(Received for publication, December 5, 1996)
From the Department of Biology, Center for Molecular
Genetics, University of California, San Diego, La Jolla, California
92093-0634, the
Department of Biological
Chemistry, Johns Hopkins University School of Medicine, Baltimore,
Maryland 21205, and the ** MRC Laboratory for Molecular Cell Biology,
University College London, Gower Street,
London WC1E 6BT, United Kingdom
The chemoattractant cAMP, acting through serpentine cAMP receptors, results in a rapid and transient stimulation of the Dictyostelium mitogen-activated protein kinase ERK2 activity (1). In this study we show that other pathways required for aggregation, including Ras and cAMP-dependent protein kinase (PKA), are important regulators of ERK2 activation and adaptation. By examining both the level and kinetics of activation and adaptation of ERK2, we show that Ras is a negative regulator of ERK2. Activated Ras or disruption of a Ras GAP gene results in reduced ERK2 activation whereas disruption of putative Ras GEF or expression of dominant negative Ras proteins have a more rapid, higher, and extended activation. CRAC, a PH domain-containing protein required for adenylyl cyclase activation, is also required for proper ERK2 adaptation. PKA overexpression results in a more rapid, higher level of activation, whereas pka null cells show a lower level but more extended ERK2 activation. Furthermore, we show that constitutive expression of PKA catalytic subunit bypasses the requirement of ERK2 for aggregation and later development, indicating that PKA lies downstream from ERK2 and that ERK2 may regulate one or more components of the signaling pathway required for mediating PKA function, possibly by directly regulating PKA R or a protein controlling the intracellular level of cAMP.
A cell's ability to respond to an extracellular signal involves
both the activation of pathways and subsequent adaptation in which the
cells are no longer fully responsive to the extracellular signal. This
allows the cells to properly regulate the level and extent of the
signaling pathway as well as adapt to changing environmental conditions. Well known examples are the pheromone pathway in yeast and
the mammalian cell -adrenergic receptor (2, 3). In Dictyostelium, aggregation is mediated by a periodic
activation and adaptation of pathways regulated by G protein-coupled
cARs1 that bind extracellular cAMP as the
ligand (4-6). In response to extracellular cAMP, guanylyl and adenylyl
cyclase activities are rapidly stimulated and then adapted. If cAMP
levels are kept constant the cells remain nonresponsive, whereas
removal of the cAMP, which in vivo occurs through its
hydrolysis by an extracellular phosphodiesterase, allows the pathways
to de-adapt within ~5 min.
MAP kinase cascades regulate a variety of intracellular responses
through the activation of cell surface receptors (1, 2, 7). During the
preaggregation and aggregation stages of Dictyostelium
development, the MAP kinase ERK2 is activated through cAMP and folate
chemotactic receptors (1).2 Stimulation
with extracellular cAMP results in a >40-fold increase in ERK2
activity that peaks at ~50 s and thereupon adapts, reaching its basal
level within 5-8 min. ERK2 activation requires the G protein-coupled
cAMP receptors that mediate cAMP-stimulated adenylyl and guanylyl
cyclase activation, chemotaxis, and gene expression. cAMP-mediated
activation of ERK2 is partially independent of G2, the G
subunit
required for cAMP stimulation of adenylyl and guanylyl cyclases, all
other identified G
subunits, and the only known G
subunit (1).
erk2 null cells are unable to aggregate due to a defect in
cAMP stimulation of adenylyl cyclase (8). Adenylyl cyclase (ACA) and
other signaling components known to be required for the activation of
adenylyl cyclase are found at normal levels in erk2 null
cells. ERK2 is also required for cell type differentiation and
morphogenesis during the multicellular stages of development as
determined by the analysis of an ERK2 temperature-sensitive mutation
(9).
