The effects of inhibition of protein glycosylation on the aggregation of Dictyostelium discoideum

The molecular basis of cellular communication in embryonic development in higher eukaryotes has been difficult to study because of the close packing and adhesion of the cells comprising the tissue. Aggregation of the cellular slime mould Dictyostelium discoideum, however, starts from a field of apparently identical amoebae (Bonner, 1944) and cell movements continue until the amoebae are associated into multicellular aggregates (Bonner, 1949 and 1963). It therefore offers an attractive system for the study of intercellular communication. The molecular basis of aggregation is now well understood (Loomis, 1975; Robertson & Grutsch, 1981) and the possibility has been raised that this system may be responsible for cellular communication and movement during the later multicellular stages of the organism’s development (Newell, 1978). Indeed, it has been a subject of conjecture that the mode of D. discoideum aggregation may be found among the as yet uncharacterized cellular communication systems of higher eukaryotes (Robertson & Grutsch, 1981; Gingle, 1977; Guillermet & Mandaron, 1980).

Aggregation of D. discoideum is initiated by certain amoebae secreting pulses of cyclic AMP and the surrounding amoebae responding by positive chemotaxis and the secretion themselves of cyclic AMP pulses (Konijn, 1972). This signal relay system entrains amoebae over a large area into each aggregation centre into which cells move in synchronous waves or spirals (Gerisch, 1968; Alcantara & Monk, 1974; Newell, 1978). The enzymes which control the cyclic AMP signal are developmentally regulated. Prior to aggregation, adenyl cyclase activity is present (Klein & Darmon, 1977; Darmon & Klein, 1978) and cyclic AMP phosphodiesterase and its inhibitor protein are secreted by the amoebae (Pann-backer & Bravard, 1972; Gerisch et al. 1972). Parallel with the acquisition of the ability to chemotact toward cyclic AMP pulses at the onset of aggregation there is a rapid increase in cell surface cyclic AMP phosphodiesterase (Malchow et al. 1972). As aggregation proceeds the synchronous moving waves of amoebae break up into ‘streams’ (Bonner, 1971), stream formation being mediated through end-to-end cell contact via cell surface contact sites A which also increase in activity during this period (Beug et al. 1973). There is evidence that the cell surface aggregation functions (cyclic AMP receptors, cyclic AMP phosphodiesterase and contact sites A) are synthesized de novo during aggregation (Sampson et al. 1978) as are a large number of secreted enzymes which increase in activity during early development (Loomis, 1975).

Cell surface cyclic AMP phosphodiesterase (Eittle & Gerisch, 1977), the secreted form of the enzyme (Orlow et al. 1981), its inhibitor (Franke & Kessin, 1981) and contact sites A (Muller & Gerisch, 1978) are all glycoproteins as are many membrane components which change during development (Toda et al. 1980). In order to study the role of de novo protein glycosylation during aggregation in D. discoideum we have used the antibiotic tunicamycin which inhibits the first step in the glycosylation of proteins via N-asparaginyl links (Takatsuki et al. 1971; Tkacz & Lampen, 1975). We report that inhibition of protein glycosylation by tunicamycin (TM) prevented the appearance of cyclic AMP phosphodiesterase and contact sites A but that cell surface cyclic AMP receptor activity developed normally. TM caused a reversible block in development but small aggregates did form in the presence of TM at the normal time of aggregation. TM prevented the development of the ability of amoebae to chemotact toward cyclic AMP but the addition of cyclic AMP phosphodiesterase to TM-inhibited cells restored this activity. The implications of these results are discussed.

Tunicamycin was the kind gift of Drs Harper and Herman, Biotechnology Department, Glaxo Research Limited, Greenford, Middlesex, England. It was stored either desiccated as a powder or as a 1–5 μgml^-1 solution in DMSO at −20 °C. DMSO at concentrations up to 0·44 M had no observed effect on development of Dictyostelium or on any of the cellular or biochemical functions measured in this study. Radiochemicals were obtained from Radiochemicals International, Amersham, England. All other chemicals and reagents were of the highest purity available.

