Characterization of the cell adhesion molecule gp24 in Dictyostelium discoideum. Mediation of cell-cell adhesion via a Ca(2+)-dependent mechanism.

Dictyostelium discoideum cells express EDTA-sensitive cell-cell adhesion sites soon after the initiation of development and EDTA-resistant adhesion sites later at the aggregation stage. A glycoprotein of M(r) 24,000 (gp24) has been implicated in the mediation of the EDTA-sensitive type of intercellular cohesiveness (Knecht, D. A., Fuller, D. L., and Loomis, W. F. (1987) Dev. Biol. 121, 277-283). In this study, a relatively simple procedure was developed to purify gp24 to homogeneity. A highly specific rabbit antiserum was raised against gp24, and the localization of gp24 at the cell surface was shown by quantitative binding of the anti-gp24 antibodies to intact cells. To demonstrate the cell binding activity of gp24, the binding of solubilized gp24 to intact cells was examined. 125I-Labeled gp24 bound to cells in a dose-dependent and saturable manner, and the binding was displaced specifically by unlabeled gp24. Purified gp24 was capable of inhibiting the reassociation of dispersed cells previously undergoing EDTA-sensitive aggregation. Moreover, precoating cells with anti-gp24 IgG and Fab fragments blocked the binding of 125I-labeled gp24 to cells. Collectively, these in vitro assays provide direct evidence that gp24 is a cell adhesion molecule that most likely functions through a homophilic mode of interaction. The binding of gp24 to cells was sensitive to EGTA, suggesting that the activity of gp24 may involve calcium ions. Binding studies showed that 45Ca2+ could bind to gp24 blotted onto nitrocellulose membrane. In addition, preincubation of the native protein with calcium ions resulted in a shift in its gel mobility. It is therefore likely that gp24 mediates cell-cell interactions via a Ca(2+)-dependent mechanism, rendering gp24 the first cell adhesion molecule in D. discoideum to utilize a Ca(2+)-based adhesion system.

283). In this study, a relatively simple procedure was developed to purify gp24 to homogeneity. A highly specific rabbit antiserum was raised against gp24, and the localization of gp24 at the cell surface was shown by quantitative binding of the anti-gp24 antibodies to intact cells. To demonstrate the cell binding activity of gp24, the binding of solubilized gp24 to intact cells was examined. 12sI-Labeled gp24 bound to cells in a dosedependent and saturable manner, and the binding was displaced specifically by unlabeled gp24. Purified gp24 was capable of inhibiting the reassociation of dispersed cells previously undergoing EDTA-sensitive aggregation. Moreover, precoating cells with anti-gp24 IgG and Fab fragments blocked the binding of 1261-labeled gp24 to cells. Collectively, these in vitro assays provide direct evidence that gp24 is a cell adhesion molecule that most likely functions through a homophilic mode of interaction. The binding of gp24 to cells was sensitive to EGTA, suggesting that the activity of gp24 may involve calcium ions. Binding studies showed that 4sCa2+ could bind to gp24 blotted onto nitrocellulose membrane. In addition, preincubation of the native protein with calcium ions resulted in a shift in its gel mobility. It is therefore likely that gp24 mediates cell-cell interactions via a Ca2+-dependent mechanism, rendering gp24 the first cell adhesion molecule in D. diecoideum to utilize a Ca2+-based adhesion system. Cell-cell adhesion plays an important role in the regulation of cell proliferation, motility, differentiation, and morphogenesis. The cellular slime mold Dictyostelium discoideum provides a n excellent model for the study of cell-cell interactions. Cells are triggered to enter their developmental cycle by starvation. Cellular differentiation takes place, and cells begin to undergo chemotactic migration, leading to the formation of multicellular aggregates called slugs or pseudoplasmodia. In the Medical Research Council of Canada. The costs of publication of this * This work was supported in part by Operating Grant MT-6140 from article were defrayed in part by the payment of page charges. This with 18 U.S.C. Section 1734 solely to indicate this fact. article must therefore be hereby marked "aduertisement" in accordance $ Supported in part by an Ontario graduate studentship.
