CELL ADHESION DURING THE MIGRATORY SLUG STAGE OF DICTYOSTELIUM DISCOIDEUM

Cell Biology International 2002, Vol. 26, No. 11, 951–962doi:10.1006/cbir.2002.0947, available online at http://www.idealibrary.com on

CELL ADHESION DURING THE MIGRATORY SLUG STAGE OFDICTYOSTELIUM DISCOIDEUM

VIVIENNE M. BOWERS-MORROW*, SINAN O. ALI and KEITH L. WILLIAMS†

Department of Biological Sciences, Macquarie University, Sydney, NSW, 2109 Australia

Received 14 February 2002; accepted 18 June 2002

Prespore-specific Antigen (PsA) is selectively expressed on the surface of prespore cells at themulticellular migratory slug stage of Dictyostelium discoideum development. It is a develop-mentally regulated glycoprotein that is anchored to the cell membrane through a glycosylphosphatidylinositol (GPI) anchor. We present the results of an in vitro immunologicalinvestigation of the hypothesis that PsA functions as a cell adhesion molecule (CAM), and of aligand-binding assay indicating that PsA has cell membrane binding partner(s). This is the firstevidence to implicate a direct role for a putative CAM in cell–cell adhesion during themulticellular migratory slug stage of D. discoideum development. Cell–cell adhesion assays werecarried out in the presence or absence of the monoclonal antibody (mAb) MUD1 that has asingle antigenic determinant: a peptide epitope on PsA. These assays showed specific inhibitionof cell–cell adhesion by MUD1. Further, it was found that a purified recombinant form of PsA(rPsA), can neutralize the inhibitory effect of MUD1; the inhibitory effect on cell–cell adhesionis primarily due to the blocking of PsA by the mAb. The resistance of aggregates to dissociationin the presence of 10 mM EDTA (ethylenediamintetraacetic acid) indicates that PsA mediatesEDTA-stable cell–cell contacts, and that PsA-mediated cell adhesion is likely to be independentof divalent cations such as Ca2+ or Mg2+. � 2002 Elsevier Science Ltd. All rights reserved.

K: development; cell–cell adhesion; monoclonal antibody inhibition; Prespore-specific Antigen;Dictyostelium discoideum; migratory slug.

*To whom correspondence should be addressed: E-mail: [email protected]†Present address: Proteome Systems Ltd, North Ryde, Sydney, NSW,2113 Australia

INTRODUCTION

Cellular adhesiveness and cell motility representtwo key, interrelated properties crucial to morpho-genetic processes. A number of families of celladhesion molecules (CAMs) mediate cell adhesion.For example, cadherin/catenin complexes aremediators of cell–cell interactions (Takeichi, 1991;Hatzfeld, 1999), and integrins mediate cell-substrate interactions (Hynes, 1992). CAMs playroles in both the physical connection of cells to oneanother, or to the extracellular matrix (ECM), andin signalling fundamental to differentiation andtissue formation.

1065–6995/02/$-see front matter

CAMs must have been crucial to the origins ofmulticellularity, and to the evolution of develop-ment. Investigation of the molecular eventswhich govern the processes of cellular adhesion inDictyostelium discoideum (cellular slime molds),and other relatively simple model systems such asPorifera (sponges), will facilitate an understandingof the origins of cellular adhesive processes (includ-ing associated signalling) and the roles theyplay in development, as well as the evolution ofmulticellularity and development (Bowers-Morrowet al., 2002).

D. discoideum represents an excellent modeleukaryotic system for the study of the molecularbasis of cell–cell interactions. It has a well-defineddevelopmental life-cycle which is characterizedby the formation of multicellular forms, and cell-type differentiation. Regulated cell–cell adhesionis an important component of D. discoideum

� 2002 Elsevier Science Ltd. All rights reserved.

952 Cell Biology International, Vol. 26, No. 11, 2002

morphogenesis (Coates and Harwood, 2001).While it is not often acknowledged, studies of D.discoideum have played a key role in the study ofcell adhesion. Contact sites A (gp80) was the firstCAM to be identified using an immunologicalapproach (Muller and Gerisch, 1978; Siu et al.,1985; Kamboj et al., 1989); the application of thisapproach to ‘higher’ eukaryotes led to the discov-ery of CAMs, and to contemporary studies ofthe molecular basis of cell contact phenomena inmetazoans (Brackenbury et al., 1977).

The asexual life cycle of D. discoideum has uni-cellular and multicellular phases; it begins with ahomogeneous population of individual, non-associating amoebae (Bonner, 1967; Loomis, 1975;Raper, 1984). Depletion of food resources triggerscAMP-mediated oscillatory cell–cell signalling, andchemotactic migration of amoebae cells. Cellstreams are formed via transient cell–cell contacts,and cells move towards an aggregation centre.Following the formation of a three-dimensionalmound of cells, is the construction of a series ofprogressively more complex multicellular struc-tures; these include the migratory slug which movesforward in a polar fashion toward light, and thespore-bearing fruiting body. The cells of the migra-tory slug differentiate and rearrange to form thesessile fruiting body. While much is known aboutcell contact phenomena at the chemotactic, stream-ing, and aggregation phases of the D. discoideumlife cycle, knowledge about the existence and roleof CAMs at the migratory slug (and later stages)has remained scant.

Two types of cell–cell adhesion systems havebeen distinguished during D. discoideum develop-ment based on the EDTA resistance of cell associ-ation in in vitro cell adhesion assays. Soon after theinitiation of development, cells acquire EDTA-sensitive binding sites (Garrod, 1972). Vegetativeamoebae cells contain EDTA-sensitive cell-bindingsites that mediate the formation of loose cell aggre-gates in liquid cultures; these aggregates are com-pletely dissociated into single cells in 1 or 2 mMEDTA (Gerisch, 1968). A putative cadherin (Ca2+-dependent CAM) of Mr 24,000, DdCAD-1 (gp24),is involved in this cell binding activity. DdCAD-1mediates cell–cell adhesion via homophilic protein-protein interactions (Knecht et al., 1987; Brar andSiu, 1993; Wong et al., 1996). Recently reportedgene disruption studies have indicated thatDdCAD-1 plays a role in cell type proportioningand pattern formation (Wong et al., 2002).

