G-protein-linked cAMP receptors play an essential role in Dictyostelium development. The cAMP receptors are proposed to have seven transmembrane domains and a cytoplasmic C-terminal region. Overexpression of the receptor in cells, when the endogenous receptor is not present, results in a 10- to 50-fold increase in cAMPbinding sites. Antisense cell lines, which lack cAMP receptors, do not enter the developmental program. Ligand-induced phosphorylation is proposed to occur on serine and threonine residues in the receptor C-terminus. The kinetics of receptor phosphorylation and dephosphorylation correlate closely with the shift of receptor mobility and the adaptation of several cAMP-induced responses. Two α-subunits, G-α-1 and G-α-2, have been cloned and specific antisera developed against each. Both subunits are expressed as multiple RNAs with different developmental time courses. The mutant Frigid A has a functional defect in G-α-2 which prevents it from entering development. We propose that G-protein-linked receptor systems will be a major component in the development of many organisms.

Dictyostelium is an elegant system for studying the role of cell-cell signaling in development. During vegetative growth, this organism exists as single amoebae that phagocytose bacteria. However, upon starvation, aggregation centers begin to secrete cAMP spontaneously at six minute intervals. Concentric or spiral waves of cAMP propagate from the centers and direct the formation of multicellular aggregates (Tomchik and Devreotes, 1981). As aggregates form ‘slugs’, cell-type-specific differentiation occurs and a pattern appears. Prestalk cells are found in the anterior region while prespore cells arise from the posterior of the cell mass (see chapter by Williams el al.). The prestalk cells become vacuolated to form a stalk, which holds a sporehead aloft in the fruiting body (Gerisch, 1987).

Aggregation is organized by closely coordinating cellcell signaling and chemotaxis. As a wave of cAMP approaches, randomly oriented cells become polarized and move up the chemical gradient. As they move, the cells secrete additional cAMP which relays the signal outward. Cells continue to respond until the peak of the cAMP wave passes. They do not react to cAMP on the declining edge of the wave since they have become adapted. Hence secretion ceases and migration becomes random. During the interlude between cAMP stimuli, the system deadapts in preparation for the next wave. Approximately 30 waves are needed to complete aggregation (Devreotes, 1983).

Extracellular cAMP also plays a critical role in gene regulation. Expression of early genes such as the surface cAMP receptor, membrane-bound phosphodiesterase, and CSA (a cell adhesion protein) are induced by intermittent, but inhibited by persistent, application of cAMP (Figure 1). Transcripts of these early genes begin to accumulate within two hours and peak between six to eight hours of development (Mann and Firtel, 1987). In contrast, prestalk and prespore gene expression require continual, rather than intermittent, stimulation by cAMP. In addition, the morphogen DIF (Differentiation-Inducing Factor, see chapter by Kay et al.) is needed for prestalk gene expression in the tip where it also inhibits the induction of prespore genes (Schaap, 1983). Most prestalk genes are induced late in aggregation, while prespore gene expression begins somewhat later (Gerisch, 1987, see chapter by Williams et al.).

Fig. 1.

Gene expression is controlled by cAMP stimulation. Many early genes are induced by intermittent stimulation by cAMP. In later development, prestalk gene markers are induced by persistent exposure to cAMP with prespore expression beginning a few hours later.

Fig. 1.

Gene expression is controlled by cAMP stimulation. Many early genes are induced by intermittent stimulation by cAMP. In later development, prestalk gene markers are induced by persistent exposure to cAMP with prespore expression beginning a few hours later.

The molecular components involved in this signaling system include cAMP receptors, G-proteins, adenylate cyclase, and phosphodiesterase (Figure 2). These developmentally regulated proteins allow the establishment of the cAMP oscillator. Exogenous cAMP binds to surface membrane receptors which activate G-proteins and stimulate the formation of intracellular cAMP by adenylate cyclase. Each cell secretes cAMP which activates its own receptors as well as those of nearby cells. After several minutes of stimulation, ligand occupancy triggers desensitization (Janssens and Van Haastert, 1987). Removal of cAMP by extracellular and cell surface phosphodiesterase allows the system to resensitize and return to its prestimulus state (Gerisch, 1987).

