treated cells

(1)

TheEMBOJounal Vol.1 No.8 pp.1011-1016, 1982

Monoclonal antibodies against

contact

sites A of

Dictyostelium

discoideum: detection of modifications of the glycoprotein in tunicamycin-

treated cells

H. Ochiail, J. Stadler, M. Westphal, G. Wagle, R. Merkl, and G. Gerisch*

Max-Planck-Institut fur Biochemie, D-8033 Martinsried, FRG Communicated byG. Gerisch

Received on 22 July 1982

Tunicamycin acts on cell aggregation in Dictyostelium discoideum by changing cell movement and by inhibiting the EDTA-stable type of intercellular adhesion. Tunicamycin- treatedcells show unco-ordinated pseudopodial activity such that pseudopods are simultaneously extended from all parts ofthecell surface, and the cells are unable to move in straight paths. Concurrent with the inhibition of formation of EDTA-stable contacts, N-glycosylation of a glycoprotein specific for aggregation-competent cells is inhibited. This

gly-

coprotein, previously called contact site A, has an apparent mol. wt. of 80kilodaltons (kd). In membranes of tunicamycin-treatedcells, two components are detected that react with certain monoclonal antibodies against contact sites A: one component of 66 kd, the other of 53 kd apparent mol. wt.

Another group of monoclonal antibodies reacts only with the 80-kd glycoprotein and the 66-kd component. These results arein accordwith the assumption that the glycoprotein carries twocarbohydrate chains, and that the antibodies differ in their requirement for glycosylation of the antigen. Despite the coincidence between blockage of EDTA-stable cell adhesion and inhibited glycosylation of contact sites A, direct involvement ofthe carbohydrate moieties of this glycoprotein in intercellular adhesion seems questionable. EDTA-stable cell adhesion hasnotbeen blocked by Fab fragments from antibodies that specifically react with the glycosylated protein.

Key words: glycoprotein/tunicamycin/cell adhesion/monoclonal antibodies/Dictyostelium

Introduction

Aggregating cells ofDictyostelium discoideum adhere to each other.Theyformstream-like aggregatesand,under the influence of thechemo-attractant cyclicAMP,theymoveto-

wards aggregation centers. There, the cells assemble into a

multicellularmass, which eventually developsintoa fruiting body (for reviews on aggregation, see van Haastert and Konijn, 1982; Gerisch, 1982).

Aggregating cells aredistinguishable from those at earlier stages ofdevelopment bythe resistance of their intercellular adhesion toEDTA(Gerisch, 1961). Attempts toidentifythe cell surface molecules involvedinthe adhesionofaggregating cells have ledtotheidentificationofaglycosylated (Mulleret

al., 1979)andphosphorylated (Coffmanetal., 1981;Schmidt andLoomis, 1982) integralmembraneprotein (Stadleretal., 1982), calledcontactsiteA (Beugetal., 1973b). Thisglyco- proteinhas been showntoneutralize univalentantibody fragments(Fab)thatblock the EDTA-stabletypeofcelladhesion 'Present address: Department of Botany, Faculty of Science, Hokkaido

University,060Sapporo, Japan.

*To whomreprint requestsshould besent.

IRL Press Limited, Oxford, England. 0261-4189/82/0108-1011$2.00/0.

(Muller and Gerisch, 1978). It has an apparent mol. wt. of

- 800 kd is absent from growth phase cells, and isexpressed on the cell surface simultaneously with the acquisition of EDTA-stable contact formation (Beug etal., 1973b; Murray etal., 1981; Ochiai et al., 1982).

Tunicamycin blocks N-glycosylation of proteins by inhibiting the transfer of N-acetylglucosamine-l-phosphate from UDP-N-acetyl-glucosamine to polyprenolphosphate (Takatsuki et al., 1975; Tkacz and Lampen, 1975; Lehle and Tanner, 1976). Earlier findings that tunicamycin inhibits cell aggregation of D. discoideum (Ochiai et al., 1981) have prompted us to study the inhibition of EDTA-stablecell contact formation in relation to inhibition of contact site A glycosylation. Monoclonal antibodies reacting with different portions of the glycoprotein have been used to identify two incomplete forms orfragments ofthe glycoprotein.