Here we examine the regulation of ERK2 activation and adaptation. Using mutants in the Ras pathway, including constitutively active and dominant negative Ras mutations, a Ras GAP and a Ras GEF null mutation, we first show that Ras acts as a negative regulator of ERK2 activity. Mutations in the cytosolic regulator of adenylyl cyclase, CRAC, result in a loss of proper ERK2 adaptation. Examination of cells in which the cAMP-dependent protein kinase PKA catalytic subunit is disrupted or overexpressed indicates that PKA also plays an important role in these processes. Furthermore, we show that overexpression of PKA in erk2 null cells bypasses the requirement of ERK2 for aggregation and that these cells form normal fruiting bodies. Our results suggest complex pathways involving Ras, adenylyl cyclase and coupled components, and PKA affecting the level and extent of ERK2 activation and that ERK2 functions to control the level of activation of PKA by regulating the activation of adenylyl cyclase and other components in this pathway.
Activation of ERK2 was measured in cells competent to aggregate (aggregation-competent cells), the stage at which the stimulation of adenylyl cyclase is maximal as described previously (1), using in gel assays to quantitate ERK2 activity. Cells were pulsed for 4 h with 30 nM cAMP prior to harvesting, washing, and then activation by cAMP. Samples are normalized for the amount of protein. Because of the possibility that the level of renaturation and activity may not be identical between gels, each gel contains four normalization samples, which allow gels to be compared and quantitated internally. In addition, we quantited the amount of ERK2 protein in samples using the anti-ERK2 antibody (1). The levels of ERK2 were not detectably different in any samples except the ERK2 gene disruption (erk2 null cells) (data not shown).
Cell Transformation and Cell CultureDictyostelium cells were transformed, grown vegetatively, and developed in suspension culture as described previously (1).
The putative Ras GEF
Aimless (AleA) is required for proper aggregation of
Dictyostelium cells (10). aleA null cells show normal activation of guanylyl cyclase but have a significantly reduced
level of adenylyl cyclase activation in response to cAMP (10), as has
been shown for erk2 null cells (8). To determine whether
Aimless, and by implication a Ras protein, was involved in the
regulation of ERK2 activation, we examined the kinetics and level of
activation of ERK2 in an aimless null strain using an in gel
assay in which an ERK2 substrate, myelin basic protein (MBP), is added
into the SDS gel. After size fractionation, the proteins are denatured
and renatured and then assayed for MBP-phosphorylating activity in the
presence of [-32P]ATP (Ref. 1; see "Materials and
Methods"). As shown in Fig. 1 and as described
previously (1), wild-type cells show a rapid activation of the 42-kDa
ERK2 that peaks at ~50 s and then rapidly decreases to basal levels
within 5-8 min. In aimless null cells, ERK2 shows a higher
basal level of ERK2 activity and a more rapid, higher, and more
extended activation of ERK2 than that observed in wild-type cells.
Exogenous expression of Aimless in these cells, which
complements the aggregation defect, also complements the effects of the
Aimless loss-of-function mutation on the kinetics and level
of ERK2 activation (Fig. 2A).
To further investigate the possible role of Ras in regulating the activity of ERK2, we examined the effect of overexpressing dominant interfering or dominant active forms of the RasD protein, one of five known Dictyostelium Ras proteins (see "Discussion"), in stably transformed strains from the actin (Act15) promoter, which is expressed throughout growth and development. Expression of the dominant negative RasD(S17N) on an integrating vector, which gives a very high level of expression, results in a more rapid and extended level of activation of ERK2 than observed in wild-type cells (Fig. 1). This is qualitatively similar to that observed in the aimless null strain, except that the profile is even more extended, with a dramatic effect on adaptation. These observations are strengthened by experiments in which expression of RasD(D57Y) protein (11) from an extrachromosomally replicating vector, which gives a lower level of expression, also results in a more rapid and higher level of activation of ERK2 (Fig. 2B; legend to Fig. 2). These results are similar to those of the aimless null cells and suggest that Ras negatively regulates ERK2 activity. As would be expected from this conclusion, overexpressing the dominant active RasD(Q61L) results in a substantially reduced and slightly delayed activation of ERK2 compared with wild-type cells (Fig. 1). Disruption of the gene encoding a Ras GAP, which should result in an increase in the level of Ras·GTP in these cells (ddrasgap1 null cells) (12), also yields a reduced activation of ERK2 when compared with wild-type cells (Fig. 2B), consistent with the effect of RasD(Q61L) on ERK2 activation. These results suggest that the Ras signaling pathway is a negative regulator of ERK2 MAP kinase activity (see "Discussion"). Overexpression of RasD(Q61L) in the aimless null strain results in a suppression of the high level of ERK2 activation in aimless null cells and a delay in the kinetics of activation as observed in wild-type cells expressing RasD(Q61L) (Fig. 1).
