Abstract
The stalk cell differentiation inducing factor (DIF) has the properties required of a morphogen responsible for pattern regulation during the pseudoplasmodial stage of Dictyostelium development. It induces prestalk cell formation and inhibits prespore cell formation, but there is as yet no strong evidence for a morphogenetic gradient of DIF. We have measured DIF accumulation by monolayers of isolated prestalk and prespore cells in an attempt to provide evidence for such a gradient. DIF is accumulated in the largest quantities by a subpopu lation of prestalk cells that specifically express the DIF-inducible genes pDd56 and pDd26. Since it has been shown recently that cells that express pDd56 are localized in the central core of the prestalk cell region of the pseudoplasmodia, our current results suggest a morphogenetic gradient generated by this region.
Introductlon
One of the major unsolved areas of biological research is the mechanism of pattern formation during embryonic development. Numerous theoretical models have been formulated for the establishment of gradients of diffusible molecules or morphogens, but direct evidence for the existence of such morphogenetic gradients has only recently been obtained (Thaller and Eichele, 1987; Driever and Nüsslein-Volhard, 1988). Much work will be required to determine how such gradients are regulated. The simplicity of the developmental process in the lower eukaryote Dictyostelium discoideum makes it an attractive system for studies on the role of morphogenetic gradients in pattern formation.
During vegetative growth, Dictyostelium amoebae are solitary, but upon starvation, they aggregate and eventually form a migrating pseudoplasmodium. Within the pseudoplasmodium there are two distinct cell types, prestalk cells and prespore cells, that are spatially segregated in a simple developmental pattern (Loomis, 1975). During subsequent morphogenesis, the prestalk and prespore cells give rise to the stalk and spore cells of the mature fruiting body. A number of molecules have been suggested as potential morphogens since they influence the relative proportions of the two cell types (Schaap, 1986; Williams, 1988). However, the detailed molecular mechanisms that are responsible for the establishment of this developmental pattern have yet to be elucidated.
The best studied Dictyostelium morphogen is the stalk cell differentiation inducing factor (DIF). In high cell density monolayers of the wild-type variant V12M2, amoebae differentiate into stalk cells in the presence of cyclic AMP, but at low cell density insufficient DIF is produced and stalk cell formation is totally dependent upon the addition of both cyclic AMP and DIF (Town and Stanford, 1979; Gross et al. 1981; Brookman et al. 1982; Sobolewski et al. 1983). Furthermore, monolayers of sporogenous mutants of V12-M2 generate both stalk and spore cells and DIF markedly stimulates formation of the former and inhibits formation of the latter (Kay and Jermyn, 1983). The expression of the prespore cell specific gene, D19 is markedly inhibited by DIF (Early and Williams, 1988).
Several developmentally regulated genes have now been isolated that are dependent upon DIF for their induction (Williams et al. 1987) and their expression is restricted to the prestalk cell population (Jermyn et al. 1987). Recently studies on the expression of two of these genes, pDd63 and pDd56, have provided confirmatory evidence for the existence of subpopulations of prestalk cells (Jermyn et al. 1989; Williams et al. 1989). One cell population, termed pstB, that expresses both pDd56 and pDd63 is localized in the central core of the anterior portion of the prestalk cell region. Another cell population, termed pstA, only expresses pDd63 and is localized around the periphery of the anterior portion of the prestalk cell region.
It has been suggested that DIF is the activator molecule in the Gierer and Meinhardt (1972) model of developmental pattern regulation and is generated by the prestalk cells of the developing organism (Gross et al. 1981). It was, therefore, predicted that the levels of DIF would be highest in the prestalk cell region of the migrating pseudoplasmodium (Gross et al. 1981). In an attempt to test this prediction, Brookman et al. (1987) dissected migrating pseudoplasmodia but found significantly higher levels of DIF in the prespore region. However, these measurements are complicated by the large amounts of DIF that are extruded into the pseudoplasmodial sheath during the migration of the pseudoplasmodia (Neave et al. 1986). In order to circumvent this problem, it was imperative to devise alternative methods to determine which cells are responsible for DIF production.
