ABSTRACT
To determine the function of the Dictyostelium Gα1 subunit during aggregation and multicellular development, we analyzed the phenotypes of gα1 null cells and strains overexpressing either wild-type Gα1 or two putative constitutively active mutations of Gα1. Strains overexpressing the wild-type or mutant Gα1 proteins showed very abnormal culmination with an aberrant stalk differentiation. The similarity of the phenotypes between Gα1 overexpression and expression of a putative constitutively active Gα1 subunit suggests that these phenotypes are due to increased Gα1 activity rather than resulting from a non-specific interference of other pathways. In contrast, gα1 null strains showed normal morphogenesis except that the stalks were thinner and longer than those of wild-type culminants. Analysis of cell-type-specific gene expression using lacZ reporter constructs indicated that strains overexpressing Gα1 show a loss of ecmB expression in the central core of anterior prestalk AB cells. However, expression of ecmB in anterior-like cells and the expression of prestalk A-specific gene ecmA and the prespore-specific gene SP60/cotC appeared normal. Using a Gα1/lacZ reporter construct, we show that Gα1 expression is cell-type-specific during the multicellular stages, with a pattern of expression similar to ecmB, being preferentially expressed in the anterior prestalk AB cells and anterior-like cells. The developmental and molecular phenotypes of Gα1 overexpression and the cell-type-specific expression of Gα1 suggest that Gα1-mediated signaling pathways play an essential role in regulating multicellular development by controlling prestalk morphogenesis, possibly by acting as a negative regulator of prestalk AB cell differentiation.
During the aggregation phase of development, gα1 null cells display a delayed peak in cAMP-stimulated accumulation of cGMP compared to wild-type cells, while Gα1 overexpressors and dominant activating mutants show parallel kinetics of activation but decreased levels of cGMP accumulation compared to that seen in wild-type cells. These data suggest that Gα1 plays a role in the regulation of the activation and/or adaptation of the guanylyl cyclase pathway. In contrast, the activation of adenylyl cyclase, another pathway activated by cAMP stimulation, was unaf fected in gα1 null cells and cell lines overexpressing wild-type Gα1 or the Gα1(Q206L) putative dominant activating mutation. However, the Gα1(G45V) putative constitutively active mutation showed significantly reduced adenylyl cyclase activity in response to cAMP. All Gα1 mutant cell lines aggregated normally; however, aggregates of cells expressing Gα1(G45V) developed ring-like structures that then developed a polarity and a small mound-like structure before forming a slug. Immunoprecipitation results suggest that the G45V phenotypes may be due to altered properties of this subunit and its association with the βg subunit.
INTRODUCTION
Multicellular development in Dictyostelium is initiated upon starvation and controlled through a number of extracellular signaling molecules that include cAMP, DIF, ammonia, adenosine, protein factor CMF and a proposed folate analog (Williams, 1988, 1991; Gomer et al., 1991; Kimmel and Firtel 1991; Schaap, 1991; Hadwiger et al., unpublished data). These molecules regulate the formation of the multicellular aggregate through chemotactic aggregation to extracellular cAMP, as well as subsequent cell differentiation, gene expression, cell patterning and morphogenesis (Mehdy et al., 1983; Devreotes, 1989; Seigert and Weijer, 1989; Berks and Kay, 1990; Kimmel and Firtel, 1991; Williams, 1991; Firtel, 1991,Traynor et al., 1992). The cellular responses to extracellular cAMP and folate are known to be regulated through independent cell surface receptors and are coupled to heterotrimeric G proteins that contain different Gα subunits (Kumagai et al., 1989, 1991; Kesbeke et al., 1990; Hadwiger et al., unpublished data).
So far, eight developmentally regulated Gα subunits have been identified from Dictyostelium (Pupillo et al., 1989; Kumagai et al., 1989, 1991; Hadwiger et al., 1991; Wu and Devreotes, 1991; Cubitt et al., 1992) and a single β subunit has been identified that is constitutively expressed throughout growth and development and required for aggregation (Lilly et al., 1993). The Gα subunit Gα2 has been the most intensively studied. It couples to multiple cyclic AMP receptors (cARs), is essential for aggregation and all cAMP-mediated responses during this stage except the activation of Ca2+ influx and has been shown to play a role in regulation of stalk cell differentiation during culmination (Kesbeke et al., 1988; Kumagai et al., 1989, 1991; Pupillo et al., 1992; Milne and Devreotes, 1993; Carrel et al., 1994). Gα4 is required for proper morphogenesis and the differentiation of the prespore pathway and is known to couple to at least two classes of folate receptors during growth and development (Hadwiger and Firtel, 1992; Hadwiger et al., unpublished data). While it is expected that the other Gα subunits play important regulatory functions during the growth and development of Dictyostelium, little is known about individual α subunits other than Gα2 and Gα4. During the aggregation stage, binding of cAMP to the cell surface receptors triggers activation of adenylyl cyclase (AC), guanylyl cyclase (GC), phospholipase C (PLC), Ca2+ uptake, actin polymerization, and aggregation-stage gene expression (Europe-Finner and Newell, 1987; Devreotes, 1989; Firtel et al., 1989; Kumagai et al., 1989; Mann and Firtel, 1989; van Haastert et al., 1989; Condeelis et al., 1990; Milne and Coukell, 1991). The cAMP produced via stimulated AC activity is secreted into the extracellular medium and can activate additional receptors on the same cell, creating a positive feedback loop, or activate receptors on adjacent cells and thus relaying the signal (Devreotes, 1989). Increases in IP3 and cGMP are detected within 3 seconds of cAMP receptor occupation, peak by 7 and 10 seconds, respectively, and then decrease to basal levels shortly thereafter, indicating an adaptation of these pathways (van Haastert et al., 1986; Europe-Finner and Newell, 1987). Adaptation of several of these pathways is also thought to be mediated by G-proteins (van Haastert et al., 1986; Small et al., 1987; Snaar-Jagalska and van Haastert, 1990; Okaichi et al., 1992), while the desensitization of the cAMP receptor is due to phosphorylation (Devreotes and Sherring, 1985; Klein et al., 1985; van Haastert, 1987). Recent analysis suggests that PLC and GC are directly regulated through the Gα subunit Gα2, while AC is probably activated by βg (Firtel et al., 1989; Kumagai et al., 1991; Okaichi et al.,1992; Lilly et al., 1993).
