Re-expression of 117 antigen, a cell surface glycoprotein of aggregating cells, during terminal differentiation of Dictyostelium discoideum prespore cells

Previous experiments have implicated the 117 antigen, a 69–72 kd glycoprotein, in the cohesion of Dictyostelium discoideum amoebae during the aggregation phase of their life cycle (Brodie et al. 1983). Glycoprotein 117 is expressed as cells develop aggregation competence and reaches its maximal level on cells as they enter into aggregates. The antigen disappears from the surface of cells as they progress from tight aggregates to slugs. At that time, two distinct cell types, prespore and prestalk are readily observed. These cells eventually give rise to the spore and stalk cells, respectively, of the fruiting body, the final stage of the developmental cycle.

As part of their terminal differentiation, some cells re-express 117 antigen on their surface. As slugs enter the culmination phase of the developmental cycle, and cells begin to lift themselves from the substratum, 117 antigen reappears on the surface of a subpopulation of cells, but does so for only a brief period of time. No 117 antigen is detectable on the surface of spores in the mature fruiting body (Sadeghi et al. 1987).

Although we have documented an unusual pattern of surface expression of the 117 antigen, both the reasons for the different levels and the nature of the cell type reexpressing 117 antigen remain unclear. During the aggregative period, the changes in the levels of surface 117 antigen expression correlate well with changes in its rate of synthesis. Limitations in the ability to incorporate radiolabel into early culminants did not allow us to determine if new protein synthesis was occurring at that stage in development. To overcome those limitations, we have analysed the developmental regulation of the 117 antigen at the molecular level. cDNA clones coding for 117 antigen were isolated and used to examine the levels of its mRNA during fruiting body formation. In addition, the population of cells which express 117 antigen during culmination were characterized according to their expression of cell-type specific components. As discussed here, 117 antigen and two other surface markers allow us to identify two stages in the terminal development of spores, which were previously undefined in biochemical terms.

AX2 amoebae were grown in HL-5 medium (Watts & Ashworth, 1970). NC4 and WS380B cells were grown on Escherichia coli B/r or Klebsiella aerogenes bacteria (Sussman, 1966). Development was initiated by washing cells with 20 mm-potassium phosphate buffer, pH 6·4 and plating them on millipore filters. For strains AX2 and NC4, interphase (2h), aggregation stage (5–6 h), loose aggregates (5–8 h), tight aggregates (10h), slugs (12 h), early culminants (15 h) late culminants (18 h) and fruiting bodies (22 h) were harvested from filter pads.

Poly (A+) mRNA was prepared, separated on formaldehydeagarose gels and transferred to cellulose nitrate filters (Oyama & Blumberg, 1986). Quantitation of the RNA was performed by optical density measurement and verified by ethidium bromide staining (Oyama & Blumberg, 1986). Filters were probed with nick-translated cDNA probes and the amount of hybridization quantitated by densitometry using the BioRad Model 620 video densitometer.

Western blots of aggregation stage cells solubilized in SDS were as described previously (Sadeghi et al. 1987).

A Agtll cDNA library obtained from aggregation competent amoebae of strain AX2 (Watts & Ashworth, 1970) was screened using polyclonal antibody raised against purified 117 antigen (Sadeghi et al. 1987). After screening 1·5×10⁵ plaques, we isolated thirteen clones. The clone with the largest insert, 1·7 kb, was plaque purified and used for further analysis. To verify that this clone contained the cDNA which coded for 117 antigen, the antibody which bound to the fusion protein was affinity purified (Weinberger et al. 1985) and used to probe a western blot of proteins obtained from D. discoideum cells starved in the presence or absence of tunicamycin until they became aggregation competent (Sadeghi et al. 1987). In the latter case, the N-linked glycosylation of 117 is inhibited and the protein .migrates on SDS-PAGE as a 60kDa protein (Sadeghi et al. 1987). Both the affinity purified and polyclonal 117 antisera recognized the glycosylated 69 and 72 kDa doublet present in control cells as well as the 60 kDa form of 117 antigen present in tunicamycin treated cells. These experiments indicate that the cDNA clone codes for the 117 antigen which has been characterized previously using monoclonal and polyclonal antibodies. A 1-7 kb cDNA insert from the Âgtll clone was subcloned into the Bluescribe vector. This construct is referred to as B-117.

