ABSTRACT
Extracellular cAMP induces expression of several classes of developmentally regulated genes in Dictyostelium. Four highly homologous surface cAMP receptors (cARs) were identified earlier, but involvement of specific cARs in gene regulation has not been clarified. Cells lacking the chemotactic receptor, cAR1, neither aggregate nor express developmentally regulated genes. Expression of aggregative genes is in wild-type cells induced by nanomolar cAMP pulses and repressed by persistent micromolar cAMP stimuli, which induce expression of prespore and prestalkenriched genes during the postaggregative stages of development. We show here that in cell lines carrying a cAR1 gene disruption, nanomolar pulses cannot induce aggregative gene expression. Remarkably, micromolar cAMP can induce expression of aggregative genes in car1− cells as well as expression of prespore and prestalk-enriched genes, and furthermore restores their ability to form normal slugs and fruiting bodies. These data indicate that cAR1 mediates aggregative but not postaggregative gene expression and morphogenesis, and suggest that after gene disruption, its function is partially taken over by a lower affinity receptor that is not subjected to desensitization. The absence of another early cAMP receptor, cAR3, does not affect development. However, in a car1−/car3− double mutant, cAMP stimulation cannot restore any developmental gene expression, indicating that cAR3 may have substituted for cAR1 in car1− cell lines.
INTRODUCTION
During Dictyostelium development extracellular cAMP induces chemotaxis and regulates expression of several classes of genes. In preaggregative cells, nanomolar cAMP pulses enhance expression of aggregative genes, coding for, e.g., surface cAMP receptors (cARs), and adhesive contact sites A (csA) (Darmon et al., 1975; Gerisch et al., 1975; Klein et al., 1988). After aggregation, micromolar cAMP levels induce the expression of prespore genes and of a subclass of prestalk genes, independent of stimulus modulation (Kay, 1982; Barklis and Lodish, 1983; Mehdy et al., 1983; Schaap and Van Driel, 1985). Studies using cAMP derivatives have indicated that the effects of cAMP on chemotaxis and gene expression are mediated by cARs (Van Haastert and Kien, 1983; Schaap and Van Driel, 1985; Oyama and Blumberg, 1986; Schaap et al., 1993), although additional activation of cAMP-dependent protein kinase is also required for spore and stalk cell differentiation (Harwood et al., 1992; Mann et al., 1992; Hopper et al., 1993).
At present, genes coding for four highly homologous cARs have been characterized; (Klein et al., 1988; Saxe, 1991; Saxe et al., 1991). cAR1 is expressed shortly after starvation and is responsible for virtually all cAMP binding activity at the aggregative phase of development. Expression of cAR1 is directly followed by expression of cAR3, which is at its optimum at the mound stage. cARs 2 and 4 are expressed during slug and fruiting body formation, and are only found in a subset of cells. In Dictyostelium, interaction of cAMP with cARs results in a transient activation of adenylyl cyclase (Roos et al., 1975; Dinauer et al., 1980), guanylyl cyclase (Mato et al., 1977; Würster et al., 1977) and phospholipase C (Europe-Finner and Newell, 1987; Van Haastert et al., 1989), as well as displacements of Ca2+, H+ and K+ ions (Milne and Coukell, 1991; Malchow et al., 1978; Aeckerle et al., 1985).
To investigate the role of cAR1 in signal transduction and development, cAR1 antisense cells have been constructed (Sun et al., 1990). These cells do not differentiate and show neither chemotaxis nor second messenger accumulation in response to cAMP. Since the four cARs are highly homologous, a consequence of the cAR1 antisense mutagenesis may be the loss of other cAR mRNAs besides cAR1. To overcome this, the cAR1 gene was disrupted by homologous recombination (Sun and Devreotes, 1991). The car1−cells display a similar phenotype to the antisense lines; they neither aggregate nor develop into fruiting bodies. However, it is at present not clear if the block in development is due to defects in signal production or in transduction of the signal to gene regulation. We investigated whether car1−cells can express developmentally regulated genes in response to cAMP stimulation and present evidence that cAR1 mediates aggregative, but not postaggregative gene induction. Cells carrying a double disruption of both the cAR1 and cAR3 genes are completely unresponsive to cAMP as an inducer of gene expression.