During development, cells form EDTA-resistantbinding sites on their surfaces at the aggregationstage (Beug et al., 1973). Cells form tight aggre-

gates which are resistant to EDTA dissociation upto a concentration of 15 mM. The expression of theGPI-anchored glycoprotein gp80 (contact sites A,csA) that mediates EDTA-resistant (Ca2+/Mg2+-independent) cell–cell adhesion via homophilicoctapeptide interactions coincides with the aggre-gation stage (Muller and Gerisch, 1978; Siu et al.,1985). Expression of gp80 ends when aggregation iscomplete, and the concentration of gp80 moleculeson the cell surface decreases steadily (Siu et al.,1988). Since cells continue to maintain EDTA-resistant cell–cell adhesion (Lam et al., 1981), it isthought that other CAMs are probably responsiblefor EDTA-resistant cell–cell adhesion after theaggregation stage of development. Gp150/LagChas been shown to be a CAM involved in EDTA-resistant adhesion in a postaggregation phase(Geltosky et al., 1979; Gao et al., 1992; Wang et al.,2000). Gp95 has been suggested to play a role inpost-aggregation adhesion although it has not yetbeen proven to be a CAM (Fontana, 1995).

The migratory slug is composed of about100,000 cells with several cell types and sub-types.The slug has an external layer of epithelial-like cells(Fuchs et al., 1993), and a ‘slime trail’ of extra-cellular material is secreted through which theorganism moves (Wilkins and Williams, 1995).Within the slug, the main subpopulations of cellsare prespore and prestalk cells. There is a consider-able degree of heterogeneity in prestalk cell popu-lations (Williams, 1997). The eventual fate of allmembers of the prestalk cell family is identical; theyform non-viable, vacuolated stalk and basal disccells that have cellulosic cell walls. Prespore cellsoccupy the posterior region within the slug, repre-senting about 80% of all the slug cells; this region isnot, however, heterogeneous, having ‘anterior-likecells (ALC)’ scattered throughout it. The presporecells ultimately become the spore cells of the fruit-ing body. The internal organization of the slugis dynamic since the cells move relative to oneanother (Bonner, 1998).

Prepore-specific antigen (PsA) is a developmen-tally regulated glycoprotein, expressed selectivelyon the surface of prespore cells (Browne et al.,1989; Voet et al., 1984; Browne and Williams,1993). The reduced form of PsA from the wild-typestrain WS380B has an apparent molecular weight28 kDa (Zhou-Chou, 1992). PsA is encoded by thepspA or D19 gene that has been sequenced (Earlyet al., 1988). PsA carries a unique epitope recog-nized by the mAb MUD1 (Gregg et al., 1982;Krefft et al., 1983). MUD1 (IgG 2a) is mono-specific for PsA, and recognizes a disulphide-dependent protein epitope in the N-terminal

Cell Biology International, Vol. 26, No. 11, 2002 953

domain (Alexander et al., 1988). MUD1 hasbeen established as a marker for prespore cells.Immunogold labelling experiments have indicatedthat PsA is mainly located at the surface of pre-spore cells (Fuchs, 1993). The level of PsA on thesurface of prespore cells increases progressivelyalong the length of the slug so the rear cells havethe highest surface density of PsA (Browne andWilliams, 1993). Considerable investigation hasbeen carried out on the structural features of PsA;it is anchored to the prespore cell membranethrough a GPI anchor, and it has an O-glycosylated stem region which projects the extra-cellular non-glycosylated protein domain outwardsfrom the cell surface (Early et al., 1988; Hayneset al., 1993; Gooley et al., 1992; Zhou-Chou et al.,1990). Both the GPI anchor (Haynes et al., 1993),and O-glycosylated domains (Zachara et al., 1996;Zachara, 1998) have been characterized.

In spite of developmental, genetic, and structuralstudies, the function of PsA has remained elusive.To date, no role for PsA in cell adhesion has beendemonstrated (Coates and Harwood, 2001). Wehave developed a relatively simple cell adhesionassay for the multicellular slug stage, and presentthe results of an in vitro immunological investi-gation of the hypothesis that PsA is a CAM. Wealso present the results of a ligand-binding assaythat indicates that PsA has binding partner(s) onslug cell membranes. Our data provide the firstevidence in support of a role for PsA in mediatingcell–cell adhesion.

MATERIALS AND METHODS

Bacterial cell culture

E. coli B/r grown on SM agar was used as a foodsource for the development of D. discoideum slugs.

SM agar was prepared using (per litre): 12 gagar; 10 g bacteriological peptone; 10 g glucose; 1 gyeast extract; 1 g MgSO4 ( · 7H2O); 1 g K2HPO4;2.2 g KH2PO4.

Bonners Salt Solution was prepared using, perlitre: 0.6 g NaCl; 0.75 g KCl; 0.4 g CaCl2.

E. coli B/r bacteria were maintained by streakingregularly on sterile SM agar containing the anti-biotic G418 (final concentration 0.01 �g/ml). E. coliB/r streak plates were maintained in a 21�C incu-bator. E. coli spread plates were prepared by takinga sterile sample of the bacteria from a streak plate,and making a suspension of this in sterile BonnersSalt Solution, then spreading the suspension onsterile SM agar in petri dishes; these plates were

kept in a 21�C incubator for 2–3 days beforestorage in a refrigerator. Plates were generally usedwithin a week of storage.

Maintenance of D. discoideum for the developmentof migrating slugs

The wild-type WS380B haploid strain of D. dis-coideum (Erdos et al., 1973) was used for all celladhesion experiments. The vegetative amoebaewere grown in association with E. coli B/r on SMagar at 21��1�C.

Development of migratory slugs

Water agar (1% w/v) was prepared with 20 g agar,0.5 g dihydrostreptomycin sulphate, and 2 l Milli-Qwater. Following autoclaving, the agar was allowedto cool before pouring into sterile petri dishes(90�15 mm). In a laminar flow cabinet, using asterile toothpick, a small volume (5–10 mm3) ofbacteria was scraped from an E. coli plate, andplaced (in a heap) on the surface of a water agarplate. Two or three such heaps were placed on theagar surface, each about 1 cm from the edge, andabout 1 cm apart. Then, at least 3–5 fruiting-bodyheads were collected on the end of a toothpick andmixed into the heaps of E. coli. The plate was thensealed in a black PVC box, lightproof except for a3 mm hole drilled in one side, and placed in atemperature-controlled room (21��1�C). Thespores germinate, amoebae grow and divide, andslugs emerge from the heaps. After 4–5 days theslugs have migrated halfway across the platetowards light entering the hole in the PVC box.

