(Received for publication, March 31, 1997, and in revised form, April 24, 1997)
From the Vollum Institute, Oregon Health Sciences University, Portland, Oregon 97201
The A kinase-anchoring protein AKAP79 coordinates
the location of the cAMP-dependent protein kinase (protein
kinase A), calcineurin, and protein kinase C (PKC) at the postsynaptic
densities in neurons. Individual enzymes in the AKAP79 signaling
complex are regulated by distinct second messenger signals; however,
both PKC and calcineurin are inhibited when associated with the
anchoring protein, suggesting that additional regulatory signals must
be required to release active enzyme. This report focuses on the
regulation of AKAP79-PKC interaction by calmodulin. AKAP79 binds
calmodulin with high affinity (KD of 28 ± 4 nM (n = 3)) in a
Ca2+-dependent manner. Immunofluorescence
staining shows that both proteins exhibit overlapping staining patterns
in cultured hippocampal neurons. Calmodulin reversed the inhibition of
PKCII by the AKAP79(31-52) peptide and reduced inhibition by the
full-length AKAP79 protein. The effect of calmodulin on inhibition of a
constitutively active PKC fragment by the AKAP79(31-52) peptide was
shown to be partially dependent on Ca2+.
Ca2+/calmodulin reduced PKC coimmunoprecipitated with
AKAP79 and resulted in a 2.6 ± 0.5-fold (n = 6)
increase in PKC activity in a preparation of postsynaptic densities.
Collectively, these findings suggest that Ca2+/calmodulin
competes with PKC for binding to AKAP79, releasing the inhibited kinase
from its association with the anchoring protein.
Protein phosphorylation of intracellular substrates by kinases and phosphatases controls many aspects of cellular function (1). As the individual components of many signaling pathways have been identified, it has become apparent that the regulation of phosphorylation events is achieved at many levels. Although soluble second messengers control the activity state of kinases and phosphatases, other factors influence where and when these enzymes have access to their substrates. Localization of kinases and phosphatases adds a measure of selectivity to their action as it restricts which phosphorylation events occur in response to a particular stimulus. Consequently, several prominent classes of serine/threonine protein kinases and phosphatases are compartmentalized through interactions with anchoring or targeting proteins (2-4). For example, protein phosphatase 1 associates with targeting subunits that localize the catalytic subunit and adapt catalytic activity to preferentially dephosphorylate certain substrates (5, 6).
An emerging family of proteins called AKAPs1 (A Kinase Anchoring Proteins) binds to the regulatory subunit of PKA, localizing the kinase to particular cellular locations, primed for activation by cAMP (for review, see Ref. 7). Some AKAPs bind more than one signaling enzyme. For example, AKAP79 not only associates with PKA but also binds protein phosphatase 2B, calcineurin, and protein kinase C, whereas another anchoring protein, gravin, binds PKA and PKC (8-11). These multivalent binding proteins serve as scaffolds for multienzyme signaling complexes. We have proposed that these signaling scaffolds preferentially control the phosphorylation of selected substrates such as ion channels and cytoskeletal components through integration of signals from distinct second messengers such as Ca2+ and cAMP (12). However, regulation of these signaling complexes is not fully understood. Although cAMP releases the catalytic subunit of PKA from AKAP79, it is evident that additional regulatory mechanisms must be involved to release inhibited calcineurin and PKC from their association with the anchoring protein.
