T. D. Butters, V. Devalia, J. D. Aplin and R. C. Hughes

INTRODUCTION

Fibronectin is a high molecular weight glycoprotein present in plasma in highconcentration and in association with many cells in monolayer culture (Yamada &Olden, 1978; Vaheri & Mosher, 1978). Fibronectin mediates the attachment andspreading of a variety of cell types to various substrata, including collagen or poly-styrene, in the short term after removal by proteolysis of the surface-bound fibronectin(Grinnell, 1978).

In previous publications (Pena & Hughes, 1978 a, 6; Hughes, Pena, Clark &Dourmashkin, 1979 a; Hughes, Pena & Vischer, 19796) we reported that the sub-stratum attachment and spreading of trypsinized baby hamster kidney (BHK) cellsrequires a minimum surface density of substrate-adsorbed fibronectin molecules.Below a critical concentration, insufficient contacts can be made and all cells remainweakly attached to the substratum in a rounded morphology with a highly convolutedsurface (Hughes et al. 1979a). Several ricin-resistant mutants of BHK cells deficientin surface carbohydrates (Meager, Ungkitchanukit & Hughes, 1976) due to deficienciesin specific glycosyl transferases involved in assembly of iV-glycosidically linked glycanchains of glycoproteins (Vischer & Hughes, 1979) showed abnormal responses toserum (Edwards, Dysart & Hughes, 1976) or fibronectin (Pena & Hughes, 19786) inthe adhesion system employed. These mutant cells adhered quantitatively less (about3- to 4-fold) to collagen films in the presence of fibronectin compared to the parentalBHK cells or ricin-resistant cells carrying a normal complement of lectin receptors anda normal distribution of glycosyl transferases (Pena & Hughes, 19786; Hughes et al.19796). Furthermore, surface carbohydrate-deficient mutant cells were unable tomaintain a stable spreadout configuration on a fibronectin lattice at, or just above, thethreshold established (Hughes et al. 1979 a) for the parental BHK cells. The mutantcells did spread out on a substratum saturated with adsorbed fibronectin, indicatingthat the mutations had altered interactions between the cell surface and adsorbedfibronectin, raising substantially the threshold number of contacts required to maintaina spread-out configuration. Collateral evidence for an altered ability of the mutant cellsto interact with endogenous fibronectin came from surface-labelling studies (Pena,Mills & Hughes, 1979) showing the absence of cell-associated fibronectin in thecarbohydrate-deficient mutant cells. This loss was not correlated with any defect infibronectin synthesis in these cells since similar amounts were recovered from thespent culture fluids of parental or mutant cells and these products were equally activein promoting the adhesion of parental BHK cells (Pena & Hughes, 19786).

On the basis of these experiments, we proposed (Pena & Hughes, 19786; Hugheset al. 19796) that cell surface carbohydrates play some role in mediating the associa-tion of BHK cells with fibronectin and that altered complex carbohydrates' metabo-lism in ricin-resistant mutant cells contribute to their abnormal adhesive characteristicswhen assessed in fibronectin-containing systems. To explore further the role of cell-surface carbohydrates in fibronectin-mediated adhesion, we have studied the effectsof tunicamycin. This drug blocks the glycosylation of newly synthesized proteins, byinhibition of iV-acetylglucosamine-containing intermediates involved in assembly of

Inhibition of fibronectin-mediated adhesion 35

the core regions of iV-glycosidically linked glycan chains of glycoproteins (Waechter& Lennarz, 1976) and keratan sulphate (Hart & Lennarz, 1978). Synthesis of collagenand proteoglycans, which have affinity for fibronectin (Yamada & Olden, 1978) andcould be involved in cellular interactions with fibronectin, are much less inhibited bytunicamycin (Duskin & Bornstein, 1977 a, b; Hart & Lennarz, 1978) and the drugprovides a relatively straightforward means to probe the role of surface glycoproteinsin various processes (Leavitt, Schlesinger & Kornfeld, 1977; Rosen, Chia, Fung &Rubin, 1978; Duskin, Holbrook, Williams & Bornstein, 1978; Damsky, Levy-Benshimol, Buck & Warren, 1979; Chatterjee, Kwiterovich & Sekerke, 1979),including the adhesive phenomena described here. Although we conclude from theseexperiments that surface carbohydrates play an important role in the cellular responseto fibronectin, the complexity of the system is indicated by the inhibitory effect ofblocking free cell surface amino groups or sulphydryls on the adherence of BHK cellsto fibronectin-coated substrata.

EXPERIMENTAL PROCEDURES

Cells

Baby hamster kidney (BHK) fibroblasts clone C13 from Flow Laboratories, Irvine, Scotland,were grown at 36 CC in Glasgow modified Eagle's medium supplemented with 10% tryptosephosphate, 10 % foetal calf serum and gentamycin. Subculturing was carried out after trypsin-ization of confluent monolayer cultures as described previously (Meager et al. 1976). Cells usedfor adhesion assays (Pena & Hughes, 19786) were dispersed by brief treatment with trypsin,suspended in Eagle's medium containing 10 % foetal calf serum previously depleted of fibro-nectin by 2 passages through gelatin-Sepharose (Pena & Hughes, 1978 a), washed twice withEagle's medium alone and finally suspended at 8 x 10' cells ml"1 in Eagle's medium alone forthe tests.

Labelling of cells

Metabolic labelling of monolayer cultures in 35-mm diameter tissue culture dishes used2 ml Eagle's medium plus tryptose phosphate and serum supplemented with 1-2-5 /tCi/ml"1 ofL-[4,5-'H]leucine (53 Ci/mmol), D-[2-3H]mannose (12 Ci/mmol) or D-[6-3H]glucosamine(38 Ci/mmol). In typical experiments one set of dishes contained tunicamycin (a gift from.Dr W. J. Cuthbertson, Glaxo-Allenburys Research, Stoke Poges) at 025 fig ml"1 final concen-tration (diluted from a 1 mg/ml tunicamycin solution in dimethyl sulphoxide) while controldishes received an equivalent amount of dimethyl sulphoxide. At various times duplicatedishes were removed, and the monolayers washed 3 times with phosphate-buffered saline(PBS) followed by a phosphotungstic acid:perchloric acid mixture (Hughes, Meager & Nairn,1977). The residues remaining were dissolved in 1 ml 0-5 M NaOH and samples were analysedfor protein by the Lowry procedure and for radioactivity using a Packard Tri-carb liquidscintillation spectrometer and a silica gel-dioxan based scintillation fluid. In other experi-ments nearly confluent monolayers were treated for 16 h with various concentrations oftunicamycin. The medium was then removed and changed for one containing a radioactiveamino acid or sugar precursor as before. After a further 2-h pulse the cells were processed asdescribed above.