Ligand binding assay

Principle of the assay. The mAb MUD50 (IgG3) recognizes O-linked oligosaccharide chainsattached to each threonine residue in the proline-threonine-valine-threonine (PTVT) repeat regionof the native PsA molecule (on the stem). MUD50also recognizes numerous other glycoproteinsassociated with D. discoideum cells and sheath(Grant and Williams, 1983; Murray et al., 1984;Grant et al., 1985; Alexander et al., 1988; Gooleyet al., 1992). HU2421 is a strain carrying the modBmutation; the modB gene is involved in post-translational modification. modB slugs express PsAmolecules which have no oligosaccharides carryingthe MUD50 determinant. Further, no other modBslug membrane proteins carry the MUD50 deter-minant. These molecules are, therefore, non-reactive to MUD50. Thus, MUD50 binds to PsA


from wild-type WS576 slugs, but not to PsA (orother membrane glycoproteins) from the mem-branes of HU2421 slugs. In this assay, then,designed to show that PsA has slug membranebinding partner(s), native PsA which has bound toHU2421 membrane proteins (nitrocellulose-bound)is detected. (Note that WS576 PsA is recognized bythe mAb MUD1, and has an identical amino acidsequence to the modB PsA).

Method. HU2421 membrane extracts, pureHU2421 PsA, or pure WS576 PsA (all preparedaccording to Zhou-Chou et al., 1990) were solu-bilized in phosphate-buffered saline (PBS)/0.1%Triton-X-100 (PBS (per litre: 9 g NaCl; 0.2 gKH2PO4; 2.9 g Na2HPO4; 2 g KCl, pH 7.2)).Samples were dotted onto dry nitrocellulose, andthe spots allowed to dry at room temperature for1 h. The blot was then washed with PBS for 10 minat room temperature, and blocked with PBS/5%skim milk (w/v) for 20 min at room temperature.Following brief rinses with PBS, pure WS576 PsA(at the same concentration as used for the firstdotting) was dotted over the previously applieddots, and the blot left on damp paper for 5 min atroom temperature. The blot was then washed withPBS at room temperature (5�3 min), and blockedwith PBS/5% skim milk (w/v) at room temperaturefor 10 min. Undiluted mAb MUD1 or mAbMUD50 supernatant was then dotted over thepreviously applied dots, and the blot left on damppaper for 5 min. Following washing in PBS(5�3 min) at room temperature, the blot wasblocked with PBS/5% skim milk (w/v) for 10 min atroom temperature. The blot was then incubatedin PBS/5% skim milk (w/v) containing sheep anti-mouse IgG conjugated horseradish peroxidase(HRP) (Silenus), for 30 min at room temperature.Following washing in PBS (5�3 min), colourdevelopment was achieved using 4-chloronaphthol.

Cell–cell adhesion assays

Collection, and dissociation of slug cells. It haspreviously been shown that PsA is rapidly lost fromthe cell surface on disaggregation of slugs, and thatthis can be reduced at 4�C (Browne and Williams,1993); for this reason all solutions and cell suspen-sions were kept at 4�C, and all cell adhesionexperiments were carried out at 4�C.

Seventy-five slugs were collected from one or twoplates of migrating slugs. Using platinum wire eachslug was transferred into 1.5 ml ice-cold dissoci-ation buffer (20 mM phosphate buffer, 40 mMEDTA, pH 6.5) in an Eppendorf tube. The slugs

were pelleted in a bench Eppendorf centrifuge forabout 10 s, and the supernatant removed. Cellswere resuspended in 1.5 ml of the same ice-coldbuffer, centrifuged, and resuspended. This step wasrepeated 5 times. After the 6th spin and resuspen-sion cycle, the cell suspension was vortexed gently,and then filtered once through buffer-washed20 �m Nytal mesh (Australian Filter Specialists).This filters out large sheets of cells (thought likelyto be epithelial cell sheets), sheath (ECM material),and debris. The collected filtrate was gentlyvortexed, and checked with a haemocytometerunder a light microscope; this generally indicatedthat there was a relatively homogeneous popu-lation of single cells plus a few cell clumps up toabout 10 cells.

Reaggregation of slug cells with or without EDTA.For the experiments without EDTA, the slug cellsuspension was centrifuged (as above), and the slugcells finally resuspended in 1.5 ml ice-cold 20 mMphosphate buffer, pH 6.5. This step was repeated.After gentle vortexing, concentration of single cells/doublets was adjusted to 2–2.5�106 cells/ml usinga haemocytometer.

Five hundred �l suspensions of cells wereallowed to reassociate (a) in the presence orabsence of MUD1 (in PBS, pH 7.4; final concen-tration 100 �g/ml); (b) in the presence or absenceof Sheep IgG polyclonal antibody (Silenus; finalconcentration 100 �g/ml).

For each reassociation experiment a pair ofsamples were prepared, and put in Agilent 2 mlclear glass vials with teflon-lined screw-top caps,and vortexed gently. The samples were then incu-bated on ice for 10 min to allow time for molecularinteractions (e.g involving MUD1) with cell surfacecomponents to take place. Note that cell aggre-gation begins during this incubation period. Ahaemocytometer count of single cells/doublets wasthen recorded for each sample. (Vials were invertedtwo or three times before counts were done, and thecells were allowed to settle on the haemocytometerslide before counting.) This represents the zeromixing time count. The vials were then placedhorizontally in crushed ice in a plastic tray on arotary orbital mixer (Ratek Instruments Australia),and samples were allowed to mix gently for 45 minat 70 r.p.m. Counts of single cells/doublets weremade at 15, 30, and 45 min after the start ofmixing. The percentage of cells recruited intoaggregates was calculated relative to the startingcount recorded before incubation of samples onice, and the mean values from 3–5 independentexperiments plotted against time.


For the experiments with EDTA the methodsused were similar, except that the reassociationbuffer was 20 mM phosphate buffer containing10 mM EDTA, pH 6.5. The starting count rangewas 2.5–3.0�106 cells per ml.

Statistics. The following statistical tests were car-ried out to determine significant changes withrespect to time within each treatment group, andto determine significant differences between treat-ments: (i) two-sample t-tests assuming unequalvariances; and (ii) single factor One Way Analysisof Variance (ANOVA).

Inhibition of cell reassociation by MUD1monoclonal antibody (with EDTA): dosage effect

For the cell adhesion assays, the materials andmethods used were similar to those for the experi-ments described above. The reassociation bufferused was 20 mM phosphate buffer with 10 mMEDTA, pH 6.5, final volume 250 �l, containingstarting cell densities in the range 2.5–3�106 cellsper ml. Cell suspensions were incubated on ice for10 min with final concentrations of MUD1 in therange 0.1–100 �g/ml, or with Sheep IgG polyclonalantibody (either 1 or 10 �g/ml Sheep IgG). Cellswere allowed to mix and reassociate for a period of15 min, then haemocytometer counts of single cells/doublets were taken.

