Involvement of cell-cell adhesion in the expression of the cell cohesion molecule gp80 in Dictyostelium discoideum

The cellular slime mold Dictyostelium discoideum is an excellent model system for studies in the areas of cell-cell signalling and gene regulation. This organism has a simple and well-defined life cycle (for a review, see Loomis, 1975). Cells survive as unicellular amebae in the presence of sufficient nutrients. Upon starvation, the cells undergo a developmental program where cells migrate chemotactically towards regions of higher concentrations of cAMP, resulting in the formation of multicellular aggregates. Within 24 hours, morphogenesis takes place and the cells eventually form a fruiting body consisting of two major types of cells, spore cells and stalk cells. Early in the developmental cycle, the cells acquire EDTAsensitive binding sites that are disrupted by low concentrations (≥1 mM) of EDTA. A surface-associated glycoprotein of M_r 24,000 (gp24) has been implicated in the formation of the EDTA-sensitive binding sites (Knecht et al., 1987; Brar and Siu, 1993). The cells acquire EDTA-resistant cell binding sites a few hours later, at the early aggregation stage of development (for reviews, see Gerisch, 1986; Siu, 1990). These cell adhesion sites are stable even in the presence of 10-15 mM EDTA and they are mediated by the homophilic interaction of an 80 kDa glycoprotein (gp80) (Muller and Gerisch, 1978; Brodie et al., 1983; Siu et al., 1985; Kamboj et al., 1988, 1989, 1990; Da Silva and Klein, 1990; Faix et al., 1990). While there is a good correlation between the level of gp80 and the extent of EDTA-resistant cell cohesion, a threshold concentration of gp80 appears to be required for these binding sites to function (Siu et al., 1988a). A second type of EDTA-resistant cell binding activity appears later in development when gp80 expression has decreased (Siu, 1990; Gao et al., 1992).

cAMP plays several important roles in the development of Dictyostelium. Early in development, some cells begin to secrete pulses of cAMP. The surrounding cells respond by undergoing chemotactic migration towards these cells. Extracellular cAMP binds to a cell-surface receptor and this results in the formation of intracellular cAMP, which is then secreted into the medium eliciting cAMP responses in other cells further away from the aggregation center. This cAMP signalling is mediated through a cell surface cAMP receptor that transiently activates adenylate cyclase via a G-protein-dependent pathway (for reviews, see Firtel, 1991; Van Haastert et al., 1991). In this fashion, cells are able to form mounds of ∼10⁵ cells. cAMP is also an important regulator of gene expression. At the early aggregation stage of development, the expression of a number of genes responds to pulses of low concentrations of cAMP, by either an increase or a decrease in their mRNA levels (Mann et al., 1988; Mann and Firtel, 1987). During the lateor postaggregation stages of development, high constant levels of cAMP are required for the expression of several presporespecific and prestalk-specific genes (Schaap and Van Driel, 1985; Oyama and Blumberg, 1986).

Cell-cell contact, as well as cAMP, plays a key role in cellular differentiation and gene expression in Dictyostelium. Several studies have suggested that cell-cell contact has a major influence on the expression of many developmentally regulated genes. In the early phase of development, the gene encoding discoidin I appears to be regulated by cell-cell contact. Here, cell-cell contact serves as a signal to down-regulate discoidin I gene expression (Berger and Clarke, 1983). Later during development, the expression of UDP-glucose pyrophosphorylase is inhibited when cells at the mound stage are dissociated and prevented from re-forming contacts by EDTA or low-density plating (Newell et al., 1971). Cell-cell contact may also be involved in regulating some cell type-specific genes, which are expressed at the post-aggregation stage. Whereas cAMP is required for the induction of prestalk-specific gene expression, induction of prespore-specific gene expression requires cell-cell contact as well as cAMP (Landfear and Lodish, 1980; Chung et al., 1981; Blumberg et al., 1982; Lodish et al., 1982; Mehdy et al., 1983; Chisholm et al., 1984). Other studies, however, have suggested that multicellularity may not be required for prespore-specific gene expression (Mehdy and Firtel, 1985; Gomer et al., 1986). Cell-cell contact and cAMP also play roles in maintaining the levels of prespore- and prestalk-specific gene expression. When cells, previously developed on filter pads for 15 hours, are disaggregated, and kept in a disaggregated state, the expression of several post-aggregative stage genes is drastically reduced, suggesting the requirement for cell-cell contact. The addition of cAMP to the disaggregated cells rescues, to varying degrees, the expression of most of these genes (Chung et al., 1981; Landfear et al., 1982; Lodish et al., 1982; Mangiarotti et al., 1982, 1983; Haribabu et al., 1986).