Fig. 2.

The cAMP oscillator controls development. cAMP binds to cell surface receptors to activate adenylate cyclase and cause secretion of cAMP. Frigid A, which lacks a functional G cr-subunit, is defective in chemotaxis and induction of gene expression. In Synag 7, receptors are unable to couple to adenylate cyclase.

Fig. 2.

The cAMP oscillator controls development. cAMP binds to cell surface receptors to activate adenylate cyclase and cause secretion of cAMP. Frigid A, which lacks a functional G cr-subunit, is defective in chemotaxis and induction of gene expression. In Synag 7, receptors are unable to couple to adenylate cyclase.

Studies in our laboratory have focused on defining and characterizing the components involved in signal transduction. Recently, we have cloned surface cAMP receptors and proposed a model of their structure. We have also investigated ligand-induced phosphorylation. Two o’- and one β-subunit of G-proteins have been cloned as well. Mutation and overexpression of these genes have helped to establish their essential roles in development.

A structural model of the cAMP receptor was derived from its primary sequence (Figure 3). It is postulated to span the membrane seven times with the N-terminus facing the extracellular and C-terminal domain facing the cytoplasmic side of the membrane (Klein et al. 1988). The model is topologically similar to other G-protein linked receptors such as rhodopsin (Hargrave, 1986) and the βadrenergic receptor (Dolhman et al. 1987). Although the A-terminus has a potential site for A’-linked glycosylation, there is no biochemical evidence that the protein is glycosylated. The C- terminal region contains multiple stretches of serine residues which are the proposed sites of ligand-induced phosphorylation.

Fig. 3.

The cAMP receptor consists of seven transmembrane domains with the iV-terminus facing the extracellular and the C-terminus facing the cytoplasmic side of the membrane. Multiple stretches of serine residues in the C-terminal region are the proposed sites of phosphorylation.

Fig. 3.

The cAMP receptor consists of seven transmembrane domains with the iV-terminus facing the extracellular and the C-terminus facing the cytoplasmic side of the membrane. Multiple stretches of serine residues in the C-terminal region are the proposed sites of phosphorylation.

Multiple criteria demonstrate that this cDNA encodes the cAMP receptor. Antisera against the purified receptor identify several fusion proteins derived from the cDNA; RNA blot analysis, using the receptor cDNA as a probe, identifies a 2 kb message that is developmentally regulated in the same manner as the receptor protein; complementary receptor RNA specifically hybrid arrests in vitro translation of a 37 kD protein that is immunoprecipitated by receptor antiserum (Klein et al. 1988). Most conclusively, vegetative cells, which do not contain endogenous cAMP receptors, acquire cAMP binding sites when transformed with the cDNA. Transformed cells have seven to fifty times the amount of cAMP-binding compared to wildtype (1–8×105 vs 1 ·5× 104 sites per cell, Johnson et al. in prep.).

To examine the function of the cAMP receptor in development, an antisense cell line that blocked production of receptor was created. A construct was designed to transcribe the receptor cDNA in the antisense orientation using a strong constitutive promoter. Antisense cell lines, which failed to express receptor RNA or protein, did not enter the developmental program. As shown in Figure 4, antisense cells remain as a smooth monolayer after eight hours of starvation while wild-type cells have developed into aggregates. By 36 hours, however, the antisense cells begin to show signs of weak aggregation (Sun et al. in Prep.).

Fig. 4.

Antisense cell lines, which lack cAMP receptors, do not enter the developmental program. The left panel shows wild-type cells forming aggregates after eight hours of starvation while antisense cells, shown in the right panel, are a smooth monolayer.

Fig. 4.

Antisense cell lines, which lack cAMP receptors, do not enter the developmental program. The left panel shows wild-type cells forming aggregates after eight hours of starvation while antisense cells, shown in the right panel, are a smooth monolayer.