The Gα1 subunit is encoded by a single gene that differentially expresses multiple mRNAs during development (Pupillo et al., 1989; Kumagai et al., 1989, 1991). A major Gα1 transcript is expressed at moderate levels in vegetative cells. The levels increase several-fold during early development, peaking at the loose aggregate stage. Two additional mRNAs are then induced at the aggregation stage and are expressed throughout culmination (Pupillo et al., 1989). Preliminary investigation of the function of Gα1 by antisense and gene disruption did not show a gross morphological change. However, cells overexpressing Gα1 from the non-cell-type-specific actin promoter, which is expressed at high levels during growth and early development and at lower levels during the multicellular stages, show large multinucleated cells during growth and abnormal development (Kumagai et al., 1989, 1991). To evaluate further the role of Gα1 during Dictyostelium development, we examined the effect of Gα1 deletion, overexpression and constitutive activation on signal transduction and morphogenesis. We show that Gα1 exerts an effect on signal transduction during early development by altering the GC pathway. lacZ reporter constructs indicate that Gα1 is preferentially expressed in a group of anterior cells, designated prestalk AB cells, which are important for regulating stalk formation, and in anterior-like cells. Overexpression of wild- type Gα1 or putative constitutively active Gα1 mutations from the Gα1 promoter results in abnormal culmination. The apparent role of Gα1 in regulating morphogenesis and stalk differentiation is discussed.
MATERIALS AND METHODS
Construction of expression vectors
Gα1 promoter, the Gα1 coding region (Kumagai et al., 1991) and the 2H3 terminator from Dictyostelium SP70/cotB gene (Crowley et al., 1985) were cloned into a plasmid with a pAT153L plasmid backbone (Haberstroh and Firtel, 1990), containing the origin of replication and the trans-acting factor from pnDeΔ1 [clone from endogenous plasmid DdP2 (Leiting and Noegel, 1988)]. This vector (Gα1.DdP2; Fig. 1A) replicates extrachromosomally with a uniform copy number of about 15 per cell (Okaichi et al., 1992). An integrating vector (Gα1.EXP4+) (Dynes et al., 1994) was constructed with a plasmid containing a PATSP backbone, neomycin resistance cassette, Gα1 full-length promoter driving the Gα1 coding region and the 2H3 terminator from the SP70/cotB gene (Crowley et al., 1985; Fig. 1B). The integrating vector Gα1.lacZ was constructed by replacing the SP60/cotC promoter with the Gα1 promoter in SP60/lacZ vector described in Haberstroh and Firtel (1990; Fig. 1C).
Maps of Gα1 vector constructs. (A) Extrachromosomal vector used for overexpression of Gα1: wild-type, G45V and Q206L mutations (Gα1.DdP2). A 1.6 kb fragment of the Gα1 promoter sequence (Gα1 prom.) and a 1.3 kb fragment containing the full-length Gα1 wild-type coding sequence, or Gα1 coding sequence containing either the amino acid substitution G45V or Q206L (Gα1 coding), was inserted into the Xho.Gα2.DdP2 vector (Okaichi et al., 1992) replacing the ClaI/SpeI fragment containing the Gα2 promoter and Gα2 coding sequence. The ATG and TAA indicates the translation initiation and termination codons of the Gα1 gene. pnDeΔ1 represents the DdP2 extrachromosomal element (Leiting and Noegel, 1988). (B) Integrating vector used for overexpression of the Gα1 subunit (Gα1.EXP4+). A SalI/HindIII fragment containing the Gα1 promoter (Gα1 prom.) and Gα1 coding sequence (Gα1) was cloned into the EXP4+ vector (Dynes et al., 1994). (C) Integrating vector containing the lacZ gene driven by the Gα1 promoter (Gα1/lacZ). The Gα1 promoter (Gα1 prom.) was inserted as a BglII/SpeI fragment to replace the BamHI/SpeI fragment containing the SP60/cotC promoter in SP60/lacZ construct from Haberstroh and Firtel (1990). The open box represents 42 amino acids from the 5′ untranslated region of the SP60/cotC gene that contains the SpeI site and the ATG translation initiation codon. The lacZ ATG has been deleted (Haberstroh and Firtel 1990). For all the constructs, ‘term.’ represents the 2H3 terminator from the Dictyostelium SP70/cotB gene (Crowley et al., 1985). G418r represents the pACT6::Neor gene fusion that confers resistance to the drug, G418. B, BamHI; X, XbaI; Sa, SalI; H, HindIII; S, SpeI; Sc, ScaI; C, ClaI; E, EcoRI; K, KpnI; P, PstI; Xh, XhoI; N, NotI; Sm, SmaI; Bg, BglII.