The availability of a monolayer assay for DIF accumulation (Kwong and Weeks, 1989) suggested an alternative approach. Isolated prestalk and prespore cells were allowed to differentiate in monolayers and the amount of DIF that accumulated was assayed at various times. The results show that the prestalk cell population accumulates more DIF rapidly than the prespore cell population. In addition preliminary evidence is presented for the fractionation of prestalk cells into distinct sub-populations.
Materials and methods
Organism and culture conditions
D. discoideum, wild type strain V12-M2 and the sporogenous mutants, HM29 and HM18, were grown on nutrient agar in association with Enterobacter aerogenes. Cells were harvested and washed free of bacteria by differential centrifugation. Pseudoplasmodia of strain V12-M2 were obtained by allowing the vegetative cells to differentiate for 16 h on the surface of 2% non-nutrient agar plates, containing 5% Bonner’s salts (Bonner, 1947) for 16 h. To obtain pseudoplasmodia of strains HM18 or HM29, vegetative cells were developed for 16 h on the surface of 1.5 % non-nutrient agar plates, containing 10% Bonner’s salts and 0.05% activated charcoal.
Separation of prestalk and prespore cells on Percoll gradients
The separation of prestalk and prespore cells was performed using either the Percoll continuous gradient method of Ratner and Borth (1983) or a modification of this method that utilizes a Percoll step gradient (Kwong et al. 1988). For both methods pseudoplasmodia were harvested and filtered through a fine nylon mesh to remove undifferentiated cells. Cells retained on the filter were collected and disrupted by mixing vigorously in phosphate buffer (20 mM K2HPO4, KH2PO4 pH 6.0). Cells were then pelleted twice at 1000 g for 5 min to remove slime sheath. The cells were resuspended at 4 ×108cellsml−1 5mM 2-[N-morpholino] ethanesulfonic acid (MES) pH6.2, 0.06% β-mercaptoethanol, lmgml−1 Sigma Type XIV protease and incubated at room temperature for 15 –20 min with constant trituration through a 23 G needle.
In the first cell separation method (Ratner and Borth, 1983), aliquots (0.25 ml) were layered onto the surface of a preformed Percoll gradient that had been generated by centrifuging 58% Percoll, 20 mM MES (pH 6.2) and 20 mM EDTA (pH 7.0) in a Sorvall SS34 rotor at 15 000 revs min−1 for 40 min at 4 °C. The cell separation was achieved by centrifuging the gradient at 13 000 revs min−1 for 10min. The low density (top) and high density (bottom) fractions were collected using a syringe attached to a 20G needle.
In the second method (Kwong et al. 1988), aliquots (0.25 ml), containing approximately 108 cells, were layered onto the surface of an ice-cold discontinuous Percoll gradient and centrifuged in a Beckman SW 41 rotor at 12000revsmin−1 for 5 min at 4°C. The discontinuous gradients contained 2.5 ml 45% Percoll in the bottom layer; 2.5 ml of 30 % Percoll in the middle layer and 2.5 ml of 15 % Percoll in the top layer, each diluted in 20 mM MES, pH6.2, 20 mM EDTA. Low density cells were recovered as a band at the interface between the top and middle layers of the gradient, whereas high density cells were recovered between the middle and bottom layers.
Cells recovered from the gradients were diluted with icecold Bonner’s salts and centrifuged at 1000g for 5 min. Cell pellets were washed once in Bonner’s salts by resuspension and recentrifugation.
Monolayer differentiation
Washed cells from the gradients were resuspended in MES-Salts (10mM MES pH 6.2, 20 mM KC1, 20 mM NaCl, ImM CaCl2, ImM MgCl2, 200 μgml−1 streptomycin sulfate and 15 μgml−1 tetracycline, supplemented with ImM cyclic AMP). The cell density was adjusted to 2 ×106cellsml−1 and 2 ml aliquots of the suspension were added to 5 cm Nunc tissue culture dishes. Cells rapidly settled to the bottom of the dish to establish a monolayer density of 2 ×105 cells cm−2. Plates were incubated at 22°C. At the indicated times, supernatants were harvested for DIF extraction and RNA was extracted from the cells.
