|Home | About | Journals | Submit | Contact Us | Français|
Protein kinase C-δ (PKCδ) is a Ser/Thr kinase that regulates a wide range of cellular responses. This study identifies novel in vitro PKCδ autophosphorylation sites at Thr141 adjacent to the pseudosubstrate domain, Thr218 in the C1A–C1B interdomain, Ser295, Ser302, and Ser304 in the hinge region, and Ser503 adjacent to Thr505 in the activation loop. Cell-based studies show that Thr141 and Thr295 also are phosphorylated in vivo and that Thr141 phosphorylation regulates the kinetics of PKCδ downregulation in COS7 cells. In vitro studies implicate Thr141 and Thr295 autophosphorylation as modifications that regulate PKCδ activity. A T141D substitution markedly increases basal lipid-independent PKCδ activity; the PKCδ-T141D mutant is only slightly further stimulated in vitro by PMA treatment, suggesting that Thr141 phosphorylation relieves autoinhibitory constraints that limit PKCδ activity. Mutagenesis studies also indicate that a phosphorylation at Thr295 contributes to the control of PKCδ substrate specificity. We previously demonstrated that PKCδ phosphorylates the myofilament protein cardiac troponin I (cTnI) at Ser23/Ser24 when it is allosterically activated by lipid cofactors and that the Thr505/Tyr311-phosphorylated form of PKCδ (that is present in assays with Src) acquires as additional activity toward cTnI-Thr144. Studies reported herein show that a T505A substitution reduces PKCδ-Thr295 autophosphorylation and that a T295A substitution leads to a defect in Src-dependent PKCδ-Tyr311 phosphorylation and PKCδ-dependent cTnI-Thr144 phosphorylation. These results implicate PKCδ-Thr295 autophosphorylation as a lipid-dependent modification that links PKCδ-Thr505 phosphorylation to Src-dependent regulation of PKCδ catalytic function. Collectively, these studies identify novel regulatory autophosphorylations on PKCδ that serve as markers and regulators of PKCδ activity.
Protein kinase Cδ (PKCδ) is a member of the PKC family of serine/threonine kinases that sit at the crossroads of signal transduction pathways implicated in a wide range of cellular responses (1). PKCδ’s structure consists of a highly conserved C-terminal catalytic domain (consisting of motifs required for ATP/substrate-binding and catalysis) and an N-terminal regulatory domain containing tandem C1A–C1B domains that bind lipid cofactors, a C2 domain that acts as a protein-protein interaction module, and an autoinhibitory pseudosubstrate domain (a sequence lacking a serine/threonine phosphoacceptor site, but otherwise resembling a PKC substrate) that maintains the enzyme in an inactive conformation Like other PKC isoforms, PKCδ is activated by growth factor receptors that promote the accumulation of lipid cofactors such as diacylglycerol (DAG) or by tumor-promoting phorbol esters such as phorbol 12-myristate 13-acetate [PMA]) that anchor the enzyme in an active conformation to membranes. However, recent studies expose phosphorylation at a highly-conserved threonine residue in the activation loop (Thr505) as an additional component of the PKCδ activation mechanism (2). For most PKC isoforms, activation loop phosphorylation is a stable ‘priming’ modification that is mediated by PDK-1 and serves (along with additional phosphorylation events at conserved turn and hydrophobic motifs in the C-terminus) to lock the enzyme in a mature catalytically-competent conformation, poised to be allosterically activated by lipid cofactors (3). In contrast, PKCδ is catalytically active even without Thr505 phosphorylation. Rather, PKCδ-Thr505 phosphorylation has emerged as a dynamically regulated autocatalytic reaction that ‘fine-tunes’ PKCδ’s catalytic function toward selected substrates (4).
While in vitro intramolecular autophosphorylations of brain PKC were described over 20 years ago (5), conventional models of PKC isoform activation do not consider a possible role for autophosphorylation reactions that regulate PKC enzyme activity. Nevertheless, an early study mapped PKC-βII autophosphorylation to Ser16 and Thr17 in the N-terminus, Thr314 and Thr324 in the hinge region, and Thr641 in the C-terminus (6). Thr641 resides in the turn motif, a highly conserved phosphorylation motif in all PKC isoforms. Other PKCβII autophosphorylation sites in the hinge region and N-terminus are not conserved in other PKC isoforms; since these autophosphorylation sites were not identified in subsequent studies that mapped in vivo PKCβII phosphorylation to Thr500 (the activation loop), Thr641 (the turn motif) and Ser660 (the hydrophobic motif (7)), N-terminal and hinge region phosphorylation sites in PKCs were largely ignored in subsequent literature. The single exception is a modeling study from the Newton laboratory which considered a possible role for PKCβII-Ser16/Thr17 autophosphorylation as a mechanism that favors activation by lowering pseudosubstrate domain binding affinity for the catalytic pocket (6;8). Interest in PKC autophosphorylation has recently been rekindled by studies from the Parker laboratory showing that human PKCδ and PKCε autophosphorylate at sites in their variable hinge regions (9;10); autophosphorylation sites in the hinge region of human PKCδ were mapped to Ser299, Ser302, and Ser304 (Fig 1). PKCδ and PKCε autophosphorylation reactions have generally been viewed as markers of enzyme activation. The notion that hinge region autophosphorylation reactions might regulate the catalytic properties of PKCδ has never been considered.
