Dictyostelium amoebae can differentiate into spores without cell-to-cell contact

In the slime mould Dictyostelium discoideum, differentiation of the two major terminal cell types produced in development (stalk and spore cells) is induced by a number of different cell interactions. During aggregation the amoebae generate and experience periodic extracellular pulses of cyclic-AMP in the nano-molar range (Shaffer, 1975). These are necessary both for directing the cells towards the cell or group of cells which are the source of the signals and for stimulating expression of particular aggregative genes (Darmon, Brachet & Pereira da Silva, 1975; Gerisch, Fromm, Huesgen & Wick, 1975). The effects of these signals on gene expression can be mimicked by the continuous application of much higher concentrations of cyclic-AMP (Klein, 1975; Sampson, Town & Gross, 1978) which also have the effect of blocking normal morphogenesis. By taking advantage of these effects of high cyclic-AMP concentrations we have devised in vitro systems in which both stalk and spore differentiation can occur Bonner, 1970; Town, Gross & Kay, 1976; Kay, Garrod & Tilly, 1978). Stalk differentiation requires cyclic-AMP and a low MW diffusible factor (DIF, Town et al. 1976; Town & Stanford, 1979) whereas spore formation requires cyclic-AMP (Kay et al. 1978; Kay, 1979) and some other cell interaction. This interaction appears to be of short range and so might be dependent for transmission on cell-to-cell contact.

The purpose of the present work was to determine whether spore cell differentiation does demand cell contact or not as well as to examine the patterning of stalk and spore cells differentiating within aggregates in vitro.

The sporogenous mutants HM18 (tsg-900, acr-900, cob-900, sci-907) and HM28 (tsg-901 and/or tsg-903, cyc-900, whi-900, sci-909) derive from V12M2. They were obtained from their immediate parents by N-methyl-N’-nitro-N-nitrosoguanidine mutagenesis and selection by detergent lysis for the ability to make spores on cyclic-AMP-containing agar (Town et al. 1976). Since in the selection conditions the wild type will make prespores but not spores, we consider that these mutants are affected in spore maturation (Kay et al. 1978). In normal developmental conditions HM18 usually makes a fruit having a central stalk with spores at the base and at the top; HM28 makes a fruit consisting of a mound of stalk and spore cells.

Cells were grown on SM-agar (per 1: Difco Bacto peptone 10 g, Difco yeast extract 1 g, glucose 10 g, MgSO₄.7H₂O 1 g, KH₂PO₄ 2·2 g, Na₂HPO₄ 1 g, agar 15 g) in association with Klebsiella aerogenes. Plates were harvested at the first sign of clearing and the slime-mould cells freed of bacteria by four centrifugal washes in KK₂ (16·6 mM-KH₂PO₄, 3·8 mM-K₂HPO₄, 2mM-MgSO₄, pH ≈ 6·1) and a last wash in either 5% Bonners salts (0·5 mM-NaCl, 0·5 mM-KC1, 0·1 mM-CaCl₂) or in NS (20 mM-NaCl, 20 mM-KCl, 1 mM-CaCl₂) as appropriate. The cells were then resuspended in more of the final wash solution and counted with a Coulter Counter; development was timed from this last resuspension.

The standard medium for spore induction consisted of 10 mM 2-(N-morpho-lino) ethanesulphonic acid (MES) Na⁺), 5 mM cyclic-AMP (Na⁺), NS, 200 μg/ml streptomycin SO₄, 15 μg/ml tetracycline (added from a stock of 7·5 mg/ml in ethanol) pH 6·2. 4 ml of this medium was normally used per petri dish (Sterilin 5 cm diameter bacteriological plastic, code no. 122). Washed cells at the densities indicated in the text were pipetted into each petri dish of medium, where they settled onto the plastic. In some experiments (Table 1) cells were plated on 2 ml of 1·5% Oxoid L28 agar made up in the above medium and again in 5 cm dishes. This agar was pre-washed where indicated in Table 1 with H₂0 and ethanol.