DIF extraction and assay
MES-Salts medium was collected from 2 plates, combined and 12.5 μg butylated hydroxytoluene (BHT) added. The samples were extracted twice with an equal volume of heptane and the pooled extracts were concentrated in vacuo and resuspended in 100 μ1 absolute ethanol.
For the DIF assay, washed V12-M2 cells were plated at a cell density of 103 cells cm−2 in 5 cm Nunc tissue culture plates. Each dish contained 2 ml assay medium (10 mM NaCl, 10 mM KC1, 2mM CaC12, 5mM MES pH6.2, 100 μgml−1 streptomycin sulfate, 5 mg ml−1 BHT, ImM cyclic AMP) and 2 to 10 μl of the concentrated DIF extracts. Stalk cells were scored after 48h incubation at 22°C by phase contrast microscopy. Cells were scored positive for stalk cell formation if more than 50% of the cell volume was vacuolated. DIF activity was determined from the values that yielded a linear response to added DIF. 1 unit DIF activity was defined as the amount of DIF necessary to form 1 % stalk cells (Brookman et al. 1982).
RNA isolation and Northern blotting
Total RNA was isolated from the monolayer cells by a modification of a previously published method (Bimboim, 1988). Briefly, the MES-Salts medium was removed from the monolayers and replaced with 2 ml of RES-1 buffer (0.5 M LiCl, IM urea, 1% SDS, 0.02M sodium citrate, 2.5mM trans-1,2-diaminocyclohexane-N,N,N′, N′-tetra-acetic acid, pH6.8) to lyse the.cells. Proteinase K was added to a final concentration of 50 μg ml−1 and the mixture was incubated at 50°C for 30min. After ethanol precipitation, the pellet containing both RNA and DNA was suspended in 100 μl of RES-I buffer and phenol/chloroform extracted. The RNA was selectively precipitated by the addition of an equal volume of 5 M LiCl/95% ethanol (3/2) and the mixture was stored overnight at 4°C to allow complete precipitation. The supernatant was removed by centrifugation for 15 min in a microfuge and the pellet was washed several times with 75 % ethanol.
RNA samples for Northern blot analysis were adjusted to 50% formamide, 40 mM 3-(N-morpholino) propanesulfonic acid pH7.0, 10mM sodium acetate, ImM EDTA, 6% formaldehyde. The samples were heat denatured for 10 min at 65 °C and then size fractionated on 1 % formaldehyde-agarose gels. The gels were then transferred to nitrocellulose filters and hybridized with randomly labelled cDNA probes as described previously (Robbins et al. 1989). The filters were then washed with 0.5 ×SSC, 0.1% SDS either at 65°C (IG7 and D19) or 60°C (pDd63, pDd56, pDd26 and Dll).
Results
(1) DIF accumulation and developmental gene expression by prestalk and prespore cell monolayers of strain V12-M2
Prestalk and prespore cells of strain V12-M2 were separated by the procedure described by Ratner and Borth (1983) and plated in monolayers. Cell exudates were analyzed periodically for DIF accumulation (Fig. 1). The results showed clearly that DIF was accumulated more rapidly by isolated prestalk cells than by isolated prespore cells. The differences shift with time, but this can be explained by the fact that prespore cells of V12-M2 eventually differentiate into stalk cells in monolayers (Town and Stanford, 1977; Kwong et al. 1988).
Accumulation of DIF by prestalk and prespore cells of strain V12-M2. Pseudoplasmodial cells of strain V12-M2 were fractionated on Percoll continuous gradients as described by Ratner and Borth (1983). The isolated low density fraction (presumptive prestalk cells) (▫) and high density fraction (presumptive prespore cells) (▪); were washed and set up in monolayers as described under Materials and methods. At the indicated time points, supernatants were decanted, heptane extracted and assayed for DIF activity as described under Materials and methods. The plotted values are for a single experiment.