Surface labelling

Lactoperoxidase-catalysed iodination was carried out using nearly confluent monolayercultures and standard procedures described in full elsewhere (Pena et al. 1979). Alternatively,the washed monolayers were incubated at room temperature with pyridoxal 5-phosphate and


labelled with sodium boroPH]hydride (250 /tCi, The Radiochemical Centre, Amersham) asdescribed in full by Rifkin, Compans & Reich (1972). After lactoperoxidase labelling, the cellswere washed with phosphate-buffered potassium iodide followed by PBS and immediatelydissolved by heating at 90 °C for 3 min in 2 % SDS plus or minus 1 % 2-mercaptoethanol.The samples were stored at — 20 °C until used for electrophoresis.

SDS-polyacrylamide gel electrophoresis

Samples disrupted in 1 % SDS-Tris HC1 buffer, pH 67, and supplemented when indicatedwith 1 % 2-mercaptoethanol were adjusted to final concentrations of 10 % glycerol and o-oi %bromophenol blue and heated at 90 °C for a further 2-5 min. Electrophoresis was carried outusing 7-5 % polyacrylamide slab gels run in o-i % SDS, 25 HIM Tris, 192 IDM glycine at pH 8-3.Suitable marker proteins of known molecular weight included /?-galactosidase (116 K),phosphorylase (94 K), catalase (60 K), aldolase (40 K), carbonic anhydrase (29 K) and myo-globulin (17 K). Gels were fixed and stained for 1-2 h in 50% methanol containing 0-25 %Coomassie blue and then destained overnight by occasional shaking at room temperature in7 % (v/v) acetic acid-5 % (v/v) methanol circulated continuously through a column (2 x 30 cm)of activated charcoal. Slab gels were dried and radioactive bands located by autoradiographyon Fuji RX film using Ilford Fast Tungstate intensifying screens or by fluorography (Bonner &Laskey, 1974).

Immunofluorescence staining of cells

Cells were grown in complete Eagle's medium to near confluence in 35- or ioo-mm diametertissue culture dishes containing coverslips previously sterilized by dipping in ethanol andflaming. The coverslips were washed twice by dipping in PBS and fixed at room temperaturefor 10-15 n^11 in 4% formaldehyde in PBS. After rinsing vigorously in PBS, the coverslipswere incubated in 35-mm diameter dishes containing 0-5 ml anti-hamster plasma fibronectinserum diluted 1 :s with PBS. The antiserum. was raised in rabbits against SDS-polyacrylamidegel electrophoretically purified fibronectin and had been adsorbed with foetal calf serum toremove cross-reacting antibodies to calf fibronectin (Pena, Mills, Hughes & Aplin, 1980).After 30 min at room temperature the coverslips were rinsed vigorously in PBS and stained byincubation at room temperature for 30 min in goat anti-rabbit Ig conjugated to fluorescein(Miles Laboratories Ltd, Stoke Poges). The coverslips were again washed thoroughly andexamined under a fluorescence microscope. Cells were photographed using Ektachrome ASA400film.

Cell adhesion assays

BHK cells were lightly trypsinized and washed free of serum proteins, suspended in Eagle'smedium at approximately io7 cells ml"1 and used within 1 h for adhesion assays. Whenlabelled metabolically with ["SJmethionine, aliquots of the cell suspensions were taken todetermine total cellular radioactivity. Total numbers of cells were estimated similarly using ahaemocytometer. The assays used measure the adhesion of cells to a fibronectin-coated sub-stratum and have been fully described elsewhere (Pena & Hughes, 19786; Hughes et al. 1979 a,6). Briefly, various amounts of a solution (0-25 mg ml"1) of hamster plasma fibronectin or insome experiments BHK cells fibronectin isolated from spent culture fluids (Pena et al. 1980)were added to 35-mm tissue culture dishes or to glass scintillation vials containing 1 ml Eagle'smedium minus serum or tryptose phosphate. After incubation for 1-2 h at 36 °C in a humidifiedatmosphere of 95% air/5% COj an aliquot of cell suspension (0-1-0-2 ml, approximately1-2 x ioe cells) was added to each vessel and incubation was continued at 36 °C. At intervalsthe cells were examined microscopically; non-adherent cells were removed by repeated gentlewashings with warm PBS and when indicated the attached cells were dissolved in I N NaOH(1 ml). Aliquots were then counted for radioactivity. Collagen-coated dishes were producedusing rat skin collagen (Sigma Chem. Co. Ltd, Poole) solution in dilute acetic acid as described(Pena & Hughes, 19786).


Chemical methods

Radioactive hexosamines were isolated from, confluent cells labelled for 1 h with ["C]-glucosamine as follows: cells were scraped from the growth surface into PBS, an equal volumeof 8 N HC1 added and hydrolysis was carried out at 100 °C for 4 h. The solutions were dried,dissolved in water and samples (20 fil) were chromatographed on thin-layer silica plates(5 cm x 20 cm) using propan-1-ol:ethylacetate:water:25 % aqueous NH8 (6:1:3:1, byvolume). Adequate separation of standard glucosamine (6-4 cm from origin) and galactosamine(5-2 cm from origin), detected by ninhydrin, was obtained after 5-8 h development. Theappropriate areas of the chromatogram were scraped off the plate, packed into small columns,and radioactive hexosamines were eluted by a small volume of the developing solvent mixtureinto scintillation vials for radioactive counting.

Ricin toxicity

Ricin, the toxic component of castor beans (Ricinus communis), also called RCAJQ or RCA II,was dissolved in PBS and added (Meager et al. 1976) to confluent monolayers of control orTM-treated cells grown at 36 °C in complete medium. After 2 h at 36 CC the medium wasremoved and the cells pulsed with Eagle's medium minus the usual concentration of leucinebut containing L-[4,5-'H]leucine (1 fiCi/ml) for 1 h at 36 °C. The washed monolayers werethen treated with phosphotungstic acid-perchloric acid and the insoluble residues weredissolved in 0-5 N NaOH. Samples were removed for quantitation of protein and for radioactivecounting.

Lectin binding

The binding assays using f^IJConcanavalin A or ricin (0-1-0-2 mg/ml, 1-2 x io* cpm/ml)were performed on confluent cell monolayers as described (Meager et al. 1976). The totalnumbers of lectin-binding sites for control cells or cells previously treated with tunicamycin(0-25 fig/mi) at 36 °C for 24 h were calculated from Scatchard plots of the binding data.