Inst., University of Toronto, 112  these stages, cells are held together by cell-cell adhesion molecules expressed on the cell surface. The slugs eventually culminate in the formation of fruiting bodies consisting of spores and stalk cells (1). There are at least two different types of cell-cell adhesion sites expressed by Dictyostelium cells during development, and they have been operationally defined by their ability to resist dissociation by EDTA (for reviews, see Refs. 2 and 3). The EDTA-resistant cell-cell binding sites are stable in the presence of 10-15 m~ EDTA, whereas the EDTA-sensitive sites are easily disrupted in the presence of 1-2 m~ EDTA. A cell-surface glycoprotein of M, 80,000 (gp80) has been identified as an adhesion molecule mediating EDTA-resistant cell-cell binding (4-6). gp80 is expressed specifically at the aggregation stage of development (6-121, and it mediates cell-cell binding by homophilic interaction (13-15). In addition to maintaining the integrity of cell aggregates (161, gp80 appears to have a role in regulating the size of slugs (17). The cell binding function of gp80 is replaced by other cell adhesion molecules in the postaggregation stages (18-221, with gp150 being the best characterized of this group (23-25). The EDTA-sensitive cell-cell binding sites appear soon after the initiation of development and persist throughout the developmental cycle (26-28). It has been suggested that the EDTA-resistant binding sites are involved in end-to-end contacts during chemotactic migration (291, whereas the EDTA-sensitive binding sites mediate side-to-side interactions among cells (2).
A glycoprotein of M, 24,000 (gp24) has been implicated in EDTA-sensitive cell-cell adhesion (27). gp24 is expressed maximally 3 4 h after the initiation of development. It is a major component solubilized by butanol extraction of cell particulates. Polyclonal antibodies raised against gel-purified gp24 have been reported to inhibit the EDTA-sensitive cell-cell binding sites (27). Morphogenesis is arrested when cells are developed in the presence of these antibodies (28). Blocking of the EDTA-sensitive binding sites results in the formation of loose aggregates at later stages of development and the loss of the compacted morphology (30).
The sensitivity to EDTA or EGTA exhibited by the early cell-cell adhesion sites suggests a requirement for divalent cations for gp24 function. In vertebrates, two mechanisms of intercellular adhesion have been identified on the basis of their requirement for Ca2+: the Ca2+-dependent type, as exemplified by members of the cadherin family (for a review, see Ref. 311, and the Ca2+-independent type, characteristic of members of the immunoglobulin superfamily (for reviews, see Refs. 32 and 33). It is therefore of interest to note that the EDTA-resistant contact sites mediated by gp80 and gp150 of D. discoideum operate by a Ca2+-independent adhesion mechanism (31, whereas the EDTA-sensitive ones, believed to be mediated by gp24, most likely belong to the Ca2+-dependent category.
To ascertain the direct involvement of gp24 in cell adhesion and to investigate its mechanism of cell binding, it was crucial to first demonstrate that gp24 purified to homogeneity had cell binding activity. A procedure involving the use of HPLC' was developed to purify gp24 to homogeneity. Both cell binding and Ca2+ binding activities of gp24 have been characterized using several different in vitro assays. Results of this study support the idea that gp24 is a cell adhesion molecule mediating intercellular adhesion through a Ca2+-dependent mechanism.
EXPERIMENTAL PROCEDURES Materials-Bio-Gel P-10 and alkaline phosphatase-conjugated goat anti-rabbit IgG antibodies were obtained from Bio-Rad. The DEAE 5PW column (manufactured by Toyo Soda Kaisha, Tokyo, Japan) was purchased through Mandel Scientific (Toronto, Ontario, Canada). Carrierfree 1261 was obtained from Amersham Canada Ltd. Bovine calmodulin was purchased from Sigma. Protein A-Sepharose, the bicinchoninic acid protein determination kit, and papain were purchased from Pierce Chemical Co. Goat anti-rabbit Fab fragments were obtained from Organon Teknika Inc. (Scarborough, Ontario, Canada).
Cell Strains and Culture Conditions-The axenic strain KAX3 of D. discoideum was used for the purification of gp24 and in most experiments. KAX3 cells were cultured in HL-5 medium (34); collected at a cell density of 5 x 10" celldml; washed twice; resuspended at 1.5 x lo7 celldml in 17 m~ sodium phosphate buffer, pH 6.4; and shaken at 180 rpm for development. In certain experiments, KAX3 cells were cultured in association with Klebsiella aerogenes on agar plates as described by Sussman (35). In this case, cells were collected from a partially cleared bacterial lawn; washed free of the bacteria; and resuspended in 17 m~ phosphate buffer, pH 6.4, for development.