Maps of Gα1 vector constructs. (A) Extrachromosomal vector used for overexpression of Gα1: wild-type, G45V and Q206L mutations (Gα1.DdP2). A 1.6 kb fragment of the Gα1 promoter sequence (Gα1 prom.) and a 1.3 kb fragment containing the full-length Gα1 wild-type coding sequence, or Gα1 coding sequence containing either the amino acid substitution G45V or Q206L (Gα1 coding), was inserted into the Xho.Gα2.DdP2 vector (Okaichi et al., 1992) replacing the ClaI/SpeI fragment containing the Gα2 promoter and Gα2 coding sequence. The ATG and TAA indicates the translation initiation and termination codons of the Gα1 gene. pnDeΔ1 represents the DdP2 extrachromosomal element (Leiting and Noegel, 1988). (B) Integrating vector used for overexpression of the Gα1 subunit (Gα1.EXP4+). A SalI/HindIII fragment containing the Gα1 promoter (Gα1 prom.) and Gα1 coding sequence (Gα1) was cloned into the EXP4+ vector (Dynes et al., 1994). (C) Integrating vector containing the lacZ gene driven by the Gα1 promoter (Gα1/lacZ). The Gα1 promoter (Gα1 prom.) was inserted as a BglII/SpeI fragment to replace the BamHI/SpeI fragment containing the SP60/cotC promoter in SP60/lacZ construct from Haberstroh and Firtel (1990). The open box represents 42 amino acids from the 5′ untranslated region of the SP60/cotC gene that contains the SpeI site and the ATG translation initiation codon. The lacZ ATG has been deleted (Haberstroh and Firtel 1990). For all the constructs, ‘term.’ represents the 2H3 terminator from the Dictyostelium SP70/cotB gene (Crowley et al., 1985). G418r represents the pACT6::Neor gene fusion that confers resistance to the drug, G418. B, BamHI; X, XbaI; Sa, SalI; H, HindIII; S, SpeI; Sc, ScaI; C, ClaI; E, EcoRI; K, KpnI; P, PstI; Xh, XhoI; N, NotI; Sm, SmaI; Bg, BglII.
Mutagenesis
Mutations within the coding regions were made by oligonucleotide-directed mutagenesis, using a mutagenesis kit (code RPN.1532, Amersham Corp.) and a single-stranded DNA template. The template was made by cloning the Gα1 coding region into M13 and isolating single-stranded DNA from purified phage particles as described in Okaichi et al. (1992). Mutant clones were sequenced to confirm changes in the nucleotide sequence. The mutant Gα1 coding regions were cloned into the Gα1.DdP2 vector (Fig. 1A) by replacing the wild-type Gα1 coding region with Gα1 cDNA containing either G45V or Q206L substitutions.
Cell transformation and gene disruption
Wild-type axenic strain KAx-3 and gα1 null strain were transformed by electroporation (Dynes and Firtel, 1989). Stable transformants were selected using G418. Co-transformation was done using Ca2PO4 precipitates (Nellen and Firtel, 1985).
gα1 null cells were created by homologous recombination in the wild-type-developing, thymidine auxotroph (strain JH10) that is isogenic with the parental strain KAx-3 except at the Pyr5-6 and Thy1 loci (Mann and Firtel, 1991; Hadwiger and Firtel, 1992) as previously described (Kumagai et al., 1991). The previous gα1 null strain described in Kumagai et al. (1991) was not isogenic with KAx-3 and thus it was difficult to ascribe differences in signaling pathways to the gα1 null mutant versus the parental background. Breifly, a 3 kb fragment containing the Thy1 gene, which complements the thymidine growth requirement in JH10 (Dynes and Firtel, 1989), was inserted into the central most BglII site in the Gα1 cDNA clone. The resulting construct was digested with EcoRI to excise the fragment containing Gα1 cDNA with the THY1 insertion and was transformed into JH10 strain, and selected for thymidine prototrophy.
AC and GC activity
Cells were starved for 16 hours at 5°C on non-nutrient agar and pulsed for 10 seconds with 50 nM cAMP for 1.5 hours at 6 minute intervals, at room temperature to increase responses to cAMP by autogenous induction of the receptor (Chisholm et al., 1987; Okaichi et al., 1992). Cell lysates following 2′deoxy cAMP addition were prepared as previously described (Kumagai et al., 1991). Cyclic AMP and cGMP concentrations were determined using a radioimmunoassay kit (Amersham Corp.).
Cytological staining and RNA assays
KAx-3 cells were cotransformed with expression vectors containing the Gα1 coding region and the lacZ gene as previously described (Haberstroh and Firtel, 1990; Esch and Firtel, 1991). Cells were plated for development in Millipore filters, fixed in 0.5% glutaraldehyde at appropriate times and stained with X-gal (BRL) according to Dingermann et al. (1989).
RNA blots were run as previously described (Mann and Firtel, 1987) using the 1.9 and 4.1 kb small and large rRNAs as size markers.
Immunoprecipitation and western blotting
For coprecipitation of Gα1 and β subunits, membrane fractions were prepared as previously described (Gunderson and Devreotes, 1990). The membranes were then solubilized in 1.5% octylglucoside, extracted on ice for 10 minutes and centrifuged at 100,000 revs/minute, 30 minutes in a TL-100. The octylglucoside concentration in the supernatants was reduced to 0.75% in econo R buffer (Gunderson and Devereotes, 1990) and incubated with anti Gα1 coupled to Protein A-agarose for 1 hour at 4°C. The resulting immunoprecipitates were recovered after a brief spin, washed twice in econo R buffer containing 0.75% octylglucoside and washed twice more in econo R buffer with no detergent. The immunoprecipitates were dissolved in sample buffer and subjected to SDS-PAGE and western blotting according to Carboni and Condeelis (1985). Western blots were stained with anti-Gα1 and anti-β subunit (kindly provided by Drs Robert Gunderson and Peter Devreotes). Production of the anti-Gα1 is described in Kumagai et al. (1989) and the anti-β is described in Lilly et al. (1993).
RESULTS
Developmental effects of Gα1 deletion and overexpression
Gα1-overexpressing clones were isolated by transformation of wild-type KAx-3 cells with an extra-chromosomal expression vector containing the Gα1 coding region driven by its own promoter (Gα1.DdP2; Fig. 1A). Previously reported Gα1 overexpression strains employed the Act15 (Actin 15) promoter using an integrating vector (Kumagai et al., 1989). The Act15 promoter is expressed in all cells at high levels during vegetative growth and early development and at lower levels during the later stages. The use of the Gα1 promoter rather than the generic Act15 promoter enabled the examination of the effect of expressing Gα1 in the same cell types as the endogenous protein.