Dictyostelium discoideum strain AX2 spores (a kind gift from Dr M. North) were stored on silica gel at 4 °C. Germination and growth were carried out in aerated suspensions of HL5 growth medium (1 %(w/v) peptone, 0·5 % (w/v) yeast extract, 20 mM-phosphate buffer, pH 6·9) containing 86 mM-glucose (Watts & Ashworth, 1970). The media contained 200 μg ml^-1 streptomycin and 5 μgml^-1 tetracycline. Flasks, at least four times the volume of suspension, were shaken at 175 r.p.m. on a New Brunswick G2 gyrotary shaker. Amoebae were harvested during log phase (2–4 × 10⁶ cells ml^-1) by centrifugation at 800g for 10 min at 22°C, washed three times in KK₂ buffer (20mM-potassium phosphate, pH6·2) containing 0·5 mM-CaCl₂ and 2 mM-MgSO4, and finally resuspended in this same buffer at a concentration of 1 × 10⁷ml^-1. Washed amoebae were shaken at 125 r.p.m. and 22°C in chromic acid washed, siliconized flasks and the onset of development (t₀) was measured from the time of final resuspension of the washed amoebae. To observe aggregation, washed amoebae were plated onto 1·2% agarose in KK2 buffer at a density of 5 × 10⁵ amoebae cm^-1.55 mm tissue culture dishes were used, each containing 3 ml agarose. Photomicrography was performed using a Wild dissecting microscope fitted with a Nikon cameraback.

The small population test was used (Konijn et al. 1969). 1 μl drops of amoebae (2 × 10⁴ cells) were placed on the surface of 0-5 % agarose in KK2 buffer at a distance of 0·5–1·5 mm from an equal volume of cyclic AMP of the stated concentration. Plates were incubated at 22 °C for 3h. The rate of chemotaxis was then measured in mmh^-1 using an ocular micrometer.

5 μCi L(4,5-³H)leucine (40–60 Ci-mmole¹) and 2 μCi D(U-¹⁴C)mannose (200-300 mCi mole^-1) were added to 10 ml of developing amoebae at t2. Tunicamycin in DMSO, DMSO alone or KK2 was added at t3 then duplicate 0·5 ml samples were removed at t4 and thereafter as stated. Nonidet P40 was added to each sample to a final concentration of 0·2%, samples were mixed vigorously for 5 sec and then an equal volume of 10 % TCA containing 1 mg ml^-1 each of leucine and mannose was added. TCA precipitates were collected onto Whatman GF/C filters and washed in 5 % TCA containing Imgml^-1 each of leucine and mannose. The filters were dried and counted in PPT scintillant (15gPPO, 18·8 mg dimethyl POPOP per 2·5 litre toluene). The relative rates of incorporation of [¹⁴C]mannose and [³H]leucine into TCA precipitates were determined by double isotope analysis (Neame & Homewood, 1974) and were used respectively as a measure of protein glycosylation and protein synthesis.

Total cell adhesion was measured by resuspending amoebae at a concentration of 10⁶ cells per ml in KK2 buffer. Cell clumps were disaggregated by three passages through a 25G × 5/8′ needle. Total cell number was determined by haemocytometer counting. The cell suspension was then shaken at 250r.p.m. for 30 min on a New Brunswick gyrotary shaker (model G2, 3/4′ circular orbit). The number of single cells remaining after this period was then determined. Total cell adhesion was then calculated by subtracting the number of single cells left after 30 min from the total number of cells and expressing this figure as a percent of total cells. Contact sites A were measured by performing the above assay in KK₂ lacking divalent cations and in the presence of 10mM-EDTA (Rosen et al. 1973). A subtraction of the adhesion due to contact sites A from total cell adhesion provided an estimate of adhesion due to contact sites B (Marin et al. 1980). This estimate is only valid at low contact sites A activity. Cyclic AMP phosphodiesterase was assayed as described (Henderson, 1975). The cell-associated enzyme activity was taken as a measure of the enzyme on the cell surface (Malchow et al. 1972) and was expressed as milli International units (μmoles cyclic AMP hydrolysed per minute per mg protein at 37°C). The activity of the extracellular enzyme was expressed as μmoles cyclic AMP hydrolysed per minute per ml of extracellular fluid at 37 °C. Cell surface cyclic AMP receptors were assayed as described (Green & Newell, 1975).