In this report we have focused on the regulation of PKC targeting by AKAP79. Although the PKC family of at least 11 phospholipid-dependent enzymes is activated in response to the generation of diacylglycerol and in some cases Ca2+, most of the isoforms have nearly identical substrate specificities (13, 14). Differential localization may contribute to the specificity of PKC action, as a combination of subcellular fractionation and immunocytochemical studies have demonstrated that certain PKC isoforms are found in different cellular compartments (15, 16). Although localization of PKC primarily involves protein-lipid interactions, it is now apparent that PKC-targeting proteins are also important in directing the location of the kinase to particular parts of the cell (14). There are several classes of PKC-targeting proteins: substrate-binding proteins that bind PKC in the presence of phosphatidylserine (17); receptors for activated C-kinase which are not necessarily substrates for PKC and bind at site(s) distinct from the substrate binding pocket of the kinase (18); and proteins that interact with C-kinase which have been cloned in two hybrid screens using the catalytic core of the kinase as bait (19). We have recently demonstrated that PKC binds AKAP79 in what appears to represent a unique class of PKC-binding proteins. PKC binds AKAP79 in a phosphatidylserine-dependent manner through an amino-terminal basic and hydrophobic sequence and is inhibited by the anchoring protein (10). We now show that Ca2+/calmodulin antagonizes this interaction presumably by competing for association with this region on AKAP79. This provides a Ca2+-dependent regulatory mechanism for releasing the inhibited kinase from its association with the anchoring protein.
Recombinant AKAP79 protein
was expressed in Escherichia coli as described (8). AKAP79
(5 µg) was incubated with calmodulin-agarose (Sigma) (20 µl of
packed beads) in Buffer A (20 mM Tris, pH 7.0, 1 mM imidazole, 1 mM magnesium acetate, 0.05%
w/v Nonidet P-40, 15 mM -mercaptoethanol, 1 mM benzamidine, 2 µg/ml pepstatin, 2 µg/ml leupeptin,
and 100 mM 4-(2-aminoethyl)-benzenesulfonyl fluoride
hydrochloride) containing 0.2 mM CaCl2 for
2 h at 4 °C. The resin was washed five times with Buffer A
containing 5 M NaCl and 0.2 mM
CaCl2 and then three times with Buffer A containing 0.2 mM CaCl2. AKAP79 was eluted following
incubation with 2 mM EGTA for 1 h at 4 °C. The
eluted proteins were boiled for 5 min in SDS-sample preparation buffer,
separated by SDS-polyacrylamide gel electrophoresis (10% gel) (20),
and immunoblotted (21). AKAP79 was detected with affinity-purified
rabbit polyclonal antibody 918I. This procedure was repeated in the
presence of Buffer A containing 0.2 mM EGTA, and proteins
were eluted with Buffer A containing 5 mM
CaCl2.
Biotinylated calmodulin (Life Technologies, Inc.) was coupled to a carboxymethyl dextran IAsys cuvette (Affinity Sensors) via NeutrAvidinTM (Pierce) using standard 1-ethyl-3(3-dimethylaminopropyl) carbodiimide, N-hydroxysuccinimide coupling chemistry (22). Briefly, the cuvette was activated by treating with 0.4 M 1-ethyl-3(3-dimethylaminopropyl) carbodiimide (Pierce), 0.1 M N-hydroxysuccinimide (Pierce) for 15 min and washed extensively with PBST (PBS containing 1 mM Ca2+ and 1 mM Mg2+ (Life Technologies, Inc.) + 0.05% Tween 20 (Amersham)). Excess NeutrAvidin was coupled to the activated cuvette in 10 mM sodium acetate buffer, pH 4.5, for 30 min at room temperature. Uncoupled protein was washed away with PBST and free amines blocked with 1 M ethanolamine (Affinity Sensors). After washing extensively with PBST, 1 µM AKAP79 recombinant protein was added to check that it did not bind nonspecifically to the cuvette surface. After washing with PBST, 2 µg of biotinylated calmodulin was then coupled to the cuvette via the NeutrAvidin. The calmodulin cuvette was washed with 5 M NaCl and PBST and a stable base line was established for 10 min before data collection. All binding experiments were performed with AKAP79 recombinant protein over a range of concentrations from 5 to 100 nM in volumes of 200 µl in PBST. The binding surface was regenerated with short (1 s) pulses of 6 M guanidinium HCl with no decrease in extent measurements for the duration of the experiments which were completed within 1 day. Data were collected over 3-s intervals and were analyzed using the FastfitTM software which was provided with the IAsys instrument. The KD value was confirmed with analysis of extent data plotted versus AKAP79 concentration which yielded an equilibrium constant that was in good agreement with the KD obtained from the rate data.
