Autophosphorylation suppresses whereas kinase inhibition augments the translocation of protein kinase Calpha in response to diacylglycerol.

We have seen that protein kinase Calpha (PKCalpha) is transiently translocated to the plasma membrane by carbachol stimulation of neuroblastoma cells. This is induced by the Ca2+ increase, and PKCalpha does not respond to diacylglycerol (DAG). The unresponsiveness is dependent on structures in the catalytic domain of PKCalpha. This study was designed to investigate if and how the kinase activity and autophosphorylation are involved in regulating the translocation. PKCalpha enhanced green fluorescent protein translocation was studied in living neuroblastoma cells by confocal microscopy. Carbachol stimulation induced a transient translocation of PKCalpha to the plasma membrane and a sustained translocation of kinase-dead PKCalpha. In cells treated with the PKC inhibitor GF109203X, wild-type PKCalpha also showed a sustained translocation. The same effects were seen with PKCbetaI, PKCbetaII, and PKCdelta. Only kinase-dead and not wild-type PKCalpha translocated in response to 1,2-dioctanoylglycerol. To examine whether autophosphorylation regulates relocation to the cytosol, the autophosphorylation sites in PKCalpha were mutated to glutamate, to mimic phosphorylation, or alanine, to mimic the non-phosphorylated protein. After stimulation with carbachol, glutamate mutants behaved like wild-type PKCalpha, whereas alanine mutants behaved like kinase-dead PKCalpha. When the alanine mutants were treated with 1,2-dioctanoylglycerol, all cells showed a sustained translocation of the protein. However, neither carbachol nor GF109203X had any major effects on the level of autophosphorylation, and GF109203X potentiated the translocation of the glutamate mutants. We, therefore, hypothesize that 1) autophosphorylation of PKCalpha limits its sensitivity to DAG and 2) that kinase inhibitors augment the DAG sensitivity of PKCalpha, perhaps by destabilizing the closed conformation.

The protein kinase C (PKC) 1 isoforms constitute a family of closely related serine/threonine kinases that are involved in pathways regulating a large number of cellular processes such as proliferation, apoptosis, differentiation, migration, and neuronal signaling. The PKC family is generally subgrouped in classical, novel, and atypical isoforms based on structural similarities and sensitivity to different activators. The classical PKCs (PKC␣, -␤I, -␤II, and -␥) are regulated by both Ca 2ϩ and by diacylglycerol (DAG), whereas the novel PKCs (PKC␦, -⑀, -, and -) are insensitive to Ca 2ϩ and considered to be primarily regulated by DAG. Atypical PKCs (PKC and -) are neither affected by Ca 2ϩ nor by DAG (1)(2)(3).
When classical or novel PKCs are activated, they frequently translocate to the plasma membrane. The translocation is primarily mediated via two classes of PKC domains, C1 and C2 domains. These domains are targeted to the membrane by DAG and Ca 2ϩ , respectively. Structural and mutational analyses of PKC␣ in vitro have led to the proposal of a model in which an initial Ca 2ϩ increase leads to the interaction of the PKC␣ C2 domain with Ca 2ϩ and a loose binding of the enzyme to lipids. This interaction leads to a disruption of binding of the C2 domain to the C1a domain, which thereby becomes available for DAG and can be inserted into the membrane (4,5). According to this model the C1 domain is inaccessible for ligands unless the PKC structure has been at least partially loosened by the interaction of Ca 2ϩ with the C2 domain. We have recently seen that the catalytic domain contributes to the inaccessibility of the C1 domain in the absence of Ca 2ϩ (6). Thus, a tethering of the C1a domain by the C2 domain is conceivably also dependent on intramolecular bindings that involve the catalytic domain.
There are also other mechanisms that may contribute to the regulation of PKC␣ translocation. PKC␣ is frequently autophosphorylated on two C-terminal sites, the turn motif (Thr-638 in human PKC␣) and the hydrophobic site (Ser-657 in human PKC␣). The phosphorylation of these sites has been suggested to be part of the regular maturation process of PKC molecules. Newly synthesized PKC molecules are initially phosphorylated on a threonine residue in the activation loop, a step that is necessary for the catalytic activity of the enzyme. Thereafter, the autophosphorylation sites are phosphorylated, completing the maturation of the enzyme (7,8). The phosphorylation of the turn motif has been shown to be crucial for PKC␤II (9) but not for PKC␣ (10) activity, whereas phosphorylation of the hydrophobic site is not required for PKC activity (11,12). The autophosphorylation has also been suggested to be of importance for the stability of the PKC molecule (13,14).