Accumulation of DIF by prestalk and prespore cells of strain V12-M2. Pseudoplasmodial cells of strain V12-M2 were fractionated on Percoll continuous gradients as described by Ratner and Borth (1983). The isolated low density fraction (presumptive prestalk cells) (▫) and high density fraction (presumptive prespore cells) (▪); were washed and set up in monolayers as described under Materials and methods. At the indicated time points, supernatants were decanted, heptane extracted and assayed for DIF activity as described under Materials and methods. The plotted values are for a single experiment.
These results suggest that prestalk cells generate DIF. However, since prestalk and prespore cells of strain NC4 undergo extensive dedifferentiation in shake suspension (Weijer and Durston, 1985), it is possible that prestalk cells undergoing dedifferentiation into prespore cells are responsible for DIF accumulation in monolayers. To address this possibility, RNA was isolated at various times during the incubation period and the levels of the prespore-specific D19 mRNA and prestalk-specific pDd63 mRNAs (Jermyn el al. 1987) were measured. Initially the prestalk cells expressed relatively little D19 mRNA, and the prespore cells expressed very little pDd63 mRNA, (Fig. 2), confirming that the gradient separated prestalk and prespore cell populations were relatively pure, as previously documented (Ratner and Borth, 1985; Jermyn et al. 1987). The prespore cell population expressed some Dll mRNA (Fig. 2), confirming that Dll is enriched in prestalk cells, but not specific for this population (Jermyn et al. 1987). As monolayer differentiation progressed the prestalk cells rapidly accumulated D19 mRNA and the prespore cells rapidly accumulated pDd63 mRNA. Thus, at the level of prespore and prestalk specific gene expression there is clear evidence for dedifferentiation during monolayer incubation. Since the dedifferentiation of both populations is apparent before DIF accumulation is detectable, the possibility that prespore cells generate DIF cannot be eliminated.
Northern blot analysis of mRNA levels in monolayers of prestalk and prespore cells of strain V12M2. Monolayers were established as described in the legend for Fig. 1. Total RNA was extracted from the low density fraction (presumptive prestalk cells) at Oh (lane 1), 2h (lane 3) 4h (lane 5) and 6h (lane 7) and from the high density (presumptive prespore cells) at Oh (lane 2) 2h (lane 4); 4h (lane 6) and 6h (lane 8). RNA (20/tg) was size fractionated on 1 % formaldehyde-agarose gels. The gels were then transferred to nitrocellulose filters and hybridized with the indicated randomly labelled cDNA probes.
Northern blot analysis of mRNA levels in monolayers of prestalk and prespore cells of strain V12M2. Monolayers were established as described in the legend for Fig. 1. Total RNA was extracted from the low density fraction (presumptive prestalk cells) at Oh (lane 1), 2h (lane 3) 4h (lane 5) and 6h (lane 7) and from the high density (presumptive prespore cells) at Oh (lane 2) 2h (lane 4); 4h (lane 6) and 6h (lane 8). RNA (20/tg) was size fractionated on 1 % formaldehyde-agarose gels. The gels were then transferred to nitrocellulose filters and hybridized with the indicated randomly labelled cDNA probes.