ImmunocytochemistryCultured neonatal rat hippocampal
neurons grown on coverslips were rinsed with PBS, fixed in 3.7%
formaldehyde (5 min), and extracted in 20 °C absolute acetone for
1 min. Cells were rehydrated in PBS containing 0.2% BSA for 1 h
and then incubated with a mixture containing affinity-purified rabbit
anti-AKAP79 antibody 2503 at 1.5 µg/ml and mouse anti-calmodulin
antibody (Upstate Biotechnology, Inc.) at 1.2 µg/ml in PBS containing
0.2% BSA for 1 h. Coverslips were washed three times with PBS
containing 0.2% BSA and incubated with a mixture of fluorescein
isothiocyanate-conjugated anti-rabbit (1:500) (Molecular Probes) and
Texas Red-conjugated anti-mouse (1:250) (Molecular Probes) secondary
antibodies in PBS containing 0.2% BSA for 1 h. The coverslips
were then washed three times in PBS containing 0.2% BSA and mounted
with Slow Fade Antifade (Molecular Probes). Cells were analyzed with a
Leitz Fluovert FU confocal photomicroscope with a Leitz 40/1.6 N.A.
lens. Specific staining was not detected in control experiments with
secondary antibody alone.
PKCII, from a baculovirus
expression system, was purified as described (23) and was a generous
gift of Dr. Alexandra Newton (University of California, San Diego). PKC
activity was assayed as described (24) in a 40-µl reaction containing
40 mM HEPES, pH 7.5, 10 mM MgCl2, 1 mM dithiothreitol, 100 µM
[
-32P]ATP (500 cpm/pmol) with epidermal growth factor
receptor peptide (VRKRTLPRL) or PKC(19-31S25) peptide (RFARKGSLRQLNV)
as substrates at 30 °C. Assays were performed with or without
activators: in the presence of either 300 µM
CaCl2 and Triton X-100 (0.1% w/v) micelles containing 20 mol % phosphatidylserine (Avanti) and 5 mol % 1,2-dioleoylglycerol
(Avanti), or 2 mM EGTA and Triton X-100 (0.1% w/v), or in
the presence of 300 µM CaCl2 and
phosphatidylserine (20 µg/ml), or 2 mM EGTA. Triton
X-100:phosphatidylserine:diacylglycerol-mixed micelles were prepared as
described (25). PKC
II was diluted in 20 mM Tris, pH 7.9, 1 mg/ml BSA, and 1 mM dithiothreitol. Time course
experiments were performed over 15 min with 10 µM AKAP79 peptide plus or minus 10 µM calmodulin. Purified bovine
calmodulin was a generous gift of Dr. Roger Colbran (Vanderbilt
University). Dose-response curves were generated over an inhibitor
concentration range of 0.1-10 µM AKAP79 protein and
AKAP79(31-52) peptide in the presence or absence of 10 µM calmodulin. PKC activity was assayed in a preparation
of rat forebrain postsynaptic densities, a generous gift of Dr. Roger
Colbran (26), following incubation with or without calmodulin for 15 min. Assays were performed in the presence of PKC activators: 300 µM CaCl2 and Triton X-100 (0.1% w/v)
micelles containing 20 mol % phosphatidylserine and 5 mol % 1,2-dioleoylglycerol and kinase inhibitors PKI(5-25) and KN62. The
assays were performed in the presence and absence of PKC(19-36)
pseudosubstrate peptide (1 µM).