There are also indications that autophosphorylation is a means to directly regulate PKC function. There are findings showing that kinase inhibition or mutation of an autophosphorylation site prolongs the plasma membrane localization of PKC (15,16), and a recent paper demonstrates that only non-phosphorylated PKC␣ activates phospholipase D1 (17). A regulatory role for autophosphorylation is further supported by several studies showing that autophosphorylation is influenced by a number of exogenous stimuli (18 -20).
Thus, it is becoming clear that autophosphorylation of PKC can be a regulated event and that it has significant impact on PKC function. However, it is far from clear how autophosphorylation influences the properties of PKC. We have previously seen that the catalytic domain of PKC␣ limits the durability of its translocation upon carbachol stimulation and its sensitivity to DAG (6). This led us to raise the novel hypothesis that regulating the DAG sensitivity may be the mechanism whereby autophosphorylation influences PKC function. Our data demonstrate that this is indeed the case; autophosphorylation does act as a regulatory switch determining the DAG sensitivity of PKC␣.

EXPERIMENTAL PROCEDURES
Plasmids-Expression vectors encoding full-length PKC␣, PKC␤I, PKC␤II, PKC␦, and kinase-dead PKC␦, all fused to EGFP, have previously been described (21)(22)(23). PKC␣, PKC␦, and PKC⑀ were mutated in the autophosphorylation sites using the QuikChange mutagenesis kit (Stratagene). The primers were designed to alter threonine or serine in the turn motif to glutamate (␣T638E and ⑀T710E) or alanine (␣T638A and ␦S643A) and to alter serine in the hydrophobic site to glutamate (␣S657E and ⑀S729E) or alanine (␣S657A and ␦S662A). Similarly, both autophosphorylation sites were mutated to glutamate (␣EDM and ⑀EDM) or alanine (␣ADM and ␦ADM). The primers are listed in Table  I. All constructs were sequenced to confirm that they contained the right mutation.
Confocal Microscopy of PKC Translocation-Cells were examined by confocal microscopy on the day after transfection. The coverslips were washed twice with buffer H (20 mM Hepes, 137 mM NaCl, 3.7 mM KCl, 1.2 mM MgSo 4 , 2.2 mM KH 2 PO 4 , 1.6 mM CaCl 2 , 10 mM glucose, pH 7.4) and mounted on the heated stage of a Nikon microscope. The cells were examined with a Bio-Rad Radiance 2000 confocal system using a 60ϫ lens (numerical aperture 1.4) with an excitation wavelength at 488 nm and emission filter 515HQ30. Images were captured every 5 or 10 s, and five images were taken before the addition of one of the following: 1 mM carbachol, 100 or 5 M 1,2-dioctanoylglycerol (DOG), or 2 M GF109203X or Me 2 SO as a control. DOG and GF109203X were solubilized in Me 2 SO at 100 and 10 mM, respectively. Carbachol was solubilized in water. The final Me 2 SO concentration was 1.2 l/ml for experiments with the combination of 100 M DOG and 2 M GF109203X and respective control. In other experiments the Me 2 SO concentration was lower. LaserPix software was used to analysis the fluorescence intensity in the cytosol. Because the cells changed shape during the experiment, only a limited section of the cytosol could be quantified. The area to be quantified was carefully chosen to not contain any plasma membrane or intracellular components throughout the entire experiment. The average of the intensities of the five images before the addition of stimulus was used as a base-line value and set to 100%.
Cell Culture and Transfection-Human neuroblastoma cells SK-N-BE(2)C were maintained in minimal essential medium supplemented with 10% fetal bovine serum and 100 IU/ml penicillin and 100 g/ml streptomycin. All cell culture reagents were from Invitrogen. For confocal imaging, cells were trypsinized and seeded with a density of 300,000 cells/35-mm cell culture dish on glass coverslips. 24 h after seeding, cells were transfected in serum-free medium with 3 l of LipofectAMINE 2000 (Invitrogen) and 1.6 g of DNA according to the supplier's protocol. For Western blot analysis, 100-mm cell dishes were seeded with 1.5 ϫ 10 6 cells/dish and transfected with 9 l of Lipo-fectAMINE 2000 and 4.8 g of DNA.