The plates were incubated in a moist atmosphere at 22°C to allow cell differentiation to occur. Stalk and spore cells were identified by phase-contrast microscopy at the end of the incubation period or as indicated in the filming section. Most spores formed at high density were oval shaped as in normal fruit, but at low density a considerable proportion of round refractile cells differentiated. These were resistant to detergent lysis and were stained by an anti-spore antibody so we counted them as spores.

Time-lapse movies were made at 1 frame per 20–30 sec on PanF film (Ilford) using an inverted phase-contrast microscope with a Bolex Hl6 reflex camera and Bolex-Wild MBF-B and C control units. Low-power films were made at a cell density of 2·5 × 10³ cells cm⁻² with a × 10 phase objective and a final magnification of × 30. High-power films were made using a × 20 phase objective (final magnification x 100 or × 200) and at a cell density of 2 × 10⁵ cells cm⁻². At this density the cells aggregated to form small balls in which it was impossible to distinguish many individual cells. It was however possible to follow cells in small aggregates by holding these flat against the bottom of the dish with cellophane held in a perspex stretcher; this procedure was still inadequate for large aggregates. In all the films the image of a clock was projected onto the film, so as to allow accurate timing of the events filmed. At 20–28 h of low-power filming the detergent Cemulsol was carefully added to the plate to a final concentration of 0·3% and filming contained for another hour or more. In this time all amoebae were lysed by the detergent but spores were completely unaffected.

Films were analysed using a Specto Motion Analysis Mk 3 projector. The optical resolution in the low-power films was such that fine processes projecting from the cell surface may have been invisible. To allow for this possibility (and since our interest is in cells that did not make contact) cells whose surfaces were separated by less than one cell diameter were considered to be in contact. In fact many of the cells which are listed in Table 2 as having made no cell contact did not even approach to within two cell diameters of another cell.

We have shown previously that amoebae of sporogenous strains of Dictyo-stelium can differentiate into spores, without normal morphogenesis, when they are plated on agar containing cyclic-AMP (the conditions of Table 1, line 1; Town et al. 1976; Kay et al. 1978). The efficiency of spore formation can be improved to that obtained in the earlier work by increasing the concentration of salts and by submerging the agar (Table 1, lines 2 and 3) A further improvement is obtained when the agar is omitted and the cells plated directly on plastic with an overlay of buffered salts plus cyclic-AMP (J Gross, personal communication; Table 1, line 4). It seems that agar contains an inhibitor of spore differentiation, which can be removed by washing (compare Table 1, lines 3 and 5).

With these improved conditions, the efficiency of spore and stalk induction is dependent on the density at which the cells are plated, being very inefficient at low density for both strain HM18 and strain HM28 (Fig. 1). At high cell density HM18 produces approximately equal numbers of stalk and spore cells but HM28 gives mainly spores. Indeed on submerged agar this strain produces more than 95% spore cells with most of the remainder being amoebae. This situation should be convenient for examining spore differentiation virtually in the absence of stalk cell differentiation.

Preliminary experiments showed that amoebae of strain HM28 moved less actively than those of HM18 in the filming conditions. We therefore used the former strain for these experiments so as to minimize collisions between cells and to increase the chances that any particular cell would remain in view throughout the film. Filming started at between 20 and 40 min of starvation and continued for about 24 h; spores normally formed after 10 to 20 h. Cell-to-cell contacts were rare and of brief duration in the period preceding spore formation; on average 1·6 contacts per cell in the first 11–12 h of development. Indeed some cells avoided all contact with other cells throughout the period filmed. Some of these cells became spores as judged by their morphology, lack of movement and resistance to detergent lysis (Table 2). This proves that in the period filmed cell contact is not essential for spore differentiation. The films also showed that cell division during development is not required for spore formation, since some spores were derived from amoebae that did not divide (Table 2).

In conditions suitable for normal development of the wild type, cells of strain HM18 produce aberrant, though clearly patterned, fruit. We wondered whether this ability to produce a pattern of differentiated cell types might persist in aggregates formed in vitro. Observation of many such aggregates shows that large coherent patterns of stalk and spore cells do not form when the cells are constrained as a monolayer. Given this we sought, without success, for rules operating on a smaller scale that might relate a cell’s fate to its position within an aggregate. For example we noted that clusters of one cell type do sometimes form but so do mixtures of the two and that stalk and spore cells can be derived from cells that have held internal or external positions in an aggregate (e.g. Fig. 2, cells 2,4, 5 and 6). Thus it seems that in the few hours preceding terminal differentiation the position of a cell bears little relationship to its eventual fate.