To show that the cloned Gα1 promoter has a similar temporal pattern of expression as the endogenous promoter, we performed RNA blot hybridization experiments using wild-type KAx-3 cells and gα1 null cells expressing the Gα1.DdP2 vector (described above) in which we did not observe a Gα1 overexpression phenotype. gα1 null cells were used to eliminate endogenous Gα1 expression for the RNA blots. We observe an ∼2.1 kb transcript that is expressed throughout growth and development in wild-type cells, similar to what has been previously described (Pupillo et al., 1989). At ∼4 hours, a ∼3.1 kb transcript is first expressed that continues through remainder of development. A weaker, ∼3.7 kb RNA is seen that is highest during the early multicellular stages. In the gα1 null cells expressing Gα1 from the Gα1 cloned promoter, a similar pattern is seen, except that the middle-sized transcript is accumulated to higher levels in this strain than in the wild type (data not shown). For analysis of developmental pheno-types resulting from Gα1 overexpression, clonal isolates that showed about ∼20-fold increase in Gα1 levels as identified by western blotting with an anti-Gα1 antibody were selected for further study (Fig. 2).
Western blot analysis of Gα1. Western blot analysis of Gα1 protein levels is shown in Dictyostelium wild-type (KAx-3) cells and transformants: gα1 null; gα1 null comp. (gα1 null cells complemented with the extrachromosomal vector Gα1.DdP2); Gα1(G45V), KAx-3 (KAx-3 cells transformed with Gα1.DdP2 vector containing the G45V amino acid substitution); Gα1(G45V), gα1 null (gα1 nulls transformed with same vector as previous cell line); Gα1(Q206L), gα1 null (gα1 null cells transformed with the Gα1.DdP2 vector containing the Q206L amino acid substitution); Gα1(Q206L), KAx-3 (KAx-3 cells transformed with the same vector as previous cell line); Gα1+, KAx-3 (KAx-3 cells transformed with Gα1.DdP2 vector containing the wild-type Gα1 coding region). Whole-cell lysates containing equal amounts of protein were fractionated on SDS-PAGE, blotted and immunoprobed for Gα1 as described in Materials and methods.
Western blot analysis of Gα1. Western blot analysis of Gα1 protein levels is shown in Dictyostelium wild-type (KAx-3) cells and transformants: gα1 null; gα1 null comp. (gα1 null cells complemented with the extrachromosomal vector Gα1.DdP2); Gα1(G45V), KAx-3 (KAx-3 cells transformed with Gα1.DdP2 vector containing the G45V amino acid substitution); Gα1(G45V), gα1 null (gα1 nulls transformed with same vector as previous cell line); Gα1(Q206L), gα1 null (gα1 null cells transformed with the Gα1.DdP2 vector containing the Q206L amino acid substitution); Gα1(Q206L), KAx-3 (KAx-3 cells transformed with the same vector as previous cell line); Gα1+, KAx-3 (KAx-3 cells transformed with Gα1.DdP2 vector containing the wild-type Gα1 coding region). Whole-cell lysates containing equal amounts of protein were fractionated on SDS-PAGE, blotted and immunoprobed for Gα1 as described in Materials and methods.
Wild-type cells aggregate into discrete mounds that elongate to form a finger and then differentiate into migrating slugs containing prespore cells in the posterior 85% of the slug and prestalk cells in the anterior 15% of the slug (Fig. 3). Slugs may migrate until conditions are favorable for development and then develop into a mature fruiting body containing spores held several millimeters above the substratum by a vacuolated stalk, ∼24 hours after starvation. Gα1-overexpressing cells aggregated normally and initially formed normal-looking slugs. The slugs then developed a bulge followed by a constriction in the anterior. These slugs culminated early (at about 20-22 hours) after starvation into fruiting bodies with stalks that were abnormally thick and flat at the bottom, a thick cellular region extending from the posterior towards the apex and a small sorus. In most cases, only the very anterior of the culminant rose from the substratum (Fig. 3).
Developmental phenotype of wild-type strain KAx-3 and the wild-type strain KAx-3 overexpressing Gα1. (A) Wild-type slug; (B) wild-type culminant; (C) Gα1 overexpressor slug; (D) Gα1 overexpressor culminant.
A gα1 null cell line was created in a wild-type background to evaluate the effect of Gα1 on signal transduction. As reported previously, ga1 null cells developed normally (Kumagai et al., 1991) except that the stalks of the culminants were longer and thinner than those of wild-type strains and had a tendency to droop, presumably due to the weight of the sorus on the thinner stalk (Fig. 4).
Spatial pattern of β-gal expression from the ecmB promoter in wild-type and Gα1 mutant strains. Whole organisms of wild-type KAx-3 strain (WT), KAx-3 strain overexpressing Gα1 (OV) and gα1 null strain (NU) were histochemically stained for β-gal activity as described in Materials and methods. Top panel, slugs; middle and bottom panels, culminants. The slug in the Gα1 (OV) panel was stained for three times longer than the adjacent panel showing the wild-type slug (WT).
Spatial pattern of β-gal expression from the ecmB promoter in wild-type and Gα1 mutant strains. Whole organisms of wild-type KAx-3 strain (WT), KAx-3 strain overexpressing Gα1 (OV) and gα1 null strain (NU) were histochemically stained for β-gal activity as described in Materials and methods. Top panel, slugs; middle and bottom panels, culminants. The slug in the Gα1 (OV) panel was stained for three times longer than the adjacent panel showing the wild-type slug (WT).