Inhibition of the Kinase Activity Makes PKC␣ Respond with a Sustained
Translocation to the Plasma Membrane after Stimulation with Carbachol-We have previously seen that stimulation of SK-N-BE(2)C neuroblastoma cells with the acetylcholine analogue carbachol gives rise to a transient translocation of PKC␣ (6). The relocation to the cytosol was dependent on the presence of the catalytic domain of PKC␣. The isolated regulatory domain of PKC␣ responded with a sustained translocation upon carbachol stimulation.
To explore whether inhibition of the catalytic activity of PKC influences the translocation of PKC␣, SK-N-BE(2)C cells expressing PKC␣-EGFP were treated with 2 M GF109203X before stimulation with carbachol. The localization of PKC␣-EGFP was thereafter examined by confocal microscopy. In cells pretreated with GF109203X for 30 s, there was an immediate and sustained translocation of PKC␣-EGFP to the plasma membrane after stimulation with 1 mM carbachol ( Fig. 1, B, E, and H). This contrasts the effect in control cells pretreated with Me 2 SO, in which stimulation with carbachol gave a transient membrane translocation of PKC␣-EGFP ( Fig. 1, A, D, and G).
The effect of GF109203X led us to hypothesize that the kinase activity of PKC␣ suppresses its plasma membrane localization and leads to a relocation of the protein to the cytosol. To further investigate the role of the kinase activity, the translocation of a kinase-dead PKC␣-EGFP, i.e. PKC␣ with a mutation of Lys-368 in the ATP-binding site to an arginine, was investigated. This PKC␣ mutant responded with a translocation that was sustained upon carbachol stimulation ( Fig. 1, C, F, and I). To estimate the time course of the translocation, the fluorescence intensity of EGFP in the cytosol was quantified. Because the cells change their shape during the experiment, it was impossible to quantify the intensity in the plasma membrane and only possible to quantify a limited section of the cytosol in each cell. These areas were carefully chosen not to  PKC␣-T638A  GACAGCCCGTCTTAGCACCACCTGATCAGCTGG  PKC␣-T638E  GGACAGCCCGTCTTAGAACCACCTGATCAGCTGG  PKC␣-S657A  GTCTGATTTTGAAGGGTTCGCGTATGTCAACCCCC  PKC␣-S657E  CCAGTCTGATTTTGAAGGGTTCGAGTATGTCAACCCCC  PKC␦-S643A  GGCGCGCCTCGCCTACAGCGACAAGAACC  PKC␦-S662A  GCATTCGCTGGCTTCGCCTTTGTGAACCCC  PKC⑀-T710E GGGAAGAGCCGGTACTCGAGCTTGTGGACGAAGC PKC⑀-S729E GGAGGAATTCAAAGGTTTCGAGTACTTTGGTGAAGACCTGATGCC contain any plasma membrane or nuclear components during the entire time course of the experiment. The analysis confirmed that PKC␣WT-EGFP pretreated with GF109203X and that the kinase inactive variant PKC␣KD-EGFP responded with a sustained translocation after carbachol stimulation, whereas carbachol induced a transient translocation of PKC␣WT-EGFP to the plasma membrane in cells pretreated with Me 2 SO ( Fig. 1, J and K). Treatment with the PKC Inhibitor GF109203X after a Period of Stimulation with Carbachol Induces a Relocation of PKC␣ to the Plasma Membrane-Pretreatment with GF109203X suppressed the relocation of PKC␣-EGFP to the cytosol after a carbachol-stimulated plasma membrane translocation. We next wanted to investigate whether GF109203X also could reverse the relocation to the cytosol once it had occurred. To do this, SK-N-BE(2)C cells expressing PKC␣-EGFP were stimulated with carbachol, and after the translocation, when PKC␣-EGFP had returned to the cytosol, 2 M GF109203X was added (Fig. 2). Stimulation with carbachol led to a translocation of PKC␣-EGFP to the membrane within 10 s (Fig. 2, A and B), and when the protein had returned to the cytosol after 90 s (Fig. 2C), GF109203X was added to the cells. This induced a re-translocation of PKC␣-EGFP to the plasma membrane, where it also remained throughout the rest of the experiment (Fig. 2D).