MATERIALS AND METHODS
Culture and incubation conditions
The Dictyostelium discoideum control cell lines AX3 and PJK1:AX3 as well as car1−cell line JS14, the car1−/car3− line RI-4 (Insall et al., 1994) and aca− cells (Pitt et al., 1992) were grown in HL-5 medium (Watts and Ashworth, 1970) and harvested at the late log phase of development. Cells were washed once with 10 mM Na/K phosphate buffer, pH 6.5 (PB) and subsequently shaken at 150 rpm in PB at 5×106cells/ml and 22°C. Cells were challenged with different regimes of stimulation with cAMP or adenosine 3′:5′-monophosphorothioate, Sp-isomer (cAMPS) (Boehringer, FRG) as indicated in the figure legends.
RNA isolation and analysis
Total cellular RNA was isolated from 2.5×107 cells as described by Nellen et al. (1987), size fractionated on 1.5% agarose gels containing 2.2 M formaldehyde and transferred to Gene Screen membranes. Northern transfers were hybridized to [32P]dATP-labeled DNA probes according to standard procedures and exposed to X-ray films. The optical density of specific mRNA bands was quantitated using an LKB Ultrascan densitometer.
Membrane preparation and western blotting
Membranes from car1− and AX3 control cell lines were prepared by the method of Klein et al. (1988) and dissolved in 100 μl SDS sample buffer. 30 μl aliquots of membrane preparation were size-fractionated on a 10% acrylamide gel, electroblotted onto nitrocellulose and probed with cAR3 specific antiserum. Antibody binding was visualised using an enhanced chemiluminescence (ECL) western blotting detection system (Amersham).
RESULTS
Induction of aggregative gene expression in car1− cells
During Dictyostelium development several classes of cAMPregulated genes are expressed, which vary in requirements for signal modulation. Following starvation, cAMP pulses in the nanomolar range strongly accelerate the expression of aggregative genes, which include cAR1 itself and the gene coding for contact sites A (csA), a glycoprotein mediating cell adhesion. We compared the effect of different concentrations of cAMP pulses on expression of the csA gene in wild-type and car1− cells. Fig. 1 shows that pulses of 30 nM cAMP at 6 minutes intervals optimally induce the synthesis of csA mRNA in the control cell line (AX3); at higher cAMP concentrations no further increase in csA mRNA levels is observed. In car1− cells, 30 nM pulses cannot induce csA gene expression, but increasingly higher cAMP concentrations induce a significant level of csA expression.
Fig. 2 represents pooled data from two experiments. These data emphasize that whereas csA expression in AX3 cells is as efficiently induced by 30, 100 or 300 nM cAMP pulses, only 300 nM pulses induce an appreciable level of csA expression in car1−cells. One possible explanation would be that low cAMP concentrations are amplified by relay in wild-type, but not car1− cells (Insall et al., 1994). However, 30 nM cAMP pulses can induce optimal csA expression in an adenylyl cyclase null mutant (aca−), which also cannot amplify the cAMP signal (Pitt et al., 1992). It therefore appears that the requirement for high concentrations in car1− cells is not a consequence of defective signal amplification, but of defective signal perception. In AX3 cells, a continuous micromolar cAMP stimulus cannot induce csA gene expression, most likely because persistent stimulation causes desensitization of this response. Quite remarkably, this stimulation regime is highly effective in inducing csA expression in car1− cells. This suggests that in car1− cells the response is mediated by a lower affinity receptor, which is not subjected to desensitization mechanisms.