Since very little is known about the role of cell-cell adhesion in early gene expression, we have chosen to determine its role in the regulation of expression of the gene encoding gp80, the protein responsible for EDTA-resistant cohesion at the aggregation stage of development. This study allows us to address specifically the role of EDTA-sensitive cell cohesion in gene regulation. Similar to that of several early genes, the regulation of the gp80 gene has been shown to occur via a two-phase regulatory pathway. gp80 expression is first turned on to a low basal level soon after the onset of development, and is then greatly augmented by cAMP pulses at the early aggregation stage (Noegel et al., 1986; Wong and Siu, 1986). Exogenous pulses of cAMP cause a precocious and enhanced level of gp80 expression at the transcription level (Siu et al., 1988a; Mann and Firtel, 1989; Ma and Siu, 1990).

To investigate the role of cell-cell contact in the regulation of gp80 expression, the accumulation of gp80 mRNA and protein was monitored after cell-cell contact was blocked by different methods. The disruption of the earlier EDTAsensitive binding sites resulted in an inhibition of the cAMPinduced, high level expression of gp80. This inhibition could not be rescued by exogenous pulses of cAMP. These findings demonstrate that cell-cell contact is part of the cAMP-mediated signal transduction pathway involved in the induction of high level gp80 expression.

D. discoideum strain NC4 was cultured on agar plates in association with Klebsiella aerogenes (Sussman, 1987). After 40 hours, cells at their exponential growth phase were collected, washed free of bacteria, and then resuspended at 1×10⁷ cells/ml or 5×10⁵ cells/ml, in 17 mM sodium phosphate buffer, pH 6.4. Development was carried out in liquid cultures, which were rotated at 180 rpm on a platform shaker at 23°C. To examine the effects of cAMP on cell cohesion and gp80 expression, exogenous cAMP was added to cell cultures at 7-minute intervals to give a final concentration of 10⁻⁸ M, beginning at 2 hours of development. To block EDTA-sensitive cell cohesion, cells were developed in the presence of EDTA or L(−)carnitine (Boehringer Mannheim). Cell samples were collected at different developmental time points for cell cohesion assay and gel electrophoresis.

Cells were developed in liquid cultures under different experimental conditions. At different development times, cell samples were collected simultaneously for immunoblot analysis and for cell cohesion assays. Total cell aggregation was determined by taking samples of developing cells and diluting them to 3×10⁶ cells/ml in the same development medium. The number of unaggregated cells (singlets and doublets) in these samples was determined using a hemocytometer and the percentage of cells that had been recruited into aggregates was estimated.

To assess the EDTA-resistant cell-cell adhesion, cell samples collected at different times during development were resuspended in the presence of 5 mM EDTA, to inhibit EDTA-sensitive cohesion. Aggregates were dissociated by brief vortexing, and samples (0.2 ml) were placed in plastic tubes and rotated vertically at 180 rpm at room temperature for cell reassociation. The percentage of cell aggregation was monitored at regular intervals for 60 minutes. Values obtained at the 40 minute point were used in comparative analysis. Experiments were generally repeated three or more times.

Cells were collected from the culture medium at different time points, and RNA was extracted using a modified method of Chirgwin et al. (1979) as described by Wong and Siu (1986). The RNA was electrophoresed on a 0.8% agarose gel containing formaldehyde. RNA was then transferred to nitrocellulose and hybridized to the gp80 cDNA probe λDdgp80c-19 (Wong and Siu, 1986), at 42°C in 50% formamide, 5× Denhardt’s (1% Ficoll, 1% polyvinylpyrrolidone, 1% bovine serum albumin), 5× SSC, 0.1% SDS. Washes were performed at 65°C in 0.1× SSC and 0.1% SDS. The filter was air dried and subjected to autoradiography using Kodak X-ray films.