Inhibition of PKC␤I, PKC␤II, and PKC␦ Kinase Activity Leads to a Sustained Translocation of the Proteins to the Plasma Membrane after Stimulation with Carbachol-We have previously seen that carbachol induces a transient translocation to the plasma membrane of PKC␤II and to a lesser extent of PKC␦ and a sustained translocation of PKC⑀ in neuroblastoma cells (6). To investigate if kinase inhibition also influences the translocation of other PKC isoforms, SK-N-BE(2)C cells expressing PKC␤I, PKC␤II, or PKC␦ fused to EGFP were treated with 2 M GF109203X and thereafter stimulated with carbachol. The localization of EGFP-tagged PKC isoforms was thereafter followed by confocal microscopy (Fig. 3). As seen before, carbachol elicited a transient translocation of PKC␤I and PKC␤II, and this response was not altered by pretreatment with vehicle, Me 2 SO. On the other hand, exposure to GF109203X made both PKC␤ isoforms respond with a sustained translocation after carbachol stimulation, although in a few cells a small amount of PKC␤II relocated to the cytoplasm. The novel isoform PKC␦ showed a weaker response to carba- chol stimulation, and some cells did not respond at all. However, treatment with GF109203X before stimulation with carbachol potentiated the magnitude of the PKC␦ translocation.
To further investigate the contribution of kinase inhibition to the response, the carbachol-stimulated translocation of kinasedead PKC␤I, PKC␤II, and PKC␦ was also analyzed. The kinase-dead PKC␤I, PKC␤II, and PKC␦ all responded to carbachol and translocated to the plasma membrane and remained there throughout the experiment. Although full-length PKC␦ sometimes did not respond to carbachol, kinase-dead PKC␦ translocated in all cells. The results were supported by a quantification of the fluorescence intensity of EGFP in the cytoplasm. Thus, a potentiation and prolonging of a translocation by kinase inhibition is an effect that is common for several PKC isoforms.
Stimulation with Carbachol Leads to a Sustained Translocation of Non-phosphorylated PKC␣-The potentiating effect of kinase inhibition on PKC translocation has previously been observed (15,16), and this has raised the hypothesis that autophosphorylation of PKC would be a negative feed-back signal and lead to a relocation of the protein to the cytosol. We, therefore, set out to investigate if the effects obtained by kinase inhibition could be due to absent autophosphorylation. The C-terminal autophosphorylation sites, the turn motif (PKC␣Thr-638) and the hydrophobic site (PKC␣Ser-657), were therefore mutated to alanines or glutamates to mimic the nonphosphorylated and phosphorylated states, respectively. SK-N-BE(2)C cells were transfected with vectors encoding PKC␣WT-EGFP, PKC␣KD-EGFP, and constructs encoding PKC␣ with mutated autophosphorylation sites. Cells were stimulated with 1 mM carbachol, and the localization of the PKC␣ variants was examined by confocal microscopy (Fig. 4).
Wild-type PKC␣ and the glutamate mutants (PKC␣T638E, PKC␣S657E, and PKC␣EDM, the mutant with both sites mutated) all either responded with a transient translocation to the plasma membrane or did not respond at all. There were some variations in terms of the proportion of cells in which PKC␣ translocated, but a sustained translocation was not observed except for two cells expressing PKC␣S657E.
On the other hand, in 100% of the cells that expressed the kinase-dead construct, PKC␣KD, carbachol induced a sustained translocation of the protein to the plasma membrane. The same effect was also seen for PKC␣ADM, a mutant where both autophosphorylation sites were mutated to alanine and, thus, mimicked the non-phosphorylated PKC␣. PKC␣ mutated to alanine either in turn motif, PKC␣T638A, or hydrophobic site, PKC␣S657A, consistently responded to carbachol, and translocated to the plasma membrane after stimulation. Contrasting the glutamate mutants and wild-type PKC␣, a sustained translocation of these mutants was frequently observed, but unlike PKC␣ADM these mutants, sometimes returned to the cytosol after stimulation with carbachol. Thus, autophosphorylation of the hydrophobic site and the turn motif conceivably contributes to the transient nature of PKC␣ translocation. However, mutation of the autophosphorylation sites to glutamate does not make PKC␣ resistant to translocation signals.