Gα1 overexpressor strains show abnormal expression of the prestalk gene ecmB
To determine whether the aberrant culmination morphology was associated with a defect in cell patterning, we used lacZ expressed from cell-type-specific promoters (Haberstroh and Firtel, 1990; Ceccarelli et al., 1991) to examine the spatial patterning of the prestalk and prespore cells within the aggregates through multicellular development. These constructs were co-transformed with a vector (Gα1.EXP4+) that contains the Gα1 coding region expressed from the Gα1 promoter and the neomycin resistance drug selection marker (Fig. 1B). Cell lines exhibiting the Gα1 overexpression phenotype and co-expressing the lacZ marker were used for further analysis.
Wild-type cells expressed ecmB in a central core of cells (prestalk AB cells) present in the very anterior tip of the slug and in the anterior-like cell (ALC) population that is randomly scattered throughout the slug (Fig. 4) as has been previously described (Jermyn et al., 1989). During culmination, ecmB was expressed in the basal disc, stalk, and upper and lower caps of the sorus. gα1 null strains showed a pattern of ecmB expression similar to that of wild-type (Fig. 4). In contrast, strains over-expressing Gα1 did not show any detectable expression of ecmB in the central core of prestalk AB cells at the tip in first fingers or in migrating slugs, even after staining for three times longer than was used to stain the wild-type aggregates (for migrating slugs, see Fig. 4; data for first finger stage not shown). Staining for ecmB was, however, observed in the scattered cell population throughout the slug that are presumably ALCs. The apparent increased intensity of ecmB/lacZ staining in the ALC population seen in Fig. 4 is due to the longer staining time used to show lack of staining in the anterior. The level of endogenous ecmB expression was examined in Gα1 overexpressors by RNA blot analysis and was not significantly different (within a factor of two) from that observed in wild-type slugs (data not shown). In the culminant, ecmB/lacZ staining was visible in what were clearly the stalk and caps on the small sorus. However, only some cells stained within the large basal region of the culminant that remained adhered to the substratum (Fig. 4). When the spatial patterns of expression of the prestalk gene ecmA and the prespore gene SP60/cotC were examined using lacZ reporter constructs in the Gα1-overexpressor and null strains and compared to the staining pattern seen in wild-type cells, no difference was observed at the slug stage. In culminants, the expression patterns of ecmA/lacZ (staining in the stalk and cups of the small sorus) and SP60/lacZ (staining in the sorus) were also the same in both wild-type and Gα1 mutant strains. However, the portion of the posterior region of the fruiting body of the Gα1-overexpressor that remained attached to the substratum showed little staining with any of the markers (data not shown). In some even more severely mutant Gα1-overexpressor culminants, most of the cells of the culminant remained attached to the substratum with irregular patches of cells observed below the tip of the spore head. These patches of cells did express SP60/lacZ but not the prestalk reporters ecmB/lacZ (Fig. 4) or ecmA/lacZ (data not shown), suggesting that these cells represented the spore mass (data not shown).
Gα1 is expressed in prestalk AB cells and ALCs in wild-type slugs in a pattern similar to that of ecmB
To determine the spatial pattern of expression of Gα1 in developing wild-type Dictyostelium, wild-type cells were transformed with a Gα1 promoter/lacZ reporter construct (Gα1/lacZ; Fig. 1C). β-gal expression was observed to be diffused throughout early aggregation mounds. As development progressed, β-gal expression became enriched to the tip of the tight aggregate and developing first finger as well as in a cell population randomly distributed throughout the length of the finger. In the slug, staining is clearly seen in the central core of apparently prestalk AB cells and in scattered cells, presumably ALCs in a pattern that is indistinguishable from that of ecmB/lacZ (Jermyn and Williams 1991) or lacZ driven by Gα4 or the phosphotyrosine phosphatase PTP2 promoter Hadwiger and Firtel, 1992; Howard et al., 1994). In culminants, Gα1/lacZ also showed a pattern of β-gal expression that was similar to that of ecmB/lacZ with staining in the stalks, and the upper and lower caps of the sorus (Fig. 5).
Spatial pattern of β-gal expression from the Gα1 promoter in wild-type KAx-3 strain. β-gal expression was detected histochemically as described in Materials and methods. (A) Early mound (8-9 hours after starvation); (B) tipped mound (10-12 hours); (C) early finger (12-14 hours); (D) slug showing localized staining in anterior (15-18 hours); (E) slug showing β-gal expression in both anterior tip and diffused throughout organism in anterior-like cells; (F) higher magnification of anterior portion of slug in D; (G) culminant (24 hours).
Spatial pattern of β-gal expression from the Gα1 promoter in wild-type KAx-3 strain. β-gal expression was detected histochemically as described in Materials and methods. (A) Early mound (8-9 hours after starvation); (B) tipped mound (10-12 hours); (C) early finger (12-14 hours); (D) slug showing localized staining in anterior (15-18 hours); (E) slug showing β-gal expression in both anterior tip and diffused throughout organism in anterior-like cells; (F) higher magnification of anterior portion of slug in D; (G) culminant (24 hours).
Phenotypes of putative constitutively active mutations of Gα1
The mutant phenotype of Gα1 overexpressors suggests that Gα1 is directly or indirectly involved in the regulation of Dictyostelium development. To further investigate the potential effect of excess Gα1 function, we examined the effect of putative constitutively active Gα1 mutant subunits on physiological responses and development. This was achieved by amino acid substitutions in the highly conserved α subunit domains known to reduce the intrinsic GTPase function in mammalian Ras and Gαs (Bourne et al., 1991). Accordingly, the glycine at position 45 was changed to a valine (GAGES®GAVES) and the glutamine at position 206 was separately converted to a leucine residue (GGQRS®GGLRS), mutations that are analogous to G12V and Q61L in Ras, and G49V and Q227L in Gαs, respectively (Graziano and Gilman 1989; Masters et al., 1989; Bourne et al., 1991). Gα1 coding regions containing these point mutations were expressed from the Gα1 promoter on an extrachromosomal expression vector in both wild-type and gα1 null cells (Fig. 1A). Clones that expressed the mutant protein to a level 15- to 20-fold higher than endogenous Gα1 were identified by Western blot analysis (Fig. 2).