Whole cell lysates were obtained by freezing the cells at −70°C, followed by thawing in the presence of 7 M urea, 2% SDS, 4% βmercaptoethanol. Total cellular proteins were electrophoresed on 8 or 10% SDS/polyacrylamide gels according to the method of Laemmli (1970). Proteins were transferred electrophoretically to nitrocellulose filters (Towbin et al., 1979). Filters were blocked for 1 hour at room temperature using 5% skim milk, and then incubated for 3 hours to overnight at 4°C with the monoclonal antibody to gp80, 80L5C4 (Siu et al., 1985). 80L5C4 recognizes a linear peptide squence that overlaps partially with the homophilic binding site of gp80 (Kamboj et al., 1989; Wu et al., 1992). After several washes in 5% skim milk, filters were incubated with alkaline phosphatase-conjugated goat antimouse IgG antibody for 1 hour. After the removal of the second antibody and several washes in 5% skim milk, color development was carried out with nitro blue tetrazolium and 5-bromo-4-chloro-3indolylphosphate as substrates.

Since cell-cell contact has been shown to be involved in regulating some genes in D. discoideum, we were interested in determining if cell-cell contact plays a role in the expression of the early gene encoding the EDTA-resistant cell adhesion molecule, gp80. Prior to gp80 expression, cell-cell adhesion is mediated by EDTA-sensitive binding sites (Garrod, 1972; Knecht et al., 1987; Siu et al., 1988b). To assess the role of cell-cell contact in regulating gp80 expression, different concentrations of EDTA were added to cell cultures to block EDTA-sensitive cohesion throughout development. Total cell aggregation, contributed by both EDTA-sensitive and EDTAresistant cell-cell binding sites, was determined at regular intervals. Cell samples were also collected and resuspended in 5 mM EDTA to monitor the levels of EDTA-resistant cell cohesion. Under normal development conditions, EDTAsensitive cell cohesion became detectable right after the initiation of development, reaching its maximum at 3 hours (Fig. 1A). On the other hand, EDTA-resistant cell cohesion was first detected at about 4-5 hours, and was maximal by 10 hours (Fig. 1B). The addition of 0.1 mM EDTA to the culture medium had a small effect on EDTA-sensitive cell cohesion, delaying its full activity by 2 hours, but it did not exert any deleterious effect on EDTA-resistant cell cohesion (Fig. 1). In contrast, development in the presence of 1 mM EDTA led to the complete loss of aggregation throughout development (Fig. 1A). Therefore, 1 mM EDTA indeed inhibited EDTA-sensitive cohesion and also resulted in the abolishment of EDTAresistant cell cohesion (Fig. 1B).

The expression of gp80, which mediates EDTA-resistant cell cohesion, was examined at both the mRNA and protein levels. Cells at different stages of development were collected for RNA and protein analyses (Fig. 2). The cDNA probe hybridized with the gp80 mRNA of ∼2.1 kb, while the monoclonal antibody recognized a major protein band at ∼80 kDa and a minor band at ∼70 kDa, which probably represented a partially glycosylated form of gp80 (Gerisch, 1986; Siu et al., 1985; Siu and Lam, 1988). In the presence of 0 or 0.1 mM EDTA, the developmental patterns obtained for accumulation of both gp80 mRNA and protein were similar (Fig. 2), and they corresponded closely to the levels of EDTA-resistant cell cohesion (Fig. 1B). However, when development was carried out in the presence of 1 mM EDTA, the expression of gp80 was abolished; even the basal level was undetectable.

The expression of gp80 is known to be greatly enhanced by cAMP secreted by cells at the aggregation stage of development. In addition, administration of exogenous cAMP pulses to developing cells leads to a much higher and precocious expression of gp80 and EDTA-resistant cell cohesion (Siu et al., 1988a; Mann and Firtel, 1989; Ma and Siu, 1990). It was therefore of interest to determine if exogenous cAMP pulses could reverse the inhibitory effects of EDTA. When cAMP pulses were applied to cells developing in the presence of 1 mM EDTA, EDTA-resistant cohesion remained negligible (Fig. 3A). Prolonged administration of cAMP pulses was unable to elevate the level of EDTA-resistant cell aggregation. Protein blots showed that gp80 expression was not detectable throughout development, whether cells were developed in the presence or absence of exogenous cAMP (Fig. 3B). In contrast, expression of gp80 and EDTA-resistant cell cohesion in control cells was greatly enhanced by cAMP pulses. It is therefore evident that cAMP is unable to restore the expression of gp80 and EDTA-resistant cell cohesion in cells developing in the presence of 1 mM EDTA. In the presence of 1 mM EDTA, the expression of gp24, which mediates the EDTAsensitive cohesion, was retained, though at a substantially reduced level (Fig. 3C).