Mutation of the Autophosphorylation Sites in PKC␦ and PKC⑀ Do Not Alter Their Sensitivity to Carbachol-We have previously seen that PKC⑀ responds with a sustained translocation and that PKC␦ responds with a weak and transient translocation upon carbachol stimulation (6). To see if the effect of autophosphorylation influences the responsiveness of these isoforms as well, we created vectors encoding PKC␦ with both autophosphorylation sites mutated to alanine (PKC␦ADM) and PKC⑀ with both sites mutated to glutamate (PKC⑀EDM). However, the mutations did not augment the weak response of PKC␦ or suppress the sustained PKC⑀ response (data not shown). Thus, in contrast to PKC␣, the sensitivity of PKC␦ and PKC⑀ to carbachol-generated second messengers is not regulated by autophosphorylation.
Neither Carbachol nor GF109203X Has Major Effects on the Amount of Autophosphorylated PKC␣-The previous experiments could have the implication that autophosphorylation of PKC␣ upon its activation at the plasma membrane is a mechanism to cause its relocation to the cytosol and thereby to turn off the PKC signal. If this hypothesis is true, it would be expected that the levels of autophosphorylation would increase after carbachol stimulation and decrease by treatment with GF109203X. We, therefore, analyzed by Western blots the levels of phosphorylation of the turn motif and the hydrophobic site. Stimulation of cells with carbachol for different time points did not have a major impact on the phosphorylation of either the hydrophobic site (Fig. 5A) or the turn motif (Fig. 5B). There was a slight tendency to increased phosphorylation of the hydrophobic site. The same result was obtained when the phosphorylation of overexpressed PKC␣-EGFP was analyzed (data not shown). Thus, carbachol does not increase the autophosphorylation, and this is, therefore, not likely the mechanism that triggers a relocation of PKC␣ to the cytosol during carbachol stimulation.
Similarly, we wanted to investigate whether treatment with GF109203X affected the phosphorylation. Treatment of cells with 2 M GF109203X did not decrease the phosphorylation of the hydrophobic site (Fig. 5C). Phosphorylation of the turn motif (Fig. 5D) was elevated in the presence of Me 2 SO, but it was clear that GF10203X did not influence the phosphorylation. The same pattern of GF109203X effect was seen when analyzing phosphorylation of PKC␣-EGFP (data not shown).
We also investigated the autophosphorylation of kinase-dead PKC␣ and the PKC␣ variants with autophosphorylation sites mutated to alanine (Fig. 5, E and F). As expected, both kinasedead PKC␣ and the alanine double mutant were devoid of phosphate on the two autophosphorylation sites, whereas wildtype PKC␣ was phosphorylated on both sites. The S657A mutant was at least partially phosphorylated on Thr-638, whereas the Thr-638 was phosphorylated on neither site, confirming previous results with PKC␤II, which shows that phosphorylation of the turn motif is a prerequisite for phosphorylation of the hydrophobic site (9).

PKC␣ with Autophosphorylation Sites Mutated to Glutamate Responds to Carbachol with a Sustained Translocation when
Pretreated with GF109203X-The Western blot analyses demonstrated that GF109203X does not influence the autophosphorylation of PKC␣, suggesting that an inhibition of autophosphorylation is not the mechanism through which GF109203X potentiates the translocation of PKC␣. To further confirm that the GF109203X effect is independent of the phosphorylation status of the turn motif and the hydrophobic site, we investigated the effect of GF109203X on the translocation of PKC␣ with both sites mutated to glutamate. Stimulation with carbachol induced an immediate translocation of PKC␣EDM to the plasma membrane followed by relocation to the cytosol (Fig. 6,  A, D, and G), and this was not influenced by Me 2 SO (Fig. 6, B,  E, and H). On the other hand, pre-exposure to 2 M GF109203X (Fig. 6, C, F, and I) made PKC␣EDM respond with a sustained translocation upon carbachol stimulation. Thus, the GF109203Xinduced block of PKC␣ relocation to the cytosol also takes place when the turn motif and hydrophobic site are mutated to glutamate, mimicking the phosphorylated state.