Purification ofgp2PSince the original protocol described by Knecht et al. (27) did not allow the purification of gp24 to homogeneity, a modified protocol including a new HPLC step was developed. gp24 was purified from KAX3 cells after 3 h of development in liquid culture. Frozen cells (-5 x 1O'O cells) were thawed and homogenized in 17 m M sodium phosphate buffer, pH 6.4. The homogenate was centrifuged at 16,000 x g for 10 min at 4 "C. The particulate fraction was resuspended in 60 ml of 10 m M sodium citrate buffer, pH 5.5, and then 45 ml of water-saturated butanol were added dropwise with constant stirring. The suspension was stirred at 4 "C for another 10 min, followed by centrifugation at 6000 x g for 20 min. This allowed the separation of the mixture into the butanol (upper) and aqueous (lower) phases. The aqueous phase was collected; concentrated; and dialyzed extensively against 20 m M sodium phosphate buffer, pH 6.2, containing 10 m M EDTA and 0.1% (w/v) octyl glucoside. The sample was subjected to fractionation on a Bio-Gel P-10 column (1.7 x 60 cm) equilibrated in 20 m M sodium phosphate buffer, pH 6.2, containing 10 m M EDTA. Fractions were subjected to SDS-gel electrophoresis, and those containing gp24 were pooled and dialyzed extensively against 10 m M Tris-HC1, pH 7.6. The protein sample was injected onto a DEAE 5PW column (7.5 x 75 mm) and subjected to separation by HPLC. The column was eluted with a linear salt gradient of 0-0.2 M NaCl at a flow rate of 1 mumin. The eluate was monitored at 280 nm, and the fractions containing gp24 were pooled and collected. gp24 obtained after this step usually achieved near homogeneity. In the case of larger samples, the material was routinely subjected to another cycle of HPLC separation on the same DEAE 5PW column.
Protein Determination-The bicinchoninic acid assay kit was used to determine protein concentrations (36). Crystalline BSA was used as the standard.
For immunostaining, proteins were transferred electrophoretically from slab gels onto a nitrocellulose membrane (39). The protein blot was blocked with 5% skim milk in phosphate-buffered saline and then incubated with an appropriate dilution of an anti-gp24 rabbit antiserum. The bound antibody was detected using an alkaline phosphatase-conjugated goat anti-rabbit IgG second antibody. matography; BSA, bovine serum albumin.
The abbreviations used are: HPLC, high-performance liquid chro-Analysis ofAmino Acid Composition-Protein samples were dialyzed against H,O to remove salt, lyophilized, and then hydrolyzed in 6 M HCl for 24 h and derivatized for amino acid analysis using the Waters PICO-TAG system.
Preparation of Anti-gp24 Antiserum-Polyclonal antibodies directed against gp24 were raised in New Zealand White female rabbits by subcutaneous immunization with gp24 prepared in complete Freund's adjuvant. This was followed by two consecutive boosts using incomplete Freund's adjuvant at 2-week intervals and a final boost 1 month later. A total of 300 pg of gp24 were used for the immunization schedule. Blood obtained from the rabbits was allowed to clot and shrink overnight at 4 "C. The serum was collected by centrifugation at 10,000 x g for 10 min at 4 "C and stored in small aliquots at -70 "C. Total IgG was isolated from the antiserum on a protein A-Sepharose column. The serum was diluted 10-fold with 10 m~ Tris-HCI, pH 7.5, before passing through the column. After extensive washing, the bound IgG was eluted using 100 m~ glycine, pH 2.5. Anti-gp24 Fab fragments were prepared by overnight digestion with papain at a 1:5 (w/w) ratio of anti-gp24 IgG to enzyme. Separation of Fab from the contaminating IgG and Fc fragments was carried out on a protein A-Sepharose column.
Binding of 1251-Labeled Anti-gp24 IgG to Cells-KAX3 cells developed for 3 h were washed and resuspended in 17 m M phosphate buffer, pH 6.4, at 6 x lo6 celldml. Anti-gp24 IgG was purified on a protein A-Sepharose column and labeled with lz5I using chloramine T. The labeled IgG was separated from the free iodine by gel filtration. Various amounts of 1251-labeled anti-gp24 IgG were added to 0.1-ml samples of cell suspension. Cell samples were briefly vortexed and then rotated on a platform shaker at 180 rpm for 45 min at 4 "C. The cells were washed twice in cold phosphate buffer and then counted in a y-counter.
Cell Cohesion Assay-Intercellular cell cohesion was assayed using a modified method (18) of the original roller tube assay of Gerisch (40).