Wild-type and gα1 null cells expressing the Gα1(Q206L) exhibit a phenotype similar to that of Gα1 overexpressors and formed slugs that culminated into fruiting bodies with thick elongated basal discs and stalks of uneven thickness (Fig. 6, data for the gα1 null cells not shown). In contrast, expression of Gα1(G45V) from the Gα1 promoter on either an extra-chromosomal (Fig. 1A) or an integrating expression vector (Fig. 1B) resulted in an aberrant aggregation phenotype, in transformed populations and clonal cell lines. The severity of the phenotype depended on the level of expression of the Gα1 protein as determined by western blotting using an anti-Gα1 antibody (Fig. 2). The Gα1(G45V)-expressing cells exhibited normal early aggregation into streams, but then formed circular streams instead of distinct mounds with a significant delay in further development, suggesting that the Gα1(G45V) mutation was altering cAMP-mediated signaling at this stage of development. These rings became polarized with a distinct thickening in one side that eventually became the anterior of a slug. These abnormal aggregates culminate in small thick based fruiting bodies after about 72 hours after starvation (Fig. 7).
Developmental phenotype of Gα1(Q206L) mutation. (A,C) developing slugs (18-20 hours); (B) culminant (20-22 hours).
Developmental phenotype of Gα1(G45V) mutation. (A) Early mounds 10 hours after starvation; (B) doughnut-shaped aggregates, 15-18 hours; (C) ring shaped aggregates, 20-24 hours; (D) migrating slugs, 24-30 hours; (E) early fruiting bodies, 34-38 hours; (F,G) culminants, 42-48 hours.
Effects of Gα1 on cAMP activation of adenylyl and guanylyl cyclases
The GC activity in wild-type cells is activated immediately upon stimulation, peaks around 10 seconds, and is down to basal levels by 30 seconds after receptor stimulation. gα1 null cells display a reproducible delay in the time at which the cGMP concentration reaches maximum levels, by peaking at 20 seconds; however, the peak activation, as determined by the amount of cGMP accumulated, was unaltered. Interestingly, we reproducibly observed that the kinetics of cGMP accumulation appeared to follow the wild-type curve as best as can be measured up to ∼5-7 seconds, but the rate of increase in cGMP accumulation decreased after this point in time, suggesting that there is a biphasic activation or adaptation of the response (Fig. 8A).
Activation of guanylyl cyclase in whole cells. cGMP levels of whole cells at various times after cAMP addition were quantitated as described in Materials and methods. Error bars in A and B represent s.e.m., n=6. (A) Guanylyl cyclase activity of wild-type KAx-3 and gα1 null strains. (B) Guanylyl cyclase activity of gα1 null and gα1 null cells complemented by an extrachromosomal vector containing Gα1 cDNA driven by the Gα1 promoter. Inset, western blot of whole cell lysates, prepared as described in Materials and methods, stained for Gα1. In the insert, A, wild-type cells; B, complemented gα1 null cells; C, gα1 null cells. (C) GC activation in cell lines expressing mutant Gα1 proteins. See text and Materials and methods for details on the cell lines. n=6-12, s.e.m. for all cell lines in C was less than 10% of each data point.
Activation of guanylyl cyclase in whole cells. cGMP levels of whole cells at various times after cAMP addition were quantitated as described in Materials and methods. Error bars in A and B represent s.e.m., n=6. (A) Guanylyl cyclase activity of wild-type KAx-3 and gα1 null strains. (B) Guanylyl cyclase activity of gα1 null and gα1 null cells complemented by an extrachromosomal vector containing Gα1 cDNA driven by the Gα1 promoter. Inset, western blot of whole cell lysates, prepared as described in Materials and methods, stained for Gα1. In the insert, A, wild-type cells; B, complemented gα1 null cells; C, gα1 null cells. (C) GC activation in cell lines expressing mutant Gα1 proteins. See text and Materials and methods for details on the cell lines. n=6-12, s.e.m. for all cell lines in C was less than 10% of each data point.
To determine whether this delay in adaptation of GC can be complemented by Gα1 protein, gα1 null cells were transformed with a Gα1 expression vector. Clones that had normal phenotype and a level of expression of Gα1 similar to that of wild-type cells, as detected by western blotting, were selected. These complemented cells displayed similar kinetics of GC activation to wild type (Fig. 8B), indicating that the change in kinetics of the GC response is due to the loss of Gα1 function. Expressing Gα1(Q206L) and Gα1(G45V) mutant subunits result in a GC activation profile with parallel kinetics to those of wild-type cells. However, the maximum level of cGMP produced was ∼35% lower in overexpressors and Gα1(Q206L) cells and 50% lower in Gα1(G45V) cells, when compared to wild-type cells (Fig. 8C). This is in contrast to the increase in cAMP-mediated stimulation of GC seen in Gα2 overexpressors (Okaichi et al., 1992).
The effect of Gα1 mutations on the activation of AC in response to cAMP stimulation at the aggregation stage was also examined (Fig. 9). Wild-type and gα1 null cells showed similar kinetics and amount of cAMP accumulation. Overexpression of Gα1(Q206L) did not significantly affect the activation of AC. The cells overexpressing wild-type Gα1 showed a slightly slower initial kinetics of activation. In the Gα1(G45V) cells, AC activation was delayed, with the response peaking around 3 minutes after cAMP addition, and the total accumulation of cAMP was significantly reduced.
Activation of adenylyl cyclase in whole cells. Levels of cAMP in whole cell lysates were measured at various times after stimulation with 2′dcAMP as described in Materials and methods. n=6-12, s.e.m. for all cell lines was less than 10% of each data point.