Neutralization of MUD1 inhibition by rPsA

In order to neutralize the inhibitory effect ofMUD1, rPsA was used; this is a recombinant(expressed in E. coli) truncated form of PsA, asoluble monomer of apparent molecular weight27 kD, lacking a GPI anchor (Zhou-Chou et al.,1995; Zachara et al., 1996; Slade et al., 1997). rPsAis bound by MUD1, but ligand binding assayssimilar to those described above for native PsAhave indicated that rPsA does not bind to nativemembrane proteins, and is, therefore, not likely todirectly interfere with PsA/binding partner inter-actions. Thus, rPsA will preferentially bind toMUD1, and neutralize its inhibitory effect onaggregation.

For the cell adhesion assays, the materials andmethods used were similar to those for the experi-ments described above. The reassociation bufferused was 20 mM phosphate buffer with 10 mMEDTA, pH 6.5 (final volume 250 �l), and containedstarting cell densities in the range 3.5–4�106 cellsper ml. Samples used were cell suspensions incu-bated on ice (a) in the presence or absence of

5 �g/ml MUD1; (b) in the presence or absence of amixed solution containing 5 �g/ml MUD1 and9.0 �g/ml rPsA (10� molar excess). Followingincubation on ice, the cells were allowed to mix andreassociate for a period of 15 min; haemocytometercounts of single cells/doublets were done at the endof this mixing period.

The results are presented as a histogram usingdata representing the mean of three independentexperiments.

RESULTS

Ligand binding assay

The results of the ligand binding dot assay shownin Figure 1 show that mAb MUD1 binds to bothwild-type PsA (WS576) and modB PsA (HU2421)(rows (a) and (d)). It is also shown that, while mAbMUD50 binds to wild-type PsA, it does not bindto modB PsA (rows (b) and (c) respectively). WhileMUD1 binds to protein (PsA) extracted frommodB membranes, MUD50 does not bind to pro-teins extracted from these membranes (rows (f) and(g) respectively). Under the experimental con-ditions used here, pure wild-type PsA did not bind toimmobilized pure modB PsA, while wild-type PsAdid bind to immobilized protein(s) extracted frommodB membranes (rows (e) and (h) respectively).

Inhibition of cell aggregation by MUD1 (withoutEDTA): tracking of recruitment of cells intoaggregates

When dissociated slug cells were allowed to mixand reassociate in 20 mM phosphate buffer,pH 6.5, cells rapidly adhered together to form largeamorphous cell aggregates (Fig. 2A). After 15 minmixing, observation of samples on a haemocytom-eter slide indicated that relatively few single cellsremained in the suspension. The aggregates insuspension were visible by eye in the mixing vial.(Note that the process was very rapid; clumpingwas visible by eye after only 5 min mixing.) Whenviewed under a light microscope (using phase con-trast), large aggregates appeared to be multilayeredi.e. 3D structures rather than 2D sheets; the cellsappeared to be closely packed together. In manyaggregates, individual cells were not discernible; itappeared they had been masked by slime sheathmaterial that had been secreted by cells followingreassociation.

By contrast, such aggregation of dissociated cellswas inhibited by the incubation (before mixing) of


Fig. 1. Ligand binding dot assay. HU2421 proteins extractedfrom delipidized membranes, pure wild-type (WS576) PsA, orpure modB (HU2421) PsA were dotted onto nitrocellulose(2 �l per dot), and probed with mAb MUD1, mAb MUD50(2 �l/dot), or pure WS576 PsA (2 �l/dot) followed by MUD50(2 �l/dot), according to the principle and protocol described inMaterials and Methods. (a) WS576 PsA dotted in a dilutionseries (1, 1/3, 1/9, 1/27 etc) [2 �l/dot] and probed with MUD1(2 �l/dot); binding is indicated. (b) WS576 PsA dotted in adilution series, and probed with MUD50; binding is indicated.(c) HU2421 PsA dotted in a dilution series, and probed withMUD50; no binding evident. (d) HU2421 PsA dotted in adilution series, and probed with MUD1; binding is indicated.(e) HU2421 PsA dotted in a dilution series, and probed firstwith WS576 PsA followed by MUD50; no binding is evident.(f) HU2421 membrane extract dotted in a dilution series, andprobed with MUD1; binding is indicated. (g) HU2421 mem-brane extract dotted in a dilution series, and probed withMUD50; no binding is evident. (h) HU2421 membrane extractdotted (2 �l dots undiluted), and probed first with WS576 PsAfollowed by MUD50; binding is indicated. Note that, in asimilar assay [results not shown], a control was included tocheck for non-specific binding; PBS/5% (w/v) bovine serumalbumen was probed with WS576 PsA, then MUD50. In thiscase no binding was evident.

Fig. 2. Differential Interference Contrast (DIC) photomicro-graphs of slug cells allowed to aggregate following pre-incubation with or without MUD1. Slugs were dissociated,and the slug cells allowed to mix and aggregate followingpre-incubation with or without MUD1, according to theprotocol described in Materials and Methods. DIC photo-micrographs were taken after 30 min of mixing. (A) Large cellaggregates formed when slug cells not pre-incubated withMUD1 were allowed to aggregate. Bar=100 �m. (B) Small cellclumps and non-aggregated cells, indicating lack of recruit-ment into aggregates following pre-incubation of cells with100 �g/ml MUD1, and 30 min mixing.

dissociated cells with 100 �g/ml of mAb MUD1.After 15 min of mixing, only single cells/doublets,and small clumps of cells (consisting of up to 10cells in number) were present. No large aggregateshad been formed. See Figure 2B.

The plots shown in Figure 3 for the cases withand without MUD1 indicate that, during the45 min mixing period, there was a significantincrease in percentage cell recruitment of cells intoaggregates with time (ANOVA; P-values <0.05).There was a significant difference between meanpercentage cell recruitment at 0 and 15 min (t-tests;P-values <0.05), after which there was no signifi-cant increase in cell recruitment. Thus, maximumcell recruitment into stable aggregates had occurredafter 15 min mixing.