Results

Inhibition ofEDTA-stable celladhesion by tunicamycin Tunicamycin inhibited cell aggregation in D. discoideum without inhibiting cell motility. Cells were shaken in non- nutrient buffer to which 0.5 or 2.0 ytg tunicamycin/ml was addedduringthe third and fourth hour ofdevelopment.The cellswerethenwashed, andshakingwascontinued.Samples, takenafter 8, 13, and 16 h ofdevelopment, werediluted and plated onto the Teflon film of Petriperm dishes for video- recording. Recording of the 16-h sample of cells treated with 2 ytg tunicamycin/ml was continued until 26 h ofdevelopment. At both concentrations of tunicamycin, aggregation was inhibited but the cells remained motile. The movement differed, however, from normal movement. The tunicamycin-treated cells actively extended pseudopods in all direc- tions, in contrast to normal cells for which polarized pseudopod formation istypical.

Aggregation-competentcells of D. discoideum arecharac- terized by their ability to form EDTA-stable contacts, in additiontoEDTA-sensitiveones whicharealreadyapparent duringthegrowth phase (Gerisch, 1961). Normally, aggrega- tioncompetenceisacquiredinsuspensioncultures ofD. dis- coideumstrain AX2-214within6-8hafter thebeginningof starvation.

The effect oftunicamycinoncelladhesionwasinvestigated under the same conditions used to

study

its effect on cell movement.

Samples

weretaken 8,

16,

and24 h afterstarva- tioncommenced,and celladhesionwasrecordedinan

agglu-

tinometer withorwithoutaddition of 10mM EDTA

(Figure

1).TheEDTA-sensitivecontacts werenotaffected

by

tunicamycin, but the EDTA-stablecontacts wereinhibitedat both concentrations.

Inhibited

expression of

thecontactsiteA

glycoprotein

Concurrent withthe

acquisition

of EDTA-stable cell adhesion, the 80-kd

glycoprotein,

called contact siteA, accumu-

latesintheplasmamembranesof D.discoideum cells. To in- vestigate whether

tunicamycin

affects either the

synthesis

or

expression

of this

glycoprotein

on the cell surface, fluores- cencewas monitored after labeling developingcells with the 1011

(2)

E/E 1.0

f.

-1'

0.5

--

CD

Lu

li

0 0 4 8 12 16 26 24

hours of stirvition

Flg.1.Celladhesion in thepresenceand absence of EDTA.Cells washed free ofnutrientmediumwereshakeninphosphatebuffer.Atthetime pointsindicated,sampleswereremovedfromthesuspension,thecells washed and transferredtoanagglutinometer (BeugandGerisch, 1972).

Opensymbols:agglutinationinthe absence ofEDTA;closedsymbols:in the presence of10mM EDTA. El,E,cells treated with 2.0ygof tunicamycin/ml; A,Awith 0.5Ag/ml;0, 0,untreated controlcells.

Tunicamycinwasadded2h,andremovedbywashing4h,after thebegin- ning ofstarvation(hatched area). Cellagglutinationwasdeterminedby recording unscatteredlight.Eistheapparentopticaldensityin thesample, Eoin anidenticalsampletowhichadhesionblockingFabwasadded for complete dissociationofthecells.Thus,E/Eovalues closetoIindicate mostlysingle cellsintheexperimentalsample,lowE/Eovalues indicate largeaggregates.

monoclonalantibody 12-120-94,which

recognizes

thecontact siteA

glycoprotein (Ochiai

^et

al.,

1982).

Control cellsexpressed

binding

sites forthe

antibody

dur- ingthe first 6 h of

development (Figure

2aand

b).

Incom-

parison, the intensity of

labeling

was reduced within cell populations treated during the third and fourth hour of developmentwith0.1 g of

tunicamycin/ml (Figure

2c). Cells that had been treated with 0.5

itg tunicamycin/ml

showed almost no fluorescence

(Figure 2d).