KAX3 cells were grown to a density of -5 x loE celldml in HL-5 medium. Cells were collected; washed; and resuspended in 17 m M phosphate buffer, pH 6.4. After 3 h of development in 17 m M phosphate buffer, pH 6.4, at 1 x lo7 celldml, cells formed EDTA-sensitive aggregates. They were centrifuged and then resuspended at 2.5 x 106 celldm1 in 17 m M phosphate buffer. Cell aggregates were dispersed by vortexing for 15 s. Cells were allowed to re-form aggregates by rotating at 180 rpm on a platform shaker at room temperature. At regular time intervals, the number of nonaggregated cells, including both singlets and doublets, was counted-under an Olympus inverted phase microscope using a hemocytometer. Phase micrographs were obtained using a Nikon Type 104 phase-contrast microscope. To determine the effects of soluble gp24 on cell reassociation, different amounts of purified gp24 were added to 100-pl aliquots. The cell samples were briefly vortexed, and the assay was carried out as described above.
Binding of 1251-Labeled gp24 to Intact Cells-Radioiodination of purified gp24 was performed using chloramine T, and 1251-labeled gp24 was separated from the free iodine on a Sephadex G25 column (0.8 x 20 cm) equilibrated in 75 m M phosphate buffer, pH 7.2, containing 150 m M NaCl, 1% (w/v) BSA, and 0.1% (w/v) octyl glucoside. KAX3 cells developed for 3 h at 2 x lo7 celldml were resuspended at 6 x 10" celldml in 17 m M sodium phosphate buffer, pH 6.4. Different amounts of lz5Ilabeled gp24 were added to 100-pl samples of cell suspension. The final concentration of detergent was kept below 0.005% in all cases, and the concentration of BSA was kept between 200 and 500 pdml. Incubation was carried out at 4 "C for 45 min on a platform shaker rotating at 180 rpm. The cell samples were centrifuged and washed twice with phosphate buffer to remove unbound 12sI-labeled gp24. The cell pellet was counted in a y-counter, and the amount of gp24 bound was determined. For comparison, cell binding studies were also carried out using veg- Competition experiments were carried out by incubating cell samples with 20 ng of '261-labeled gp24 and various amounts of unlabeled gp24. The assay was carried out under the same conditions as described above. The amount of 1251-labeled gp24 bound was expressed as a percentage of the amount bound in the control sample.
To determine the effects of anti-gp24 IgG on lz5I-labeled gp24 binding to cells, 3-h KAX3 cells were first incubated with different concentrations of anti-gp24 rabbit IgG. The incubation was carried out at 4 "C for 15 min. Cells were washed once with phosphate buffer, and the cell pellet was resuspended at 6 x lo6 celldml in the same buffer containing 0.25 mdml goat anti-rabbit IgG Fab at 4 "C for 15 min. Secondary Fab was included to minimize IgG-mediated cross-linking of cells and binding of labeled gp24 to the membrane-bound IgG. The sample (100 pl) was briefly vortexed before the addition of 20 ng of '261-labeled gp24. Subsequent steps were carried out as described above.
To examine the effects of metal chelation on gp24 binding activity, different concentrations of EGTA were added to cell samples containing 20 ng of lZsI-labeled gp24. The cell binding assay was carried out at 4 "C for 45 min on a platform shaker rotating at 180 rpm. The amount of labeled gp24 bound relative to the control was determined.
Ca2+-dependent Shift in Electrophoretic Mobility-Samples containing gp24 were incubated with 50 mM CaClz or 10 mM EDTA in 10 mM Tris-HC1, pH 7.5, at 22 "C for 30 min (41). The protein samples were subjected to electrophoretic separation on a 12.5% SDS-polyacrylamide gel and subsequently electrotransferred onto a nitrocellulose membrane. The protein blot was incubated with the anti-gp24 rabbit antiserum, and color development was achieved using a goat anti-rabbit alkaline phosphatase-conjugated IgG.
Binding of 45Ca2+ to g p 2 P P r o t e i n fractions from various stages of purification were subjected to SDS-polyacrylamide gel electrophoresis, followed by electroblotting onto a nitrocellulose membrane. Binding studies were carried out as described by Maruyama et al. (42). The membrane was washed with an overlay buffer containing 60 mM KCl, 5 mM MgCl2, and 10 mM imidazole, pH 6.8, for 1 h with four changes. The blot was then incubated with 2 pCi/ml 4sCa2+ in the overlay buffer for 10 min a t room temperature on a platform shaker, followed by four 5-min washes with deionized water. Autoradiography was carried out by exposure of the air-dried 4"Ca-labeled membrane to Kodak x-ray film (X-Omat A R ) for 24 h.