There is recent evidence that PKCδ activity also is controlled through tyrosine phosphorylation. As many as 8 highly conserved tyrosine residues in various regions of PKCδ’s structure have been implicated as targets for regulatory phosphorylations. Our recent studies focused on PKCδ phosphorylation at Tyr311 and Tyr332, residues unique to the hinge region of PKCδ that are phosphorylated in vitro by Src or related Src family kinases (2;4;11). We showed that PKCδ-Tyr311 phosphorylation increases in vivo in PMA- and H2O2-treated cardiomyocytes (2). While PKCδ phosphorylation at tyrosine residues and at the activation loop (at Thr505) generally are viewed as independently regulated events, our recent experiments exposed a novel form of cross-regulation, showing that Src and PKCδ-Tyr311/Tyr332 phosphorylation enhances PKCδ-Thr505 autophosphorylation (2). We also demonstrated that phosphorylations at Tyr311 and Thr505 cooperate to ‘fine-tune’ PKCδ’s enzymology towards cardiac troponin I (cTnI), a physiologically important PKCδ substrate in the heart (4). cTnI is the “inhibitory” subunit of the troponin complex that is critical for Ca2+-dependent regulation of myofilament function (12). cTnI contains three phosphorylation clusters (Ser23/Ser24, Ser43/Ser45, and Thr144) that exert distinct effects on cardiac function. We showed that (1) allosterically-activated PKCδ (in assays with PS/PMA) phosphorylates cTnI at Ser23/Ser24 and that this modification leads to depressed tension at sub-maximum but not maximum [Ca2+] in detergent-extracted single cardiomyocytes, (2) PKCδ becomes a cTnI-Thr144 kinase (and phosphorylates cTnI at both Ser23/Ser24 and Thr144) and that PKCδ depresses maximum tension and cross-bridge kinetics in single detergent-extracted cardiomyocytes when it is tyrosine phosphorylated by Src, and (3) the effect of Src to convert PKCδ into a cTnI-Thr144 kinase is abrogated by PKCδ-Y311F or T505A substitutions (4). These results implicate PKCδ-Tyr311/Thr505 phosphorylation as dynamically-regulated modifications that alter PKCδ enzymology and allow for stimulus-specific control of cardiac mechanics during growth factor stimulation and oxidative stress. Studies reported herein identify additional novel PKCδ autophosphorylation sites that regulate PKCδ activity.
Antibodies were from the following sources: PKCδ-Thr(P)505, PKCδ-Tyr(P)311, Troponin I-Ser(P)23/Ser(P)24 and anti-pTXR, Cell Signaling Technology; PKCδ, Santa Cruz Biotechnology; mouse monoclonal anti-GFP antibody 3E6, Invitrogen. Recombinant human PKCδ was from Calbiochem. The human recombinant full-length His-tagged Src enzyme, expressed in Sf9 cells and purified by sequential chromatography (purity >95% by Coomassie blue staining of the SDS-PAGE gel) was obtained from Invitrogen. Previous studies further validated the purity of this enzyme preparation, showing that  the kinase activity in this preparation is completely blocked by the Src inhibitor PP1 and  this enzyme preparation is detected a single radioactive band corresponding to the autophosphorylated Src protein when it is subjected to in vitro kinase assays with 32P-ATP and then run on a gel (i.e., no other enzymes/substrates are detected even with long exposures of the gel (4)). PMA was from Sigma. 1,2-Dioleoyl-sn-glycerol was from Avanti Polar Lipids, Inc. Other chemicals were reagent grade.
Cardiomyocytes were isolated from hearts of 2 day-old Wistar rats by a trypsin dispersion procedure that uses a differential attachment procedure followed by irradiation to enrich for cardiomyocytes (11). Cells were plated on protamine sulfate-coated culture dishes at a density of 5 × 106 cells/100-mm dish and grown in MEM (Gibco, BRL) supplemented with 10% fetal calf serum for 4 days and then serum-deprived for 24 hr prior to experiments.