PKCII (30 nmol) was digested
with trypsin (300 ng/ml) (Sigma) in a 15-µl reaction in 20 mM HEPES, pH 7.0, and 1 mM dithiothreitol for 8 min at 30 °C as described (25). The reaction was terminated with
excess soybean trypsin inhibitor (Sigma) and the digested material
placed on ice. Proteolytic products were then assayed for PKC activity
in the presence of CaCl2 (300 µM) or EGTA (2 mM) plus or minus calmodulin over a range of concentrations
of 0.1 to 10 µM AKAP79(31-52) peptide.
A partially purified
preparation of PKC was prepared from rabbit brain as described (27).
AKAP79 protein (10 µg) was incubated with rabbit brain PKC (2 µg)
in the presence or absence of bovine calmodulin in hypotonic buffer (10 mM HEPES, pH 7.5, 1.5 mM MgCl2, 10 mM KCl, Nonidet P-40 (0.5% w/v), 1 mM
benzamidine, 2 µg/ml pepstatin, 2 µg/ml leupeptin, and 100 mM 4-(2-aminoethyl)-benzenesulfonyl fluoride hydrochloride)
containing 1 mM CaCl2 for 1 h at 4 °C. Samples were then incubated with either affinity-purified anti-AKAP79 918I antibodies (4 µg) or preimmune IgG (4 µg) at 4 °C for
18 h. Immune complexes were isolated by the addition of 100 µl
of 10% v/v protein A-Sepharose CL-4B (Sigma) which had been
equilibrated in hypotonic buffer containing 1 mM
CaCl2. Following incubation at 4 °C for 60 min, the
beads were washed three times with hypotonic buffer containing 1 mM CaCl2 and 1 M NaCl and three
times with hypotonic buffer containing 1 mM
CaCl2. The immunoprecipitation experiment was repeated in
the presence of hypotonic buffer containing 0.5 mM EGTA.
For immunoblot analysis, precipitated proteins were eluted by boiling
the washed beads in SDS-sample preparation buffer for 5 min and
separated by SDS-polyacrylamide gel electrophoresis on a 10%
denaturing gel. Proteins were transferred to nitrocellulose and
analyzed by Western blot with an antibody to PKC/
(Transduction Laboratories). For measuring PKC activity associated with the beads,
the washed beads were resuspended in 60 µl of hypotonic buffer
containing 1 mM CaCl2 and assayed for PKC
activity in duplicate. PKC activity was expressed as fold increase over
PKC activity associated with preimmune complexes.
We
have shown previously that PKC associates with AKAP79, and the
principal site of contact lies between residues 31 and 52 on the
anchoring protein (10) (Fig. 1A). This
stretch of alternating basic and hydrophobic residues also resembles a
calmodulin-binding domain (28). As previous studies have shown that
AKAP79 and the bovine homolog AKAP75 are calmodulin-binding proteins
(8, 29), we wanted to determine if binding was
Ca2+-dependent. Immunoblot analysis using
antibodies to AKAP79 demonstrates that AKAP79 binds calmodulin-agarose
in the presence of 0.2 mM Ca2+, and
binding is abolished in the presence of 2 mM EGTA (Fig. 1B). This suggests that AKAP79 binds calmodulin in a
Ca2+-dependent manner.
To investigate further the interaction between AKAP79 and calmodulin
in vitro, the binding affinity of recombinant AKAP79 to
biotinylated calmodulin was measured by surface plasmon resonance (Fig.
2). Binding of AKAP79 to immobilized calmodulin in the
presence of excess Ca2+ was measured over a range of AKAP79
concentrations (5-100 nM) (Fig. 2A). Analysis
of the association rate data showed that the binding was biphasic,
which may result from steric problems or may indicate more than one
site of binding. The association rate constant
(ka 203,029 ± 24,805 M1 s
1, n = 3)
was obtained by plotting the measured kon
versus AKAP79 concentration (Fig. 2B). The
dissociation rate constant (kd
0.0055 ± 0.00043 s
1 n = 3) was
obtained directly from dissociation experiments. These values were used
to calculate a KD of 28 ± 4 nM
(n = 3) for the AKAP79-calmodulin interaction. The
observed dissociation rate value obtained from the single phase fit of
the data is in good agreement with the dissociation rate constant
extrapolated from the plot of kon
versus AKAP79 concentration (Fig. 2B). The equilibrium constant of 28 nM is within the physiological
concentration range for these proteins and is consistent with the idea
that AKAP79 and calmodulin may associate in the cell.
