Interacting signalling pathways regulating prestalk cell differentiation and movement during the morphogenesis of Dictyostelium

When their bacterial food source is depleted Dictyostelium cells undergo a remarkable transition, from unicellularity to multicellularity. The initially isolated cells aggregate together to form a simple, but precisely patterned structure, the culminant or fruiting body (Fig. 1). It consists of a mass of spores supported by a stalk, composed of dead vacuolated cells. About four-fifths of the cells that enter development differentiate to form spore cells while the remaining one fifth form stalk cells.

In common with higher organisms, development involves both cellular differentiation and morphogenetic cell movement and the questions that arise when considering the process are those which face all developmental biologists interested in pattern formation. Specifically, why do stalk cells become stalk cells and spore cells become spore cells and how do they come to occupy their final relative positions within the mature culminant?

One great boon to the analysis of morphogenesis is the existence of an intermediate developmental stage, the migratory slug. It can be used to investigate the signals that induce cellular differentiation and is of interest in its own right, as a multicellular organism formed by the co-aggregation of individual cells. The slug has a patterned structure that foreshadows that of the mature culminant but cells are not irreversibly committed to their fates. The precursors of the stalk cells, the prestalk cells, occupy the front one-fifth of the migratory slug. Prespore cells are located in the rear four fifths of the slug but at culmination the two cell types reverse their relative locations in a well orchestrated series of cell movements.

This review will concentrate almost entirely on the prestalk-stalk cell pathway, because it is the differentiation and movement of prestalk cells that shapes the Dictyostelium culminant. The key to understanding differentiation along this pathway is a small, diffusible molecule called DIF (Differentiation Inducing Factor).

DIF is a chlorinated hexaphenone which is synthesized during Dictysostelium development (Brookman et al., 1982; Morris et al., 1987; Town et al., 1976) and which will, in an in vitro assay, divert cells from spore to stalk cell differentiation (Kay and Jermyn, 1983). The ecmA and ecmB genes are rapidly induced by DIF and dependent upon it for their expression (Jermyn et al., 1987). They encode closely related, extracellular matrix proteins composed of tandem repeats of a cysteine-rich, 24 amino acid sequence (Williams et al., 1987; Ceccarelli et al., 1987).

At the slug stage the EcmA protein is localized to the slime sheath, the extracellular matrix that surrounds the slug and which is deposited as a trail marking the slug’s migration path along the substratum (McRobbie et al., 1988). In the mature culminant the EcmB protein forms part of the stalk tube, the rigid cellulose-conlaining matrix that surrounds mature stalk cells (McRobbie et al., 1988).

The ecmA and ecmB genes are the only known, definitive markers of prestalk and stalk cell differentiation and they have given several new insights into Dictyostelium morphogenesis. The first of these was the discovery of multiple prestalk cell types within the migratory slug.

When the promoters of the ecmA and ecmB genes were fused to a reporter gene and introduced into Dictyostelium cells, by transformation, a hitherto unsuspected degree of complexity in slug structure was revealed. Three kinds of prestalk cells were defined by differences in level of expression of the ecmA and ecmB markers (Jermyn et al., 1989). Prestalk A (pslA) cells express the ecmA gene al a high level, pstO cells express the ecmA gene at a lower level and pstAB cells express both the ecmA and ecmB genes (Jermyn et al., 1989; Jermyn anti Williams, 1991; Gaskell et al., 1992) In the newly formed slug. pslA cells occupy approximately the front one third of the prestalk region. pslO cells occupy the rear two thirds of the prestalk region and there is a small core of pstAB cells in the very front of the slug tip (Figs 2. 3).

These quantitative differences in the level of ecmA and ecmB gene expression raise two obvious questions: how is the difference in level of expression acheived anti, more fundamentally. why should there be these three cell types? The answer to the first question came from an analysis of the promoter of the ecmA gene. Fusion of sub-fragments of the ecmA promoter to a lacZ reporter shows it to contain at least two, apparently redundant elements that direct expression in pstO cells but not in pslA cells (Fig. 4 anti Early et al., 1993). This observation proves that pstA and pstO cells are discrete prestalk cell sub-types, which must differ in the efficiency with which they transcribe the ecmA gene. These two cell types presumably fulfil different functions within the slug and the ecmA promoter analysis yielded a clue as to the possible function of the pstO cells.

There are cells in the rear, prespore region of the slug that share many of the properties of the prestalk cells (Sternfeld and David. 1981; Devine. 1985) and which are therefore called anterior-like cells (ALC). The smallest sub-fragment of the ecmA promoter which retains the ability to be expressed in pstO cells, a fragment of 164 nucleotides, is also expressed strongly in ALC. This suggests that there may be a close relationship, perhaps even an identity, between these cell types. This notion is strengthened by analysis of cell movement within the slug.