(2) DIF accumulation by prestalk and prespore monolayers of sporogenous mutants
Although the wild-type strain V12-M2 differentiates into only stalk cells in monolayers, prestalk and prespore cells of the sporogenous mutants, HM18 and HM29, differentiate into stalk and spore cells respectively, under these conditions (Tsang and Bradbury, 1981). It is therefore possible that these cells do not undergo extensive dedifferentiation under monolayer conditions. We isolated prestalk and prespore cells from a sporogenous mutant strain HM29 using the Percoll continuous gradient method of Ratner and Borth (1983). The possibility of prestalk cell dedifferentiation in monolayers was assessed as before. The presumptive prestalk cell fraction (low density) contained the prestalk specific pDd63, pDd56, pDd26 and Dll mRNAs and no prespore specific D19 mRNA, indicating that the preparation contained negligible prespore cell contamination. During the monolayer incubation period, D19 mRNA did not accumulate (Fig. 3), suggesting no extensive dedifferentiation of prestalk cells into prespore cells. The levels of pDd63, pDd56, pDd26 mRNAs remained relatively constant whereas the amounts of Dll mRNA increased during the incubation period. At the end of the monolayer incubation period over 90 % of the cells had differentiated into stalk cells (data not shown).
Northern Blot analysis of mRNA levels in monolayers of pseudoplasmodial cells of strain HM29. The experimental protocol was the same as for Fig. 2 except that strain HM29 was used instead of V12-M2.
In contrast to the pure prespore fraction that has been routinely isolated from wild type strains (Ratner and Borth, 1983; Jermyn et al. 1987; Fig. 2), the presumptive prespore cell fraction (high density) from HM29 contained D19 mRNA as anticipated, but in addition considerable amounts of pDd63 and Dll mRNA (Fig. 3). This fraction, however, contained no pDd56 or pDd26 mRNA. These results suggested that the prespore cell preparation was contaminated with a subpopulation of prestalk cells that did not express pDd56 or pDd26. During the monolayer incubation period, the level of D19 mRNA gradually declined whereas the level of pDd63 mRNA remained relatively constant and the level of Dll mRNA increased. Since the prespore cell population is contaminated with a subpopulation of prestalk cells, it is not possible to determine the extent of prespore cell dedifferentiation under these conditions. Detectable levels of pDd56 and pDd26 began to be expressed by this high density population during monolayer incubation and levels increased as the monolayer incubation proceeded. By the end of the monolayer incubation period approximately 70 % of the cell population were spore cells and 20% were stalk cells (data not shown).
These results indicated that the prestalk cells had been fractionated into two distinct subpopulations. In both HM29 and a second sporogenous mutant, HM18, DIF was accumulated more rapidly by monolayers of the low density prestalk cells than by monolayers of the mixture of prespore cells and high density prestalk cells. In strain HM29, DIF was first detectable after 4h of incubation of the low density prestalk cell monolayers and levels steadily increased during the remainder of the incubation (Fig. 4A). In contrast, DIF was not detectable in exudates of the monolayers of the combined high density prestalk and prespore cell fraction until 24 h and then only at considerably lower levels (Fig. 4A). Similar results were obtained with strain HM18 although the differences in the levels of the accumulated DIF were not as pronounced (Fig. 4B). These results suggest that DIF is generated by a subpopulation of prestalk cells that specifically express pDd56 and pDd26. It should be noted that although DIF is not detectable in exudates from prespore cells after 6h of incubation in monolayers (Fig. 4A), both pDd56 and pDd26 genes are expressed at this time (Fig. 3). These results indicate that very low concentrations of DIF are sufficient for pDd56 and pDd26 gene expression. Similarly, it was shown recently that DIF is not detectable in cell exudates from monolayer cells, under conditions where pDd63 gene expression is induced (Kwong and Weeks, 1990).
Accumulation of DIF by pseudoplasmodial cells of sporogenous mutants. The experimental protocol was the same as for Fig. 1, except that (A) cells of HM29, and (B) cells of HM18 were used instead of V12-M2. Plotted values are the means of two experiments for A and one experiment for B.
Williams and co-workers have recently shown that the expression of pDd56 gene is specific to a central core of cells in the anterior region of the prestalk region. They have called these cells pstB, to distinguish them from the cells in the periphery of the anterior region, termed pstA (Jermyn et al. 1989; Williams et al. 1989). Our results indicate that pstA and pstB cells of the sporogenous mutant, HM29, are separated by the continuous Percoll gradient, with the pstA cells fractionating along with the prespore cells. These results suggest that pstA cells are more dense than pstB cells in the mutant. The difference between the fractionation of the wild type cells and that of the sporogenous mutant cells on identical continuous gradients are illustrated in Fig. 5.