The IgG fractions of the following mouse monoclonal antibodies were prepared from ascitic fluid: MUD1, which identifies prespore surface glycoprotein, PsA (Krefft et al. 1984); MUD3, which identifies a spore coat glycoprotein of ∼100 kDa (Voet et al. 1985) that is most probably Sp96 (Orlowski & Loomis, 1979); and 117, which identifies cell surface glycoprotein, 117 antigen (Brodie et al. 1983). The IgG fractions were conjugated with fluorescein isothiocyanate (FTTC, Sigma) using the method described by Goding (1983), or with R-phycoerythrin (PE) purified from Griffithsia monilis (Hiller et al. 1983) using the cross-linking reagent SPDP (Carlsson et al. 1978).

D.discoideum strain WS380B was used for flow cytometry experiments (Sadeghi et al. 1987). Migrating slugs were obtained by plating amoebae on water agar and inducing them to migrate toward a lateral light source. Slugs (approx. 2 days old) were picked up by their slime trails and induced to culminate by transferring them to buffer soaked millipore filters. Culminating aggregates were individually selected on the basis of the morphological stage at which the 117 antigen was previously shown to be present and disaggregated using the enzyme papain (Sadeghi et al. 1987). The separated cells were double labelled using FITC- and PE-coupled antibodies diluted in PBS (Bernstein et al. 1988). Samples were incubated on ice for at least 1 h before measurement. They were analysed in an Epics V flow cytometer under conditions for two colour measurement (Bernstein et al. 1988). Three variables were collected for each cell: forward angle light scatter, integrated green fluorescence due to FTTC and integrated red fluorescence due to phycoerythrin. Data were displayed as two-parameter histograms, 64×64 channels with gating to eliminate measures for small particles, debris, etc., falling into the first 4 channels of forward angle light scatter. Frequency of cell counts is represented as a density map in the xy-plane. Only data falling within normal single cell range of fluorescence were considered when calculating percentages, i.e. a small subset of data (∼5–10%) reflecting measures of cell aggregates were ignored.

Developmental changes in the levels of the mRNA that encodes the 117 antigen have been determined by Northern blot analysis (Fig. 1). This mRNA was not detectable in vegetative cells. The mRNA first appeared as cells developed aggregation competence, and reached a maximum by the time cells formed loose mounds. After this stage, the 117 mRNA levels decreased and by the time cells formed slugs, no 117 mRNA was detectable. Prolonged exposure of the autoradiogram did not reveal any 117 mRNA present at that stage. The beginning of culmination marked the reappearance of 117 mRNA. Densitometer scanning of the autoradiogram indicated that the level present in culminating cells was approximately 5–10% of that found in cells which had formed loose aggregates. In other experiments, late culminants and fruiting bodies were examined for 117 mRNA. None was detectable at those stages. In all cases where 117 mRNA was observed, it appeared as a single band of approximately 1·8 kbp, consistent with the fact that the two size forms of the mature 117 antigen contain a single protein of approximately 52 000 (Sadeghi et al. 1987).

The brief reaccumulation of 117 mRNA (and its protein product) late in development indicated that 117 defines a transitional developmental stage, which had not been previously defined biochemically. To obtain insights into the nature of cells at that stage, we probed an identical Northern blot for the levels of the prespore and prestalk enriched transcripts, PL3 and Dll respectively (Barklis & Lodish, 1983; Oyama & Blumberg, 1986). As shown in Fig. 2, PL3 was first detected when cells formed tight mounds. Levels of PL3 decreased thereafter. The prestalk enriched mRNA, Dll, on the other hand, was present at significant levels in cells which had formed loose mounds. Its level decreased after slug formation and when culmination began, Dll was no longer detectable.