Carnitine has been found to be a relatively potent inhibitor of both EDTA-sensitive and EDTA-resistant cell cohesion and it can prevent the aggregation of several types of vertebrate cells in suspension (Fritz and Burdzy, 1989). We have previously shown that carnitine is capable of reversibly inhibiting cell-cell cohesion in Dictyostelium without deleterious effects on cell morphology and subsequent development (Siu et al., 1992). We therefore used carnitine to disrupt the EDTA-sensitive binding sites from the beginning of development, and then examined its effect on gp80 expression and EDTA-resistant cell cohesion. Cells developed in the presence of 20 mM carnitine showed very little aggregation (<20%) throughout development (Fig. 4A). Therefore, EDTA-sensitive cell cohesion was inhibited even though the expression of gp24 was similar to that of the control (Fig. 4D). It was of interest to note that cAMP stimulated an approximately 2-fold increase in the level of gp24 under these conditions. In the presence of carnitine, cells only expressed a basal level of EDTA-resistant cell cohesion during development, while control cells reached ∼50% EDTA-resistant aggregation by 9 hours (Fig. 4B). Immunoblots showed that the expression of gp80 was also reduced to a basal level in cells developed in carnitine (Fig. 4C). The data thus suggest that prior cell-cell contact via EDTA-sensitive binding sites is required for the full induction of gp80 expression during the aggregation stage of Dictyostelium development.

Further studies were carried out to determine if exogenous cAMP pulses could stimulate the expression of gp80. Control cells responded by expressing gp80 and EDTA-resistant cohesion in an enhanced and precocious manner. In the presence of carnitine, however, cAMP pulses were unable to rescue the expression of EDTA-resistant cohesion or gp80. EDTA-resistant cohesion and gp80 expression remained at a low basal level (Fig. 4B,C).

Since EDTA and carnitine may have effects other than the inhibition of EDTA-sensitive cohesion, the inability of cells to express gp80 might not be a direct consequence of the loss of cell-cell contact. As an alternative approach, high shearing forces were used to disrupt cell-cell contacts in liquid cultures. To achieve complete dissociation of cell aggregates, cells were developed at low density (5×10⁵ cells/ml), in 17 mM phosphate buffer containing 0.1 mM EDTA, and fast shaking at 260 rpm. Under these conditions, gp24 was expressed but total cell aggregation remained at the 20% basal level throughout development, indicating that most of the EDTA-sensitive binding sites were effectively disrupted (data not shown). When EDTA-resistant cell cohesion was assayed, only 20% cell aggregation was achieved during development. Protein blots showed that gp80 was barely detectable in these cells (Fig. 5). Therefore, disruption of the EDTA-sensitive cell-cell binding sites resulted in a reduced basal level of gp80 expression.

Exogenous cAMP pulses were administered to examine if cAMP could reverse the inhibitory effects of cell disaggregation. The addition of cAMP pulses to cells developed under the low-density/fast-shaking regime stimulated only a slight increase in the expression of gp80 and the level of EDTAresistant cohesion remained relatively low (Fig. 5). Thus, cellcell contact appears to be directly involved in the cAMPmediated pathway required for the induction of high level gp80 expression.

To assess further the relationship between cell-cell contact and cAMP signal response in the regulation of gp80 expression, cells were developed at the normal cell density of 1×10⁷/ml, but rotated on a platform shaker at 260 rpm. Under these conditions, the endogenous cAMP signal relay system was perturbed (Chisholm et al., 1984). However, the EDTAsensitive cell cohesion was not significantly affected (Fig. 6A). When EDTA-resistant cell cohesion was measured, these cultures achieved ∼30% cell aggregation (Fig. 6A). Protein blots revealed that expression of gp80 remained low and was close to the basal level (Fig. 6B). Nevertheless, expression of gp80 and EDTA-resistant cell cohesion was rapidly restored in these cells by the administration of exogenous cAMP pulses (Fig. 6A,B). It was therefore evident that the formation of cellcell contacts somehow permitted cells to respond to exogenous cAMP even when the endogenous cAMP signalling system was disrupted.