After 20 h of

develop-

menttheintensityoflabelingwasaboutonethird ofthatin 6-hcontrol cells (Figure 2e).

Incomplete form ofthe contactsite A glycoprotein

synthe-

sizedby tunicamycin-treatedcells

Blottingonnitrocelluloseof membraneproteins separated bySDS-polyacrylamide gel electrophoresis has been used to detect an incomplete form of the glycoprotein synthesized under theinfluenceoftunicamycin. Starvedcells wereculti- vated onMillipore

filters.

Underthese conditions and in the absence of tunicamycin, they passed through all stages of development. Experimentalgroups were

transferred

together with thefiltersonto

fiter

pads soaked with asolution of 8

sg

tunicamycin/ml and incubated for various periods of time (Figure3). Crude membrane preparations of the

cells

were extracted withbutanol/water for enrichment of the contact site A

glycoprotein

inthewaterphase.

In blots from control

cells

the 80-kd bandof the contact site A glycoprotein was strongly labeled with monoclonal

antibody

12-120-94.Twominorbandslabeled with the same

antibody

have beendescribed previously(Ochiai et

al.,

1982).

One ofthesebands, designated 140 kd, is seen

in

Figure 3a.

Thisband appears during cell development together with the

Fig.2.Light scattering(left)and fluorescence(right)ofcellslabeled with antibody12-120-94 andFITC-conjugatedanti-mouse-IgG.Cellpopulations wereanalyzedusing a FACS IVcellsorter.Cellnumbers are plotted on the ordinate, relativelightscattering orfluorescenceintensities on the abscissa.Light scattering isameasure forcellsize, whichdecreases slightly afterstarvation.a,cellsatthe beginningofstarvation; b, 6-h control cells;

c,6-hcellstreated with0.1 gtunicamycin/mlfrom 2 to 4 h of development; d,6-hcellstreated with 0.5ggtunicamycin/mlfrom 2to 4 h of development;e,20-hcellstreated as in d. Average relative fluorescence in- tensities(Ire werecalculated by subtracting the average fluorescence inten- sityinpanela(endogenousfluorescence and unspecific antibody binding) fromtheaverageintensitiescalculated for the other samples. Thecorrected valueforbwas set as 1.

(3)

Monodonal andbodiesagainstaDictyosteliumglycoprotein

_j _ _ ...~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~,::

_d_

Fig. 3.Slow expression ofalower mol. wt.form ofcontactsitesAintuni- camycin treated cells. Top: labeling of antigens blotted from SDS-poly- acrylamidegelswithantibody 12-120-94. Bottom: times of development afterremoval of nutrientmedium (columns), and oftreatmentwith 8 Atg tunicamycin/ml(hatchedareas),for the samples shownontop.

80-kd band and might represent a precursor of the contact site A glycoprotein. Another band, designated 106 kd, ap- pearsduring the slugstage;itappe-arsinFigure 3has aband labeled faintly. This figure also shows that,inagreementwith previous results, the 80-kd glycoprotein disappears almost completely after the aggregationstage.

None of theantigens labeled in controlcells were seen in tunicamycin-treated cells. Instead, a new labeled band ap-

peared in the 66 kd position (Figure 3b-g). Figure 3b^-e

shows thatitissufficienttoincubatethe cells withtunicamy- cinduring the thirdand fourthhours ofdevelopment in order

to suppress fully the appearance of the complete glycoprotein. Synthesis ofcontactsitesAbeginsafter 4hofdevelop- ment.Thismeansthat eitheraprecursorglycolipidis synthesized beforethecommencement ofcontactsiteA synthesis,

ortunicamycin is storedwithin thecells. The latter explana- tion appears to be correct because incubation of cells with tunicamycin during the seventh and eighth hours ofdevelop- menthas ^{an even}stronger effect (Figure3g).

Verylittle of the66-kdcomponentwaspresentafter 8hof development inmembranes oftunicamycin-treated cells, i.e., atastagewhenincontrolcellsthe80-kdglycoproteinwasful- lyexpressed(compareFigure3aandb). Substantialamounts of the 66-kd component were not found earlier than 14 h afterstarvation (Figure 3e).Thedelayintheappearanceofa

modification of the glycoprotein detectable with antibody 12-120-94 is in agreementwiththedataobtained by fluores- centlabeling ofliving cells (Figure 2).