Suspensions of developing amoebae were perfused with a solution of cyclic AMP which was delivered by an LKB microperspex pump (model 2132) at the rate of 50 μl per minute such that the concentration of cyclic AMP following the first drop was 0·2 μM. This treatment maximizes the secretion of cyclic AMP phosphodiesterase and represses the production of the protein inhibitor (Gerisch, 1979). After fifteen hours of development the suspension was cleared of cells by centrifugation at 800g for 10 min at 22 °C. The cell pellet was resuspended in 50 mM-tris-HCl, pH 7·5 (T_7·5 Buffer) and the cells lysed by the addition of NP40 to a final concentration of 0·2 % (v/v). This lysate served as the source of cellular proteins. The extracellular fluid was cleared of debris by centrifugation at 10000g for 10 min at 4 °C. Protein was precipitated by the addition of ammonium sulphate to 90 % saturation. The precipitate was collected, resuspended in T7.5 buffer and dialysed against 200 volumes of the same buffer. This solution provided the source for the analysis of extracellular proteins and, after passage through a column of concanavalin A Sepharose as previously described (Kessin et al. 1979), provided a partially purified extract of cyclic AMP phosphodiesterase, 30i.u. of cyclic AMP phosphodisterase were obtained from 100 ml of an amoebal suspension.

Protein determinations were carried out by the method of Lowry (Lowry et al. 1951) and the secreted polypeptides were analysed using sodium dodecyl sulphate polyacrylamide gel electrophoresis (Laemmli, 1970) followed by silver staining of the separated proteins (Oakley et al. 1980).

We studied the effect of tunicamycin (TM) on the incorporation of [¹⁴C]mannose and [³H]leucine into TCA-precipitable material in order to monitor its effects on both protein synthesis and glycosylation. Mannose was chosen because it is the major sugar constituent of N-linked oligosaccharides, the synthesis of which is inhibited by TM. [¹⁴C]mannose and [³H]leucine were added to cell suspensions 2h after the onset of development (t2). TM was then added at t3 to final concentrations ranging from 1μgml^-1 to 40μgml^-1. The amounts of radioisotopes incorporated were then determined. The results (Table 1) show that TM up to a concentration of 40μ gml^-1 had no significant effect on the incorporation of [³H]leucine into TCA-precipitable material. However, increasing concentrations of the antibiotic markedly inhibited the incorporation of [¹⁴C]mannose. A concentration of 5μgml^-1 inhibited the incorporation of [¹⁴C]mannose by 50 %, near maximal inhibition being obtained at concentrations of 10μgml^-1 or greater. We take these data to show that TM inhibits protein glycosylation in Dictyostelium but has little or no effect on total protein synthesis. Thus, it appears that TM affects D. discoideum in exactly the same way that it effects other eukaryotic cells. The effect of TM on the time course of protein synthesis and glycosylation is shown in Fig. 1. There was no appreciable effect of TM on the incorporation of [³H]leucine (Fig. 1A), but, under the conditions of the assay, TM reduced the rate of [¹⁴C]mannose incorporation to 4 % of the control rate. This low level incorporation of [¹⁴C]mannose into TCA-precipitable material in the presence of TM may reflect the general metabolism of mannose or the incorporation of the sugar into oligosaccharide chains that are attached to serine or threonine residues in proteins via O-glycosidic bonds. Such incorporation would be expected to be TM insensitive. Treatment of TCA precipitates with 25 % 2-methoxyethanol (v/v) to solubilize large molecular weight carbohydrate did not reduce this low level, TM insensitive [¹⁴C]mannose incorporation (data not shown).