For studies of PKCδ phosphorylation in vivo in cardiomyocyte cultures, cardiomyocytes were infected with an adenoviral construct that drives expression of mouse PKCδ. 2 days later, the cultures were incubated for 4–5 h at 37°C in phosphate-free MEM containing [32P]orthophosphate (0.15–0.26 mCi/ml; 4 ml per dish). Cardiomyocytes were then treated for 10 min with vehicle or 5 mM H2O2 and experiments were terminated by aspirating the radioactive medium and rinsing culture dishes with a buffer containing 133 mM NaC1, 4.7 mM KC1, 1.8 mM CaC12, 16.5 mM dextrose, and 20 mM HEPES. Cellular proteins were harvested in homogenization buffer (20 mM Tris-Cl, pH7.5, 0.05 mM EDTA, 0.5 mM DTT, 0.2% Triton X-100, 5 μg/ml aprotinin, 5 μg/ml leupeptin, 5 μg/ml benzamidine, 1 mM PMSF, 5 μM pepstatin A) and PKCδ was immunoprecipitated, separated from other cellular proteins by SDS-PAGE, and subjected to phosphopeptide mapping analysis.
pPKCδ-EGFP (pGFP-PKCδ) was obtained as a generous gift from Dr. Mary Reyland (University of Colorado Health Sciences Center, Denver). The pPKCδ-EGFP construct expresses mouse PKCδ with enhanced GFP fused to its C-terminus. pPKCδT141A-EGFP, pPKCδT141D-EGFP, and pPKCδT295A-EGFP were generated by site-directed mutagenesis according to the manual for the Quick-ChangeSite-directed Mutagenesis Kit (Stratagene). PKCδ expression plasmids were introduced into COS-7 cells by Effectene Transfection Reagent (Qiagen) according to the instruction manual. Cells were grown for 24 hours, lysed in homogenization buffer. Cell extracts were subjected to immunoprecipitation with mouse monoclonal anti-GFP antibody 3E6 (Invitrogen).
In vitro kinase assays were performed with 0.032 units (32.4 ng) of recombinant human PKCδ (Calbiochem) for assays of cTn complex phosphorylation or 0.4 units for peptide mapping studies. Assays were performed in 80 μl of a reaction buffer containing 30 mM Tris-Cl, pH 7.5, 5 mM MgCl2, 10 mM MnCl2, 0.9 mM EDTA, 0.9 mM EGTA, 3 mM DTT, 0.1 mM sodium vanadate, 76 mM NaCl, 23.5% glycerol, 0.006% Brij-35, 0.023% Triton X-100, 0.04 mM phenylmethylsulfonyl fluoride, 0.2 mM benzamidine, 83 μg/ml phosphatidylserine (PS), 175 nM PMA, 4 μg of troponin complex, and [γ-32P]ATP (25 μCi, 97 μM, unless indicated otherwise). Incubations were for 30 min at 30°C in the absence or presence of Src (0.66 units) or lipid cofactors. PS/PMA is included in assays to allosterically activate PKCδ and render it a better substrate for Src-dependent tyrosine phosphorylation (11).
Immunoblotting was performed on cell extracts according to methods described previously or manufacturer’s instructions (11). In each figure, each panel represents the results from a single gel (exposed for a uniform duration); detection was with enhanced chemiluminescence and quantification was by laser scanning densitometry. All results were replicated in at least 3 experiments.
For peptide mapping studies, kinase reactions (see above) were stopped by adding 27 μl of 4X SDS-PAGE sample buffer. Proteins were separated by SDS-PAGE, transferred to nitrocellulose, and the band corresponding to PKCδ was excised from the membrane, cut into small pieces, and treated for 30 min at 37°C with polyvinylpyrrolidone (0.5%, w/v) in acetic acid (100 mM), followed by 5 water washes (to remove the acid) and a 10 min incubation at room temperature in the dark with 100 mM iodoacetate to carboxymethylate PKCδ. Membrane pieces were then washed three times with water and twice with 50 mM ammonium bicarbonate and incubated overnight at 37°C in 60 μl of a buffer containing 42 mM ammonium bicarbonate, 17 μM HCl, and 10 μg sequencing grade trypsin. Digested peptides were eluted from the membrane by sonication, lyophilized, and the residue was reconstituted in 0.1% trifluoroacetic acid and fractionated by RP-HPLC on a Vydac semimicro C18 column (2.1 × 250 mm). Peptides were eluted with a linear gradient from 0.1% trifluoroacetic acid in water to 0.1% trifluoroacetic acid in acetonitrile over 140 min at a flow rate of 1 ml/min. The eluant was monitored at 220 nm and fractions were collected every 30 sec for Cherenkov counting. Fractions containing radioactivity were subjected to LC-MS/MS analysis on a Micromass Q-Tof hybrid quadrupole/time-of-flight mass spectrometer with a nanoelectrospray source. Capillary voltage was set at 1.8kV and cone voltage 32V; collision energy was set according to mass and charge of the ion, from 14eV to 50eV. Chromatography was performed on an LC Packings HPLC with a C18 PepMap column using a linear acetonitrile gradient with flow rate of 200 nl/min. Raw data files were processed using the MassLynx version 4.0 ProteinLynx software with the MaxEnt 3 algorithm. Phosphorylated peptides were verified by manual inspection of MS/MS spectra.