Substitution of cell surface reactive groups

Pyridoxal phosphate (PLP) (Sigma) in PBS pH 7-5 was added to suspensions of trypsinizedcells (5-8 x 10') at 8-5 mM final concentration. After incubation at room temperature for varioustimes, usually 6 min, Eagle's medium was added in excess, the cells were recovered by centri-fugation and resuspended in Eagle's medium. (0-2 ml). Other treated samples in PBS beforecentrifugation were reduced by addition of either excess sodium borohydride (NaBHj) orsodium cyanoborohydride (NaCN BHS) (Aldrich Chem. Co. Gillingham, Dorset) at approxi-mately 10 mM final concentration. Reduction was at room temperature for 5 min or at 2 °C for10 min, after which the cells were centrifuged and suspended in Eagle's medium. (0-2 ml). Thesuspension was immediately added to fibronectin-coated 35-mm dishes containing i-mlEagle's medium for adhesion assays. In some experiments PH]borohydride (The RadiochemicalCentre, Amersham) was used. The labelled cell samples were extracted with methanol/chloro-form and the extracts and insoluble residues counted for the radioactivity contained in the lipidand non-lipid material.

Trinitrobenzene sulphonate (TNBS) treatment of cells was at 8-5 mM final concentration atpH 7 for 6 min at room temperature in PBS. The treated cells were greater than 90 % viable asshown by exclusion of trypan blue. Extent of modification with TNBS was assayed using theabsorbance at 335 nm of chloroform: methanol extracts and residues, compared with those of anunlabelled control.

Fluorescein 5-isothiocyanate (FITC) (Molecular Probes Inc., Piano, Texas, USA) wasdissolved in PBS, pH 7-4 and added to trypsinized cells (5-8 x 10s) at various concentrationsup to 1 mM. After incubation at room temperature for 10-12 min the cells were diluted into3 ml Eagle's medium, suspended in Eagle's medium (0-2 ml) for adhesion assay. To quantitatecell-bound fluorescein, samples were washed twice with PBS (5 ml), extracted with chloroform-methanol (2:1, v/v) and the residues dissolved in 2 % SDS. Relative intensity of fluorescence


was measured on a Perkin Elmer Fluorescence spectrophotometer MPF-4 using an excitationwavelength of 492 nm and an emission wavelength 520 nm, and using a calibration curveconstructed from fluorescence-intensity measurements obtained with a series of standardsolutions containing FITC bovine serum albumin in the concentration range of interest.Using fluorescence, the partition coefficient of FITC between water and chloroform: methanol(2:1 v/v) was found to be 3-3 x io3.

The sulphydryl reagent, p-chloromercuribenzene sulphonate, dissolved in PBS pH 7-0 wasadded to trypsinized cells (5-8 x io8) at up to 0-4 mM final concentration. Reaction was at roomtemperature for 10 min and the cells prepared subsequendy for adhesion assays as previously.The cells were metabolically pre-labelled with ["SJmethionine.

RESULTS

Tunicamycin inhibits BHK cell protein glycosylation reversibly

Damsky et al. (1979) have shown that tunicamycin (TM) (1 /tg/ml) inhibits theincorporation of fucose into total BHK cell glycoproteins by more than 80 % after4I1 while leucine incorporation is decreased by only 30-40%. Fig. IA shows thatsomewhat lower concentrations of TM (0-25 /ig/ml) abolish mannose incorporationin a 2-h pulse following TM treatment for 16 h while leucine incorporation is hardlyaffected. The effect is very rapid, is seen within 2 h (Fig. 2 A) and is readily reversedby withdrawing TM from the cells, when mannose incorporation rises to controllevels after about 24 h (Fig. 1 B). By contrast, TM inhibits glucosamine incorporationinto total BHK cell glycoproteins (Fig. 1 A) by only a small (40-45 %) but significantextent (Fig. 2B) under the most extreme conditions examined, that is 0-5 /ig/ml TMfor 16-24 n- We can suggest 2 probable reasons for these findings: glucosamine ispresent in iV-glycosidically linked glycan chains in the core region and the primaryeffect of TM appears to be on the formation of iV-acetylglucosamine pyrophosphoryldolichol, the intermediate donor of these residues. However, the addition of moreperipheral iV-acetylglucosamine and sialic acid residue in the Golgi would proceednormally during run-off of nascent glycoproteins in the first hours of TM treatmentbecause these reactions do not involve formation of dolichyl phosphate intermediates(Hughes, 1976; Schachter, 1978). Hence, the flow of radioactivity from a glucos-amine precursor into the peripheral iV-acetylglucosamine and sialic acid residues wouldcontribute to the residual incorporation.

Analysis of the glycoproteins labelled with glucosamine in the presence and absenceof TM provides another explanation for these findings. In control cells radioactivity isfound to be distributed between glucosamine (approximately 55 %) and galactosamine(approximately 45 %) residues as expected since the respective sugar nucleotides aremetabolically interconvertible. In the TM-treated cells, by contrast considerably more(78%) of the radioactivity is recovered as galactosamine. Galactosamine is a majorconstituent of O-glycosidically linked glycan chains, the assembly of which is notdependent on polyprenoid-linked sugar derivatives and would be expected to beinsensitive to TM (Schachter, 1978). These chains evidently contribute a major partof the total carbohydrate constitution of BHK cell glycoproteins since TM inhibitioninduces little significant change in the mobilities of glycoproteins during SDS-polyacrylamide gel electrophoresis (Fig. 2B). The cellular glycoprotein, fibronectin,


is readily identified on such gels as a subunit of 220 K (Fig. 25). It is notable that TMinhibition of incorporation of radioactive glucosamine into fibronectin is not signifi-cantly different to the inhibition into other major cellular glycoproteins (Fig. zB).

0-1 0-2 0-3 0-4Tunicamycin, ug/ml

•i 08 12 16 20Time after TM, h

24

Fig. 1. A. Effect of various concentrations of tunicamycin on the incorporation of3H-precursors into acid-precipitable glycoproteins of BHK cells. Cells were treatedwith the drug at each concentration for 16 h at 36 °C before addition of the radio-active substance and incorporation was measured after a 2-h pulse. O, [*H]leucine;A, PHJglucosamine; # , pjTJmannose. B. Reversal of the inhibition by tunicamycin(TM) of PHJmannose incorporation into BHK cell glycoproteins. Cells were treatedwith 025 /tg/ml drug at 36 °C for 16 h. The cells were then incubated in fresh mediumwithout drug for various periods. At times the extent of incorporation of radioactiveprecursor during a 2-h pulse was measured (O). Control cells (#) had received notunicamycin.