Developing cells in a solid tissue might interact with each other by two distinct sorts of mechanism: diffusion-mediated, in which diffusible signal molecules would be released by the cells and contact-mediated, in which the signals would remain cell bound. We are analysing the cell interactions required in vitro (Kay, Town & Gross, 1979) to trigger Dictyostelium amoebae into spore differentiation. In principle we can disrupt either type of interaction by sufficient dilution of the cells: in the one case the signal molecules would become too dilute to be effective and in the other, the cells would not make contact. By the criterion of dilution, spore formation by amoebae of sporogenous mutants in the presence of cyclic-AMP is stimulated by some cell interaction. In this paper we have been able to prove that amoebae can differentiate into spores without cell-to-cell contact (except possibly in the first 20 minutes of development, which we have been unable to film). Therefore the cell interactions necessary for spore induction are not contact-mediated and must instead be diffusion-mediated.

This conclusion prompted a renewed testing of the effects of conditioned medium, which we can now show will greatly stimulate spore differentiation at low cell density (Kay, unpublished observations). Tn earlier work (Kay et al. 1978) in which cells were plated at low density on cellophane overlaying ‘helper’ cells at high density, we found that spore differentiation in the low density population correlated strongly with cell-to-cell contact. Although this result apparently contradicts our present conclusions, in the original work less than 4% of the cells at low density became spores. Hence the effective concentration of conditioning factor may have been so low that it only exceeded the threshold concentration for induction very close to the cells which were its source, giving an apparent contact dependence of differentiation.

Many workers have shown that physical separation of cells during later development prevents further cell differentiation (Newell, Longlands & Sussman, 1971; Gregg, 1971; Takeuchi & Sakai, 1971). This result points to the existence of some sort of cell interaction that drives cell differentiation. But in these experiments the cells were both disaggregated and diluted and so it is not possible to decide whether disruption of a diffusion-or a contact-mediated interaction was responsible for the inhibition of development. Other workers have reported that certain post-aggregative gene products are not made in shaken cell suspensions where the cells are largely separate, despite being at a high density (Rickenberg, Tihon & Güzel, 1977; Okamoto & Takeuchi, 1976; Landfear & Lodish, 1980). Preliminary experiments suggest that this outcome could be due to other aspects of the conditions used by these workers, rather than to the absence of cell contact (M. Peacey and J. Gross, personal communication). On the other hand we cannot exclude the possibility that our sporogenous mutants lack a contact-dependent interaction normally present in the wild type. An analysis of the requirements for prespore cell differentiation in the wild type and a comparison with other sporogenous isolates (Wilcox & Sussman, 1978; Ishida, 1980) should help to resolve this matter.

Although slime-mould cells develop in the absence of exogenous nutrients a certain amount of cell division normally occurs (Bonner & Frascella, 1952) so that by the completion of morphogenesis there can be approximately a doubling in cell numbers (Zada-Hames & Ashworth, 1978). These divisions are found mainly during aggregation and at the slug stage where they are restricted to the prespore zone (Durston & Vork, 1978). In most other systems development is accompanied by extensive cell division and it has been suggested that division may somehow be essential for the expression of new developmental genes (Holtzer, Weintraub, Mayne & Mochan, 1972). This idea is intrinsically unlikely in Dictyostelium, given the small amount of cell division that does occur and the continuous changes in gene expression throughout development. Sussman & Sussman (1960) have shown that the mutant Fty-1 can fruit without cell division and Cappuccinelli, Fighetti & Rubino (1979) found that inhibitors of mitosis delay, but do not prevent development of the wild type. Likewise our results show that spores can differentiate without cell division. Thus with three different ways of perturbing development it is possible to obtain cell differentiation without cell division, making it very unlikely that normally the former is dependent on the latter.

We are most grateful to Julian Gross, Shuji Ishida and Jenny Brookman for discussions and improvements in the manuscript.