In Dictyostelium, the activation of AC has been proposed to be regulated by βg rather than the α subunit Gα2 in response to cAMP stimulation (Okaichi et al., 1992; Lilly et al., 1993). To determine whether the Gα1(G45V) phenotype might result from a sequestering of βg subunits by excess Gα1(G45V), crude membrane fractions were isolated from wild-type, Gα1(G45V) and Gα1(Q206L) strains 60 seconds after cAMP stimulation (at the time of peak AC activity), and immunoprecipitated using an anti-β antibody (Lilly et al., 1993). About 98% of the total β subunit and ∼50% of the total Gα1 subunit in cells were recovered in the membrane fraction, with the remainder of the Gα1 subunit present in the supernatant. When membranes containing equal amounts of β subunit were immunoprecipitated with the anti-β antibody, ∼2.3 and 1.3 times more Gα1 subunit were found in the immunoprecipitate when membranes were used from Gα1(G45V) and Gα1(Q206L) cells, respectively, compared to wild-type cells (Table 1). These results indicate that, within the limitations of our experimental resolution, approximately twice as much β subunit was associated with Gα1 subunit in Gα1(G45V)-expressing cells compared to Gα1(Q206L) or wild-type cells. Since the anti-Gα1 antibody (Kumagai et al., 1989) used to quantitate Gα1 cannot distinguish between the wild-type and mutant forms of Gα1, we could not determine the relative ratio of the wild-type and mutant subunits co-precipitated with β subunit.
DISCUSSION
Regulatory role of Gα1 during morphogenesis
Our results suggest that Gα1 plays an important function in controlling Dictyostelium development. The morphological phenotype of the gα1 null strain, although subtle, suggests that Gα1 plays a regulatory role in ecmB cell regulation in the slug and stalk differentiation during culmination. This conclusion is emphasized by the spatial expression of Gα1 during development, where it displays a remarkable similarity to the distribution of the prestalk-specific gene ecmB, especially in the tips of slugs. Furthermore, the strains overexpressing wild-type Gα1 and Gα1(Q206L) subunits showed very aberrant morphogenesis starting at the slug stage. These strains also produced culminants in which part of the slug remains undifferentiated in that it does not express the ecmA and ecmB prestalk- and the SP60/cotC prespore-specific markers. Moreover, Gα1-overexpressors do not detectably express ecmB in the apical central core of prestalk AB region of the slug, while scattered expression indicative of ALCs is observed. Since the morphology of cells expressing the putative constitutively acting mutation Q206L is similar to cells overexpressing the wild-type subunit, we expect that the observed phenotypes were probably due to increased Gα1 activity rather than an indirect effect such as sequestering βg subunits (see below). In addition, overexpression of the Gα subunit Gα2 on either the Act15 promoter or the Gα2 promoter (which is expressed in prestalk and prespore cells from different promoters) does not have a visible effect on development (Okaichi et al., 1992), suggesting that the observed phenotypes for overexpressing Gα1 are probably not due to a general excess of Gα subunits that might compete for components with another Gα subunit. We cannot exclude the possibility that while Gα2 does not appear to compete with other G protein functions, Gα1 might. We also note that overexpression of either Gα4 or Gα7 gives completely distinct developmental phenotypes, again suggesting that the phenotypes of Gα1 and Gα1(Q206L) over expressors are specific for Gα1 (Hadwiger and Firtel, 1992; Wu et al., 1994).
A significant effect of Gα1 overexpression is the apparent loss of expression of ecmB in the anterior cone of cells that have been designated prestalk AB cells without the loss of ecmA expression in the anterior of the slug (prestalk A cells) and without the apparent loss of ecmB in the ALC population or in the stalk tube. The apparent co-localization of Gα1 expression to that of ecmB (they both independently stain the central core of prestalk AB cells) suggests a direct effect of Gα1 signaling pathways on stalk cell differentiation. Work from J. Williams and colleagues has shown that ecmB promoter regulation is complex, with a negative regulatory sequence that restricts expression to the prestalk AB population by inhibiting expression in prestalk A cells (Ceccarelli et al., 1991; Harwood et al., 1992). Deletion of this sequence element from the ecmB promoter releases the inhibition and the promoter is then also expressed in prestalk A cells as well. Insertion of this element into the ecmA promoter, which is expressed in prestalk A and prestalk AB cells, prevents expression of ecmA in prestalk A cells but allows expression in prestalk O cells. It has been postulated that the negative regulation is released in prestalk AB by PKA activation, possibly via high levels of intracellular cAMP, in prestalk AB but not prestalk A cells (Harwood et al., 1992, 1993). Inhibition of PKA function in these cells from the expression of a dominant negative regulatory subunit blocks culmination and ecmB expression. The pathways regulating ecmB expression and prestalk AB cell differentiation are complicated since overexpression of the catalytic subunit of PKA from either the ecmA or ecmB promoters from the time these promoters are first induced in the aggregate results in a block of development, suppresses rather than stimulates ecmB expression and results in an inhibition rather than stimulation of stalk cell differentiation (Mann and Firtel, 1993; Hopper et al., 1993). The apparent block of or significant reduction in ecmB expression in prestalk AB cells in Gα1 overexpressors suggests that Gα1 may be a negative regulator of this pathway and prestalk AB cell differentiation and is not required for the activation of this pathway. This is a model that is consistent with the absence of a strong null phenotype. It is possible that Gα1 could be modulating the activation of AC and cAMP production or the regulation of PKA activity. Alternatively, Gα1 may be a regulator of the ammonia signaling pathway, which has been postulated to be the repression signal in ecmB expression (Harwood et al., 1993). Since ammonia levels are high during early culmination and decrease during the later stages, this could explain the negative effect of Gα1 overexpression on ecmB expression in prestalk AB cells. The absence of a strong null phenotype may indicate that pathways other than the one controlled through Gα1 may be controlling prestalk AB differentiation.