In this report, we have provided evidence that cell-cell adhesion plays a role in the regulation of gp80 expression. The expression of gp24, which mediates EDTA-sensitive cell cohesion, is induced soon after the initiation of development (Knecht et al., 1987) and is regulated by the prestarvation factor (Clarke et al., 1987; Rathi and Clarke, 1992). gp80, on the other hand, is induced a couple of hours later and shows a biphasic pattern of expression (Mann and Firtel, 1989; Ma and Siu, 1990). The first phase results in a low basal level of gp80 expression very early in development. The mechanism involved in the induction of this low level expression of gp80 is unknown. It does not seem to depend on cAMP (Ma and Siu, 1990) or the prestarvation factor (Rathi and Clarke, 1992). Conditioned medium also has no effect on the expression of gp80 (unpublished results). The second phase requires cAMP pulses and results in high levels of gp80 expression. The role of cell-cell adhesion in the regulation of gp80 expression has been investigated using several different methods to disrupt the EDTA-sensitive binding sites. In all cases, the disruption of EDTA-sensitive cell cohesion resulted in drastically reduced levels of gp80 expression and EDTA-resistant cell cohesion. When cells were developed at low density with fast shaking or in the presence of carnitine, gp80 expression remained at a low basal level (Figs 4, 5). Apparently, cells were unable to respond to cAMP and they failed to accumulate the normal level of gp80. Under the same conditions, gp24 expression was only slightly depressed. On the other hand, when 1 mM EDTA was used to disrupt EDTA-sensitive cohesion, cells were unable to activate the basal level of gp80 expression. Since the level of gp24 is substantially reduced in the presence of 1 mM EDTA, EDTA probably has pleiotropic effects on Dictyostelium development. The inhibition of the initial phase of gp80 expression may result from the blockage of an earlier developmental event, other than intercellular cohesion.

Our observations suggest that the loss of cell-cell contact affects primarily the second phase of gp80 expression, when cAMP stimulates a rapid increase in gp80 level, but not its first phase of low basal level expression. Exogenous cAMP pulses, which are capable of inducing an enhanced and precocious expression of gp80 under normal circumstances, cannot stimulate high gp80 expression in the absence of EDTAsensitive cell cohesion (Figs 1, 2, 4, 5). This provides further support that cell-cell contact is required for the activation of the cAMP-mediated pathway resulting in high gp80 expression.