Summary of the Percoll gradient results. (A) V12M2 cells separated on a Percoll continuous gradient (Ratner and Borth, 1983); (B) HM29 cells separated on a Percoll continuous gradient (Ratner and Borth, 1983). (C) V12-M2 cells separated on a Percoll step gradient (Kwong et al. 1988).
Summary of the Percoll gradient results. (A) V12M2 cells separated on a Percoll continuous gradient (Ratner and Borth, 1983); (B) HM29 cells separated on a Percoll continuous gradient (Ratner and Borth, 1983). (C) V12-M2 cells separated on a Percoll step gradient (Kwong et al. 1988).
(3) Separation of prestalk and prespore cells of strain V12-M2 by Percoll step gradients
One disadvantage of the Ratner and Borth (1983) procedure for separating prestalk and prespore cells is that the two populations band close together on the gradients, which makes recovery of uncontaminated populations difficult and tedious. We therefore developed a Percoll step gradient to allow a better separation of the two cell populations (Kwong et al. 1988). The new procedure yielded a low density (presumptive prestalk) cell population that was only minimally contaminated with the prespore specific enzyme UDP-galactosyl-polysaccharide-transferase and a high density (presumptive prespore) cell population that was only minimally contaminated by the prestalk cell marker acid phosphatase-II (Kwong et al. 1988). However, when pseudoplasmodial cells of V12-M2 were separated by the step gradient and the extracted RNA was probed with pDd63 and pDd56 as more selective and quantitative markers of prestalk cells, it was found that the presumptive prespore fraction was heavily contaminated by cells expressing pDd63, but not pDd56 (Fig. 6). The RNA samples were also probed with Ddras, a gene that is enriched but not totally specific for prestalk cells (Jermyn et al. 1987). The majority of the Ddras mRNA was localized in the high density fraction. These results suggest that the cell separation procedure devised by Kwong et al. (1988) fractionates the prestalk cells into a high density population that expresses Ddras and pDd63 (pstA) and a low density population that expresses pDd63 and pDd56 (pstB). The high density population sediments along with the prespore cells. This fractionation is also illustrated in Fig. 5.
mRNA levels in pseudoplasmodial cells of V12-M2. Total RNA was extracted from low density fraction (lane 1) or high density fraction (lane 2) of pseudoplasmodial cells separated on Percoll step gradients by the method of Kwong et al. (1988). RNA (20 μg) was extracted, size fractionated, blotted and hybridized as described under Methods.
mRNA levels in pseudoplasmodial cells of V12-M2. Total RNA was extracted from low density fraction (lane 1) or high density fraction (lane 2) of pseudoplasmodial cells separated on Percoll step gradients by the method of Kwong et al. (1988). RNA (20 μg) was extracted, size fractionated, blotted and hybridized as described under Methods.
Monolayers of low density cells isolated by the Kwong et al. (1988) procedure accumulated more DIF than those of the high density cells (Fig. 7). Thus the prestalk cells that sediment at low density accumulate more DIF than the mixture of the high density prestalk cells and prespore cells. Again this result suggests that DIF is preferentially generated by the subpopulation of prestalk cells that express pDd56.
DIF accumulation by pseudoplasmodial cells of strain V12-M2. The experimental protocol was the same as that described for Fig. 1 except that the low density cell fraction (▫) and the higndensity cell fraction (▪) were separated on Percoll step gradients by the method of Kwong et al. 1988. Plotted values are the means of 4 experiments.
DIF accumulation by pseudoplasmodial cells of strain V12-M2. The experimental protocol was the same as that described for Fig. 1 except that the low density cell fraction (▫) and the higndensity cell fraction (▪) were separated on Percoll step gradients by the method of Kwong et al. 1988. Plotted values are the means of 4 experiments.