Induction of postaggregative gene expression and development in car1− cells
After cells have entered the aggregation phase of development, micromolar cAMP concentrations can induce expression of prespore genes and of a class of genes that are more strongly expressed in prestalk cells. We compared expression of the two prestalk-enriched genes, D14 and CP2 (Barklish and Lodish, 1983; Pears et al., 1985), and prespore genes D19 and SP96 (Barklis and Lodish, 1983; Fosnaugh and Loomis, 1989) in car1− and AX3 cells. Fig. 3 shows that both in car1− and AX3 cells, expression is absent or barely detectable in the absence of exogenous cAMP. In both cell lines the prestalk-enriched genes are induced within 2 hours after addition of cAMP, whereas the prespore genes gain optimal expression levels after 4 to 6 hours of stimulation. No significant difference between gene induction in AX3 and car1− cells was observed.
After 8 hours of stimulation with 300 μM cAMP, cells have formed tight aggregates in suspension. When these aggregates are placed on agar, they form tips, slugs and fruiting bodies in a similar fashion to wild-type cells. The same results are obtained when cells are stimulated with a single 5 μM dose of the non-hydrolysable cAMP derivative, adenosine 3′:5′monophosphorothioate, Sp-isomer (cAMPS) (Fig. 4). The use of this derivative obviates the need for the nonphysiological, high concentrations of cAMP that are required to counteract degradation by phosphodiesterases.
To conclude, it appears that induction of postaggregative genes by micromolar cAMP concentrations occurs normally in car1− cells. Furthermore, once transcription of postaggregative genes has been induced in suspension, car1− cells have acquired full potency to complete all subsequent stages of development.
cAR3 expression in car1−cells
The data presented in Fig. 2 indicate that a lower affinity receptor may have taken over the function of cAR1 in aggregative gene expression. The most obvious candidate for a receptor substituting for cAR1 is cAR3, which is expressed somewhat later during development than cAR1. We first measured whether appreciable amounts of cAR3 protein are expressed during treatments inducing gene expression in car1− cells. AX3 cells and the car1− cell line JS14 were stimulated with 300 nM cAMP pulses or with a single dose of 5 μM cAMPS. A western blot of membrane proteins isolated during stimulation shows that both cAMP pulses and 5 μM cAMPS induce a detectable amount of cAR3 protein in car1− cells, although in both cases cAR3 expression remains lower than in AX3 cells (Fig. 5).
Gene expression and development in car1−/car3− cells
To investigate whether cAR3 can substitute for cAR1 in development and gene induction experiments, a double car1−/car3−cell line was constructed (Insall et al., 1994). Similar to car1− cells, car1−/car3− cells neither aggregate nor differentiate, when deposited on solid substratum. In a first attempt to test whether the developmental defect in the car1−/car3− cells is cell autonomous, we measured whether these cells can go through development in synergy with wild-type cells. 107 cells of either the G418 resistant control cell line pJK1:AX3, the car1− cell line or the car1−/car3− cell line were mixed with an equal amount of AX3 cells and allowed to develop. After 4 days, spore heads were picked into HL5 medium and serial dilutions were plated in duplicate. After allowing 2 days for spores to hatch, the medium in one duplicate set was replaced with medium supplemented with 20 μg/ml G418. 10 days later, after each spore had given rise to a colony of cells, the number of G418 resistant colonies was compared with the total number in the unselected plates. From the AX3-pJK1:AX3 mixture, 2400 out of 8700 (28%) spores were G418 resistant; out of 26000 spores from the car1−-AX3 mix, 250 spores (1%) were resistant and from the car1−/car3−-AX3 mix 0 out of 20000 spores were G418 resistant. The car1− cells already show very poor efficiency to synergize with wild-type cells as was also reported by Sun and Devreotes (1991), which is most likely due to their chemotactic defect. car1−/car3− cells cannot synergize at all, suggesting that they are completely insensitive to signals emitted by the wild-type cells.
We next measured the effects of cAMP on expression of aggregative, prestalk-enriched and prespore genes in car1−/ car3− cells. Fig. 6 shows that neither pulses of 30, 100 or 300 nM cAMP nor high doses of 300 μM cAMP can increase expression of the aggregative csA gene in car1−/car3− cells. In both AX3 and car1−/car3− cells a low level of csA mRNA is detectable without cAMP stimulation, which is probably induced by CMF, an 80×103Mr glycoprotein (Mann et al., 1988, 1989; Gomer et al., 1991).