Our previous flow cytometry studies indicated that 117 antigen present on aggregating cells is gone from the surface of cells at the time a well established prespore marker, the PsA glycoprotein, is first detected on prespore surfaces (Sadeghi et al. 1987). This result was confirmed at the mRNA level by probing the Northern blot with the D19 cDNA clone (Barklis & Lodish, 1983), which codes for the PsA glycoprotein (Early et al. 1988). D19 mRNA was detected when cells formed tipped mounds, the time that 117 mRNA was lost. Maximal levels of D19 mRNA were expressed in slugs and, interestingly, declined at early culmination, the time at which 117 mRNA was again observed.

Clearly, the developmental changes in the levels of Dll, PL3, and D19 mRNAs are distinct from that of 117 mRNA, suggesting that the cells which re-express 117 antigen during culmination may define a new developmental stage. Experiments described below in which cell surface appearance of the 117 antigen is compared with that of two other well characterized cell surface glycoproteins are consistent with this suggestion.

To characterize the cells which re-express 117 antigen, two colour flow cytometry was used to analyse culmination stage cells using paired combinations of the monoclonal antibodies (Mabs) MUD1, MUD3 and 117 which identify PsA glycoprotein, a spore coat glycoprotein (probably Sp96) and 117 antigen respectively (see Table 1). This procedure allowed levels of expression of two cell surface antigens to be studied simultaneously on a cell by cell basis. A careful study of the results indicates that 117 antigen is re-expressed on the surface of prespores shortly before they mature into spores, and two new stages in spore development are apparent, which we have named pdspl and pdsp2 (partially differentiated spore type 1 and type 2).

A well characterized marker of mature spores is the spore coat glycoprotein recognized by Mab MUD3 (Table 1). In migrating slugs this glycoprotein is only detected inside prespore cells and is absent from prestalk cells (Devine et al. 1983; Bernstein et al. 1988). During culmination MUD3 antigen begins to appear in the extracellular spore coat. Analysis of a group of culminating cells using PE-coupled MUD3 and FITC-coupled 117 Mabs (Fig. 3A) showed that the population of cells which expressed 117 antigen (∼39%) also carried the MUD3 antigen at the surface. This firmly demonstrated that 117 antigen is expressed only on the surface of cells in the spore differentiation pathway. A second population of cells (∼40%) labelled with neither antibody was a mixture of prestalk (pst) and prespore cells which are not yet expressing the MUD3 antigen at the surface (psp). A third population (∼21%) was observed which carried MUD3 antigen only and not 117 antigen. Figs 3B & 3C will demonstrate that this population was in fact a mixture of partially differentiated spores and mature spores.

When the two colour flow cytometry studies were expanded to include a prespore marker, PsA glycoprotein detected by the MUD1 Mab, a more complex picture of maturing spore cells emerged. Four distinct subpopulations were observed when separated culmination stage cells were reacted with PE-coupled MUD3 and FITC-coupled MUD1 (Fig. 3B). One population (∼19%) consisted of prestalk cells (pst) which expressed neither PsA nor MUD3 antigen. A second group (∼11%) consisted of ‘classical’ prespore cells (psp) as found in migrating slugs, which expressed PsA but not MUD3 antigen on their surfaces. A third group (∼46%) was a mixture of mature spores and partially differentiated spores which expressed MUD3 antigen and not PsA. The distinction between mature spores and partially differentiated spores will become clear when Fig. 3C is considered. Finally, the fourth population detected in Fig. 3B (∼24%) consisted of partially differentiated spores which expressed both prespore (PsA) and spore (MUD3 antigen) specific markers on their surface. In summary, Fig. 3B revealed, in addition to expected prestalk, prespore, and spore populations, a new class of partially differentiated spores which we have called pdspl (partially differentiated spore type 1). The prespore (psp) and pdspl subpopulations were very clearly separated; few cells were recorded in transition from psp to pdspl (Fig. 3B), indicating a possible rapid change between these two stages, perhaps as a result of secretion of internal spore coat glycoprotein (MUD3 antigen) to start forming the spore coat.