We attempted to confirm our conclusions with more homogeneous preparations of pstA, pstB and prespore cells. Pseudoplasmodial cells were initially separated on a Percoll continuous gradient and the low density fraction was then fractionated on a step gradient. Unfortunately, insufficient cells were recovered to allow determination of DIF levels.
Discussion
It has been proposed that DIF is the activator in a model of developmental pattern regulation (Gross et al. 1981) of the type proposed by Gierer and Meinhardt, 1972. This model postulates that DIF is generated by the cells in the prestalk region of the pseudoplasmodium and that higher levels of accumulation occur in this region. However, Brookman et al. (1987) dissected pseudoplasmodia into segments and found that the prespore region contained the highest levels of DIF. This finding suggests a possible alternative model in which DIF is produced by prespore cells and acts as a feedback inhibitor of the spore pathway of differentiation. Consistent with this alternative model, DIF is a potent inhibitor of spore formation (Kay and Jermyn, 1983) and a potent inhibitor of the expression of the prespore cell specific gene, D19 (Early and Williams, 1988). The dissection experiments of Brookman et al. are however, subject to the criticism that considerable amounts of DIF diffuse into the pseudoplasmodial sheath (Neave et al. 1986). Since the pseudoplasmodium crawls through its sheath, the level of DIF in the prespore region may be considerably overestimated.
In view of the importance of determining the site of DIF synthesis with regard to understanding the mechanism of pattern regulation in Dictyostelium, we devised an alternative experimental approach. Prestalk and prespore cells isolated from Percoll gradients contained very low levels of DIF (Neave et al. 1986). However, when these cells were placed in monolayers detectable levels of DIF accumulated (Figs 1,4,7). In every experiment prestalk cells accumulated DIF more rapidly than prespore cells, suggesting that the prestalk region of the migrating pseudoplasmodium is the major site of DIF accumulation.
In addition, in two separate sets of experiments, a sub-population of prestalk cells that specifically express pDd56 characteristic of the pstB cells described by Williams and co-workers (Jermyn et al. 1989; Williams et al. 1989) has been separated from the pstA cells. These are the cells that most rapidly accumulate DIF in vitro (Figs 4 and 7) and it possible that these are the cells that generate DIF in the migrating pseudoplasmodia. It must be emphasized that if cells drastically alter their DIF accumulation characteristics when they are disaggregated from the pseudoplasmodia then such a conclusion is unwarranted. For example, the ability of prespore cells to generate DIF might be severely impaired during the isolation of the prespore cell fraction. However, while it is at present impossible to experimentally measure DIF generation in vivo in the migrating pseudoplasmodia, there is evidence to suggest that DIF accumulation is not significantly altered by the disruption of the pseudoplasmodia. The amounts of DIF produced during the in vitro studies reported here are similar to the amounts of DIF produced by migrating pseudoplasmodia (Sobolewski et al. 1983). Furthermore, the amounts accumulated by pronase disrupted pseudoplasmodial cells are identical to the amounts generated by cells that were not treated by pronase (data not shown). It is clear from our studies, that neither pstA cells nor prespore cells rapidly accumulate DIF. The kinetics of DIF accumulation in the wild type strain indicate that prespore cells do eventually produce DIF (Fig. 1), but this can be explained by the fact that these cells slowly differentiate into stalk cells in monolayers (Kwong et al. 1988). Furthermore, it should be noted that the prestalk cell population exhibits some dedifferentiation in monolayers (Fig. 2) and the accumulation of DIF by this population might be even more pronounced, if it were not for this dedifferentiation. In the monolayers of the sporogenous mutants, there is no evidence for rapid dedifferentiation and the differences in DIF accumulation between the isolated fractions are more pronounced.