Under all conditions shown in Figure 3b^-g, a discrete band appearedin the 66 kd position. Nomaterial of inter- mediatemol.wt.between 66 and 80 kdwaslabeled. Thismay

beexplained in oneoftwoways: (1)carbohydrate is linked

to the protein in large blocks, so that lack of one block alreadyshifts themolecule^tothe66 kdposition;or(2)carbohydrate is linkedto theprotein in several smaller units, but inour experiments, thetunicamycin concentration was suf- ficiently highto fullyinhibitN-glycosylation. If(2)werecor-

rect,onewouldexpectmaterial of intermediate sizetoappear

with decreasing tunicamycin concentration. Therefore, cells were incubated with either 2.0 or 0.5 jig tunicamycin/ml, under the same conditions used to determine EDTA-stable contact formation.

As showninFigure 4, nointermediateband is seenwhen

a b c d e

Fig. 4. Appearance oftwodiscretebands labeled with[mI]antibody in cells treated with low concentrations of tunicamycin. a,purifiedcontactsiteA glycoproteinas acontrol; b,waterphase from butanol/water extracted crude membrane fractions ofcellstreated from 2to4h ofdevelopment with0.5 isgtunicamycin/ml;c,thesame,butcellstreated with 2.0iLgof tunicamycin/ml;d,the butanol- and interphases of thesameextractasin b;^e,butanol- andinterphases of the^sameextractasinc.Blots fromSDS- polyacrylamide gels^werelabeledwithantibody 12-120-94.

the80-kd bandofthecompleteglycoproteinisclearlydetect- abletogetherwiththe 66-kd band. This result indicates that carbohydrate is linked in large units to the contact site A protein.

Two classes of monoclonal antibodies

For purified contact sites A, a carbohydrate content of 33 mg/100 mg protein had been determined (Muiller et al., 1979). The mol. wt. difference between thecomplete glycoprotein and the 66-kd product of tunicamycin-treated cells

seemedto us too smallto account for the fullcarbohydrate content of theglycoprotein. Therefore, wehave screened a

series of eight monoclonal antibodies for their ability to

recognize products of lowermol. wt. thatmaynotberecog-

nizedby antibody 12-120-94.

All of theseantibodieslabeled the 80-kd bandofthecom-

plete contact site Aglycoprotein. With membranes of tunicamycin-treated cells, two different labeling patterns were

observed: (1) six antibodies reacted, like 12-120-94, with a

66-kd component; (2)two antibodies reacted withtwoanti- genshavingapparent mol.wts.of53 and 66kd,asshown for 20-6-4 (Figure 5). This antibody faintly labeled afew addi- tionalbands, below the 80-kdglycoprotein, whose positions

werenotshiftedintunicamycin-treated cells. Theconclusion is that the 66-kd componentsynthesizedunder the influence oftunicamycinhasanantigenicdeterminant incommonwith thecomplete80-kdglycoprotein, andthatthis determinant is missing inthe53-kd component.

The apparent mol. wts. of66 and 53 kd weredetermined using ^100o SDS-polyacrylamide gels. Inaccordwith itshigh carbohydratecontenttheapparentmol. wts. of thecomplete contact site A glycoprotein varies with the polyacrylamide concentration(Muilleretal., 1979).Under the conditionsused here itwas 77kd.Thus,thedifferences betweentheapparent mol.wts. of the threecomponentsinvolvedare 11and 13 kd.

The 80-kdglycoproteinis knowntocontainphosphoserine (Coffmanetal., 1981; Schmidt and Loomis, 1982).In vivo

incorporation of3P showedthat the 66-kd componentsyn-

1013

a,

^...