As a control, dissociated slug cells were allowedto aggregate following incubation with pure SheepIgG polyclonal antibody; the amount of polyclonalantibody used was equivalent to the amount ofMUD1 used. As for the case without MUD1, cellsrapidly adhered together to form large cell aggre-gates. Maximum cell recruitment into aggregateshad occurred after 15 min of mixing, and no fur-ther significant change occurred during the next30 min. See Figure 3.

Statistical tests confirmed significant differencesbetween percentage mean cell recruitment for thecases with and without MUD1; the differencesbetween the means at 0, 15, 30, and 45 min werefound to be significant in all four tests (t-tests;P-values <0.05). Significant differences were also


confirmed between percentage mean cell recruit-ment for the cases with MUD1 and with sheep IgGpolyclonal at 0, 15, 30, and 45 min. Differencesbetween percentages mean cell recruitment forthe cases without MUD1 and with Sheep IgGpolyclonal were found to be not significant.

The plots presented in Figure 3 indicate that forthe case without MUD1, 50% cell recruitment hadoccurred before the start of mixing. By contrast, forthe case with MUD1, only 25% cell recruitmenthad already occurred before the start of mixing.For both cases, the greatest rate of cell recruitmentinto aggregates occurred during the first 15 min ofmixing; a further 40% of cells without MUD1 wererecruited, resulting in a total of 90% recruitmentinto aggregates. By contrast, a further 30% of cells

with MUD1 were recruited during the first 15 min,resulting in a total of 55% cell recruitment.

Fig. 3. Inhibition of cell recruitment into aggregates by mAbMUD1 IgG (without EDTA) as a function of time. Slugs weredissociated, and the slug cells allowed to mix and aggregatefollowing pre-incubation with or without MUD1, according tothe protocol described in Materials and Methods. Counts ofsingle cells/doublets were made using a haemocytometer;counts were made at 15, 30, and 45 min. The 0 min count wastaken just prior to the pre-incubation period. Per cent cellrecruitment was calculated by determining the count of singlecells/doublets at x min, subtracting this from the 0 min count,then dividing by the 0 min count. Per cent cell recruitment isplotted here versus mixing time. Data represent the meanvalues obtained from 3 to 5 independent experiments. Eacherror bar represents the standard error of the mean. Adhesioncurves are shown as follows: open square, solid line, cellsallowed to aggregate following pre-incubation with 100 �g/mlof MUD1; open triangle, dotted line, cells allowed to aggregatefollowing pre-incubation without MUD1; cross, dashed line,cells allowed to aggregate following pre-incubation with anamount of Sheep IgG polyclonal antibody equivalent to100 �g/ml of MUD1.

Inhibition of cell aggregation by MUD1 (with10 mM EDTA): tracking of recruitment of cellsinto aggregates

During development of the protocol, preliminarytests indicated that dissociated slug cells can readilyaggregate in phosphate buffer containing 2.5 mMEDTA, pH 6.5. With 40 mM EDTA, aggregationof dissociated cells is strongly inhibited. Followingdissociation with 40 mM EDTA, cells can thenaggregate in the presence of 10 mM EDTA; theaggregates resist dissociation. For the experimentsreported here, cells were dissociated in 40 mMEDTA, and allowed to aggregate in 10 mMEDTA.

The aggregation of dissociated slug cells wastracked from the beginning of the incubationperiod on ice until 75 min mixing time. The plot inFigure 4 shows the percentage of dissociated cellsrecruited into cell clumps versus mixing time for thefirst 45 min. For the case without MUD1, about18% of cells have already been recruited by the endof the incubation period. During the mixing time,the greatest rate of cell recruitment is during thefirst 15 min, with about 55% of dissociated slugcells being recruited during this time. Maximumcell recruitment has occurred by 30 min whenabout 70% of slug cells have now been recruitedinto cell clumps.

The cell clumps formed during this process inphosphate buffer with added EDTA are smaller insize than the larger aggregates formed in phosphatebuffer without EDTA. Further, as observed undera light microscope (with phase contrast), in manyclumps, consisting of up to 20 cells, the cells do notappear to be tightly packed together; the borders ofindividual cells can be easily seen (not masked bysecreted slime sheath). However there are a fewmultilayered clumps, with the clumps consisting ofmore than 50 cells.

During the 0–75 min mixing period there wassignificant change in percentage cell recruitmentwith time (ANOVA; P-value <0.05). There wasa significant difference between mean percentagecell recruitments at 0 and 15 min (t-test; P-value<0.05), with no statistical change after this time.Thus, maximum cell recruitment had occurred after15 min, as in the experiment done without EDTA.

For each of the cases (i) with MUD1; (ii) withSheep IgG polyclonal, during the 75 min mixingperiod there was significant change in per-centage cell recruitment into aggregates with


time (ANOVA; P-value <0.05). There was asignificant difference between mean percentage cellrecruitment at 0 and 15 min (t-test; P-value <0.05).However, differences between mean percentagescell recruitment at 15 and 30 min, and at 30 and45 min were not statistically significant. Thus, nosignificant cell recruitment occurred after 15 minmixing, and the aggregates already formed werestable.

For the case with MUD1, a few per cent of cellsare recruited by the end of the incubation period onice. The greatest rate of cell recruitment is duringthe first 15 min mixing time, with about 30% ofdissociated slug cells being recruited during thistime.

For the case with Sheep IgG polyclonal, morethan 30% of cells have already been incorporated

into cell clumps by the end of the incubationperiod on ice. The greatest rate of cell recruitmentis during the first 15 min of the mixing time.Maximum cell recruitment has occurred after15 min mixing time; about 75% of cells have nowbeen recruited into cell clumps.

For the cases with and without MUD1, sig-nificant differences between the correspondingmeans at 0, 15, 30, and 45 min, for percentage cellrecruitment, were found (t-tests; P-values <0.05).Significant differences between percentages meancell recruitment were also found for the caseswith MUD1, and with Sheep IgG (P-values all<0.05).

Comparison of Figures 3 and 4 indicates that,with MUD1, in either the presence or absence ofEDTA, about 30% of cells are recruited into cellaggregates during the first 15 min of mixing (25 to55% for the case without EDTA; 5 to 35% for thecase with EDTA). However, for the case withEDTA, only a few per cent of cells have beenrecruited before mixing starts.

Statistical testing indicates that, with or with-out EDTA, the plateaus are significantly differentfor the cases with MUD1, and for the caseswithout MUD1 (t-tests; P<0.05). Thus, thedownward shift apparent on comparison of Fig-ures 3 and 4 is significant. However, the fact thatEDTA seems to have some effect on cell recruit-ment, does not impinge on the fact that MUD1inhibition is significant whether EDTA is presentor not.