The cAMP receptor exists in two interchangeable forms designated R (40kD) and D (43 kD). This reversible shift of electrophoretic mobility is coupled to the spontaneous oscillations in cAMP synthesis in developing cells (Klein et al. 1985a). The R form is predominant in unstimulated cells and converts to the D form upon stimulation with cAMP (Devreotes and Sherring, 1985). Phosphorylation of both forms of receptor has been demonstrated in vitro and in vivo. Upon stimulation with cAMP, the phosphorylation increases by several-fold to a level of about four moles phosphate/mole receptor. It is likely that this phosphorylation underlies the shift in electrophoretic mobility (Klein et al. 19856).

The kinetics of receptor phosphorylation and déphosphorylation correlate closely with the shift of receptor mobility and the adaptation of several cAMP-induced responses (Figures 5 and 6). Phosphorylation is detectable in five seconds, has a half time of 30 s, and reaches a steady state within five minutes. When cells are freed of cAMP, both the shift in receptor mobility and phosphorylation decay. Déphosphorylation is detectable in 30s, has a half time of 2·5min, and is complete within 15 min (Figure 6). The kinetics of receptor phosphorylation and dephosphorylation match those of adaptation and deadaptation of cAMP-stimulated activation of adenylate cyclase and myosin phosphorylation. The half-maximal responses of these events all occur with concentrations of 5 nM-cAMP (Vaughan and Devreotes, 1988). Receptor phosphorylation is probably essential in the adaptation response as demonstrated in other G-protein-linked receptor systems (Sibley et al. 1988). Isolation of a cAMP-receptor kinase mutant would directly test if receptor desensitization requires phosphorylation.

Fig. 5.

The kinetics of phosphorylation (top panel) correlate closely with the receptor shift to the D form (bottom panel) during stimulation with cAMP. Both increase exponentially and reach a steady state by five minutes.

Fig. 5.

The kinetics of phosphorylation (top panel) correlate closely with the receptor shift to the D form (bottom panel) during stimulation with cAMP. Both increase exponentially and reach a steady state by five minutes.

Fig. 6.

The kinetics of dephosphorylation (top panel) and the shift to the R form of receptor (bottom panel) are similar after removal of cAMP.

Fig. 6.

The kinetics of dephosphorylation (top panel) and the shift to the R form of receptor (bottom panel) are similar after removal of cAMP.

Predominately serine and some threonine residues of the cAMP receptor are phosphorylated (Klein et al. 1985b). Two serines are present on the third cytoplasmic loop and an additional 18 are found in the C-terminal domain (Figure 3). When membranes containing the phosphorylated receptor are trypsinized, a soluble fragment is released. This peptide has an apparent molecular weight of 19 kD and appears to contain all of the ligand-induced phosphorylation sites. Studies are now underway to isolate this fragment and map the sites of phosphorylation (Vaughan and Devreotes, in prep.).

Multiple G-proteins and signal transduction pathways

Two G-protein <r-subunits, designated G-cv-1 and G-α-2, have been isolated from cDNA libraries using oligonucleotide probes based on the highly conserved GTP-binding and GTPase sites of the α-subunits in mammals. The nucleotide sequences of both subunits encode polypeptides of approximately 41 kD. Both subunits are 45 % identical to each other and to the mammalian tr-subunits, G” Go, and transducin, while only 35% identical to Gs and yeast GPA-1. When compared to the other α-subunits, some highly conserved regions are notable (Figure 7). Region A is almost 100% identical among all seven subunits and is the proposed GTPase site by analogy to c-Ha-ras and EF-Tu. Regions C, E and G are also highly conserved and are believed to form the GTP-binding site. A fifth domain, designated T, contains an almost completely conserved sequence of TCATDT and is near the C-terminus (Pupillo et al. 1989). This region may interact with receptor (Hamm et al. 1988).

Fig. 7.