Micromolar concentrations of cAMP effectively induce expression of the prespore gene D19 and the prestalk-enriched gene CP2 in AX3 cells, but no expression is observed in car1−/car3− cells (Fig. 7). These data show that disruption of both the cAR1 and the cAR3 genes completely obliterates cAMP-induced gene expression.
DISCUSSION
The surface cAMP receptor cAR1 has traditionally been associated with chemotaxis and chemotactic signaling. However, disruption of the cAR1 gene blocks cell differentiation as well as aggregation. We investigated whether the developmental defect is due to the fact that cells cannot produce extracellular cAMP required for gene induction, or whether cAR1 itself mediates gene induction. A signifant outcome of the present study is that postaggregative gene expression and fruiting body formation can be completely restored by stimulating car1− cells with exogenous cAMP.
car1− cells cannot express aggregative genes in response to cAMP pulses of 30 nM, which is optimal for gene induction in wild-type cells. Pulses of 300 nM cAMP induce a low level of expression, but very remarkably, a continuous micromolar cAMP stimulus is highly effective to induce aggregative gene expression in car1− cells, while inhibiting expression in wild-type cells. This strongly suggests that cAR1 mediates both activation and desensitization of pulse-induced gene expression. Apparently, in the absence of cAR1, a lower affinity cAR can restore the stimulatory but not the inhibitory component of this response. Similar results have been reported for adaptation of adenylylcyclase (ACA) activation (Pupillo et al., 1992). Whereas in wild-type cells a 5 minutes preincubation with 10 μM cAMP strongly reduces subsequent activation of ACA by GTPγS measured in lysates, preincubation with cAMP does not affect ACA activation in car1− cells, suggesting that car1− cells are defective in adaptation.
We investigated whether cAR3, the only presently known cAR that is expressed at the aggregation stage of development, has substituted for the function of cAR1. car1−/car3− double mutants show no cAMP-induced gene expression, either to nanomolar cAMP pulses or to micromolar cAMP concentrations and cannot be induced to develop by prolonged stimulation with cAMP, as is the case for car1− cells. Postaggregative genes also cannot be induced in car1−/car3− cells, which would seem to suggest that either cAR may mediate this response and the double mutation has blocked transduction by both the primary and the substituting transducer. However, this cannot be stated with certainty, since postaggregative gene expression in general requires achievement of a state of differentiation competence, which may not be reached in the car1−/car3− mutant. Involvement of cAR1 in postaggregative gene expression is unlikely, because the latter response requires high cAMP concentrations and is insensitive to adaptation.
The function of cAR3 in wild-type cells is thus far obscure; car3− cells show no aberrant phenotype and display normal expression of developmentally regulated genes. This suggests that either cAR3 has no obvious function in gene regulation and signaling, or its functions can be completely replaced by other receptors. Curiously, also the other two cARs, cAR2 and cAR4, which are expressed later in development in the prestalk region of slugs and fruiting bodies, have no obvious function in the regulation of cAMP-induced prespore or prestalkenriched genes. cAR2 and cAR4 null mutants appear to be defective in regulation of the DIF-induced prestalk genes and if anything, to overexpress prespore genes (Saxe et al., 1993). So at present, either none of the cloned cARs mediates prespore and prestalk-enriched gene expression, or groups of cAR subtypes show plasticity to replace each others function whenever required; single cAR gene disruptions may be insufficient to implicate functions of specific cARs, without careful analysis of the kinetic properties of the response.
ACKNOWLEDGEMENTS
We are grateful to Drs A. Noegel, J. G. Williams and W. F. Loomis for their kind gifts of csA, CP2 and SP96 DNA probes respectively, and to Dr R. L. Johnson for his gift of cAR3-specific antibody. This research was supported by GB-MW grant 900-546-077 from the Netherlands Organization for Scientific Research, by NIH grant GM34933 and by a SERC/NATO grant to R. H. I.