Tunicamycin induces changes in BHK cell proteins

Although TM inhibits very quickly iV-glycosidically linked glycan assembly, it takesseveral hours to register pronounced morphological changes (Fig. 3), altered cellproteins (Fig. zB) and altered cell surfaces (Fig. 4).

Lactoperoxidase-catalysed iodination of the surface of confluent BHK cell mono-layers labels fibronectin prominently as well as several other smaller glycoproteins(Fi 4)- ^ dramatic reduction in surface-exposed fibronectin is seen (Fig. 4) after

4o T. D. Butters, V. Devalia, J. D. Aplin and R. C. Hughes

6005004003002001000

5550 -

40

0 30x

1 20o

I 10

c 14o

I"oI 10

8

6

4

d

0 1 2 3 4 5 6 7 8Period of incubation in tunicamycin, h

Fig. 2. A. Time course of the incorporation of radioactivity into acid-precipitable BHKcell glycoproteins in the presence or absence of tunicamycin. Cells were incubated at36 °C with PHjleucine (A), |7H]glucosamine (B) or L3H]mannose (c) in the presence( • ) of C25 /tg/ml tunicamycin. Control cells (O) received no drug. B. SDS-75 %polyacrylamide gel of [14C]glucosamine-labelled BHK cell glycoproteins. Confluentmonolayers were treated with 0-25 /Jg/ml tunicamycin for the times (h) indicated atthe tops of the columns, the cells were washed free of the drug and pulsed for 2 h with[14C]glucosamine. Whole cells were disrupted in 2 % SDS-2 % 2-mercaptoethanoland equal amounts of protein used for electrophoresis. Top panel: Coomassie bluestaining; bottom panel: radioautography. The positions of fibronectin (FN) and2 bands enhanced in tunicamycin-treated cells are indicated (arrows).

Inhibition of fibronectin-mediated adhesion

0 1 2 4 6 8 24

FN

Fig. zB. For legend see opposite.

CEL44

T. D. Butters, V. Devalia, J. D. Aph'n and R. C. Hughes

' ' •£

B

Fig. 3. Appearance of BHK cell monolayera after treatment with tunicamycin(0-25 fig/mi) at 36 °C for 24 h (A) compared with control culture (B). X 176.

treatment of the cells with 0-25 /tg/ml of TM at 36 °C for 24 h. This reduced labellingwas detected using iodination and less clearly with a small molecular probe (PLP).Cells treated under identical conditions for 8 or 16 h (Fig. 4) exhibit nearly normallevels of surface fibronectin. This effect of TM on surface-labelling has been obtainedwith cells grown to different densities in monolayer cultures and is a reproduciblephenomenon (Fig. 5), in agreement with results obtained with other cells (Duskin &Bornstein, 1977 a). Coincidental with the disappearance of surface fibronectin inTM-treated cells, 2 new polypeptides of molecular weights 92 and 78 K appear (seeFig. 25). These bands are labelled with PLP, indicating a cell surface localization(Fig. 4)-

Inhibition of jbronectin-mediated adhmRnon 43

A

B

C

D

E

F A

B

C

D

E

F

- F

N-

Fig

. 5.

Lac

tope

roxi

dase

-cat

alys

ed i

odin

atio

n of

cel

ls a

t di

ffer

ent

stag

es o

f gr

owth

. A

, B

, to

uch

ing

con

flue

ncy;

c, D

, co

nflu

ent;

E, F

, ve

ry

conf

luen

t cel

ls. T

reat

ed f

or 2

4 h

wit

h tu

nica

myc

in (

0.25

pg/

rnl)

at

36

OC

(B,

D,

F);

con

trol

cel

ls (

A,

c, E

). T

he

Coo

mas

sie

blue

-sta

ined

ge

l is

show

n at

the

left

and

the

rad

ioau

togr

am o

n th

e ri

ght.

Th

e p

osit

ions

of

fibr

onec

tin

(FN

) an

d t

he

2 p

rote

ins

prom

inen

t in

tr

eate

d ce

lls

are

indi

cate

d.


Fig. 6. Immunofluorescence staining of confluent BHK cell monolayers with anti-fibronectin serum (A-C) and FITC-ricin (D, E). Cells were treated with 0-25 /tg/mltunicamycin at 36 °C for o h (A, D) or 24 h (B, C, E), washed and fixed either immedi-ately (A, B, D, E) or after growth for 24 h in TM-free medium (c) and stained either withFITC-ricin or for fibronectin by indirect immunofluorescence with anti-hamsterplasma fibronectin followed with FITC-conjugated anti-rabbit Ig as described inExperimental procedures, x 176.


Immunofluorescence labelling and kctin-binding properties of tunicamycin-treated cells

Control and TM-treated cell monolayers were fixed and stained with antibodyagainst hamster plasma fibronectin using indirect immunofluorescence. Control cellsshowed a dense mat of fluorescent fibres (Fig. 6) between cells and on top of themonolayer. TM-treated cultures by contrast showed a more mixed appearance withpoorly stained areas interspersed with brightly fluorescent regions. In general, theintense fibrillar arrangement seen in control cultures was replaced with a punctate or

100

0-5

< 0 4

c<S 0-3

ree

^ 0-23O

m 0-1

B

^ X\ \v .̂' \ \. \ . X.

20 30 40Ricin added,

2 4 6Con A bound,

8 10

Fig. 7. A. Ricin-mediated cytotoxicity in control and tunicamycin-treated BHK cells.Confluent control monolayers (#) or monolayers treated (Q) for 24 h at 36 °C withTM (o'25 /Jg/ml) were washed and then incubated with various concentrations ofricin for 2 h at 36 °C before addition of [3H]leucine. Radioactivity incorporated withacid-precipitable cell fractions was then measured and expressed as the % of incor-poration into cells not exposed to ricin. B. Scatchard plot showing the effect of tuni-camycin (025 /tg/ml, 36 °C, 24 h) on binding of [mI]Con A to cells. Treated cells,A; control cells, 9 .

patchy fluorescent staining. It should be emphasized that the TM-treated cell mono-layer was intact (see Fig. 3) during these experiments using non-toxic concentrationsof T M , although the majority of cells were clearly grossly distended. After removal ofT M and re-incubation in fresh growth medium for 24 h the overall appearance ofstaining was partly reversed and the patchy fluorescence was diminished in importancewith the reappearance of a more fibrillar-type of staining (Fig. 6).