In a third experiment, culminating cells were double labelled with PE-coupled MUD1 and FITC-coupled 117 Mabs (Fig. 3C). Three clear subpopulations of cells were observed. One group (∼29%) expressed neither PsA nor 117 antigen, and this was comprised of prestalk cells and mature spores (pst+sp). Second, a group of cells (∼27%) which expressed only PsA and not 117 at the surface consisted of ‘classical’ prespores and partially differentiated spores (psp+pdspl) which can be distinguished on the basis of MUD3 antigen labelling (Fig. 3B). The third group (∼44%), expressed the 117 antigen but not PsA on their surface. This is the same subpopulation revealed in Fig. 3A carrying the 117 antigen together with the MUD3 antigen at the surface. It is clear that there is no subpopulation of cells expressing both PsA and 117 antigen simultaneously. Therefore the partially differentiated spore type 1 (pdspl) cells described above expressing PsA and MUD3 antigen can not express 117 on their surface. As 117 antigen is always accompanied by the presence of MUD3 antigen at culmination, this group forms a second distinct partially differentiated spore type which we name pdsp2. These results are summarized in Table 2.

To confirm the predictions made on the basis of two colour flow cytometry in Fig. 3, the data was reanalysed in terms of forward angle scatter which correlates with cell size (Fig. 4). Prestalk, prespore and spore cells all have distinctive sizes. Of particular note is the dramatic decrease in forward angle light scatter as prespores differentiate into spores. The two classes of partially differentiated spores (pdspl and pdsp2) exhibited the progressive decrease in forward angle light scatter expected of cells differentiating from prespores to spores. The relatively large amount of material having small forward angle light scatter, of a size expected for damaged cells, in the pdspl stage and to a lesser extent the pdsp2 stage is thought to indicate the fragility of the cells during these stages of development.

Fig. 3 summarizes the results of three different experiments, yet the quantitative estimates of the percentage of cells in each of the five differentiation classes were remarkably consistent. Since a substantial number of prestalk cells had differentiated into stalk cells which cannot be disaggregated for flow cytometry (and hence have been removed from the system), the estimate of 19% for pst cells in Fig. 3B is not unexpected. In

Fig. 3B only 11% of the population is shown to be ‘classical’ prespore cells similar to those found in slugs (i.e. expressing PsA, but not MUD3 antigen at the cell surface). Taken together the results also indicate that only a small percentage (5–10%) of prespore cells had fully matured into spores. Hence ∼60% of the population consists of partially differentiated spores of which 15–20% are pdspl and 40–45% pdsp2. This is not unexpected for a culminating fruiting body. A second series of experiments identical to Fig. 3 using culminating fruiting bodies at a slightly later stage confirmed the findings shown in Fig. 3 (data not shown). The two classes of partially differentiated spores were once again observed, with ∼7% of cells expressing both PsA and MUD3 antigen (pdspl) and ∼45% of cells expressing both MUD3 antigen and 117 antigen (pdsp2).

The second appearance of 117 antigen and its mRNA suggested that 117 antigen identifies a transition stage in late development that has not been previously characterized biochemically. Such an identification requires cell by cell analysis and is only really practicable by flow cytometry. The transition stage was confirmed by examining, along with 117 antigen, the population of cells which expressed surface glycoproteins indicative of prespore and spore cells via two colour flow cytometry (Table 2). The data indicate that cells which express 117 antigen have embarked on the spore development pathway, since they also express the spore coat glycoprotein, MUD3 antigen.

It is known that slug prespore cells express PsA but not 117 nor MUD3 antigen at their surfaces, and mature spores express MUD3 antigen but neither 117 nor PsA at their surfaces. By combining this information together with the results of Fig. 3A, B, C we have constructed a developmental profile of a single cell depicting the various transition stages which lead to spore maturation (Fig. 5). A prespore cell at the slug stage expresses only PsA at its surface. The spore coat glycoprotein then appears such that the cell expresses high levels of both PsA and MUD3 antigen. We have called cells which have reached this developmental stage, partially differentiated spores type 1 (pdspl), since they express a spore coat protein. PsA then disappears from the cell surface, and when it has completely disappeared, 117 antigen appears. We have named cells at this stage partially differentiated spores type 2 (pdsp2). 117 antigen remains exposed at the surface only for a short time, as mature spores do not express it at their surface. It should be noted that our previous single label flow cytometry studies revealed both PsA and 117 glycoprotein on subpopulations of cells at early culmination (Sadeghi et al. 1987). Our results here show clearly that at the single cell level these determinants are not present concurrently. The apparent expression of both PsA and 117 antigen at the same time resulted from asynchrony in the stage of development of cells within a culminating fruiting body.