When TM (10–40 μgml^-1) was added to suspensions of developing amoebae between t₀ and t₄, and the cells then plated on agarose containing an equivalent concentration of the antibiotic, aggregation was severely impaired and all subsequent development was blocked. Amoebae failed to aggregate fully, instead, small aggregates began to form at about the same time as control amoebae aggregated (Fig. 2). The formation of these small aggregates was very rapid, proceeding from initiation to completion in 30 to 40 min (Fig. 3). These small aggregates often resembled an interconnected array of flat areas of amoebae (e.g. Fig. 2B and Fig. 4A) but further development did not occur. However, if amoebae were exposed in suspension to TM from to to tw then plated as 2 drops on agarose lacking the antibiotic, normal-looking fruiting bodies were observed 48 h later, showing that TM does not cause an irreversible block in development. TM also affected aggregation and subsequent development at concentrations below that required for complete inhibition of protein glycosylation (i.e. 10μgml^-1; Table 1). The small, abnormal aggregates were observed in the presence of TM at concentrations ranging from 40 ng ml^-1 to 40 μgml^-1. Their formation began at the same time as aggregation of control amoebae and was always complete within 60 min, although cell movements were still observed after this time and the small aggregates very often fused together into an interconnected array. This process was usually complete by t₈ and t₂₂ was most marked at TM concentrations below 5μgml^-1. No further development was observed within the flat arrays at TM concentrations above 1·25 μgml^-1 but at lower antibiotic concentrations mounds of cells were observed at t₂₂ which were larger and more numerous at the lowest TM concentrations studied (Figs 4A, B and C). These mounds had developed between t₈ and t₂₂ by apparently drawing in amoebae from the surrounding flat arrays but they did not develop further. At the lowest TM concentrations (40 ng ml^-1 to 0 ·16 μg ml^-1) nearly all the amoebae present initially in the flat arrays had collected into mounds by t₂₂ and ‘finger-like’ projections could be seen protruding from these mounds (Fig. 4D). These structures failed to develop further. Thus it appears that TM blocked development at all the concentrations studied but that the precise point of the block depended on the concentration of TM.

If TM (final concentration of 10 μgml^-1) was aded to amoebae developing in suspension they subsequently failed to acquire cell-associated cyclic AMP phosphodisterase and contact sites A activity (Fig 5A and 5B, respectively) but they did acquire near normal levels of cell surface cyclic AMP receptors (Fig. 5C) Amoebae developing on agarose containing TM also failed to accumulate cell-associated cyclic AMP phosphodiesterase and contact sites A (data not shown). Addition of TM to amoebae during the normal developmental increase in cyclic AMP phosphodiesterase and contact sites A activity prevented any further accumulation of these two activities (data not shown), indicating that the antibiotic has an almost immediate inhibitory effect on their expression.

We consistently observed a slight decrease in total cell adhesion during the first three hours of development which then increased as contact sites A (csA) activity increased (Fig. 6). Since the level of csA is very low during the first three hours, this decrease must be due to a slight reduction in contact site B (csB) mediated adhesion. Our estimate of csB activity is only valid at low csA activity therefore we are unable to plot csB activity at later stages of development, however, csB activity has been reported to remain fairly constant throughout development Beug et al. 1973). Amoebae incubated in the presence of 10 jugml^-1 TM from t₀ lost all cohesion by t₃ (Fig. 6B), the rate of loss being more rapid in the presence of TM than the early loss of adhesion of control amoebae (compare Fig. 6A and 6B). Thus de novo protein glycosylation would appear to be necessary not only for the developmental appearance of csA but also for the maintenance of csB activity throughout development.

Figure 7 shows the effect of TM on the ability of amoebae to chemotact toward cyclic AMP as measured by the ‘small population test’ (Konijn et al. 1969). Amoebae were allowed to develop from to in suspension in either the presence or absence of TM (10 μgml^-1). At t4 amoebae were removed from suspension and deposited on chemotaxis plates as described in the “Materials and Methods” section. Plates were examined 3h later for signs of chemotaxis. All chemotaxis tests for TM-treated amoebae were performed on agar containing 10μgml^-1 TM. Control amoebae responded to 50 μM cyclic AMP by positive chemotaxis at a rate of 0·73 mm h^-1 over the 3 h period (Fig. 7A). TM-treated amoebae failed to chemotact significantly toward any cyclic AMP concentration in the range 50μM to 50 mM. However, amoebae were consistently observed to respond to cyclic AMP in the range 0·1 mM to 0·5 mM by moving out slightly from the original confines of the droplet toward the cyclic AMP source (Fig. 7C).