In order to examine other alterations at the cell surface of treated cells we usedlectins. Lectin binding to the 24-h TM-treated cells was visualized using F I T C -conjugated ricin (Fig. 6) (or Con A (results not shown)) or by direct binding assays.Using P^IJCon A (Fig. 7B) the binding data indicated a small but significant reduc-tion in binding sites from 12-3 x io6 per cell for control cells to 8-9 x io6 per cell forcells treated with T M . p-25I]Ricin binding was similarly reduced after T M treatment(results not shown). Clearly T M has an effect on the expression of lectin-bindingsurface glycoproteins after 24-h treatment and this was confirmed by the smallprotective effect of T M on ricin-mediated toxicity in these cells (Fig. 7A).


Cell adhesiveness reduced after tunicamycin treatment

The effect of long-term TM treatment on the behaviour of trypsinized cells whenadded to a fibronectin-coated surface is shown in Fig. 8. Approximately 70% ofcontrol cells attached rapidly to tissue culture plastic dishes in the absence of serum or

100I

80*

(60

40

20

n

i

-

-

i

' A

1 1

A

\

i 1 I

8 12 16 20 24Time with TM, h

B

10 20Fibronectin added, /

Fig. 8. Effect of tunicamycin on BHK cell attachment. A. [MS]methionine-labelledmonolayer cultures were treated at 36 °C with either 25 /ig/ml tunicamycin (O, A)or i-8 /ig/ml cycloheximide (A) for various times. The washed cells were trypsinized,added to glass scintillation vials previously treated at 36 °C for 2 h with hamsterplasma (GO or BHK medium (A, A) fibronectin solutions (33/Jg/ml) and the %attached cells estimated after a further 90 min at 36 °C. Control cells attachment in theabsence of fibronectin is also shown (#). B. Cell attachment to scintillation vialsprecoated with increasing concentrations of hamster plasma fibronectin or withbovine serum was measured as before. Cell monolayers had been exposed to tuni-camycin (0-25 /ig/ml) at 36 °C for 24 h before trypsinization (O) while control cells(#) received no drug. 0 , O, arrowed indicate values with serum only.

fibronectin (Fig. 8 B). A serum coating reduced this attachment somewhat and fibro-nectin had only a small enhancing effect (Fig. 8 B). Cells treated with TM adheredpoorly to the plastic dishes in the presence of saturating amounts of fibronectin(Fig. 8B). Interestingly, TM-treated cells attached almost as well as control cells toserum-coated dishes, indicating that other adhesive factors are present in serum and

T. D. Butters, V. Devalia, J. D. Aplin and R. C. Hughes

^- B C

coo

• o

0-2 1 10Fibronectin, [ig/m\

Fig. 9. Effect of tunicamycin on fibronectin-induced BHK cell spreading. Plastic35-mm diameter dishes were treated with fibronectin in Eagle's medium (1 ml) at36 °C for 1 h before addition of trypsinized cells. The extent of cell spreading wasexamined 90 min later after incubation at 36 °C by photomicrography (above) orcell counting (below). Tunicamycin treatment at 36 °C was with 0-25 /ig/ml for 24 hand reversal in drug-free medium was for 38 h at 36 CC. # , control; A, TM thendrug removed; O, with TM.


TM-treatment has little effect on the response of cells to these. The TM-effect onfibronectin-mediated attachment is time-dependent and a maximum inhibition isobtained after about 24 h (Fig. 8 A). This is not due to the small inhibition of proteinsynthesis by TM since cycloheximide, an efficient inhibitor of protein synthesis inBHK cells, is without effect (Fig. 8 A).

In addition to inhibition of cell attachment, TM also diminishes the ability ofattached cells to spread out on the fibronectin-coated surface (Fig. 9 A). At concen-trations below 1 /ig/ml fibronectin, most of the attached cells are rounded but thesecells do spread out at higher lattice densities (Fig. 9B). Cells treated with TM andthen incubated for at least 24-36 h without the drug behave more normally (Fig. 9B).

Role of cell surface amines in fibronectin-induced adhesion

The agglutination of cells with chick cell fibronectin was reported to be inhibitedby primary amines (Yamada et al. 1975) and the binding of collagen by fibronectin isalso sensitive to arginine (Vuento & Vaheri, 1978). In agreement with these resultsindicating a binding site for primary amines in the fibronectin molecule, severalsimple amines (lysine, arginine, ethanolamine, glucosamine, galactosamine, spermi-dine) prevent the attachment and spreading of trypsinized BHK cells to fibronectin-coated substrata (Table 1). This is the result of an interaction between amines and

Table 1. Inhibition of BHK cell spreading on fibronectin-coated polystyrene dishes

ist treatment

PBSPLPPLP

TNBS

FibronectinFibronectinFibronectinFibronectinFibronectinFibronectinFibronectin

2nd treatment

Experiment A

NaCNBH3NaCNBH3

NaBH<

Experiment BCellsCells+ 71 mM lysineCells+ 3 5 mM lysineCellsCells+ 71 mM arginineCells + 35 mM arginineCells+ 20 mM ethanolamine

Viability,/o

> 90> 90> 90> 80> 80> 90

> 95

> 95

> 05

Degree ofspreading

+ + + +——

+ +—+

+ + + +—+ +

+ + + +—

+ + +

In experiments A, trypsinized cells were treated with the various reagents as described inExperimental procedures before addition to the fibronectin-coated dishes. Spreading assessedafter 90-180 min at 36 CC.

In experiments B, dishes were first treated with fibronectin for 1 h at 36 °C before washingwith PBS and addition of trypsinized cells with or without inhibitors. Spreading was assessedas A.

Abbreviations: PBS, phosphate-buffered saline; PLP, pyridoxal-5'-phosphate; TNBS,trinitrobenzene sulphonate; NaCNBHj, sodium cyanoborohydride; NaBHi, sodium boro-hydride.


0 0-4 08 1-2FITC added,

2 4 6 8 10 12 14 16Total FITC in lipid and protein, nmol

18

Fig. io. A. Behaviour of cells in spreading assays after labelling with FITC. Extentsof modification assayed by fluorescence as described in Experimental procedures.Inset: FITC in protein ( • ) and lipid (O) fractions of modified cells as a function ofreactant concentration. B. Phase-contrast photomicrography of cells labelled withvarious concentrations of FITC: (A, AO, O; (B, BO, O-2 mM; (c, cO, o-6 mM; (D, DO,I-I mM. A'-D', xo,6; A-D, x i 150. C. Photomicrography of cells labelled using0-2 mM FITC under A, C, phase-contrast; B, D, fluorescence illumination. X450.Several poorly spread cells exhibit weak fluorescence in the cell body which is notvisible in filipodia (arrowheads). Rounded cells fluoresce more intensely than others.

fibronectin since treatment of cells before addition to fibronectin-coated dishes did notprevent spreading.