Commitment of cells to certain developmental pathways involves the turning on or off of specific genes. It is common for cells in a given pathway to express a particular gene for only a limited period of time. The cyclic expression of genes at different stages of development is more unusual since it requires the capacity to turn the gene on and off at different developmental stages. Based on the periodic changes in 117 mRNA levels we suggest that 117 is a gene that falls into this more unusual category. We recognize that our current data do not eliminate the possibility that 117 mRNA levels may be primarily regulated at the level of degradation. In either case the components which regulate 117 mRNA levels at different developmental stages are either different or are, themselves periodically regulated. This plasticity of gene expression may well be fundamental to the morphogenesis of a multicellular organism.

In this regard, it is of interest to note that, like 117 antigen, the neuronal cell adhesion molecule, N-CAM, of higher eukaryotes is expressed in a periodic fashion on the surface of specific cells during embryogenesis (Edelman, 1984). The cyclic expression of cohesion molecules, like N-CAM and 117 antigen, underscores their importance in metazoan development. In the case of 117 antigen, it is not clear if the molecule serves the same function during culmination as in aggregation. It remains possible that the antigen present during culmination may differ, biochemically, from that in aggregating cells. 117 antigen is extensively modified both co- and post-translationally by carbohydrate residues (Sadeghi et al. 1987). Since the types of oligosaccharides synthesized by cells are developmentally regulated (Henderson, 1984), it is possible that aggregation and culmination forms of 117 antigen differ. A precedent for tissue specific differences in glycosylation patterns exists in that rat brain and thymocyte forms of Thy-1 are differently glycosylated (Parekh et al. 1987).

The re-expression of 117 antigen during a short period in culmination may reflect a requirement, at this stage of morphogenesis, for not only cohesion of a subpopulation of cells, but for cohesion in a manner qualitatively similar to that which occurs in aggregation. It is noteworthy that the cyclic AMP phosphodiesterase and the cyclic AMP-receptor show a similar developmental pattern to 117 antigen in terms of changes in surface activity (Henderson, 1975; Schaap & Spek, 1984; see Fig. 5 in Lacombe et al. 1986). Both cyclic AMP binding and cyclic AMP surface phosphodiesterase are maximum in aggregating cells, decrease upon formation of tight aggregates, and increase slightly at culmination. A precise developmental time course of this re-expression has not been performed, nor has the identity of the cell type(s) involved in this late expression been established. Based on the observation that prestalk cells, when disassociated from the slug, are more chemotactic with respect to cyclic AMP than are prespore cells, it has been hypothesised that the cyclic AMP binding and phosphodiesterase activities are associated with prestalk cells. Here we have shown that at least one aggregation marker, 117 antigen, is reexpressed later in development by a subpopulation of cells which are not prestalks, but those developing along the spore cell differentiation pathway. We are currently examining if other aggregation components, specifically the cyclic AMP receptor and cyclic AMP phosphodiesterase, are also expressed by psdp2 cells as is the 117 antigen. It may be that, for this population of cells, culmination is indeed a recapitulation of aggregation. Such a hypothesis based on morphological studies, has been presented (Rand & Sussman (1983)).

We have previously suggested that the loss of 117 mRNA and surface 117 antigen in post aggregative cells may help determine the precise time when amoebae are no longer aggregative but instead have embarked on a subsequent step(s) in their developmental program (Sadeghi et al. 1987). Similarly, the narrow window of time when 117 mRNA reaccumulates and 117 antigen is expressed at the prespore surface again precisely specifies previously undefined transitional states through which prespores pass to become spores. The re-expression of an aggregative marker during culmination provides a tool to begin understanding the events involved in this morphogenetically more complex process.

This research was supported by a NH & MRC (Australia) grant to K.L.W., NCI Intermural research program grant to D.B., and grant no. GM34561 to C.K. We thank Christine Stephenson for coupling monoclonal antibodies with FITC and PE. We thank Michelle Thorpe for typing the manuscript.