There are two distinct effects of cAMP on Dictyostelium development. First, concomitant with their acquisition of the ability to respond chemotactically to cAMP, cells respond by secreting cAMP into the medium and transcription of gp80 becomes greatly augmented by the extracellular cAMP. In this cAMP signal relay pathway, cAMP binds a cell-surface cAMP receptor, cAR1 (Klein et al., 1988; Saxe et al., 1991), and this leads to the G-protein-coupled activation of adenylate cyclase that produces intracellular cAMP (Snaar-Jagalska et al., 1988; Pitt et al., 1992). The cAMP is then secreted into the medium and binds to the cAR1 receptors on adjacent cells. Since cAR1 is desensitized soon after ligand binding, cAMP is produced and secreted in a pulsatile manner. The second pathway leads to the differential activation of certain aggregation-stage specific genes, including the regulation of gp80 expression. This begins by the transient activation of a cell-surface cAMP receptor, most likely cAR1. The receptor is coupled to the G-protein Gα2, one of eight Gα proteins identified in Dictyostelium (Newell et al., 1988; Van Haastert et al., 1989; Pupillo et al., 1989; Kumagai et al., 1989, 1991; Hadwinger et al., 1991; Wu and Devreotes, 1991), and Gα2 has been shown to be involved in this pathway (Mann and Firtel, 1989; Kumagai et al., 1989; Ma and Siu, 1990; Kumagai et al., 1991). However, adenylate cyclase is not involved in this pathway and intracellular cAMP is not required for gp80 expression (Ma and Siu, 1990). In most cases, the regulation of gene expression in Dictyostelium does not occur via cAMP-induced protein kinase A (Schaap et al., 1993), but via the phospholipase C-activated inositol cycle (Van Haastert et al., 1989; Ginsburg and Kimmel, 1989) or perhaps guanylate cyclase. While the intermediary steps remain to be elucidated, it is evident that this pathway results in the interaction of DNA-binding protein(s) with cAMP-response elements in the upstream DNA of the gp80 gene (Desbarats et al., 1992). Our results suggest that in order for the above pathway to be activated, EDTA-sensitive cell-cell adhesion must be present. Disruption of cell-cell contact leads to a repression of gp80 expression. Whereas the decrease in expression of lateand post-aggregation stage genes caused by the disruption of cell-cell contacts can be rescued by the addition of high concentrations of cAMP (Chung et al., 1981; Landfear et al., 1982; Mangiarotti et al., 1983; Haribabu et al., 1986), the expression of gp80 cannot be rescued by cAMP when EDTA-sensitive cell cohesion is blocked. This suggests that cell-cell contact has a direct effect on the ability of cells to respond to the cAMP signal. However, even in the presence of cell-cell contact, expression of gp80 remained low when cAMP signalling was perturbed by fast shaking (Fig. 6), indicating that cell-cell contact alone is not sufficient to achieve high gp80 expression. It is of interest to note that, in the presence of cell-cell contact, cells were still capable of responding to exogenous cAMP even when endogenous cAMP signalling was disrupted. In contrast, cells were unable to respond to exogenous cAMP in the absence of cell-cell contact, even though they were rotated at low speed, which does not affect endogenous cAMP signal response. Therefore, cell-cell contact via the EDTA-sensitive binding sites and cAMP pulses appear to be two closely interrelated signals required for high level gp80 expression.

The fact that cell-cell adhesion is required to achieve high gp80 expression is of interest in light of several recent reports indicating that cell-cell contact elicits cAMP secretion and may modulate cAMP signalling in Dictyostelium (Fontana and Price, 1989; Fontana et al., 1991a,b). These observations are consistent with the idea that an interdependent relationship exists between cell-cell contact and cAMP signalling. However, our results demonstrate that cell-cell contact not only has an effect on the cAMP signal relay system, but also has an effect on the pathway required for gene regulation. If cell-cell contact only affected cAMP signalling, then exogenous cAMP pulses would have been able to bypass the requirement for cAMP signalling and result in high gp80 expression. It is evident that the cAMP-mediated signal relay pathway and the gene regulation pathway are connected, at least at the level of cAR1. In addition, the activation of Gα2 may be a shared downstream event, since Gα2 has an effect on adenylate cyclase activity in vivo (Snaar-Jagalska et al., 1988; Kumagai et al., 1991). Therefore, cell-cell contact may modulate the ability of cAR1 to bind cAMP or its ability to transmit the signal for gene regulation upon ligand binding. Exactly how cell-cell adhesion affects cAR1 is unknown. It is conceivable that changes in cell-cell contact may have subtle effects on the conformation of cAR1, its mobility in the bilayer, or its adaptive response to cAMP. These changes in turn may either permit or interfere with the interactions of cAR1 with Gα2 or other as yet unidentified components of the pathway.

There is increasing evidence that cell adhesion molecules are involved not only in cell adhesion, but also in signal transduction. Several neural cell-cell adhesion molecules, including NCAM, L1 and N-cadherin, are known to promote neurite outgrowth from a variety of neuronal cells via a pertussis toxinsensitive G protein-dependent pathway (for a review, see Doherty and Walsh, 1992). We have recently found that gp24 interacts with several cytoplasmic components, which may have a role in the dynamics of the cytoskeleton as well as signal transduction (Brar and Siu, unpublished data). Further studies on the relationship between the gp24 complex and cAR1 will be important for our understanding of the role of cell-cell adhesion in signal transduction.

We thank Dr Peter Lewis for advice and discussion. This work was supported by Operating Grant MT-6140 from the Medical Research Council of Canada. L. Desbarats was supported by a Medical Research Council studentship and S. K. Brar was supported in part by an Ontario graduate scholarship.