140 kd __

antigen 80kd - antigen

E

0-Q

z

-as4)

4)0

O06kd -ntigen -n;gkd

intigen

kd 116

^-

93_-

64-

(4)

i1 elV: 6 Om

4' 64-v

.mm

45 -d

kd

Oh 8h TM

mAb 12-120-94

Oh 8-i TM

mAb 20-6-4

Fig. 5.BlotsfromSDS-polyacrylamide gels of plasma membrane-enriched fractions, labeledwith twodifferent monoclonal antibodies and[125l]anti- mouselgG.The panels showcellsat the beginning of starvation(0h), con- trolcellsat 8 hafterbeginningofstarvation, cellsat 14 h to which 0.5ytg tunicamycin/mlwas added at 3 h of development(TM). The same mem- branepreparationswereused forlabelingwiththe monoclonalantibodies mAb 12-120-94 and 20-64.Comparison of the 0 h and 8hpanelsshows the stringent developmental regulation of contact sites A.

thesized in tunicamycin-treated cells is also phosphorylatable (Figure6).32Pincorporationinto the 53-kd bandcouldnotbe detected, possiblybecause the backgroundwas too high.

Distinction ofmonoclonal antibodies by binding to living cells

The 80-kd glycoprotein is an integral membrane protein (Stadler etal., 1982). Part of themolecule is exposedtothe outermembrane surface; it can be labeled in livingcellswith antibody 12-120-94 (Figure 2). The six antibodies that failed to reactwiththe 53-kd component showed the same pattern ofbinding toliving cells: they labeled the surfaces of aggregation-competent, but not of growth-phase cells. These results indicate that determinants of the 80-kd glycoprotein which are presentonthe 66-kd component, and absent from the 53-kd component, are exposed to the outer cell surface.

These determinantsdistinguishcontactsitesAfromother cell surfaceantigens which are notdevelopmentally regulated.

No surface labeling ofaggregation-competent or growth- phase cells was observed with antibody 20-6-4. The same result was obtained with 20-63-1, the other antibody which labeled the 53-kd component of tunicamycin-treated cells.

These results indicateadeterminant within the membranesor attheir inner surface.