All of the identified G α-subunits share conserved regions. Domain A contains the proposed GTPase site. Regions C, E, and G comprise the putative GTP-binding site. The T domain is thought to interact with receptors. (G1 - Dictyostelium G-α-1. G2 - Dictyostelium G-α-2. YE - Yeast GPA1. GO - Bovine Go. GI - Rat G1. TR - Bovine transducin. GS - Bovine Gs.

Fig. 7.

All of the identified G α-subunits share conserved regions. Domain A contains the proposed GTPase site. Regions C, E, and G comprise the putative GTP-binding site. The T domain is thought to interact with receptors. (G1 - Dictyostelium G-α-1. G2 - Dictyostelium G-α-2. YE - Yeast GPA1. GO - Bovine Go. GI - Rat G1. TR - Bovine transducin. GS - Bovine Gs.

Both G-α-1 and G-α-2 are expressed as multiple RNAs with different developmental time courses (Figure 8). The predominant 1·7 kb band of G-α-l is present during vegetative growth, increases several-fold until the ‘loose aggregate’ stage and then decreases. In contrast, the major 2·7 kb RNA of G-α-2 is almost completely absent during vegetative growth and is induced during early development with peak expression in the aggregation state. The developmental regulation of G-α-2 is similiar to those genes involved in cell aggregation that require intermittent cAMP stimulation for induction.

Fig. 8.

The time course of G-α-1 and G-α-2 mRNA expression reveals multiple transcripts. The major 1·7 kb band of G-α-l is present throughout development with peak expression between 12–15 h. The predominant 2·7 kb band of G-a-2 is developmentally regulated and expressed maximally at 12 h.

Fig. 8.

The time course of G-α-1 and G-α-2 mRNA expression reveals multiple transcripts. The major 1·7 kb band of G-α-l is present throughout development with peak expression between 12–15 h. The predominant 2·7 kb band of G-a-2 is developmentally regulated and expressed maximally at 12 h.

Peptide antisera were developed that specifically recognize each n-subunit as well as a common domain of both proteins (Figure 9). When visualized by immunoblot, G-α-1 migrates as a 38 kD protein, while G-a-2 has an apparent molecular weight of 40 kD. The time course of protein expression is consistent with the transcription of the major mRNA for each subunit. G-α-1 is present in both vegetative growth and throughout development, while G-α-2 is expressed in early development, peaking around 6–8 hours in shaking cultures (Figure 9, Kumagai et al. 1989).

Fig. 9.

Antibodies specifically recognize G-α-1 (lanes 1–4) and G-α-2 (lanes 5–8). In vegetative growth, G-α-1 is present (lane 1 & 9) but G-α-2 is absent (lane 5 & 9). Both a-subunits are present at six hours of development (lanes 2, 6, and 10). A cell line overexpressing G-a-1 shows a high increase in the staining of this protein (lane 3) while G-α-2 staining is normal at six hours of development (lane 7). Frigid A allele, HC 85, lacks G-α-2 (lane 8) when developed six hours. Both antibodies were mixed to stain G-α-1 and G-α-2 (lanes 9 & 10). The two bars show where G-α-2 (top bar) and G-α-l (bottom bar) migrate.

Fig. 9.

Antibodies specifically recognize G-α-1 (lanes 1–4) and G-α-2 (lanes 5–8). In vegetative growth, G-α-1 is present (lane 1 & 9) but G-α-2 is absent (lane 5 & 9). Both a-subunits are present at six hours of development (lanes 2, 6, and 10). A cell line overexpressing G-a-1 shows a high increase in the staining of this protein (lane 3) while G-α-2 staining is normal at six hours of development (lane 7). Frigid A allele, HC 85, lacks G-α-2 (lane 8) when developed six hours. Both antibodies were mixed to stain G-α-1 and G-α-2 (lanes 9 & 10). The two bars show where G-α-2 (top bar) and G-α-l (bottom bar) migrate.