RESULTS
Purification ofgp24"Isolation of the M , 24,000 glycoprotein was achieved by a four-step purification procedure (Table I).
Proteins associated with the plasma membrane were collected as a pellet after centrifugation of the whole cell lysate. The pellet was then extracted with butanol. The polar glycoproteins in the aqueous phase of the butanoltaqueous mixture were concentrated, dialyzed against 0.1% octyl glucoside, and then loaded onto a Bio-Gel P-10 column. Most of the high molecular weight components eluted in the void volume while gp24 was included (Fig. 1). Those fractions containing gp24 were pooled and dialyzed. HPLC was employed as the final step of purification. The protein sample was loaded onto a DEAE 5PW column, and the bound material was eluted with a linear salt gradient. gp24 eluted as a single symmetrical peak at 100 m M NaCl with a retention time of 12.7 min (Fig. 2). The purified protein showed immunoreactivity with an anti-gp24 antiserum developed by Knecht et al. (27).
The pooled samples obtained at each step of purification were analyzed by SDS-polyacrylamide gel electrophoresis (Fig.  3A). Contaminating proteins were removed with each step so that the product obtained after HPLC migrated as a single silver-stained band with an apparent M , of 24,000. This purification protocol allowed the isolation of -1 mg of gp24 from 10" KAX3 cells (Table I).
Polyclonal antibody was produced by immunization of rabbits with purified gp24. At 1:1000-fold dilution, the serum still detected 5 ng of gp24 immobilized on a nitrocellulose membrane. The antiserum was highly specific for gp24 as evidenced by the staining of a single band on a Western blot of the whole KAX3 cells were developed in 17 mM phosphate buffer, pH 6.4, for 3 h and then collected for gp24 isolation. Approximately 5 x 10' O cells were homogenized, and the particubte fraction was collected for butanol extraction. The amount of protein recovered after each step was determined using the bicinchoninic acid protein assay.
The relative amounts of gp24 were estimated by densitometric tracings of immunoblots of protein samples obtained after each step. gp24 was estimated to constitute -0.15% of total cell protein.  Pooled samples from the Bio-Gel P-10 column were dialyzed against 10 m M Tris-HC1, pH 7.6. A total of 12 ml containing -1 mg of protein were loaded onto a DEAE 5PW column that was eluted with a salt gradient of 0-0.2 M NaCl in 10 mM Tris-HC1, pH 7.6 (broken line). gp24 eluted as a major peak (shaded area) at 12.7 min. cell lysate derived from the 3-h stage of development (Fig. 3B 1. Cell-cell reassociation was specifically inhibited by anti-gp24 antibodies as assayed by the method of Springer and Barondes (43) (data not shown). These results are consistent with those reported by Knecht et al. (27).
A p-mercaptoethanol-dependent shift in gp24 gel mobility was detected, suggesting the presence of one or more internal disulfide bonds within the gp24 molecule (Fig. 3 0 . In the presence of P-mercaptoethanol, gp24 was visualized by immunostaining at the M , 24,000 position, but it migrated to the M, 22,000 position in the absence of P-mercaptoethanol. The stepped appearance of the gp24 band in lane b of Fig. 3C was due to seepage of P-mercaptoethanol from the neighboring lane. Amino Acid Composition of g p 2 P T h e amino acid compositions of three different preparations of gp24 were determined  presence (lane a ) or absence (lanes b and c ) of P-mercaptoethanol was blotted onto nitrocellulose for immunostaining. The arrow indicates the position of nonreduced gp24.