To assess more directly the role of cell surface amine groups in fibronectin-mediated adhesion, cells were treated with reagents reactive with primary amines,Pyridoxal phosphate (PLP) labels cell surfaces vectorially by forming Schiff bases withfree amino groups of proteins and possibly phospholipids (Rifkin et al. 1972). TheSchiff bases may be stabilized by reduction to secondary amines with sodium boro-hydride. Cells labelled in confluent monolayer cultures in this way show stronglabelling of fibronectin together with several other components of lower molecularweights (Fig. 4). Even after 24 h treatment with TM (0-25 /tg/ml) some surface-associated fibronectin was detectable using this relatively small labelling reagent incontrast to the results already discussed for lactoperoxidase labelling (see Figs. 4, 5).We are unable at present to explain satisfactorily this discrepancy. Possibly inhibitionof protein glycosylation by TM alters in some way the orientation of the fibronectinremaining at the cell surface to restrict access by the macromolecular enzymicreagents while not excluding chemicals such as PLP. Alternatively, sodium boro-hydride may penetrate TM cells to label intracellular material non-specifically.

When PLP was added to trypsinized cells under relatively mild conditions atpH 7-5 and room temperature for up to 16 min, the treated cells failed to spread outover 3 h on a fibronectin-coated surface (Table 1). The cells were viable, excluding

Inhibition of fibronectin-mediated adhesion

Fig. 10 B. For legend see opposite.

trypan blue and over 24 h spreading out as well as control untreated cells, indicatingthat the effect of PLP is slowly reversible. In an attempt to stabilize surface-boundpyridoximine derivatives, treated cells were reduced with sodium borohydride. UsingPHJlabelled borohydride it was shown that at least 60% of the incorporated radio-activity was extractable with chloroform-methanol, representing Schiff base formationwith amino phospholipids and other non-specific incorporation into lipid, and 40%


0

L.I

0 ^ /

Fig. 10C. For legend see p. 50.

in protein. By contrast, using piTjborohydride alone without PLP we found thatmore than 95 % of the incorporated radioactivity went into the lipid fraction. Cellstreated with borohydride alone, although largely impermeable to trypan blue, showedimpaired spreading on fibronectin-coated plastic surfaces and therefore the milder,more specific, reducing agent sodium cyanoborohydride was used. When cells were


treated with this reagent spreading was less impaired, although still somewhat reducedrelative to the control cells. Cells treated with PLP then cyanoborohydride showed thegreatest impairment in fibronectin-mediated spreading (Table i). In a controlexperiment, fibronectin adsorbed to a plastic dish was treated with 10 rriM PLP atroom temperature for 25 min followed with excess 10 mM cyanoborohydride for5 min. After extensive washing, the fibronectin-coated surface was fully active ininducing spreading of untreated trypsinized cells.

/ u

60

50

40

30

20

10

n

O)

\

\

o \

r \-

-

-i

V o

o

1 1 1

0-1 0-2 0-3CMBS added, m

0-4 0-5

Fig. 11. Attachment of [MS]methionine-labelled cells to fibronectin-coated poly-styrene dishes after modification with the thiol reagent ^-chloromercuribenzenesulphonate (CMBS).

Another mild non-penetrating reagent modifying accessible cell surface amines istrinitrobenzenesulphonate (TNBS) (Gordesky & Marinetti, 1973). Cells treated withthis reagent failed to spread on a fibronectin-coated surface in the short term (3 h)but overcame this block after incubation in the adhesion assay for 16—24 b. Tomeasure the amount of cell-bound TNBS, reacted cells (io8) were extracted withchloroform: methanol overnight at 2 °C and the insoluble residue dissolved in 5 %SDS by heating at 100 °C for 15 min. Absorbance at 335 nm showed approximatelyequal amounts (4-5 x io 8 molecules/cell) in the lipid and protein fractions.

Similar results were obtained with fluorescein isothiocyanate (FITC). This reagentsubstitutes both lipid and protein fractions of trypsinized cells (Fig. 10 A) and inhibitscell spreading in a dose-dependent manner (Fig. 10A). Figs. 10B and C show phase-contrast and fluorescence photomicrographs of cells labelled with FITC. Thespreading process is clearly impaired as blocking becomes more extensive (Fig. 10B),

54 T. D. Butters, V. Devalia, J. D. ApUn and R. C. Hughes

cells observed at intermediate stages B and c extending fewer processes and exhibitingreduced surface area and (arrows) thicker margins in the absence of attachments.Among mixed populations (Fig. 10C) rounded cells tend to fluoresce more intensely,suggesting a correlation between extent of blocking and inability to spread. Weaklyfluorescent cells appear with filipodial extensions which do not themselves fluoresce(arrow); in other cases, partially spread cells show edges lacking in processes whichfluoresce more brightly than other parts of the cell (arrow).

In agreement with a previous study (Weiss, 1974), none of these amine reagentsaffected attachment of cells to fibronectin-coated or uncoated polystyrene dishes andwere in sharp contrast to the effects of TM as previously described and a sulphydrylreagent, ^>-chloromercuribenzene-sulphonic acid (CMBS). After treatment oftrypsinized cells the proportion attaching to dishes fell from 60-70% to less than 10%at the highest concentration of CMBS used (Fig. 11), in agreement with earlierfindings of Grinnell & Minter (1979). Cell viability was at least 90% in our experi-ments. It should be noted that a qualitative distinction can be made between inhibitionof adhesion by tunicamycin and by CMBS. In the former case cells were observed tosettle on the plastic surface and in some cases spread out to a bipolar configuration;adhesive contacts were, however, generally weaker than in untreated cells, the cellsbeing readily displaced by gentle washing (Fig. 8). CMBS-treated cells, in contrast,often remained in suspension, failing to form even a passive contact with substratum.

DISCUSSION

Our studies strongly suggest that interaction of BHK cells with fibronectin iscomplex and several classes of surface components are likely to be involved in theoverall mechanism leading to a normal display of surface-associated fibronectin andfibronectin-mediated adhesiveness.