In the experiments described in this report, we have measured the level of DIF accumulation which presumably represents the difference between the rate of DIF synthesis and DIF degradation. Recent experiments have indicated that DIF-1, the major species of DIF activity, is rapidly degraded by Dictyostelium cells (Traynor and Kay, unpublished observations). Thus the accumulation values reported here might be due to a greater rate of synthesis by prestalk cells or a greater rate of degradation by prespore cells. However, since it is the steady state concentration of DIF either in the cell or in the media that is presumably important in inducing differentiation, the accumulation of DIF should be the most relevant determining factor.
The fractionation of the prestalk cell of V12-M2 into two populations using the method of Kwong et al. (1988), resolves a previously published anomaly with regard to Ddras gene expression. It was shown initially that Ddras mRNA was preferentially localized in the prestalk cell population (Reymond et al. 1985) and this result has subsequently been confirmed (Jermyn et al. 1987). In contrast, Ddras protein synthesis only occurred in the prespore cell population (Weeks et al. 1987). However, since the pseudoplasmodial cells used in these latter experiments were separated by the step gradient procedure (Kwong et al. 1988), the prespore cell fraction would have been contaminated with high density prestalk cells that preferentially expressed Ddras. We have confirmed that when cells are separated by the method of Ratner and Borth (1983), Ddras mRNA is localized preferentially in the prestalk cell population (data not shown). However, when cells are separated by the step gradient (Kwong et al. 1988), Ddrov mRNA is localized predominantly in the high density prestalk cells that contaminate the prespore cell population (Fig. 6). These results suggest the possibility that Ddras is localized in the pstA cells described by Williams and co-workers (Jermyn et al. 1989).
In addition to the recent identification of pstA and pstB cells as distinct populations within the prestalk region of the pseudoplasmodia (Jermyn et al. 1989), and the separation of these populations described here, other recently published experiments have indicated that the prestalk cell population is heterogeneous. We have shown that the prestalk population can be functionally subdivided into cyclic AMP-independent and cyclic AMP-dependent cells on the basis of the requirements for stalk cell formation in monolayers (Kwong et al. 1988). The present results indicate that the cyclic AMP-independent subpopulation might correspond to the pstB fraction. In addition, differences in morphology (Kopachik, 1982) and distribution of antigen (Gomer et al. 1986) have indicated the existence of more than one cell population within the prestalk region of the migrating pseudoplasmodia.
PstA and pstB cells are first detected at the aggregation stage of development, the pstB cells as a group at the base of the aggregate and the pstA cells randomly dispersed throughout (Williams et al. 1989). It is not clear how the initial inductions are triggered, but it is tempting to speculate that the cells at the base of the aggregate generate DIF and eventually express pDd56 in response to the high concentration of DIF that would accumulate in this region. PstA cells would be formed throughout the aggregate from randomly responsive cells induced by the low concentrations of DIF diffusing from the base. We have attempted to isolate the early pstA and pstB cells from the aggregation stage, to determine their capacity to generate DIF, but have thus far been unsuccessful.
The best evidence for a gradient of a specific morphogen thus far demonstrated is during early development in Drosophila, where the bicoid gene product is retained in the anterior pole and then diffuses along the length of the egg (Driever and Nüsslein-Volhard, 1988). The accumulation of the diffusible molecule DIF by a subset of the prestalk cells would generate a similar gradient within the prestalk region of the pseudoplasmodium and might be important in maintaining the developmental pattern. Thus the prestalk cells that express pDd56 would be exposed to a higher concentration of DIF than the remainder of the prestalk cell population. Since DIF is required for the expression of pDd63 and pDd56, it will be interesting to determine if the latter gene requires a higher concentration of DIF for its expression. It has also been speculated that, in Drosophila, genes are activated by different concentrations of bicoid and that expression is therefore dependent upon position (Struhl et al. 1989; Driever et al. 1989).
ACKNOWLEDGEMENTS
We wish to thank Dr Julian Gross for suggesting the experiments with the sporogenous mutants. pDd63, pDd56 and pDd26 cDNAs were obtained from Dr J. G. Williams, Ddrascl was from Dr R. Firtel and D19 and Dll were from Dr H. Lodish. This work was supported by a grant from the Medical Research Council.