TABLE I1
Amino acid composition of gp24 gp24 was hydrolyzed in 6 M HCI for 24 h and analyzed by the Waters PICO-TAG system. Cys and Trp were not determined. Values represent results derived from three different preparations of gp24, with one preparation analyzed twice.  (Table 11). Variations among the different analyses were minimal as indicated by the small standard deviations, suggesting homogeneity among the different gp24 preparations. gp24 contained 17% Asx and 11% Glx. Gly, Thr, Ser, and Val were also present in relatively high amounts (%lo%). Polar residues constituted -58% of all the amino acids. Localization of gp24 on Cell Surfwe-One of the criteria for a cell-cell adhesion molecule is its association with the cell surface. To demonstrate the presence of gp24 on the plasma membrane, the binding of 1251-labeled anti-gp24 IgG to intact cells was examined. Cells developed for 3 h were collected, and samples were incubated with varying amounts of labeled IgG. The binding of 1251-labeled anti-gp24 IgG to these cells was dose-dependent and saturable, whereas 1251-labeled preimmune IgG did not show significant binding (Fig. 4). The number of IgG-binding sites was estimated to be -1.4 x 105/cell. Immunofluorescence labeling also demonstrated the presence of gp24 on the cell surface (data not shown). the detection limit. Cell samples were rotated on a platform shaker with different amounts of soluble gp24, and the number of single cells was monitored for each sample at 20-min intervals for 60 min. Control samples, without the addition of gp24, showed >80% cell aggregation, whereas a dose-dependent inhibitory effect was observed in cell samples reassociated in the presence of gp24 (Fig. 5). At 20 pg/ml gp24, >60% of the cells failed to re-form aggregates, and microscopic observation showed that aggregates formed by the remaining cells consisted of 3-5 cells compared with >20 celldaggregate in control samples. The inhibitory effect of gp24 became negligible at 0.6 pg/ml. As a negative control, cells were assayed in the presence of different concentrations of BSA. BSA did not exert any inhibitory effect on cell reassociation (Fig. 5B).
Binding of gp24 to Cells-To provide direct evidence that gp24 is a cell adhesion molecule, its cell binding properties were examined in a gp24-to-cell binding assay. Cells were incubated with different amounts of 1251-labeled gp24 at 4 "C for 45 min on a platform shaker. The low temperature was maintained to minimize the nonspecific uptake of gp24 by cells. The binding curves showed that gp24 bound to cells in a dosedependent manner and was capable of saturating the binding sites on the cell surface (Fig. 6). Binding studies were also carried out with vegetative (0-h) cells. The amount of gp24 bound on 0-h KAX3 cells was about one-third of the amount bound on 3-h cells. Since immunoblot analysis showed only minute amounts of gp24 in vegetative cells (data not shown), most of the radioactivity associated with 0-h cells might be due to background binding.
Competition experiments were carried out to demonstrate the specificity of gp24 binding to cells. Approximately 55% of the amount of 1251-labeled gp24 bound to 3-h cells was competed off by unlabeled gp24. It is possible that a substantial amount of the remaining radioactivity represented nonspecific binding.
Inhibition of gp24 Binding to Cells by Anti-gp24 Antibody -To investigate whether soluble gp24 bound homophilically to gp24 associated with the cell surface or heterophilically to other membrane receptors, anti-gp24 IgG was used to block cell-surface gp24 molecules from binding to 1251-labeled gp24.
Cells were precoated with anti-gp24 IgG, followed by the addition of anti-rabbit Fab fragments. Alternatively, cells were Effects of EGTA on gp24 Binding to Cells-The early cell-cell binding sites of D. discoideum are known to be sensitive to EDTA and EGTA, suggesting that the cell binding activity of gp24 is dependent on calcium ions. To test this hypothesis, different concentrations of EGTA were added to cells, followed by incubation with 20 ng of 1251-labeled gp24. The binding of gp24 to 3-h cells was inhibited by -70% in the presence of EGTA (Fig. 8). As a negative control, the binding of 1251-labeled anti-gp24 IgG to cells was assayed in the presence of EGTA.
EGTA did not have significant effects on the binding of anti-gp24 IgG to cells. The inhibitory effects of EGTA on the binding of gp24 to cells correlate well with the ability of chelating agents to cause the dispersion of 3-h cell aggregates and are consistent with the mediation of EDTA-sensitive intercellular adhesion by gp24.
Identification of gp24 as Calcium-binding Protein by 45Ca2+ Autoradiography-The binding of Ca2+ to a protein is known to have subtle effects on its conformation and function. To determine whether the sensitivity of cell binding to EGTA exhibited by gp24 is the result of the chelation of Ca2+ required for gp24 function, the ability of gp24 to interact with Ca2+ was explored. The 45Ca2+ overlay technique (42) was used to visualize calcium-binding proteins present at the initial stages of Dictyostelium development. Samples were separated on gels, and proteins were blotted onto a nitrocellulose membrane. Proteins that contained binding sites for Ca2+ were detected by equilibrating the blot with 45Ca2+, followed by autoradiography. The gp24 band was found to be labeled on the autoradiograph (Fig. 9A) under conditions that allowed the detection of only bona fide calcium-binding proteins (42,44). The specificity of gp24 for Ca2+ was evidenced by the inability of the Mg2+ cation to displace the Ca2+ cation, despite the presence of a 1000-fold higher concentration of Mg2+ in the overlay buffer.