The effect of exogenous cyclic AMP phosphodisterase (PDE) on chemotaxis was also studied. Approximately 4 milli i.u. of partially purified extracellular PDE was added to the amoebal droplet used in the chemotactic tests. In the presence of PDE, the threshold cyclic AMP concentration below which control amoebae would not chemotact was raised approximately 100-fold (Fig. 7B). Addition of the enzyme to TM-treated cells resulted in amoebal chemotaxis toward cyclic AMP concentrations ranging from 0·5 mM to 50 mM (Fig. 7D). The chemotactic rates of control and TM-treated amoebae toward a range of cyclic AMP concentrations under these different conditions are summarized in Table 2. Comparison of Figs 7C and 7D also illustrates that in the presence of PDE TM-treated cells that are left within the confines of the original droplet formed into very compact, small aggregates instead of the usual rather amorphous field of small aggregates. This same difference in morphology was observed when TM-treated cells were placed on agar in the absence of a droplet of cyclic AMP (data not shown).

Amoebae were allowed to develop in suspension for 15 h either in the presence or absence of tunicamycin. Developing cells were perfused with cyclic AMP throughout the 15 h period in order to maximize the production of extracellular cyclic AMP phosphodiesterase (see Materials and Methods). At tis cellular and secreted proteins were processed for analysis. TM-treated cells contained 10 % more protein than controls but secreted 50 % less protein than controls (data not shown). We were unable to detect any extracellular cyclic AMP phosphodiesterase activity in TM-treated cells, even after the removal of any unhydrolysed cyclic AMP from the extracellular fluid. Analysis of secreted proteins by sodium dodecyl sulphate polyacrylamide gel electrophoresis showed that TM-treated cells failed to secrete three major polypeptides under the conditions used (Fig. 8). These polypeptides had apparent relative molecular masses of 65 000 (p65), 52 000 (p52) and 50 000 (p50). No polypeptides were observed in the secreted extracts of TM-treated amoebae which were not also present in the secreted extracts of controls. There is evidence to show that p52 and p50 are closely related glycoprotein subunits of the extracellular cyclic AMP phosphodiesterase (C. J. McDonald and J. Sampson, manuscript in preparation).

In this paper we report that during the aggregation phase of development in Dictyostelium discoideum tunicamycin (TM) at concentrations above 10 μgml^-1 inhibited protein glycosylation but had little or no effect on protein synthesis. This inhibition of protein glycosylation was accompanied by a reversible block in development. These data substantiate a preliminary report that TM blocked the development of D. discoideum (Toda et al. 1980). Two recent studies have shown that TM inhibited protein glycosylation in D. discoideum although as concentrations of only 3 μg ml^-1 (Lam & Siu, 1982) and 5 μgml^-1 (Yamada et al. 1982) were used, complete inhibition was not achieved. In all other eukaryotes studied so far, TM specifically inhibited the first step of the N-asparaginyl-linked pathway of protein glycosylation (Lennarz, 1980). It is therefore reasonable to assume that TM inhibition of protein glycosylation in D. discoideum occurs via a similar mechanism.