Characterization of the mutant Frigid A has helped to establish a function for G-α-2. Frigid A mutants bind cAMP, but appear to be blocked in subsequent steps of signal transduction (Coukell et al. 1983). Activation of adenylate cyclase and early gene expression fails to occur even when cells are stimulated with cAMP. However, adenylate cyclase can be activated by GTP in vitro (Kesbeke and Van Haastert, 1988; Kumagai et al. 1989). Both RNA blot and immunoblot analysis show most alleles to have a strong reduction or absence of G-α-2 mRNA and protein at all stages of cell development. Most conclusively, one Frigid A allele, HC 85, contains a deletion in the G-α-2 gene (Kumagai et al. 1987). Hence both biochemical and genetic evidence suggest that the Frigid A allele is the G-a-2 subunit. One Frigid A mutant, HC 112, however, has near normal levels of both mRNA and protein. The G-a-2 gene in this mutant was cloned by PCR (Polymerase Chain Reaction) and sequenced. It carries a point mutation within the coding sequence. Further characterization of HC 112 is in progress (Pitt and Devreotes, in prep.).

To investigate the function of the G-α-1 subunit, this protein was overexpressed. Both G-α-1 mRNA and protein levels are approximately 20-fold higher than wild-type as detected by RNA blot and immunoblots (Figure 9). Transformed cells grown on surfaces are large and multinucleate. When these cells are developed, most fail to aggregate but those that do, produce small and abnormal fruiting bodies. This phenotype suggests a role for G-α-1 in cytokinesis during growth (Kumagai et al. 1989).

A G-protein β-subunit has also been cloned using probes based on conserved sequences of Drosophila and human β-subunits. It is expressed as a constitutive 1·9 kb mRNA. The amino acid sequence has 63% identity to the human β-2-subunit. The first 50 residues share only 30% identity, with the remainder of the protein being 70% identical. The significance of these different domains is under investigation. Work is in progress to examine the function of the β-subunit by creating antisense and overexpression cell lines (Lilly and Devreotes, in prep.)

Our studies of Dictyostelium show that G-protein-linked signal transduction pathways play a central role in the development of a multicellular organism. The cAMP receptors we have identified are essential to development; antisense cell lines do not enter the developmental program. The Gα2 subunit is critical for cAMP chemotaxis and induction of gene expression; Frigid A, which lacks a functional G-α-2, is unable to differentiate. We propose that G-protein-linked receptor systems will be a major component in the development of many organisms.

The oscillatory nature of cAMP-receptor activation and inactivation plays a critical role in cell-cell signaling, aggregation and gene expression in Dictyostelium. Positional signaling, via a cAMP oscillator, underlies cell aggregation in early development and establishes cell differentiation in later stages. Perhaps higher organisms such as C. elegans, Drosophila, or vertebrates use this type of cellular communication during embryogenesis. One could envision progenitor cells creating organization centers by establishing gradients or oscillations of a signaling molecule. These organization centers could influence local cell differentiation or act in concert to institute larger changes. Hence cAMP signaling may not just be limited to slime molds, but utilized in more complex organisms as well.