The binding of 45Ca2+ to gp24 was, however, effectively displaced by the inclusion of 16 m~ (1000-fold excess) unlabeled Ca2+ in the assay.
The majority of proteins at 3 h of development did not bind 45Ca2+. Only a few protein bands in the whole cell lysate were labeled. The most prominently labeled band had a molecular mass of -17 kDa (Fig. 9) and was identified as calmodulin using an anti-calmodulin monoclonal antibody (45). A separate blot of calmodulin was also incubated with 45Ca2+ under similar conditions (Fig. 9B) since it contained four Ca2+-binding sites (46,47) and thereby served as a positive control in this experiment. As a negative control, molecular weight markers were subjected to the same treatment, and the binding of 45Ca2+ to these proteins was negligible (data not shown).
Shifr in Electrophoretic Mobility of gp24 in Presence of Ca2+ -A shift in electrophoretic mobility is a property exhibited by a number of Ca2+-binding proteins, including calmodulin, troponin C, and parvalbumin (41, 44). It was of interest to investigate whether calcium ions are capable of inducing a similar change in the electrophoretic mobility of gp24. A small but reproducible shift in the electrophoretic mobility of gp24 was observed in the presence of CaC1, (Fig. 10). Several different Ca2+ concentrations were tested, and a single shifted species was observed in all cases (data not shown). As noted previously for other proteins containing Ca2+-binding sites (42, 481, the interaction of Ca2+ with gp24 appears to be unaffected by heat treatment in the presence of SDS. As a negative control, the electrophoretic mobilities of molecular weight markers were examined under similar conditions, and the addition of 50 m~ CaC1, did not alter the mobility of these proteins (data not shown).

DISCUSSION
In this report, we have described the purification and characterization of a Dictyostelium protein, gp24, as a calciumbinding cell adhesion molecule. A rapid isolation scheme has been developed, and the use of HPLC provides a relatively simple method for the purification of gp24 to near homogeneity. Chemical characterization revealed that gp24 has a relatively high percentage of polar residues. This is characteristic of soluble proteins, suggesting that gp24 may be largely exposed on the cell surface. The amino terminus of gp24 is apparently blocked since attempts to obtain amino-terminal sequence have been unsuccessful. An interesting feature of gp24 is the sensitivity of its electrophoretic mobility to P-mercaptoethanol. In the absence of the reducing agent, gp24 exhibited a faster gel mobility, migrating with an apparent M, of 22,000. This obser- Loomis and Fuller (49) have reported the isolation of a cDNA vation suggests the presence of intramolecular disulfide fragment by screening an expression library with anti-gp24 bondb). A more tightly folded globular conformation may result antibodies. A pair of tandemly repeated genes have been isofrom the formation of internal disulfide bridges, causing gp24 lated using this cDNA fragment as a probe. Although these two to migrate faster on SDS-polyacrylamide gels.
genes do not appear to code for gp24, transient expression of . 9. Autoradiographs of calcium-binding proteins. A, total cell protein derived from 3-h cells and purified gp24 were separated on a 12.5% SDS-polyacrylamide gel, followed by electrophoretic transfer onto a nitrocellulose membrane. The blot was incubated with 46Ca2+ ( 2 pCi/ml) for 10 min at room temperature. After washing with HzO, the blot was exposed to x-ray film for 24 h. The same blots were subsequently stained with Coomassie Brilliant Blue. antisense RNA directed against these sequences leads to a delay in the expression of gp24 and EDTA-sensitive sites (50). The role of these tandem genes in early cell adhesion is still unknown.
The role of gp24 as a cell-cell adhesion molecule during Dictyostelium development has been based largely on the effects of anti-gp24 antibodies on cell aggregation. Although this has been useful as an initial step in the identification of cell adhesion molecules, this approach does not distinguish between specific binding and nonspecific steric hindrance.  (27,30). When axenic cells are cultured in liquid medium, expression of gp24 is a cell density-dependent event. It is not detected in low density cells, but it is present in high density cells. Similar to several other early developmentally regulated enzymes, expression of gp24 can be induced by the prestarvation factor PSF in low density cultures (51). In cells grown on bacteria, gp24 is expressed soon after the initiation of development. It accumulates rapidly during the first 4-5 h of development and then remains a t more or less the same level until the culmination stage (27), suggesting a key role for gp24 in the maintenance of cell-cell adhesion throughout development.