Attemptstoblock cell adhesion by Fab against the glycosylated contact sites A

Contact sites A neutralize polyspecific Fab that blocks EDTA-stable cell adhesion(MullerandGerisch, 1978; Muller etal., 1979). The polyspecificFab had beenprepared from rabbit antisera against whole membranes of aggregation competentcells (Beugetal., 1973b)andcertainlyreactswithmore than one determinant on the contact site A molecule. The question is whether thecarbohydrate moieties of the glycoprotein are among the target structures to which adhesion blockingFab molecules bind.

~~~~~~~~~~~~~~~~~~~~~~~~~~~...

ALA'

^';

-h TM Oh '. 'TM

32",

rabbit IgG

Fig.6. In vivophosphorylation (left panel), andlabelingof membrane proteins with IgG 67/1 from a rabbit that had been immunized with heated, aggregationcompetentcells (rightpanel).[32P]Phosphatewas added tocells 3 h afterthe beginningofstarvation. At thesametime 0.5Agtunicamycin/

ml was added tothe experimental samples. Controlcellswere harvestedat 8h,tunicamycin-treatedcells at 14 hafter the beginning of starvation (TM). Plasmamembraneenrichedfractionswereextracted withbutanol/

water,thewaterphaserun onSDS-polyacrylamidegels and subjected to autoradiography. ForlabelingwithrabbitIgG,cells wereharvestedatthe beginningofstarvation (0h);at8 h for untreatedcells;or at 14hforcells towhich0.5Agtunicamycinwas added at 3 h (TM). Plasma membrane- enriched fractionswereused forSDS-polyacrylamide gel electrophoresis, blotting and labeling with 67/1 IgG and goat[125I]anti-rabbitIgG.

In earlier experiments, we have immunized rabbits with heated D. discoideum cells (Beug et al., 1970). Although Fab from these antisera efficiently bound to the surface of living cells, no inhibition of EDTA-stable cell adhesion was observed (Beug etal., 1973a). IgG from the same rabbit anti- serum K67/1 used in these earlier studies has now been applied to label membrane antigens from tunicamycin-treated and controlcells.

Incontrolcells, many bandswerelabeled (Figure 6). Dif- fuse background labeling was probably due to glycolipid- binding antibodies. Particularly the 80-kd band ofaggregation competent cells was strongly labeled. In tunicamycin- treated cells a band appeared in the same position as the 66-kd band recognized by monoclonal

antibodies,

whereas the 53-kd band was not seen. Since the antibodies of rabbit serumK67/1recognize carbohydratestructuresonglycolipids andglycoproteins (Beugetal., 1970),itislikelythatthe66-kd but not the 53-kd component of tunicamycin-treated cells containscarbohydrate.

The fact that K67/1 Fab doesnotblockEDTA-stablecell adhesion indicates that not all antibodies that bind to the 80-kd glycoprotein also block EDTA-stable cell adhesion, and that carbohydrate moieties of contact sites A are pro- bablynot the target sites ofadhesion blockingFab.

Fab from antibody 12-120-94, which showed the same H.Ochiaietal.

kd 116- 93

64-' 45-

1014

tv %3

^---w

(5)

Monodonal antibodiesaginstaDictyosteliumglycoprotein selectivityfor the 66-kdcomponentas the rabbitantibodies,

also did not block EDTA-stable cell adhesion. The monoclonal Fab was applied at concentrations up to 1 mg/ml

without a detectable effect. For comparison, 0.1 mg poly-

specific rabbit Fab per ml inhibited under the same conditions detectably, and 0.6mgFab perml inhibited completely, the

EDTA-stable adhesion.

Discussion

The inhibition by tunicamycin of

cell

aggregation in D.

discoideum is not due to an overall blockage of

cellular

func- tions; this means inhibition of protein synthesis is not the pre-

vailing

effect of tunicamycin. Tunicamycin-treated

cells

re- main motile for many hours, in contrast to cycloheximide- treated

cells.

The defect in co-ordination of pseudopodial activity seen in tunicamycin-treated

cells

resembles the defects observed in certain mutants (Gerisch, 1980) and suggests the involvement of a glycoprotein in organized

cell

movement.