Fig. 4. Inhibition of cell recruitment into aggregates by puremAb MUD1 IgG (with 10 mM EDTA) as a function of time.Slugs were dissociated, and the slug cells allowed to mix andaggregate following pre-incubation with or without MUD1,according to the protocol described in Materials and Methods.Counts of single cells/doublets were made, using a haemo-cytometer, at 15, 30, and 45 min. The 0 min count was takenjust prior to the pre-incubation period. Per cent cell recruit-ment was calculated by determining the count of single cells/doublets at x min, subtracting this from the 0 min count, thendividing by the 0 min count. Per cent cell recruitment is plottedhere versus mixing time. Data represent the mean valuesobtained from 3–6 independent experiments. Each error barrepresents the standard error of the mean. Adhesion curves areshown as follows: open square, solid line, cells allowed toaggregate following pre-incubation with 100 �g/ml of MUD1;open triangle, dotted line, cells allowed to aggregate follow-ing pre-incubation without MUD1; cross, dashed line, cellsallowed to aggregate following pre-incubation with 100 �g/mlSheep IgG polyclonal antibody.

Inhibition of cell reassociation by MUD1 (withEDTA): dosage effect

In order to check the specificity of the inhibitoryeffect of purified MUD1 IgG on cell–cell reassoci-ation (in the presence of 10 mM EDTA), celladhesion was assayed following pre-incubation ofcells in various concentrations of MUD1 rangingfrom 0.1 to 100 �g/ml. Figure 5 shows that there isa dose responsive relationship between amount ofMUD1 and per cent cell recruitment into aggre-gates. Maximum recruitment occurs with MUD1concentrations in the range 0.1 and 1.0 �g/ml, withabout 75% of cells being recruited into aggregates.Fifty per cent inhibition is brought about by about3 �g/ml MUD1, and maximum inhibition wasachieved at 10 �g/ml with about 40% cell recruit-ment into aggregates. As a control, Sheep IgGpolyclonal antibody was used; it showed no effecton cell adhesion in the concentration range 0.1 to100 �g/ml.


Neutralization of MUD1 inhibition by purifiedrPsA

If the inhibitory effect of MUD1 on cell adhesionwas due to the blocking of PsA on cell surfaces,then purified rPsA should be able to neutralize thiseffect of MUD1.

Note that ligand binding experiments similar tothose described above for native PsA, but usingrPsA as the probe instead, have indicated that thisform of PsA does not bind to proteins extractedfrom slug cell membranes [results not shown]. Thisis presumably because the recombinant form ofPsA has a different three-dimensional configurationfrom the native form. Nevertheless, rPsA is boundby MUD1, and would, therefore, be expected to bean effective neutralizer of MUD1 activity.

For the case involving no MUD1, cell clumpswere visible in the vial by eye within 5 min ofmixing. After 15 min mixing, very large cell aggre-gates were observed under a light microscope; someaggregates appeared to be multilayered. There werefew remaining single cells. The percentage cell

recruitment measured as previously described was85% (see Fig. 6). For the case with MUD1 (finalconcentration 5 �g/ml), after 15 min mixing nolarge cell aggregates were observed under a lightmicroscope; the largest clumps were comprised ofabout 20 cells. The percentage cell recruitment was60% (see Fig. 6). The difference between the meanswas significant (t-test; P-value <0.05).

For the case involving a mix of MUD1 andrPsA, cell clumps in the vial were visible by eyewithin 5 min of mixing. After 15 min mixing, verylarge aggregates were again observed under a lightmicroscope; there were few remaining single cells.The percentage cell recruitment was 87% (see Fig.6). The differences between the means for the casewith MUD1 only and the case with MUD1+rPsA,were significant (t-test; P-value <0.05). There wasno significant difference between the means for thecases without MUD1, and with MUD1+rPsA.

The inhibitory effect of MUD1 IgG can beneutralized by pre-incubation of cells with amixture containing 5 �g/ml MUD1 and 9 �g/mlrPsA (10� molar excess); the percentage cellrecruitment is 87% (see Fig. 6).

Fig. 5. Inhibition of cell reassociation by MUD1 (with 10 mMEDTA): dosage effect. Slugs were dissociated, and the slugcells allowed to mix and aggregate following pre-incubationwith various amounts of MUD1; the range of concentrationswas 0.1 to 100 �g/ml (see Materials and Methods). The dose-response curve is shown here as per cent cell recruitment versusconcentration of MUD1 (�g/ml). Data represent the meanvalues obtained from three independent experiments. Eacherror bar represents the standard error of the mean. Adhesioncurves are shown as follows: open square, solid line, cellsallowed to aggregate following pre-incubation with MUD1;cross, dashed line, cells allowed to aggregate following pre-incubation with amounts of Sheep IgG polyclonal antibodyequivalent to the concentrations of MUD1 used.

Fig. 6. Neutralization of MUD1 inhibition by rPsA. Thehistogram shows per cent cell recruitment for (i) the caseinvolving pre-incubation of cells without MUD1; (ii) the caseinvolving pre-incubation of cells with 5 �g/ml MUD1; (iii) thecase involving pre-incubation of cells with a mixed solutioncontaining 5 �g/ml MUD1, and 9 �g/ml rPsA (10� molarexcess). Slugs were dissociated, and the slug cells allowed tomix and aggregate following pre-incubation according to theprotocol described in Materials and Methods. Sampleswere allowed to mix for 15 min. Values represent the means ofthree independent experiments. Each error bar represents thestandard error of the mean.

DISCUSSION

CAMs are cell recognition molecules that areinvolved in direct cell–cell contact, and in signaltransduction. Two of the requirements for the


characterization of a glycoprotein as a CAM arelocation on the cell surface in an exposed fashion,and developmental regulation such that its exist-ence on the cell surface corresponds to the periodproposed to involve cellular adhesiveness. Flowcytometric studies have previously indicated thatPsA is exposed on the cell surface, and that it isselectively expressed on prespore cells at the maturemulticellular slug stage of development (Browneet al., 1989; Krefft et al., 1983).

For characterization of a glycoprotein as a CAMit must be demonstrated to have binding proper-ties. In vitro ligand binding assays have indicatedthat native PsA has molecular recognition andbinding properties. Assays involving the probingwith PsA, of slug cell membrane protein extracts(solubilized with the non-ionic detergent TritonX-100) dotted onto nitrocellulose, have indicatedthat PsA binds to native membrane protein(s);there is no evidence that PsA can bind to other PsAmolecules. Thus, PsA has heterophilic bindingproperties. While it is known that PsA is restrictedto the surface of prespore cells, it is not knownwhether its binding partner(s) are located on thesurface of prespore cells, or prestalk family cells, orboth.