Studies with vital dyes, which selectively stain prestalk cells, have shown there to be a considerable degree of physical interchange between the ALC and cells in the the prestalk region (Bonner. 1957; Francis and O’Day. 1971; Kakutani and Takeuchi. 1986). A novel technique of labelling living cells (Tomoaki Abe and Jeff Williams, unpublished results) indicates that the anterior band of pstO cells and the subset of ALC. which are able to utilise the pstO-specilic elements in the ecmA promoter, arc interChanging cells, so that pstO cells and ALC could be identical cell types that continually shuttle between the prestalk and prespore zones. Previous studies have suggested that the ALC may act as a kind of intermediate stale in the transdifferentiation of prestalk and prespore cells (Blaschke et al., 1986) and their ability to move between the prestalk and prespore zone may play some part in fulfilling this function.

During slug migration the core of pstAB cells is periodically discarded into the slime trail and, very quickly thereafter, they acquire the vacuolated appearance typical of stalk cells (Sternfeld, 1992). PstAB cells seem to form a kind of stalk primordium and al culmination the process of differentiation of pstA and pstO cells, into pstAB cells and thence into stalk cells, becomes continuous (Jermyn and Williams. 1991). Thus activation of the ecmA gene, to become a pstA cell or a pstO cell, marks a cell as a prestalk cell or ALC while the subsequent activation of the ecmB gene, to form a pstAB cell, appears to be the commitment step to stalk cell differentiation.

The ecmA gene has given valuable insights into the process of slug formation. Upon starvation, up to 100,000 cells move together in response to pulsatile cAMP signals that emanate from the centre of the aggregation territory. During the later stages of aggregation the cells form themselves into a mound, as they pile atop one another. The event that breaks the symmetry of the mound is the formation of the tip. a nipple shaped structure located at the apex of the mound (Fig. 1). The tip elongates until the hemispherical mound is transformed into a cylindrical structure, the first finger or standing slug. Under environmental conditions that are inappropriate for immediate culmination the first finger topples onto its side and migrates away. The slug is exquisitely thermotactic and phototactic and these sensitivities are believed to direct its movement to the upper reaches of the forest floor, where spore dispersal can occur more efficiently. If the conditions favour fruit formation, the first finger undergoes culmination in situ.

The tip at the mound stage forms the tip of the migratory slug and is composed of prestalk cells. In order to understand slug formation, therefore, it is essential to understand initial prestalk cell differentiation. There were two opposing views concerning tip formation, either that prestalk cells arise at the apex of the mound in response to a positionally localised signal or that prestalk cells arise al random positions within the aggregate and then migrate to the apex to form the tip.

Identification of the first cells to express the ecmA gene showed that prestalk cells differentiate at random positions within the aggregate and then move to the tip. i.e. that the slug is formed by a cell sorting mechanism (Williams et al., 1989). Analysis of a strain of Dictyostelium that overexpresses the extracellular form of cAMP phosphodiesterase further showed that the ecmA-expressing cells migrate to the apex of the aggregate in response to chemotactic cAMP signalling (Traynor et al., 1992). In such a .strain the rate at which pstA cells migrate to the tip is greatly retarded relative to normal development and if the aggregate is transferred to a substratum containing cAMP the pstA cells actually reverse their direction of movement and accumulate in the base.

If prestalk cells differentiate at random positions within the aggregate, how is the proportion of prestalk and prespore cells regulated? Analysis of the metabolism of DIF suggests that there may be a homeostasis mechanism in the aggregate, which limits the number of cells that differentiate as prestalk cells. DIF induces the enzyme responsible for its own degradation (DIF-1 dechlorinase) and prestalk cells are enriched in this enzyme (Insall et al., 1992; Kay et al., 1993). providing a feedback loop that could control the concentration of DIF within the aggregate.

At culmination the pstA cells and about half of the pstO cells enter the mouth of the stalk tube, activate expression of the ecmB gene and so become pstAB cells (Jermyn and Williams. 1991; Early et al., 1993). The other half of the pstO cells remain outside the stalk tube and form a structure termed the upper cup. that lies between the prestalk cells in the apical papilla and the prespore cell mass (Jermyn and Williams. 1991). These cells also activate expression of the ecmB gene but they do so using a different part of the promoter than is utilised by the cells that enter the stalk tube (Ceccarelli et al., 1992). For simplicity, and because they have the same final fate, we will henceforth in this review consider the pstA cells and those pstO cells that enter the stalk tube as a single population.