The effect of TM on aggregation was very much concentration dependent. At 10 μgml^-1, which caused maximal inhibition of protein glycosylation, aggregation was severely impaired and only very small aggregates developed; all further development was blocked. At TM concentrations below lO/igml^-1 protein glycosylation was incompletely inhibited and the small aggregates of cells were observed to collect together into larger interconnected arrays of cells. At TM concentrations below 1·25 ptgml^-1 mounds of cells developed within these interconnected arrays and at the lowest antibiotic concentrations tested (40 ng ml”¹) multiple ‘finger-like’ projections developed from these mounds by t22, but development did not progress further. The morphology of aggregation of amoebae from the flat arrays into mounds and ‘finger-like’ structures at sub-inhibitory concentrations of TM may be grossly aberrant, but the phenomenon could represent a slower version of normal aggregation caused by the slower accumulation of glycoproteins necessary for the process. Defects in the process would arise due to glycoproteins not reaching the required concentration for the formation of normal structures and normal levels of enzyme activity. We suggest that the small aggregates which form regardless of the concentration of TM may represent a transitional stage in aggregation which is blocked from further development by the absence of specific glycoproteins which are crucial for normal development.

TM prevented the acquisition of two glycoproteins implicated in normal development:- cyclic AMP phosphodiesterase and contact sites A. Indeed, the small aggregates formed in the presence of TM were reminiscent of those formed by aggregation-deficient mutants which lack cyclic AMP phosphodiesterase (Riedel et al. 1973). Cyclic AMP phosphodiesterase activity has been shown to be required for normal aggregation: an aggregateless mutant deficient in the enzyme, HPX235, can aggregate when supplied with an exogenous source of cyclic AMP phosphodiesterase (Darmon et al. 1978), and aggregation can be blocked by a specific neutralizing anti-cyclic AMP phosphodiesterase antibody (Goidl et al. 1972). Amoebae acquired cyclic AMP receptors normally in the presence of TM and when PDE was added to them they regained the ability to chemotact toward cyclic AMP, formed more compact aggregates but did not develop further. We therefore suggest that the primary block in aggregation induced by TM treatment is to prevent the synthesis of active PDE and that the subsequent block in development is perhaps caused by either a defect in regulation of the enzyme’s activity or by the lack of cell cohesion due to the inhibition of the appearance of contact sites A. The formation of small aggregates which we observed on agarose containing TM may be explained by the diffusion of secreted cyclic AMP into the agarose thus preventing the immediate saturation of cyclic AMP receptors and allowing amoebae to respond to a limited number of chemotactic signals. Similar small aggregates were not observed when amoebae were spread on nitrocellulose filters (Lam & Siu, 1982) although it is interesting that in that study amoebae were observed to ‘orientate’ pseudopodia towards cyclic AMP in chemotactic tests.

The normal acquisition of cyclic AMP receptor activity in the presence of TM implies that de novo protein glycosylation is not necessary for the proper expression of its function. This does not rule out a regulatory role for preexisting glycoproteins such as contact sites B (Marin et al. 1980) but as cell cohesion rapidly decreased during early development in the presence of TM, if such a role existed, one might expect at least a slower accumulation of receptor activity and a reduced ability to chemotact toward cyclic AMP in the presence of exogenous PDE. The results also suggest that the cyclic AMP receptor is not a glycoprotein although it has to be borne in mind that certain glycoproteins still retain biological activity in their non-glycosylated states (Speake et al. 1981; Onishi et al. 1979). The reversibility of the inhibition of development by TM may be useful in studies of the reaggregation of cells rendered non-cohesive by the TM inhibition of expression of cohesion-mediating molecules both in Dictyostelium and in other cell types.

We would like to thank the SERC for financial support in the form of a project grant and a studentship to C.J.M. We also thank Mrs Jane Drew for her excellent technical assistance.

 
• cyclic AMP

  adenosine 3′ 5′ cyclic monophosphate

 
• dimethyl POPOP

  1.4-bis-[′2-(4-Methyl-5-phenyloxazolyl)-3]-Benzene

 
• DMSO

  Dimethylsulphoxide

 
• i.u.

  International Unit of Enzyme Activity

 
• NP 40

  The non-ionic detergent Nonidet P40

 
• PPO

  2,5-diphenlyoxazole

 
• SDS-PAGE

  sodium dodecyl sulphate polyacrylamide gel electrophoresis

 
• TCA

  trichloroacetic acid

 
• TM

  Tunicamycin

 
• t_x

  time in hours (x) after the onset of development

  Enzymes quoted

 
• PDE

  cyclic AMP phosphodiesterase [E.C.3.1.4.17]