The in vitro cell reassociation assays reportedhere indicate biological relevance for PsA’s bind-ing properties. The findings implicate PsA’s rolein EDTA-resistant cell–cell adhesion. Since simi-lar changes in mean percentages cell recruitmentwere observed in both the presence and absenceof EDTA (for +MUD1 vs �MUD1), the contactsites involved in both cases must be EDTA-resistant. Thus, the assays performed in the pres-ence of 10 mM EDTA (only EDTA-stablecontacts can form under these conditions) indi-cate that PsA is involved in an EDTA-resistantinteraction.

Two major criteria for bona fide ligand-receptorinteractions are specificity, and reversibility. ThemAb MUD1, monospecific for a protein epitopeon PsA, is a specific and significant inhibitor ofEDTA-resistant slug cell–cell reassociation. MUD1shows a dosage effect in its inhibition of EDTA-resistant slug cell–cell reassociation. It is likely thatMUD1 binds to its protein epitope on cell-surfacePsA, thus blocking cell–cell adhesion. If so, solublerPsA (which does not bind to slug membraneproteins) should be able to neutralize the inhibitoryeffect of MUD. rPsA is in fact capable of neutral-izing the adhesion blocking activity of MUD1.Thus, the inhibitory effect of MUD1 on cell ad-hesion is primarily due to the blocking of PsAmolecules on cell surfaces by mAb molecules. This

represents the first evidence to implicate a directrole for a putative CAM involved in cell–celladhesion during the multicellular migratory slugstage of D. discoideum development.

The mechanism by which PsA mediates cell–cell adhesion is not yet known. The mAb MUD1has a protein epitope on the extracellular, non-glycosylated protein moiety of the PsA molecule,and specifically inhibits PsA-mediated cell–celladhesion; it is postulated that protein-protein in-teractions are involved in PsA-mediated cell–cell adhesion. Parish and his colleagues (Kelleret al., 1994) obtained a mAb (TK1) which wasfound to interfere with EDTA-stable adhesion ofslug cells, and which could involve a carbo-hydrate epitope. However, TK1 cross-reactswith many antigens, making interpretation ofthe results of their work difficult. MUD1, onthe other hand, is monospecific for a uniqueprotein epitope on PsA molecules. Nevertheless,our experiments do not rule out the possibilityof carbohydrate involvement in slug cell–celladhesion.

The results of the ligand binding assays suggestthat PsA-mediated cell–cell interactions areheterophilic. Further, since EDTA is a chelatingagent which sequesters divalent cations, the factthat MUD1 specifically inhibits EDTA-resistantadhesion may indicate that the binding mechanismby which PsA mediates cell–cell adhesion isindependent of cations such as Ca2+ and Mg2+.

ACKNOWLEDGEMENTS

We thank Dr Martin Slade for advice with cultur-ing techniques, and Dr Zhou-Chou for her kinddonation of purified WS576 and HU2421 PsA,and crude HU2421 membranes for the ligandbinding assays. We also thank Dr Cathy Mossfor her kind donation of purified recombinantPsA (rPsA). We acknowledge, with thanks, DrRichard Morrow, and Kim Shaddick for technicalassistance. Professor Duncan Veal is acknowl-edged for use of equipment in his laboratory. Wethank Jenny Norman and Ron Oldfield for assist-ance with the DIC photography, and Dr RichardMorrow for assistance with production of thefigures, and editing. We also acknowledge,with thanks, Dr Andrew Gooley, for helpfuldiscussions.

This research was funded, in part, by anAustralian Research Council Large Grant, 1998–2000, K. L. Williams, M. B. Slade, N. E. Packer.


REFERENCES

A S, S E, D L, G A, P SB, BLH, W KL, 1988. Characterization of an anti-genically related family of cell-type specific proteinsimplicated in slug migration in Dictyostelium discoideum.Differentiation 38: 82–90.

B H, K F, G G, 1973. Dynamics of antigenicmembrane sites relating to cell aggregation in Dictyosteliumdiscoideum. J Cell Biol 56: 647–658.

B JT, 1967. The Cellular Slime Moulds (2nd edn).Princeton, NJ, Princeton University Press.

B JT, 1998. A way of following individual cells in themigrating slugs of Dictyostelium discoideum. PNAS 95:9355–9359.

B-M VM, W KL, 2002. Comparativemolecular analyses of cell contact phenomena in metazoandevelopment: an exploration of universality. Submitted.

B R, T JP, R U, E GM,1977. Adhesion among neural cells of the chick embryo. I.An immunological assay for molecules involved in cell-cellbinding. J Biol Chem 252: 6835–6840.

B SK, S C-H, 1993. Characterization of the cell adhesionmolecule gp24 in Dictyostelium discoideum. J Biol Chem 268:24902–24909.

B LH, S H, B D, W KL, KC, 1989. Re-expression of 117 antigen, a cell surface glyco-protein of aggregating cells, during terminal differentiationof Dictyostelium discoideum prespore cells. Development 105:657–664.

B LH, W KL, 1993. Pure populations ofDictyostelium discoideum prespore and prestalk cellsobtained by flow cytometry have different redevelopmentcharacteristics at their cell surfaces. Cytometry 14: 660–667.

C JC, H AJ, 2001. Cell-cell adhesion and signaltransduction during Dictyostelium development. J Cell Sci114: 4349–4358.

E AE, W JG, M HE, P SB, S E,W KL, G AA, 1988. Structural characteris-ation of Dictyostelium discoideum prespore specific gene D19and of its product, cell surface glycoprotein PsA. Molecularand Cellular Biology 8: 3458–3466.

E GW, R KB, V LK, 1973. Mating types andmacrocyst formation in Dictyostelium discoideum. Proc NatAcad Sci USA 70: 1828–1830.

F DR, 1995. Dictyostelium discoideum cohesion andadhesion. In: Richardson PD, Steiner M, eds. Principles ofCell Adhesion. CRC Press, Inc. 63–86.

F M, 1993. PhD thesis, Macquarie University, Sydney,Australia.

F M, J MK, W KL, 1993. Characterisationof an epithelium-like layer of cells in the multicellularDictyostelium discoideum slug. J Cell Sci 105: 243–253.