ÇOukell
,
M. B.
,
Lappano
,
S.
&
Cameron
,
A. M.
(
1983
).
Isolation and characterization of cAMP unresponsive (frigid) aggregationdeficient mutants of Dictvostelium discoideum
.
Dev. Genet
3
,
283
-
297
.
Devreotes
,
P N.
(
1983
).
Cyclic nucleotides and cell-cell communication in Dictvostelium discoideum
.
Adv. Cvchc Nucl. Res
.
15
,
55
96
.
Devreotes
,
P. N.
&
Sherring
,
J. A.
(
1985
).
Kinetics and concentration dependence of reversible cAMP-induced modification of the surface cAMP receptor in Dictyostelium
.
J. biol. Chem
.
260
,
6378
6384
.
Dohlman
,
H. G.
,
Caron
,
M. G.
&
Lefkowitz
,
R. J.
(
1987
).
A family of receptors coupled to guanine nucleotide regulatory proteins
.
Biochem
.
26
,
2657
2664
.
Gerisch
,
G.
(
1987
).
Cyclic AMP and other signals controlling cell development and differentiation in Dictyostelium
.
Annu. Rev Biochem
.
56
,
853
879
.
Hamm
,
H. E.
,
Deretic
,
D.
,
Arendt
,
A.
,
Hargrave
,
P. A.
,
Koenig
,
B.
&
Hofmann
,
K. P.
(
1988
).
Site of G-protein binding to rhodopsin mapped with synthetic peptides from the rr-subumt
.
Science
241
,
832
835
.
Hargrave
,
P. A.
(
1986
).
Molecular dynamics of the rod cell
. In
The Retina
, (ed. R. Adler &
D.
Farber
), pp.
207
237
.
New York
:
Academic Press
.
Janssens
,
P. M. W.
&
Van Haastert
,
P. J. M.
(
1987
).
Molecular basis of transmembrane signal transduction in Dictyostelium discoideum
.
Microbiol. Rev
.
51
,
396
418
.
Kesbeke
,
F.
,
Snaar-Jagalska
,
B. E.
&
Van Haastert
,
P. J. M.
(
1988
).
Signal transduction in Dictyostelium fgd A mutants with a defective interaction between surface cAMP receptors and a GTP-binding regulatory protein
.
J. Cell Biol
.
107
,
521
528
.
Klein.
P.
,
Fontana
,
D.
,
Knox
,
B.
,
Theibert
,
A.
&
Devreotes
,
P.
(
1985b
).
cAMP receptors controlling cell-cell interactions in the development of Dictyostelium
.
Cold Spring Harbor Symp. Quant. Biol
.
50
,
787
799
.
Klein
,
P.
,
Sun
,
T. J.
,
Saxe
,
C. L.
,
Kimmel
,
A. R.
,
Johnson
,
R. L.
&
Devreotes
,
P. N.
(
1988
).
A chemoattractant receptor controls development in Dictvostelium discoideum
.
Science
241
,
1467
-
1472
. ‘
Klein
,
P.
,
Theibert
,
A.
,
Fontana
,
D.
&
Devreotes
,
P. N.
(
1985a
).
Identification and cyclic AMP-induced modification of the cyclic AMP receptor in Dictyostelium discoideum
.
J. biol. Chem
.
260
,
1757
-
1764
.
Kumagai
,
A.
,
Pupillo
,
M.
, GUNDERSON. R.,
Miake-Lye
,
R.
,
Devreotes
,
P. N.
&
Firtel
,
R. A.
(
1989
).
Regulation and function of G-n-1 protein subunits in Dictvostelium
.
Cell
57
.
265
-
275
.
MANN, S. K O. &
Firtel.
R. A.
(
1987
).
Cyclic AMP regulation of early gene expression in Dictyostelium discoideum: mediation via the cell surface cyclic AMP receptor
.
Molec. cell. Biol
.
7
,
458
469
.
Pupillo
,
M.
,
Kumagai
,
A.
,
Pitt
,
G.
,
Firtel
,
R. A.
&
Devreotes.
P. N.
(
1989
).
Multiple rr subunits of G-proteins in Dictyostelium
.
Proc. natn. Acad. Sci. (in press)
.
Schaap
,
P.
(
1986
).
Regulation of size and pattern in the cellular slime molds
.
Differentiation
33
,
1
16
.
SIBLEY. D. R.,
Benovic
,
J L.
,
Caron
,
M. G.
&
Lefkowitz
,
R. J.
(
1987
).
Regulation of transmembrane signaling by receptor phosphorylation
.
Cell
49
,
913
22
.
Tomchik
,
K. J.
&
Devreotes
,
P. N.
(
1981
).
cAMP waves in Dictyostelium discoideum: demonstration by a novel isotope dilution fluorography technique
.
Science
212
,
443
446
.
Vaughan
,
R. A.
&
Devreotes
,
P. N.
(
1988
).
Ligand-induced phosphorylation of the cAMP receptor from Dictyostelium discoideum
.
J. biol. Chem
.
263
,
14538
14543
.