Further evidence for the association of AKAP79 and calmodulin was
provided by immunofluorescence staining of these proteins in cultured
rat hippocampal neurons (Fig. 3). AKAP79 exhibits a
distinct staining pattern concentrated at two subcellular locations: at
the periphery of the cell body and in dendritic bundles within the
neurite extensions (Fig. 3A). Calmodulin staining is
similar, with staining at the periphery of the cell body and in the
neurite extensions, although the calmodulin is more uniformly
distributed throughout the neurites, and there is also staining in the
nucleus of the cell (Fig. 3B). Double labeling experiments
show considerable overlap of staining at the periphery of the cell and
in neurite extensions, demonstrating that these proteins occupy the
same focal plane (Fig. 3C). Collectively, these findings
suggest that AKAP79 interacts with calmodulin, and it is feasible that
this interaction occurs in vivo.
Calmodulin Releases PKC from Inhibition by AKAP79
The region
on AKAP79 which binds PKC resembles a calmodulin binding sequence (Fig.
1A). The idea that calmodulin and PKC may share common
binding determinants suggested the intriguing possibility that
Ca2+/calmodulin may play a role in regulating the
interaction of PKC with AKAP79. We therefore investigated the effect of
Ca2+/calmodulin on the inhibition of PKC activity by
AKAP79. AKAP79 and a peptide encompassing residues 31-52 of the
anchoring protein inhibit PKC activity (10). Addition of
Ca2+/calmodulin to the reaction releases PKC activity so
that it is no longer inhibited by AKAP79 (Fig. 4). Fig.
4A shows a time course of PKC activity. PKC activity
increases in a linear fashion over 15 min. In the presence of the
AKAP79(31-52) peptide (10 µM), PKC activity is
inhibited; however, addition of 10 µM calmodulin prevents
inhibition of PKC activity (Fig. 4A). Since calmodulin alone
does not affect the activity of PKC (data not shown), it is likely that
calmodulin competes for binding to the AKAP79(31-52) peptide.
Similarly, PKC activity is inhibited by AKAP79(31-52) peptide in a
concentration-dependent manner; but when 10 µM calmodulin is added, the peptide no longer inhibits
PKC activity (Fig. 4B). The relief of inhibition by
AKAP79(31-52) in the presence of calmodulin is specific as inhibition
by the pseudosubstrate PKC(19-36) peptide is not affected by
calmodulin (data not shown). When AKAP79(31-52) peptide (100 µM) is present in excess over calmodulin (10 µM), PKC activity is once more inhibited (not shown).
These data suggest a potential mechanism of regulation whereby
Ca2+/calmodulin competes with PKC for binding to the
AKAP79(31-52) peptide relieving the inhibition of PKC activity.
Calmodulin also affects the inhibition of PKC by the recombinant AKAP79 protein (Fig. 4C). Addition of 10 µM calmodulin causes a shift in the dose-response curve such that the inhibition is reduced. Interestingly, this effect is not as dramatic as that seen with the AKAP79(31-52) peptide, which is somewhat surprising given the high affinity interaction between AKAP79 and calmodulin. However, the binding studies with AKAP79 and calmodulin demonstrate biphasic association, suggesting the potential for binding at more than one site. Thus, calmodulin does not fully prevent inhibition under these conditions. This suggests that calmodulin partially relieves the inhibition of PKC by the AKAP by competing with PKC for binding to AKAP79. Interestingly, the anchoring protein inhibits PKC activity more potently in the presence of phosphatidylserine, and this inhibition is relieved slightly with diacylglycerol/phosphatidylserine micelles (Fig. 4D). Inhibition of PKC by the anchoring protein is relieved to a greater extent in the presence of Ca2+/calmodulin with diacylglycerol/phosphatidylserine micelles relative to phosphatidylserine (Fig. 4D). This suggests that Ca2+/calmodulin as well as the second messenger diacylglycerol may coordinate to relieve PKC from inhibition by AKAP79.