Several groups (Duskin et al. 1978; Olden, Pratt & Yamada, 1978) have shown thatthe synthesis of fibronectin is not preferentially inhibited by tunicamycin. The under-glycosylated fibronectin produced and secreted by TM-treated cells (Olden et al.1978) or by certain ricin-resistant mutant BHK cells deficient in specific glycosyltransferases (Pena & Hughes, 19786) is biologically active in mediating cell adhesion,making it unlikely that the carbohydrate portion of fibronectin plays a role in theinteractions with the cell surface. Therefore, other reasons must be considered for thefailure of TM-treated cells to retain fibronectin in a normal configuration at the cellsurface. Similar aberrations may also explain why these cells do not respond normallyto exogenously added fibronectin in adhesion tests. It could be argued that TMprevents the normal synthesis and export to the cell surface of components involvedin interaction with external fibronectin and organization of the extracellular matrix.Two possible candidates can be provisionally excluded from playing an importantpart in the present experiments. Thus, proteoglycan synthesis and secretion, includingheparan sulphate and hyaluronic acid are relatively insensitive to TM (Hart & Lennarz,1978). Secondly, TM does not block synthesis and secretion of collagen (Duskin &Bornstein, 19776). These findings would implicate other glycosylated products as


necessary in cell-fibronectin interactions, presumably glycoproteins of the plasmamembrane. This idea is corroborated by the fact that neither collagenase nor hepa-rinase pretreatment of wild type BHK cells in suspension affects their behaviour insubsequent spreading assays (J. D.A., unpublished).

The rearrangement and loss of surface-associated fibronectin from confluentmonolayers requires relatively prolonged treatment with TM, considerably longerthan the time at which an efFect on assembly of iV-glycosidically linked glycan chainsis first noted and concurrently there is a marked decrease in the ability of TM-treatedcells to adhere well to a fibronectin-coated substratum. These effects could becollectively explained by depletion from the surface of TM-treated cells of carbo-hydrate chains required for integration with fibronectin. Surface glycoprotein turnoverwith half-lives equal to or less than one cell doubling time has been demonstrated witha variety of cell types (Hughes, Sanford & Jeanloz, 1973; Buck & Warren, 1976) andappears to be prominent in confluent cell cultures such as the ones used in our presentexperiments. Certainly the wholesale replacement of highly glycosylated surfaceproteins by carbohydrate-deficient components in TM-treated BHK cells within24 h would be feasible. There is no indication from our present results, in agreementwith others, that the under-glycosylated proteins of TM-treated BHK cells fail toreach the cell surface and become accessible to surface-labelling reagents.

However, although the 2 separate fibronectin-related phenomena described abovefor TM-treated BHK cells can be discussed in terms of one mechanism involvingturnover of normal cell surface glycoproteins, another possible reason for the TM-induced loss of surface-associated fibronectin must be considered. It has been shownthat non-glycosylated fibronectin produced by TM-treated cells is sensitive to proteo-lysis. It cannot be excluded, therefore, that some of the fibronectin synthesized andsecreted by TM-treated cells is fixed transiently at the cell surface but is quicklyremoved. Indeed, Olden et al. (1978) have shown convincingly that the primary effectof TM is to accelerate the turnover of this glycoprotein. In the normal conditionfibronectin turnover in BHK cells appears to be of the order of 16-24 n> since this isthe time scale necessary to deplete TM-treated confluent monolayers of the surface-associated fibronectin, most of which was present before addition of TM to the cellsand is, therefore, glycosylated normally, and is the minimum time at which surfacefibronectin appears in treated cells after removal of the drug. Experiments are inprogress to test this possible explanation.

The observation that fibronectin-mediated BHK cell attachment and spreadingwas inhibited by primary amines (Table 1), together with the discovery (Grinnell &Minter, 1979; J.D.A., unpublished) that modification of protein carboxyl residues(though not sialic acid) in fibronectin with water-soluble carbodiimide abolished itsactivity in spreading cells led us to investigate the effect of blocking amine function-alities at the cell surface. The 3 reagents chosen for this purpose all contained anionicgroups - phosphate, sulphonate and carboxylate respectively - in order to ensureexclusion from the intracellular compartment. Such exclusion has previously beenshown for all 3 reagents in other systems with mild conditions and over short periodsof reaction as are used here.

56 T. D. Butters, V. Devalia, J. D. ApUn and R. C. Hughes

That pretreatment of cells with any one of these hydrophilic reagents inhibitedspreading, though not attachment, of BHK cells on a carpet of fibronectin is goodevidence (i) that amine groups are involved in the interaction, and (ii) that these arelocated on the extracellular face of the plasma membrane. The intense fluorescence ofFITC and its derivatives, which enabled visualization of modified cells under thefluorescence microscope as well as quantitation of the extent of modification, made it thereagent of choice, since TNBS and PLP are relatively weak chromophores. Theability to visualize modified sites is particularly important since even after labellingcells in suspension, label is not uniformly distributed over the cell population(Fig. 10 C). In the second place, it has enabled us to make a qualitative correlationbetween spreading behaviour and fluorescence intensity in different areas of single,viable cells. Where adhesive processes are extended from cell edges, fluorescence isweak or not detectable; where 'bunching* is visible, that is, where the cell cross-section is more rounded and less flattened, and process extension has not occurred,fluorescence is more intense. The existence of such partially spread cells which over24 h overcome the inhibition and spread normally provides valuable evidence for cellviability under conditions of amine blocking; such controls have not always beenpossible in surface modification studies. We hope to extend these observations tomonitor the behaviour of the microfilamentous system after FITC cell surfacelabelling.

Finally, we have been able to show, in agreement with previous evidence (Grinnell,1978) that attachment and spreading are separate processes between which a bio-chemical distinction can be made. While attachment is prevented by non-specificblocking of sulphydryl groups, again, given the anionic nature of the reagent and themild conditions and short time of reaction likely to be located at the external surfaceof the plasma membrane, it is unaffected by modification of amines, the latterprocedure in turn inhibiting subsequent spreading.

J.D.A. holds an M.R.C. Training Fellowship.

REFERENCES

BONNER, W. M. & LASKEY, R. A. (1974). A film detection method for tritium-labelled proteinsand nucleic acid in polyacrylamide gels. Eur. J. Biochem. 46, 83—88.

BUCK, C. A. & WARREN, L. (1976). The repair of the surface structure of animal cells. J. cell.Physiol. 89, 187-199.

CHATTERJEE, S., KWITEROVICH, P. O. & SEKERKE, C. S. (1979). Effects of tunicamycin on thebinding and degradation of low density lipoproteins and glycoprotein synthesis in culturedhuman fibroblasts. J. biol. Chem. 254, 3704-3707.

DAMSKY, C. H., LEVY-BENSHIMOL, A., BUCK, C. A. & WARREN, L. (1979). Effect of tunicamycinon the synthesis, intracellular transport and shedding of membrane glycoproteins in BHKcells. Expl Cell Res. 119, 1-13.

DUSKIN, D. & BORNSTEIN, P. (1977a). Changes in surface properties of normal and transformedcells caused by tunicamycin. Proc. natn. Acad. Set. U.S.A. 74, 3433-3437.

DUSKIN, D. & BORNSTEIN, P. (1977ft). Impaired conversion of procollagen to collagen by fibro-blasts and bone treated with tunicamycin, an inhibitor of protein glycosylation. jf. biol. Chem.252. 955-962.