G EN, S P, S C-H, 1992. Purification andpartial characterization of a cell adhesion molecule (gp150)involved in postaggregation stage cell-cell binding inDictyostelium discoideum. J Biol Chem 267: 9409–9415.

G DR, 1972. Acquisition of cohesiveness by slimemould cells prior to morphogenesis. Exp Cell Res 72:305.

G JE, W J, B A, L RA, 1979.Identification of a cell surface glycoprotein involved in cellaggregation in Dictyostelium discoideum. Cell 18: 391–398.

G G, 1968. Cell aggregation and differentiation inDictyostelium. Curr Top Dev Biol 3: 157–197.

G AA, M R, W KL, 1992. Size poly-morphisms due to changes in the number of O-glycosylatedtandem repeats in the Dictyostelium discoideum glycoproteinPsA. Genetics 130: 749–756.

G WN, W KL, 1983. Monoclonal antibodycharacterization of sheath matrix: the extracellular matrix ofDictyostelium discoideum. EMBO J 2: 935–940.

G WN, W DL, W KL, 1985. A polymor-phic, prespore-specific surface glycoprotein present in theextracellular matrix of Dictyostelium discoideum. Mol CellBiol 5: 2559–2566.

G JM, K M, H-K A, W KL, 1982.Antigenic differences between prespore cells of Dictyosteliumdiscoideum and D. mucoroides using monoclonal antibodies.Exp Cell Res 142: 229–233.

H M, 1999. The Armadillo family of structuralproteins. International Review of Cytology 186: 179–224.

H PA, G AA, F MAJ, R JW,W KL, 1993. Post-translational modificationsof the Dictyostelium discoideum glycoprotein PsA. Glyco-sylphosphatidylinositol membrane anchor and compositionof O-linked oligosaccharides. European Journal ofBiochemistry 216: 729–737.

H RO, 1992. Integrins: versatility, modulation andsignalling in cell adhesion. Cell 69: 11–25.

K RK, G J, S C-H, 1989. Identification of anoctapeptide involved in homophilic interaction of the celladhesion molecule gp80 of Dictyostelium discoideum. Cell59: 615–625.

K T, E E, B K, C C, P RW,1994. A monoclonal antibody that interferes with the post-aggregation adhesion of Dictyostelium discoideum cells.FEBS Letters 339: 119–123.

K DA, F DL, L WF, 1987. Surface glyco-protein, gp24, involved in early adhesion of Dictyosteliumdiscoideum. Dev Biol 121: 277–283.

K M, V L, M H, W KL, 1983.Analysis of proportion regulation in slugs of Dictyosteliumdiscoideum using a monoclonal antibody and a FACS-IV.Exp Cell Res 147: 235–239.

L UK, 1970. Cleavage of structural proteins during theassembly of the head of bacteriophage T4. Nature 227:680–685.

L TY, P G, G J, S C-H, 1981. Differentialcell cohesiveness expressed by prespore and prestalk cells ofDictyostelium discoideum. Differentiation 20: 22–28.

L WF, 1975. Dictyostelium discoideum: A Develop-mental System. New York, Academic Press.

M K, G G, 1978. A specific glycoprotein asthe target site of adhesion blocking Fab in aggregatingDictyostelium cells. Nature 274: 445–449.

M BA, W S, J T, L WF, 1984.Mutations affecting a surface glycoprotein, gp80, ofDictyostelium discoideum. Mol Cell Biol 4: 514–519.

R KB, 1984. The Dictyostelieds. Princeton, NJ, PrincetonUniversity Press.

S C-H, L TY, W LM, 1988. Expression of the con-tact site A glycoprotein in Dictyostelium discoideum:quantitation and regulation. Biochim Biophys Acta 968:283–290.

S C-H, L TY, C AHC, 1985. Inhibition ofcell-cell binding at the aggregation stage of Dicty-ostelium by monoclonal antibodies directed against the


contact site A glycoprotein. J Biol Chem 260: 16030–16036.

S MB, E KR, W KL, 1997. Expressionof recombinant glycoproteins in the simple eukaryoteDictyostelium discoideum. Biotechnology and EngineeringReviews 14: 1–35.

T M, 1991. Cadherin cell adhesion receptors as amorphogenetic regulator. Science 251: 1451–1455.

V L, K M, M H, W KL, 1984.An assay for pattern formation in Dictyostelium dis-coideum using monoclonal antibodies, flow cytometry, andsubsequent data analysis. Cytometry 5: 26–33.

W J, H L, A D, L WF, F RA, S C-H,2000. The membrane glycoprotein gp150 is encoded by thelagC gene and mediates cell-cell adhesion by heterophilicbinding during Dictyostelium development. Dev Biol 227:734–745.

W MR, W KL, 1995. The extracellular matrix ofthe Dictyostelium discoideum slug. Experientia 51: 1189–1196.

W J, 1997. Prestalk and stalk cell heterogeneity inDictyostelium. In: Dictyostelium—a Model System for Celland Developmental Biology. Universal Academy Press, Inc.and Yamada Science Foundation.

W E, Y C, W J, F D, L WF, S C-H,2002. Disruption of the gene encoding the cell adhesionmolecule DdCAD-1 leads to aberrant cell sorting and

cell-type proportioning during Dictyostelium development.Development, in press.

W EFS, B S, S H, Y C, S C-H,1996. Molecular cloning and characterization of DdCAD-1,a Ca2+-dependent cell-cell adhesion molecule, inDictyostelium discoideum. J Biol Chem 271: 16399–16408.

Z NE, P NH, T MD, S MB, JDR, K P, M CJ, M BC, C PMG,W KL, G AA, 1996. Recombinant prespore-specific antigen from Dictyostelium discoideum is a beta-sheet glycoprotein with a spacer peptide modified byO-linked N-acetylglucosamine. European Journal ofBiochemistry 238: 511–518.

Z NE, 1998. PhD thesis, Macquarie University,Sydney, Australia.

Z-C T, G AA, S MB, B VM,W KL, 1990. Purification of a membrane glycopro-tein containing an inositol-containing phospholipid anchorfrom Dictyostelium discoideum. Journal of Biotechnology 16:233–244.

Z-C T, 1992. PhD thesis, Macquarie University,Sydney, Australia.

Z-C T, S MB, W KL, G AA, 1995.Expression, purification and characterisation of secretedrecombinant glycoprotein PsA in Dictyostelium discoideum.Journal of Biotechnology 38: 137–149.

Documents

CELL ADHESION DURING THE MIGRATORY SLUG STAGE OF DICTYOSTELIUM DISCOIDEUM