It is tempting to explain the tunicamycin effect on EDTA- stable celladhesion by the inhibition of contact site A glycosylation. However, for three reasons our results do not pro- vide conclusive evidence for a function of the carbohydrate moiety of this particular glycoprotein in cell adhesion. (1) Fab fragments from antibodies directed against the glycosylated protein, and possibly against the carbohydrate moiety itself, did not block EDTA-stable cell adhesion. (2) Not only the glycosylation, but also the rate of synthesis or the lifetime of the protein might be reduced in tunicamycin-treated cells. (3) The effects of tunicamycin on other glycoproteins have not been studied.

The appearance of a 53- and 66-kd component in tunicamycin-treated cells, and the pattern of their labeling with antibodies, can be explained byassuminga contactsiteAprotein with an apparent mol. wt. of 53 kd to which two identical or similarcarbohydratechains,eachresponsible foran apparent mol. wt. increase of ^- 12 kd, are attached by N-glycosylation. These assumptions are in good agreement with a carbohydrate content of 33 g/100 g protein

(Muller

et

al..,

1979). This content would correspond to atotal mol. wt.

of

the carbohydrate of 17.5 kd, and of -9 kd per chain. Our unpublished data obtained by two-dimensional electrophoresis indicate that thecarbohydrate doesnotsupply additional charges to the molecule. Therefore, the 12kd value for one carbohydrate chain as indicated by SDS-polyacrylamide gel electrophoresis is likely to be an overestimate.

Although we think that the assumption of two N-glycosidically-linked carbohydrate chains is well- supported, two points needto beconsidered: (1) part ofthe carbohydratemight belinked by0-glycosidic bonds;and (2) lack of

glycosylation

could make theprotein highly sensitive to proteases (Olden et al., 1978, 1982).Thus, it isnotexclud- ed that the 66-kd component is the

protein

that lacks N-glycosidically linkedcarbohydrate, and the53-kd material is produced from the 66-kd component by proteolytic cleavage.

If, nevertheless, the assumption is correctthatnotonlythe 80-kd but also the 66-kd component carries a

carbohydrate

chain,antibodieslike 12-120-94canbeconsideredas

specific

for the N-glycosylated contactsites A. Such antibodies bind either to thecarbohydrateitself, or

they recognize

aconfor- mation of theproteinwhich isonly achieved after

glycosyla-

tion, as suggested for an HLA

histocompatibility antigen

(Wilson et al., 1981), and for a murine retrovirus

antigen

(Pierotti et al., 1981). Because ourantibodiesreact with antigens heated with SDS, it isunlikelyto be aspecific conforma- tion of the protein that is recognized.

Previous results have suggested that inD.discoideummore than one type of carbohydrate is transferred from lipid pre- cursors to proteins. Rossler et al. (1978)haveidentified three different precursor mannolipids. Ivatt etal. (1981)discovered a group of protein-linked oligosaccharides that remained the same throughout development up to the completion of cell aggregation. These oligosaccharides differ by their smaller sizes from the carbohydrate linked to the contact site A protein. If antibody 12-120-94 binds directly to this carbohydrate, it should be a unique carbohydrate that distinguishes contact sites A from many other glycoproteins in D. discoideum membranes, which are not labeled by the antibody (Ochiai et al., 1982).

Materialsand methods

Cells of D. discoideum strain AX2-214 were cultivated axenically at 23 k1°Cand starved in 17 mM Soerensen phosphate buffer pH 6.0 as described (Malchow etal., 1972). Except for theexperiment showninFigure 3, which was performed with cells spread onMilliporefilters (HAWP, 0.45

Ampore size), starvedcellswere shaken. For in vivo32P labeling, cellswere washed and resuspended in 10 mMimidazole/HCl bufferpH 6.5, and 30 mCi [32P]phosphatewas added to 400 mlof asuspensionof 1 x107cells/ml. The measurement of cell adhesion and itsinhibition by Fab hasbeendescribedby Beug etal.(1973b).

Monoclonal antibodies were collected from hybridomas obtained from three different BALB/c mice immunized withpartially purified contactsites A, using Al-hydroxide and Bordetellapertussis antigen as adjuvants. The code numbers of monoclonal antibodies, e.g., 12-120-94, denote thenumber of the mouse from which the spleen wasobtained, thewell fromwhich the hybridomawas cloned, and the number of theclonepickedup fromthewell.