DUSKIN, D., HOLBHOOK, K., WILLIAMS, K. & BORNSTEIN, P. (1978). Cell surface morphologyand adhesive properties of normal and virally transformed cells treated with tunicamycin,an inhibitor of protein glycosylation. Expl Cell Res. 116, 153-165.

EDWARDS, J. G., DYSART, J. MCK. & HUGHES, R. C. (1976). Cellular adhesiveness reduced inricin-resistant hamster fibroblasts. Nature, Lond. 264, 66—68.

GORDESKY, S. E. & MARINETTI, G. V. (1973). Asymmetric arrangement of phospholipids in thehuman erythrocyte membrane. Biochem. biophys. Res. Commun. 50, 1027-1031.

GRLNNELL, F. (1978). Cellular adhesiveness and extracellular substrata. Int. Rev. Cytol. 58,65-144.GRINNELL, F. & MINTER, D. (1979). Cell adhesion and spreading factor. Chemical modification

studies. Biochim. biophys. Acta 550, 92-99.HART, G. W. & LENNARZ, W. J. (1978). Effects of tunicamycin on the biosynthesis of glycos-

amjnoglycan8 by embryonic chick cornea. J. biol. Chem. 253, 5795—5801.HUGHES, R. C. (1976). Membrane Glycoproteins: a Review of Structure and Function, pp. 214-228.

London: Butterworth.HUGHES, R. C , MEAGER, A. & NAIRN, R. (1977). Effect of 2-deoxy-D-glucose on the cell

surface glycoproteins of hamster fibroblasts. Eur. J. Biochem. 72, 265-273.HUGHES, R. C, PENA, S. D. J., CLARK, J. & DOURMASHKIN, R. R. (1979a). Molecular require-

ments for the adhesion and spreading of hamster fibroblasts. Expl Cell Res. 121, 307-314.HUGHES, R. C, PENA, S. D. J. & VISCHER, P. (19796). Cell surface glycoproteins in fibroblast

adhesion. In Cell Adhesion and Motility (ed. A. Curtis & J. D. Pitts), pp. 329-356. CambridgeUniversity Press.

HUGHES, R. C, SANFORD, B. H. & JEAMLOZ, R. W. (1973). Regeneration of the surface glyco-proteins of a transplantable mouse tumour cell after treatment with neuraminidase. Proc.natn. Acad. Set. U.S.A. 69, 942-947.

LEAVITT, R., SCHLESINGER, S. & KORNFELD, S. (1977). Tunicamycin inhibits glycosylation andmultiplication of Sindbis and Vesicular Stomatitis viruses. J. Virol. 21, 375-385.

MEAGER, A., UNGKITCHANTJKIT, A., NAIRN, R. C. & HUGHES, R. C. (1975). Ricin in hamsterfibroblasts. Nature, Lond. 257, 137-139.

MEAGER, A., UNGKITCHANUKIT, A. & HUGHES, R. C. (1976). Variants of hamster fibroblastsresistant to Ridnus covtmunis toxin (ricin). Biochem. J. 154, 113—124.

OLDEN, K., PRATT, R. M. & YAMADA, K. M. (1978). Role of carbohydrates in protein secretionand turnover: effects of tunicamycin on the major cell surface glycoprotein of chick embryofibroblasts. Cell 13, 461-473.

PENA, S. D. J. & HUGHES, R. C. (1978a). Fibroblast to substratum contacts mediated by thedifferent forms of fibronectin. Cell Biol. Int. Rep. 2, 339-344.

PENA, S. D. J. & HUGHES, R. C. (19786). Fibronecrin-plasma membrane interactions in theadhesion and spreading of hamster fibroblasts. Nature, Lond. 276, 80-83.

PENA, S. D. J., MILLS, G. & HUGHES, R. C. (1979). Two-dimensional electrophoresis of surfaceglycoproteins of normal BHK cells and ricin-resistant mutants. Biochim. biophys. Acta 550,100-109.

PENA, S. D. J., MILLS, G., HUGHES, R. C. & APLIN, J. D. (1980). Polypeptide heterogeneityof hamster and calf fibronectins. Biochem. J. (in press).

RIFKIN, D. B., COMPANS, R. W. & REICH, E. (1972). A specific labelling procedure for proteinsat the outer surface of membranes. J. biol. Chem. 247, 6432-6437.

ROSEN, O. M., CHIA, G. H., FUNG, C. & RUBIN, C. S. (1978). Tunicamycin-mediated depletionof insulin receptors in 3T3-L1 adipocytes. J. cell. Physiol. 99, 37-42.

SCHACHTER, H. (1978). Glycoprotein biosynthesis. In The Glycoconjugates (ed. W. Pigman &M. Horowitz), pp. 87-181. New York: Academic Press.

VAHERI, A. & MOSHER, D. F. (1978). High molecular weight cell-surface associated glycoprotein(fibronectin) lost in malignant transformation. Biochim. biophys. Acta 213, 1-25.

VISCHER, P. & HUGHES, R. C. (1979). Glycosyl transferases of ricin-resistant baby hamsterkidney cells. In Glycoconjugates (ed. R. Schauer, P. Boer, E. Buddecke, M. F. Kramer,J. F. G. Vliegenthart & H. Wiegandt), pp. 283-284. Stuttgart: Georg Thieme.

VUENTO, M. & VAHERI, A. (1978). Dissociation of fibronectin from gelatin-agarose by aminocompounds. Biochem. J. 175, 333-336.

WAECHTER, C. J. & LENNARZ, W. (1976). The role of polyprenol-linked sugars in glycoproteinsynthesis. A. Rev. Biochem. 45, 95-112.

5 CEL44


WEISS, L. (1974). Studies on cell adhesion in tissue culture. Positively charged surface groupsand the rate of cell adhesion. Expl Cell Res. 83, 311-318.

YAMADA, K. M. & OLDEN, K. (1978). Fibronectin-adhesive glycoproteins of cell surface andblood. Nature, Lond. 275, 179-184.

YAMADA, K. M., YAMADA, S. S. & PASTAN, I. (1975). The major cell surface glycoprotein ofchick embryo fibroblasts is an agglutinin. Proc. natn. Acad. Set. U.S.A. 72, 1217-1224.

(Received 13 December 1979)

Documents

INHIBITION OF FIBRONECTIN-MEDIATED ADHESION OF HAMSTER … · 2005-09-06 · ADHESION OF HAMSTER FIBROBLASTS TO SUBSTRATUM: EFFECTS OF TUNICAMYCIN AND SOME CELL SURFACE MODIFYING