Components of the Adenylyl Cyclase/PKA Pathway Are Required for Activation and Adaptation of ERK2Because ERK2 activity is
required for cAMP receptor-mediated activation of adenylyl cyclase, we
examined components of the adenylyl cyclase pathway for possible
involvement in a regulatory role in modulating ERK2 activity. The PH
domain-containing protein CRAC is required for the activation of
adenylyl cyclase ACA and is thought to interact with G (which
activates adenylyl cyclase in Dictyostelium) and adenylyl
cyclase (13, 14). In crac null cells, the initial kinetics
of ERK2 activation in response to cAMP were more rapid than seen in
wild-type cells and the activity remained at an elevated level for >8
min, a time at which the wild-type level of ERK2 activity had returned
to the basal level (Fig. 2, C and D).
Complementation of the crac null mutation with an
Act15/CRAC expression vector restored the ERK2 activation
profile to that of wild-type cells (Fig. 2C). Interestingly,
in aca null cells, the kinetics of ERK2 activation were
indistinguishable from those observed for wild-type cells (Fig.
2B), suggesting that the ability of cells to properly
stimulate adenylyl cyclase activity is not essential for the normal
adaptation of ERK2 but that CRAC and/or a CRAC-associated protein may
be essential for the adaptation.
PKA is also known to be important in regulating different aspects of aggregation. pka null cells (cells in which the PKA catalytic subunit has been disrupted) (15) are aggregation-deficient due to an inability to relay cAMP (16-19). Previous analyses showed that pka null cells or cells expressing a dominant negative PKA regulatory subunit do not express the required aggregation-stage adenylyl cyclase ACA (19, 20); however, expression of ACA in pka null cells, while restoring chemoattractant activation of adenylyl cyclase, does not restore the ability of the cells to aggregate (19). To examine whether aspects of the activation/adaptation pathway may require PKA, cAMP stimulation of ERK2 activity was examined in pka null cells. Both the kinetics and level of stimulation were similar to those found in crac null cells (Fig. 2, C and D). The maximum level of activation was consistently slightly lower than that of wild-type cells, and the adaptation of the activity was suppressed with the activity being high for more than 8 min. In addition, the basal level of ERK2 activity in unstimulated pka null cells was higher than that in wild-type cells (Fig. 2C). In pka null:ACA cells, the kinetics of activation and adaptation of ERK2 are similar to those in pka null cells (data not shown), indicating that ACA, which is not expressed in pka null cells (19), was not the limiting factor in these cells for maximal ERK2 activation. The reduced level of ERK2 activation observed in aca null cells (Fig. 2B), is possibly due to the absence of endogenous cAMP that would be expected to result in a reduction in the level of active endogenous PKA. Because the lack of PKA activity altered ERK2 activation and adaptation, we examined the effect of overexpressing PKA in wild-type cells from the Act15 promoter, which results in a high level of constitutive PKA activity (21). As shown in Fig. 2B, these cells exhibit a more rapid increase, a higher maximum level of activation, and a more extended period of activation compared with wild-type cells. This stimulation is independent of ACA, as this pattern is observed in aca null cells expressing PKA (Fig. 2C).
Constitutive PKA Bypasses the Requirement of ERK2 for AggregationWhen plated for development, erk2 null cells are aggregation-deficient and are unable, even in response to exogenous cAMP, to proceed through development (8, 9). However, the erk2 null cells overexpressing the PKA catalytic subunit aggregate and produce fruiting bodies with wild-type morphology (data not shown), indicating that these cells are able to bypass the requirement of ERK2 for aggregation (see "Discussion").