Effect of Calmodulin on Ca2+-independent PKC ActivityAs calmodulin binding to AKAP79 requires
Ca2+, we wished to investigate whether relief of PKC
inhibition by the AKAP79(31-52) peptide was
Ca2+-dependent. Since PKCII activity also
requires Ca2+, it was necessary to generate a
Ca2+-independent form of PKC. Limited trypsin digestion of
PKC was used to liberate the constitutively active,
Ca2+-independent catalytic core of the enzyme (Fig.
5A). Inhibition of
Ca2+-independent PKC activity by the AKAP79(31-52) peptide
was the same as the intact kinase either in the presence or absence of Ca2+ (Fig. 5B). This suggests that the peptide
binds at the catalytic core of the kinase, which is consistent with
previous kinetic studies on the mechanism of inhibition by the
AKAP79(31-52) peptide (10). Addition of Ca2+/calmodulin
(10 µM) resulted in complete relief of inhibition of the
constitutively active PKC (Fig. 5B). Addition of calmodulin in the presence of 2 mM EGTA reduced the inhibition,
suggesting that the effect of calmodulin on PKC activity in the
presence of the AKAP(31-52) peptide is partially dependent on
Ca2+.
Ca2+/Calmodulin Reduces PKC Coimmunoprecipitated with AKAP79
Having shown that calmodulin reverses the inhibition of
PKC activity by the AKAP79(31-52) peptide and reduces the inhibition of PKC activity by the full-length protein, we wanted to look more
directly at the effect of calmodulin on PKC binding to AKAP79. Coimmunoprecipitation experiments were performed with an antibody to
AKAP79 where recombinant AKAP79 was incubated with PKC in the presence
or absence of calmodulin (Fig. 6). Immunoblot analysis demonstrates PKC coimmunoprecipitated with AKAP79 in the presence of
either Ca2+ or EGTA (Fig. 6A, lanes 1 and 5). Preimmune serum was used as a control (Fig.
6A, lanes 2, 4, 6, and
8). When Ca2+/calmodulin was present, PKC was no
longer coimmunoprecipitated (Fig. 6A, lane 3),
whereas in the presence of EGTA, calmodulin did not prevent PKC
coimmunoprecipitation (Fig. 6A, lane 7). Thus calmodulin competes with PKC for binding to AKAP79 in a
Ca2+-dependent manner. To obtain a more
quantitative evaluation, complementary experiments measured PKC
activity in immunoprecipitates. Ca2+/calmodulin markedly
reduced PKC activity coimmunoprecipitated with AKAP79 (Fig.
6B). This effect was Ca2+-dependent
as the presence of EGTA/calmodulin did not reduce PKC activity. These
findings support the idea that Ca2+/calmodulin regulates
the interaction between AKAP79 and PKC by displacing PKC from its
association with AKAP79.
Calmodulin Releases PKC Activity from Postsynaptic Densities
On the basis of in vitro experiments, we
developed a working hypothesis that Ca2+/calmodulin
competes with PKC for binding to AKAP79, thus releasing the kinase from
its inhibition by the anchoring protein. This idea is represented
schematically in Fig. 7A. To test this idea in a more physiological context, PKC activity was measured in a
preparation of postsynaptic densities (PSDs) (Fig. 7B). We
have shown previously that AKAP79 is enriched at the PSD (8). To ensure
that we were selectively measuring PKC activity, assays were performed
using the PKC(19-31S25) peptide as substrate in the presence of the
kinase inhibitors, PKI and KN62, to block the activities of PKA and
calmodulin kinase II, respectively. PKC activity under these conditions
was determined as counts inhibited by the PKC pseudosubstrate inhibitor
peptide PKC(19-36). The 2.6 ± 0.5-fold (n = 6)
increase in PKC activity following incubation with calmodulin (Fig.