Inhibition of the Catalytic Activity Makes PKC␣ Sensitive to Diacylglycerol-The previously mentioned experiments indicate that autophosphorylation does not make PKC␣ resistant to receptor-stimulated translocation in general. We, therefore, hypothesized that kinase inhibitors and reduced autophospho- Cells transfected with vector encoding PKC␣WT were pretreated with GF109203X for 30 s before stimulation with DOG (C-F). Images show unstimulated cells (A, C, E, G, I), PKC␣WT exposed to 100 M DOG for 300 s (B), PKC␣WT treated with GF109203X and exposed to 100 M DOG for 300 s (D) or 5 M DOG for 480 s (F), and PKC␣KD exposed to 100 M DOG for 300 s (H) or 5 M DOG for 300 s (J). rylation may make PKC␣ sensitive to DAG. We have previously seen that carbachol induces a transient increase in intracellular Ca 2ϩ levels and a sustained increase in DAG (6) and that PKC␣ is insensitive to DAG (Fig. 7, A and B). An increased sensitivity to DAG may, therefore, explain the observed findings with kinase inhibition. To investigate this we pretreated cells with 2 M GF109203X for 30 s and thereafter stimulated them with 100 M DOG. This led to a sustained translocation of PKC␣-EGFP to the plasma membrane (Fig. 7, C and D). DOG did not induce translocation of PKC␣ when cells were pretreated with Me 2 SO (data not shown). Similar to what was obtained when PKC␣WT was pretreated with GF109203X, 100 M DOG induced a rapid and sustained translocation of PKC␣KD-EGFP to the membrane (Fig. 7, G and H). Both PKC␣WT in the presence of GF109203X (Fig. 7, E and F) and PKC␣KD (Fig. 3, I and J) also responded to 5 M DOG. PKC␣KD responded to DOG within a few seconds, whereas PKC␣WT, inhibited with GF109203X, had a slower response to DOG. This was particularly evident when 5 M DOG was used.
PKC␣ with Autophosphorylation Sites Mutated to Alanine Is Sensitive to DAG-Because treatment with DOG induced a translocation of kinase-inactive PKC␣, but not of wild-type PKC␣, to the plasma membrane, we next wanted to investigate if DOG had an effect on the PKC␣ mutants in which the autophosphorylation sites were mutated to alanine (Fig. 8). In a majority of the cells, PKC␣T638A and PKC␣S657A both started to translocate within 1 min to the plasma membrane, where they remained throughout the entire experiment. Nearly 100% of the cells responded to the stimulation with 5 M DOG. PKC␣ADM responded with an immediate translocation in almost all cells, and this also persisted throughout the experiment.

DISCUSSION
This study reports a novel role for PKC␣ autophosphorylation as a regulatory switch determining the DAG sensitivity of PKC␣. We, therefore, propose a model in which the V5 domain participates in intramolecular interactions that maintain an inaccessibility of the C1a domain of PKC␣ to DAG. The interaction is dependent on autophosphorylation and can also be disrupted by direct interactions with PKC inhibitors such as GF109203X.
Our initial experiments demonstrated that inhibition of PKC␣ either with application of GF109203X or by a kinaseinactivating mutation of the ATP-binding site resulted in a sustained translocation upon carbachol stimulation. The results may be explained by a model in which an activation of PKC␣ results in autophosphorylation, which in turn induces a relocation of PKC␣ to the cytosol. This could function as a negative feedback signal to turn off the activated PKC␣. Such a hypothesis is supported by papers which have demonstrated that PKC inhibitors prolong the agonist-induced membrane localization of classical PKC isoforms (15,16). Our data demonstrate that the phenomenon is common for several PKC isoforms. Both PKC␣ as well as the PKC␤ isoforms are similarly affected by kinase inhibition, and there was also a similar, albeit smaller, effect for PKC␦.
If the hypothesis is true, it would be expected that mutation of the C-terminal autophosphorylation sites to alanine, mimicking a dephosphorylated state, would abolish the relocation of PKC to the cytosol during agonist stimulation. This is indeed what we found in our experiments. It would also be expected that mutation of the sites to glutamate, mimicking the phosphorylated state, would result in a PKC that was already desensitized and insensitive to translocation signals. However, our experiments showed that PKC␣ with both sites mutated to glutamate responded with at least equal prominence as did wild-type PKC␣, and pretreatment with GF109203X made the mutant respond with a sustained translocation. Furthermore, treatment with GF109203X did not decrease the amount of phosphorylated PKC␣ despite the fact that it abolished the relocation of PKC␣ to the cytosol. Thus, there does not seem to be a correlation between phosphorylation of the C-terminal autophosphorylation sites and insensitivity to a receptor-stimulated translocation signal. It is, therefore, clear that autophosphorylation is not a mechanism for desensitization of the PKC translocation.