The pathways required for the activation of adenylyl cyclase in
Dictyostelium are significantly more complex than the
textbook paradigm. The direct activation in vivo in response
to cAMP stimulation or in vitro in the presence of GTPS
requires the G
subunit and CRAC (13, 14, 22). In vivo
stimulation also requires ERK2 and the putative Ras GEF (8, 10). The
kinetics of ERK2 activation/adaptation are similar to those of adenylyl
cyclase, suggesting a model in which ERK2 activity is required at the
time of adenylyl cyclase stimulation (1). In contrast to the situation in metazoans where MAP kinase pathways can be stimulated by Ras (23),
our results suggest that Ras is a negative regulator of chemoattractant
receptor-mediated ERK2 activation. ERK2 activation is delayed and the
level significantly reduced in cells overexpressing RasD(Q61L), whereas
activation is enhanced in aimless null cells and cells
overexpressing RasD(S17N) or RasD(D57Y). Moreover, reduced ERK2
activation is also observed in a Ras GAP null strain, and the elevated
ERK2 activation in aimless null cells is suppressed by
RasD(Q61L). At present, we cannot reconcile the effects of Ras pathway
mutants on ERK2 activation with those on adenylyl cyclase activation.
The effects of the RasD(S17N) are more severe than those in
aimless null cells, yet aimless null cells have a
broader range of phenotypes than the RasD(S17N) overexpression cells:
RasD(S17N)-expressing cells aggregate, whereas aimless cells
do not and overexpression of RasD(Q61L) does not complement the
aggregation defect of aimless cells. We expect, therefore, that one or more Ras proteins must positively regulate other pathways required for aggregation and that differences in Ras protein function account for different effects on chemotaxis and adenylyl cyclase activation, both required for aggregation. Indeed, cells overexpressing an activated RasG do not aggregate or activate adenylyl cyclase (24).
RasD is only one of five known Dictyostelium Ras genes (25-27), some of which, including RasD, are essentially identical to mammalian Ras proteins within the highly conserved N-terminal domain, whereas others show nonconserved amino acid substitutions in this region, including the effector domain. Our results on the effect of Ras on ERK2 activation are in direct conflict with those of Knetsch et al. (11), who show that cells expressing the dominant active RasD(G12T) have a high basal level of ERK2 and that activity decreases with cAMP stimulation but also only show a <5-fold cAMP-stimulated ERK2 activation. We cannot account for this discrepancy but emphasize that our results are internally consistent with the analysis being performed on multiple mutants affecting Ras function. It is possible that the Knetsch group examined a different kinase; we have confirmed that we are examining ERK2 using a purified, in vivo activated His-tagged ERK2.3
The activation and adaptation pathways also involve CRAC. However, this requirement is not dependent on the ability to synthesize cAMP or activate cAMP-dependent protein kinase. When CRAC translocates to the plasma membrane after cAMP stimulation it may interact with upstream components of the ERK2 MAP kinase pathway to down-regulate ERK2 activation. Our results also present evidence for a role of PKA in regulating ERK2 activity. pka null cells have a higher basal activity and a lower level, but more extended time, of activation. Overexpression of PKA yields a more rapid, extended, and higher level of activation, suggesting that PKA may be essential for maintaining low basal levels and maximally stimulating ERK2 activation. The more extended activation in pka null cells may be associated with other roles of PKA, which is known to be required for multiple aspects of aggregation and later multicellular development (19, 28).
Lastly, we showed that cells overexpressing the PKA catalytic subunit,
which has been shown to lead to constitutive PKA activity (21),
suppresses the erk2 null phenotype. This suggests that ERK2
is upstream from PKA and functions in part to control the activation of
PKA, which is required for aggregation. In wild-type cells, cAMP
produced through the activation of adenylyl cyclase would activate PKA.
We expect that a component of this pathway may be a direct substrate
for ERK2. ERK2 might function to directly regulate the PKA catalytic or
regulatory subunit to mediate the interaction of these subunits or
another protein that may control the level of cAMP in cells or other
aspects of the pathway. As constitutive PKA activity suppresses the
erk2 null phenotype, we expect that such a protein would lie
downstream from ERK2 and upstream of PKA. A model depicting the
possible inter-relationships between different components in the ERK2
pathway is shown in Fig. 3. Further analyses should
continue to elucidate the mechanisms controlling chemoattractant
receptor-mediated signaling during aggregation.