7B) is consistent with the idea that calmodulin releases the
inhibited kinase from its binding to the anchoring protein.
Collectively these findings support a potential role for
Ca2+/calmodulin in regulating the release of PKC from
anchored sites at the PSD.
The anchoring protein AKAP79 coordinates the location of PKA, PKC, and calcineurin (8-10). Targeting of these three signaling enzymes to submembrane sites such as the PSDs of neurons would ensure that each enzyme is well placed to receive signals transduced across the postsynaptic membrane (Fig. 7A). Although recent mapping studies have provided information on the individual enzyme binding sites on AKAP79, less is known about the regulation of the signaling complex. Our previous studies have shown that all three enzymes are inactive when bound to the anchoring protein (8-10). Although each enzyme responds to distinct second messenger signals generated at the postsynaptic membrane, additional factors contribute to their release and activation. PKA is directly activated following stimulation by its second messenger signal, cAMP (9). In contrast, both PKC and calcineurin are inhibited by AKAP79 even in the presence of their respective activators. Although the signals that release calcineurin from its association with AKAP79 remain to be elucidated, we have found that Ca2+/calmodulin may play a role in regulating the interaction between PKC and AKAP79 (Fig. 7A). This suggests the potential convergence of Ca2+/calmodulin and PKC signaling pathways in the regulation of PKC phosphorylation events at the PSD (Fig. 7A).
In this report we have shown that Ca2+/calmodulin binds AKAP79 with high affinity and influences PKC activity by releasing the kinase from inhibition by the anchoring protein. Our evidence that excess Ca2+/calmodulin overcomes the inhibitory effect of the AKAP79(31-52) peptide on PKC implicates this region as a potential regulatory site on the anchoring protein. This sequence exhibits the hallmarks of a calmodulin binding site in that it is rich in basic and hydrophobic residues (28). This suggests that calmodulin and PKC share the same or overlapping binding determinants on AKAP79. As AKAP79 is eluted from calmodulin-agarose by EGTA this demonstrates that binding is calcium-dependent. Although we could also purify the AKAP on calmodulin-agarose from bovine brain (9), interpretation of this result is complicated as a proportion of the AKAP is likely to be present through its association with calcineurin which is also a calmodulin-binding protein. However, the nanomolar affinity constant (calculated by surface plasmon resonance) and immunocytochemical data are consistent with the idea that AKAP79 and calmodulin interact inside cells.
Kinase activity measurements suggest that calmodulin prevents inhibition of PKC activity by the AKAP79(31-52) peptide but only partially prevents inhibition by the full-length protein. Although this was surprising given the high affinity interaction between AKAP79 and calmodulin, this result may reflect the complex nature of these protein-protein interactions. Since the full-length protein inhibits PKC more potently than the AKAP79(31-52) peptide, we cannot exclude the possibility that there are additional sites of contact on the anchoring protein for PKC which are less affected by calmodulin. Furthermore, calmodulin may bind to more than one site on the anchoring protein. For example, there are three regions on AKAP79 which are rich in basic and hydrophobic residues (8). Indeed, the surface plasmon resonance binding measurements indicated a biphasic association of AKAP79 with calmodulin, suggesting that there may be more than one site of interaction. These factors may contribute to Ca2+/calmodulin being less effective when the full-length anchoring protein was used. The partial relief of inhibition by the full-length AKAP79 protein in the presence of calmodulin may also reflect the activation state of PKC when bound to the anchoring protein. PKC is activated following recruitment to the plasma membrane in response to diacylglycerol and Ca2+ (for Ca2+-dependent isoforms) (30, 31). The interaction of PKC with AKAP79 requires phosphatidylserine, which may adapt the kinase to be in a particular orientation for binding. Interestingly, the anchoring protein inhibits kinase activity more potently when phosphatidylserine is present, and this inhibition is partially reversed with diacylglycerol/phosphatidylserine micelles. Thus, it appears that PKC associates with AKAP79 by a mechanism that involves interactions with phosphatidylserine at the plasma membrane, but the presence of the second messenger diacylglycerol weakens the interaction. Although it is clear that binding of diacylglycerol is not sufficient to trigger PKC release, the combination of diacylglycerol and Ca2+/calmodulin may synergize to release fully active PKC from its association with the anchoring protein. In support of this idea, inhibition of PKC by the full-length anchoring protein is relieved to a greater extent in the presence of Ca2+/calmodulin with diacylglycerol/phosphatidylserine micelles relative to phosphatidylserine.