It is remarkable that the binding of gp24 to cells can elicit dramatic effects on the formation of multicellular aggregates. Cells at 3 h of development have acquired the EDTA-sensitive cell-cell binding sites and are fully competent to undergo cell aggregation in liquid cultures. However, when cell reassociation is camed out in the presence of gp24, cells are only capable of forming small aggregates of 3-10 cells, and more than half of them remain as single cells. The inhibitory effects of gp24 are dose-dependent, and a relatively low concentration of gp24 is required to block cell aggregation. The observation that the binding of gp24 to cells can block cell reassociation suggests that each gp24 molecule contains only one cell-binding site.
Binding of calcium to a target protein oRen induces a conformational change that triggers biological activity. It is therefore of interest to determine whether the cell binding activity of gp24 is dependent on Ca2+ binding. In support of the hypothesis that Ca2+ is required for gp24 function, the binding of 1251labeled gp24 to cells is sensitive to EGTA, suggesting the requirement for Ca2+ in the binding reaction. This finding is also consistent with the nature of the EDTA-sensitive cell-cell binding sites. In addition, the binding of 45Ca2+ to gp24 immobilized on nitrocellulose membrane provides a direct demonstration that gp24 is a Ca2+-binding protein.
Upon binding to gp24, Ca2+ induces a conformational change in the molecule that probably renders it more tightly folded. The evidence for this postulate is a shiR in the electrophoretic mobility of gp24, making it migrate faster on an SDS gel. The residues involved in Ca2+ binding are expected to be closely clustered to elicit a conformational change in the protein (52).
The intensity of the bands on the autoradiograph from the 45Ca overlay experiment reflects the amount of calcium bound to a particular protein. A comparison of the ratios of 45Ca autoradiographic signal to the amount of purified protein loaded on the gel lane between gp24 and calmodulin yielded a value of -15.4. In determining this ratio, it was assumed that both gp24 and calmodulin were equally retained and renatured on the nitrocellulose blots. Assuming further that there are four Ca2+-binding sites in the Dictyostelium calmodulin (46,47), each gp24 molecule probably binds one calcium ion. These results strongly implicate a Ca2+-dependent mechanism in gp24mediated cell-cell adhesion.
It is likely that gp24 mediates intercellular adhesion through homophilic interactions with apposing gp24 molecules present on the surfaces of adjacent cells. This notion is supported by the observation that the binding of lZ5I-labeled gp24 to cells is inhibited by the precoating of these cells with anti-gp24 Fab or IgG to block direct access to membrane-bound gp24. Although little is known about the molecular details of gp24-gp24 interaction, the involvement of electrostatic interactions has been postulated based on the sensitivity of the EDTA-sensitive sites to carnitine (53), which is known to have nonspecific inhibitory effects on the aggregation of several types of mammalian cells in culture (54). In vertebrates, members of the cadherin family are also known to mediate cell-cell adhesion via homophilic binding (55), and a region containing the His-Ala-Val sequence has been found to play a vital role in cadherin-cadherin interactions (56). Upon the successful cloning of the gp24 gene, it will be of interest to determine if a similar site is conserved and utilized in gp24-gp24 binding.
Differential cell cohesiveness has been implicated in the cell sorting phenomenon, which takes place when prestalk cells and prespore cells are randomly mixed in in vitro assays (21, 57-59). Prespore cells appear to have very different adhesive properties from prestalk cells. Results based on immunostaining and their respective sensitivity to antibody inhibition show a difference in the expression of gp150 in these two cell types (18,  21, 23). On the other hand, anterior or prestalk cells in a migrating slug exhibit a much more compact morphology in situ, whereas posterior cells, composed primarily of prespore cells, have a loosely packed morphology (60). This difference in morphology may reflect a differential expression of gp24 since the EDTA-sensitive cell-binding sites have been implicated in eliciting cell compaction (30).
The role of cell adhesion molecules during development may go beyond the mediation of intercellular adhesion. Recent studies have implicated the involvement of cell-cell and cell-substratum adhesion molecules in signal transduction and gene regulation (61-65). A close interplay between cell-cell contact and CAMP signaling has also been observed during Dictyostelium development. The formation of cell-cell contacts affects both CAMP metabolism and CAMP signaling response (66, 67), which in turn regulate the expression of a large number of genes at the aggregation stage of development. Indeed, when EDTA-sensitive cell adhesion is blocked, expression of aggregation stage-specific components, such as the cell adhesion molecule gp80, is reduced to the basal level.2 Therefore, in addition to being a Ca2+-dependent cell adhesion molecule, gp24 may be involved in the signaling process leading to gene regulation.