Our data instead point to other important regulatory functions of autophosphorylation. Both kinase-inactivated PKC␣ and PKC␣ with the autophosphorylation sites mutated to alanine responded to exogenous application of DOG. This contrasts the reaction of wild-type PKC␣, which is insensitive to DOG. We propose that this difference explains why the former mutants display a sustained translocation upon carbachol stimulation, whereas the translocation of wild-type PKC␣ is transient. We have previously seen that after the addition of carbachol there is a rapid and transient increase in intracellular Ca 2ϩ levels, but the DAG levels accumulate gradually and remain elevated for at least several minutes (6). We, therefore, propose that autophosphorylation of the C-terminal sites serves as a regulatory switch determining whether PKC␣ is sensitive to DAG or not. It is in this context interesting to note that several stimuli and extracellular factors influence the amount of classical PKC isoforms that are autophosphorylated (14, 18 -20). Our data, therefore, highlight the possibility that regulated phosphorylation of the hydrophobic site and the turn motif serves to alter the DAG sensitivity of PKC␣ and possibly also of other classical isoforms.
The question remains as how phosphorylation of the Cterminal sites alters the DAG sensitivity of PKC␣. Previous studies demonstrate that DAG primarily binds the C1a (24) domain and that this interaction is suppressed by an intramolecular interaction between the C2 and the C1a domain in the inactive PKC␣ (4, 5). Our previous study (6) showed that the catalytic domain is necessary to maintain the inaccessibility of the C1a domain, and this study indicates that the autophosphorylation sites are necessary for the effect. It is in this context interesting to note the proposed model for the intramolecular interactions in PKC␤ between a receptor for activated C-kinase-binding site and a pseudo-receptor for activated C-kinase motif (25,26). The interaction is putatively mediated via the V5 and the C2 domain. Because the autophosphorylation sites are located in the V5 domain of PKC␣, our data further support a model in which the V5 domain participates in intramolecular interactions that maintain the PKC molecule in a closed conformation. The strength of these interactions may be influenced by the presence of phosphate on the autophosphorylation sites.
Besides influencing the translocation response, phosphorylation of the turn motif and the hydrophobic site have also been suggested to increase the stability of PKC and to prevent it from accumulating in a detergent-insoluble fraction (9,12). Several studies indicate that the down-regulation of PKC that follows upon prolonged activation of the enzyme is mediated via an initial dephosphorylation of the autophosphorylation sites (14,27,28), perhaps followed by a ubiquitination of the enzyme. However, there are also studies showing that primarily the fully phosphorylated PKC molecule is degraded without an initial dephosphorylation (29). Our experiments show that the mutant with both autophosphorylation sites mutated to alanine had a similar localization pattern as wild-type PKC␣, demonstrating that the propensity of the non-phosphorylated variants to accumulate in detergent-insoluble fractions does not translate into an apparent alteration of its localization in the cell. Furthermore, when analyzing the expression of the alanine mutants, we could not detect any degradation products of PKC␣. This indicates that the rate of degradation of nonphosphorylated PKC␣ is slow in the neuroblastoma cells.
The kinase-dead PKC␣ variant was, as expected, not autophosphorylated, which may explain its DOG sensitivity and sustained response to carbachol stimulation. However, the inhibitor GF109203X potentiated the translocation response, but it did not reduce the autophosphorylation. Furthermore, it prolonged the translocation of PKC␣ with the autophosphorylation sites mutated to glutamate, and it could reverse the relocation of PKC␣ to the cytosol once it had occurred (Fig. 2). It is, therefore, likely that GF109203X enhances the translocation of PKC␣ through other means than by suppression of the autophosphorylation of PKC␣. We speculate that the inhibitor, by interacting with the ATP-binding site, destabilizes the closed conformation of PKC and thereby exposes the C1a domain and makes it available for interaction with DAG.
Both exogenous application of PKC inhibitor and overexpres-sion of a kinase-inactivated PKC mutant are widely used to suppress PKC activity in order to investigate the role for PKC in different cellular processes. Our data clearly show that both methods not only block PKC activity but also have other effects on the protein, such as increasing the amount of PKC molecules, albeit most likely catalytically inactive enzymes, at the plasma membrane. Given the fact that PKC has been shown to exert several effects independently of its kinase activity (21, 30 -34), the use of inhibitors or dominant-negative variants may actually potentiate these PKC effects instead.