In a parallel study, we showed that, upon the onset of culmination, there is a rapid, extended rise in 1,2-diacylglycerol (DG) production that is rapidly activated by overhead light, which is an inducing signal for culmination (Cubitt et al., 1993). We showed that gα1 null cells and cells overexpressing Gα1 from the Gα1 promoter have a specific effect on this rise in DG levels, with the level of DG being substantially reduced in gα1 null cells while the increase in DG occurs precociously and is observed over an extended period of time in Gα1 overexpressors. These findings suggest a possible role for Gα1 in regulating this response. The fact that DG accumulation is significantly and differentially affected in both the null and overexpressor strain is consistent with Gα1 playing an essential role in regulating morphogenesis.
At present, we do not know the mechanism by which the presumed guanine nucleotide exchange for Gα1 is activated. Since the overexpression of the wild-type and Q206L mutation leads to similar phenotypes, we presume that the form of Gα1 that is responsible for these phenotypes is Gα1GTP and that this directly activates downstream effectors. Whether a specific seven transmembrane domain regulates this process or whether it is regulated by another mechanism is not known. We note that it has been shown that Gαi3 is associated with the Golgi and can directly mediate exocytosis, suggesting that mechanisms other than classical cell surface seven-pass receptors may activate Gα subunits (Stow et al., 1991). When Gα2 is overexpressed, we do not see an activation of downstream effectors in the absence of cAMP receptor stimulation (Okaichi et al., 1992), possibly because this free α subunit may not readily exchange guanine nucleotides without upstream stimulation. It is possible that guanine nucleotide exchange for Gα1 is mediated by a process other than the interaction of a ligand with a cell surface receptor. If this were the case, then overexpression of the subunit would lead to high levels of the activated protein and a constitutive stimulation of the down-stream pathways. Identification of these putative effectors will be essential to our understanding of these processes.
There are distinct differences between the observed growth and developmental phenotypes seen for the overexpression of Gα1 on its own promoter presented here versus the very initial observations reported earlier for the overexpression of Gα1 from the Act15 promoter (Kumagai et al., 1989), which is expressed at high levels during growth and early development and at significantly reduced levels during the multicellular stage. Previously, we reported that growing cells were very large and multinucleate and that there was some abnormal morphogenesis during the later stages. Using the Gα1 promoter, we saw only minor indications of the large multinucleate phenotype, presumably due to the difference in vegetative expression between the two different promoters. Presumably, the observed differences in the developmental phenotypes are due to directing high level Gα1 expression to the same cells that express it in vivo rather than ectopically expressing the protein in all cells.
Regulatory role of Gα1 during early development
Our data indicate that Gα1 may play a role in regulating the GC pathway during aggregation. Activation and adaptation of the GC system has been shown to be regulated by G-proteins (Small et al., 1987; Schulkes et al., 1992). The Gα2 subunit appears to function as the activator of this pathway and dominant active mutants of this Gα subunit result in the constitutive adaptation of AC, GC and PLC (Kumagai et al., 1989; Okaichi et al., 1992). However, the possible function of other Gα subunits in regulating the possible activation or adaptation of GC is not known. Of interest is the shift of the activation profile observed in gα1 null cells. Whether this indicates a biphasic activation profile that is mediated by more than one pathway, one of which requires Gα1, cannot be distinguished in these experiments. Since overexpression of the wild-type and Gα1(Q206L) does not alter the kinetics of activation, but leads to a decrease in the peak cGMP response, it is possible that Gα1 is regulating the adaptation of GC. What is unclear is whether this biochemical phenotype is associated with the developmental phenotype during culmination or the effect on ecmB expression during the slug stage.
There are several possible mechanisms underlying the morphological phenotypes of the Gα1(G45V) mutant and the effect on the activation of AC. Overexpression of either wild-type Gα1 or the Gα1(Q206L) mutation only minimally reduces the activation of AC and has less of an effect than over-expression of Gα2. This effect of Gα1 overexpressors and Gα1 (Q206L) mutants could be indirect due to sequestering of βg subunits, which are thought to be the activating G protein subunit for AC (Okaichi et al., 1992; Lilly et al., 1993). Gα1(G45V), in contrast, has a much more severe effect on aggregation and the activation of AC. This phenotype is similar to that seen when cells are given high concentrations of caffeine, which inhibits intracellular cAMP signaling (Seigert and Weijer, 1989). We note that there is an increased fraction of β subunit associated with Gα1(G45V) compared to wild-type Gα1 or Gα1(Q206L), although this is only by a factor of two. Studies on mammalian Gαs shows that the equivalent mutation (G49V) is impaired in its ability to form a high affinity complex with Mg2+/GTPgS, may not undergo conformational changes and fails to dissociate from βg (Lee et al., 1992). This may account for its inability to show GTPgS and hormonal activation in membranes from cells expressing the subunit, although it does show stimulation of AC in vitro in a reconstituted system (Graziano and Gilman, 1989; Masters et al., 1989). It is possible that the effects of Gα1(G45V) on both the activation of AC and morphogenesis may be indirect and may block ligand-stimulated activation or other upstream activation of endogenous Gα1 or possibly Gα2. Alternatively, Gα1(G45V) may directly sequester βg subunits. Further clarification of the pathways leading to the activation of AC in this system and other possible pathways that are affected during aggregation are necessary to understand the effect of this mutation.
Acknowledgements
The authors would like to thank Jeanette Rusch for help with the initial stages of this project and to Alexandra M. Clark for technical assistance. We are indebted to R. Gunderson and P. N. Devreotes for generously supplying the anti-Gα1 antibody and to P. Lilly and P.N. Devreotes for supplying the anti-β subunit antibody. S. D. was supported, in part, by a USPHS post-doctoral training grant PHS HD07398. This work was supported by USPHS grants to R. A. F.