In contrast to initial pstA cell differentiation, which occurs in scattered cells within the aggregate, the formation of pstAB cells occurs in a positionally defined manner. Cells only a few diameters above the mouth of the tube do not express the ecmB gene while cells within the stalk tube express the gene at a high level (Fig. 5 and Jermyn and Williams. 1991). The ecmB gene is inactive or is expressed at only a low level in pstA cells during slug migration (Gaskell et al., 1992). At culmination the pstA cells differentiate further, firstly to become pstAB cells and then stalk cells, by a lifting of this repression. The structure of the ecmB promoter reflects this pattern of control.

The ecmB gene is under negative control that is lifted at culmination (Ceccarelli et al., 1991). One part of the promoter has the potential to be active in all prestalk cells, prior to their entry into the stalk tube (Fig. 6). This positively acting region in the ecmB promoter is, presumably, the point of action of the DIF signal transduction pathway. Downstream from this region are two repressor elements that prevent gene expression until prestalk cells have entered the stalk lube (Ceccarelli et al., 1991 and Harwood et al., 1993). Deletion of these elements effectively converts the ecmB gene into an ecmA gene, at least in so far as its behaviour at culmination is concerned (Ceccarelli et al., 1991).

Identification of the upstream signalling pathway that regulates the repressor came from analysis of the cAMP- dependent protein kinase.

At culmination intracellular cAMP levels rise rapidly (Abe and Yanagisawa, 1983: Merkle et al., 1984) and several separate pieces of recent evidence have combined to show that intracellular cAMP acts, via PKA. to induce stalk cell differentiation. The cAMP-depcndcnt protein kinase in Dictysotelium is a dimer composed of a catalytic (C) and a regulatory (R) subunit (de Gunzburg and Veron, 1982: Mutzel et al., 1987). As in mammalian cells, where the holoenzyme is an R₂C₂ tetramer, the RC complex is catalytically inactive when cAMP binds to the R subunit the RC complex dissociates and the C subunit becomes active.

When a mutant form of the R subunit (Rm) that acts as a dominant inhibitor of the C subunit of PKA is expressed under the control of the promoter of the ecmA gene, prestalk cells become arrested in their differentiation (Fig. 7 and Harwood et al.,1992). They do not express the eemli gene nor do they vacuolate and die. This suggests that, during normal development. PKA acts, either directly or as part of a kinase cascade, to modify the repressor so that il dissociates from the ecmB gene and from other genes involved in stalk cell differentiation (Fig. 6).

Since Dictyostelium ser/thre phospho-protein phosphatases seem not be developmentally regulated (Simon et al.,1992a), it is reasonable to believe that it is the control of PKA activity during development that sets the phosphorylation state of the putative repressor protein. Interestingly, characterisation of the two copies of the repressor sequence (consensus sequence TTGnCAA, where n is in one case two and in the other case four nucleotide, Harwood et al., 1993) shows there to be 110 similarity with the CRE sequences (cAMP Response Elements) which, in mammalian cells, bind the CREB protein (reviewed by Lee, 1991). Perhaps then this is a novel PKA signalling pathway.

In addition to its role in stalk cell differentiation PKA is also involved in controlling cell movement al culmination. During slug migration. pstA cells move around the long axis of the slug tip (Siegen and Weijer, 1992). but at culmination they move up to the apex and then change their direction of movement as they enter the stalk lube and move downwards towards the base (Raper and Fennell, 1952). Inactivation of PKA in pstA cells prevents their movement to the tip al culmination (Fig. 8 and Harwood et al., 1992), This suggests that PKA may play a role in the upward, directed movement of pstA cells but there is as yet no indication how the subsequent downward movement of pstAB cells is brought about.

The Dictyostelium fruiting body is a vastly simpler structure than a fly or a mouse but very similar principles are used in its construction. Furthermore, while extracellular signalling molecules such as DIF may have no direct counterpart in higher organsims, conserved, intracellular signalling components such as PKA play several essential roles in Dictyostelium morphogenesis.

As we have described, PKA is a regulator of stalk cell differentiation and is also known to be necessary for spore cell maturation (Anjard et al., 1992; Mann et al., 1992; Simon et al., 1992b) A role for PKA in transcription is not surprising given the many precedents from higher organsims (Lee, 1992). However, there has, to our knowledge, been no previous evidence for an involvement of PKA in morphogenetic cell movement. This observation, that cells in which PKA is inhibited are not competent to move to the entrance to the stalk tube, raises a number of interesting questions. Is PKA required for basal cell motility or is it involved in (he chemotactic sensing mechanism? Is it required for the transcription of a new cell component or does it act at the cytoplasmic level on pre-existing molecules?

These are major problems but they are approachable because Dictyostelium has become the organsim of choice for studying cell motility, and chemotaxis to cAMP during early development is understood in considerable detail. The challenge now is to obtain a similar understanding of cell movement during multicellular development.