Cloning, purification of monoclonal antibodies from theculturefluid, their labeling with 25I, fluorescent labeling ofliving cells with antibodies and fluoresceine isothiocyanate (FITC)-conjugated anti-mouse IgG sheep im- munoglobulins, SDS-polyacrylamide gel electrophoresis, and blotting were performed as before (Ochiai etal.,1982). Fab waspreparedfromamonoclo- nal

IgG,

antibody by incubating 5 mgIgG for 3 h at37°Cwith 0.4-0.6units of papain, bound to CM-cellulose, in 20mlphosphatebufferedsaline con-

taining2 mM EDTA and0.1 Mmercaptoethanol.

Tunicamycin was fractionated byh.p.l.c. and a fractioncorrespondingto peak B of Keenan etal.(1981)was usedforallexperiments. Thepurifiedpeak B tunicamycin waslyophilized, dissolved in ethanol, andkept inastocksolu- tion of 0.5 mg tunicamycin per ml of2007oethanol.

Crude membrane preparations wereobtained from frozen cellsbyceiitri- fugation for 20min at 13 000 g. Eithercrude orplasma membrane-enriched fractions were extracted withbutanol/wateraccordingtoMulleretal.(1979).

Plasma membrane-enriched fractions were prepared from crudemembrane preparations using the polyethyleneglycol/dextran method of Brunette and Till (1971), where the plasma membranes arecollected from theinterphase.

Goat anti-rabbit IgG was a gift fromDr.HeinzSchwarz,Tubingen.Tuni- camycin was purchased fromCalbiochem (LaJolla, CA), FITC-anti-mouse lgG immunoglobulin from the Institut Pasteur production (Paris), papain

bound to CM-cellulose from E. Merck(Darmstadt,FRG), Petripermdishes from Heraeus (Hanau, FRG).

Acknowledgements

We thank Ms. K. Opatz, B.Schonbauer, G. Blank,and Mr. G.Monok for co-operation.

References

Beug,H., Gerisch,G., Kempff,S., Riedel,V., and Cremer,G. (1970) EXp.

Cell Res.,63, 147-158.

Beug,H., andGerisch,G. (1972) J. Immunol. Methods, 2,49-57.

Beug,H., Katz,F.E., Stein,A., andGerisch,G. (1973a)Proc.Nati. Acad. Sci.

USA, 70,3150-3154.

Beug,H. Katz,F.E.,and Gerisch,G. (1973b)J. CellBiol., 56,647-658.

Brunette,D.M., andTill,J.E. (1971)J.Membr. Biol.,5,215-224.

Coffman,D.S., Leichtling,B.H., andRickenberg,H.V. (1981)J. Supramol.

1015

(6)

Struct. Cell. Biochem., 15, 369-385.

Gerisch,G. (1961) Exp. Cell Res., 25, 535-554.

Gerisch,G. (1980) Rheinisch-Westfdlische Akademie der Wissenschaften, Vortrgge N292, WestdeutscherVerlag, Opladen, FRG,pp.7-38.

Gerisch,G. (1982)Annu.Rev.Physiol., 44,535-552.

Ivatt,R.J., Das,O.P.,Henderson,E.J., and Robbins,P.W. (1981) J. Supra- mol.Struct. Cell.Biochem., 17, 359-368.

Keenan,R.W., Hamill,R.L., Occolowitz,J.L., and Elbein,A.D. (1981) Bio- chemistry(Wash.),20, 2968-2973.

Lehle,L., and Tanner,W. (1976) FEBS Lett., 71, 167-170.

Malchow,D., Nggele,B., Schwarz,H., and Gerisch,G. (1972) Eur. J. Bio- chem., 28, 136-142.

MtiIler,K.,andGerisch,G. (1978)Nature, 274,445449.

Millaer,K.,Gerisch,G., Fromme,I., Mayer,H., andTsugita,A. (1979)Eur. J.

Biochem., 99, 419-426.

Murray,B.A., Yee,L.D., and Looniis,W.F. (1981)J.Supramol.Struct.Cell.

Biochem., 17, 197-211.

Ochiai,H., Toda,K., Ono,K., Hanawa,M., Takatsuki,A., and Tamura,G.

(1981) Workshop ofEMBOandtheMax-Planck-GesellschaftonGeneEx- pression and MembraneChangesinCellularSlimeMolds,Abstract.

Ochiai,H., Schwarz,H., Merkl,R., Wagle,G., and Gerisch,G. (1982) Cell Differ., 11, 1-13.

Olden,K.,Pratt,R.M., and Yamada,K.M. (1978) Cell, 13, 461-473.

Olden,K., Bernard,B.A., White,S.L., and Parent,J.B. (1982)J. CellBio- chem., 18, 313-335.

Pierotti,M., DeLeo,A.B., Pinter,A.,O'Donnell,P.V., Hammerling,U., and Fleissner,E. (1981) Virology,112,450460.

ROssler,H., Peuckert,W., and Risse,H.-J.(1978)Mol.Cell. Biochem.,20,3- 15.

Schmidt,J.A., and Loomis,W.F.(1982) Dev.Biol.,91,296-304.

Stadler,J., Bordier,C., Lottspeich,F., Henschen,A., andGerisch,G. (1982) Hoppe-Seyler'sZ.Physiol. Chem.,363,771-776.

Takatsuki,A., Kohno,K., and Tamura,G. (1975)Agric. Biol. Chem., 39, 2089-2091.

Tkacz,J.S., andLampen,J.O.(1975) Biochem. Biophys. Res. Commun.,65, 248-257.

vanHaastert,P.J.M.,andKonijn,T.M.(1982) Mol. Cell.Endocrinol., 26,1- 17.

Wilson,B.S., Glassy,M.C., Quaranta,V., Ng,A.-K., andFerrone,S. (1981) Scand.J.Immunol., 14, 201-205.

1016