The integration of Ca2+/calmodulin and PKC signaling pathways is reminiscent of regulation of the MARCKS protein (32, 33). MARCKS binds Ca2+/calmodulin and cross-links actin. Calmodulin binding is dependent on Ca2+ and is prevented by PKC phosphorylation (34). Cross-linking of actin is disrupted by both phosphorylation and Ca2+/calmodulin (35). There are several other examples of integration between PKC and calmodulin, such as neuromodulin (GAP43), neurogranin, and adducin. In each case phosphorylation by the kinase prevents association with calmodulin (36, 37). In contrast, calmodulin binding to AKAP79 does not appear to involve phosphorylation. The PKC binding site on AKAP79 does not contain a phosphorylation site, and phosphorylation does not affect PKC binding to the anchoring protein (10). It therefore seems unlikely that the interplay between Ca2+/calmodulin and PKC is the same as for MARCKS. However, it is striking that there is such a precedence for convergence between Ca2+/calmodulin and PKC signaling events.
Our coimmunoprecipitation experiments clearly show that Ca2+/calmodulin reduces PKC coprecipitating with AKAP79. Presumably Ca2+/calmodulin competes with PKC for binding to the anchoring protein. But does this regulation occur in vivo? The high concentration of calmodulin at the PSD (26) and our evidence for a high affinity interaction between Ca2+/calmodulin and AKAP79 favor a model for Ca2+/calmodulin-dependent antagonism of the AKAP79-PKC interaction. Consistent with this hypothesis are experiments measuring PKC activity from PSD preparations. The idea that the regulatory signal for release comes from a Ca2+-dependent signaling event that does not directly activate the kinase is intriguing. This implies that Ca2+ acts at several places in the release and activation of anchored PKC. For example, calcium influx through channels at the postsynaptic membranes could dictate the release of anchored PKC through calmodulin, independently of kinase activation by lipid signaling, thereby introducing another level of organization for control of PKC phosphorylation events. Analogous to this idea is the finding that calmodulin binding to Ras-related GTP-binding proteins, Kir and Gem, inhibited binding of GTP, leading the authors to suggest that calmodulin-binding motifs may represent an important module regulating protein-protein interactions in signal transduction pathways (38).
In conclusion, the data in this article present evidence for Ca2+/calmodulin regulating the protein-protein interaction between PKC and AKAP79. These studies show that one component of the AKAP79 signaling complex, PKC, is regulated by the concerted action of two different second messengers: Ca2+/calmodulin to release the kinase from the anchoring protein and Ca2+/phospholipid, which is required to stimulate enzyme activity. This type of regulation represents another example of synergism between calmodulin and PKC signaling events. An intriguing aspect of this model is that Ca2+/calmodulin is also involved in activating the calcineurin that is also a component of the AKAP79 complex. Future experiments will be designed to test this model inside the cell.
We are grateful to Dr. A. Newton for
providing purified PKCII, Dr. R. Colbran for preparations of PSDs
and purified calmodulin, J. Engstrom for assistance with confocal
microscopy, and Dr. Z. Hausken for help with surface plasmon resonance.
We thank Dr. M. Dell'Acqua for helpful discussions and colleagues in
the Vollum Institute for a critical evaluation of this manuscript.