Recombinant PKCδ (human sequence) was phosphorylated in vitro in buffers containing [γ-32P]ATP and PS/PMA. Since these studies were performed as part of an ongoing effort to identify sites for PKCδ phosphorylation by Src, assays were performed without and with active Src. Radiolabeled PKCδ from the in vitro kinase assays was purified by SDS-PAGE, blotted to nitrocellulose, excised from the membrane, and then subjected to digestion with trypsin and fractionation by reverse phase-HPLC. Figure 2 shows that RP-HPLC chromatograms of phospho-peptide fragments derived from in vitro kinase assays with PKCδ, Src, and [γ-32P]ATP contained two distinct radioactive peaks that are not present in the chromatograms from kinase assays with PKCδ alone (Peaks 1 and 4). Sequencing of peptides in these peaks by electrospray ionization (ESI)-LC/MS/MS analysis identified phospho-Tyr313- and phospho-Tyr334-containing fragments (corresponding to phospho-Tyr311 and phospho-Tyr332 in rodent PKCδ, see Figure 1). Peptides containing Tyr64 and Tyr52 also were recovered in peak 4; while these tyrosine residues also have been identified as sites for regulatory phosphorylation, these peptide fragments were not phosphorylated. Sequencing of these and other peaks by ESI-LC/MS/MS analysis achieved ~40% coverage of the PKCδ sequence and identified six novel sites for PKCδ autophosphorylation: (1) Peak 2 contains a peptide fragment that is phosphorylated at Thr218; this residue resides in a phosphorylation motif that is evolutionarily conserved in the C1A–C1B interdomain regions of PKCδ and PKCθ, but not other PKC isoforms (13). (2) Peak 3 contains a peptide fragment that is phosphorylated at Thr141. The Thr141 phosphorylation motif is conserved in PKCη, but not PKCθ or other PKC isoforms. Of note, Thr141 is positioned N-terminal to the pseudosubstrate domain of PKCδ, similar to the position of Ser16Thr17 relative to the pseudosubstrate domain (22RKGALR27) of PKCβII. These results suggest that PKCδ-Thr141 and PKCβII-Ser16Thr17 autophosphorylation reactions might subserve similar regulatory functions. (3) Peak 4 contains a peptide fragment that is phosphorylated at Ser503. Ser503 is adjacent to the activation loop phosphorylation site in the kinase domain of human PKCδ (NIFGES503RAST507); its significance is uncertain, since it is not conserved in mouse PKCδ (NIFGEG501RAST505) or rat PKCδ (NIFGEN501RAST505). (4) Peak 5 contains peptide fragments phosphorylated at three hinge region residues, Thr295, Ser302 and Ser304.
Table 1 provides a summary of the PKCδ autophosphorylation sites detected in these experiments. Several aspects of the results deserve comment. First, the S302DS304ASSEPVGIY313QGFEK peptide fragment was recovered in three separate phospho-peptide mapping experiments as a serine-phosphorylated peptide in peak 5 or as a tyrosine-phosphorylated peptide in peak 4. In each case, the monophosphorylated peptide was detected; we never recovered a dually Ser/Tyr phosphorylated peptide. We also never recovered a dually Ser302/Ser304-phosphorylated peptide. Second, we detected PKCδ phosphorylation at the turn motif (Ser645), but the trypsin digest failed to yield peptides containing either the activation loop (Thr505) or hydrophobic motif (Ser662) phosphorylation sites. While these phosphorylation reactions are readily detected by immunoblot analysis, these regions of PKCδ were not captured by our sequencing methods that achieved only ~40% coverage of the PKCδ sequence. Other PKCδ autophosphorylation sites may also have evaded detection and deserve further analysis. Third, PKCδ autophosphorylation at Thr295 and Ser304 was detected in kinase assays both without and with Src. These results indicate that PKCδ-Tyr313/Tyr334 phosphorylation by Src is not required for PKCδ autophosphorylation at these sites. In contrast, PKCδ autophosphorylation at Thr141, Thr218, Ser503, and Ser645 was detected exclusively in kinase assays with Src. The significance of this finding is uncertain. Our previous studies using immunoblot analysis showed that Src increases PKCδ autophosphorylation at Thr505 and Thr645 (Ref (14) and data not shown). Therefore, an effect of Src to increase PKCδ-Thr505 autophosphorylation might lead to a general increase in PKCδ autophosphorylation throughout the protein. Alternatively, the failure to detect certain phospho-peptides in assays without Src may simply result from a bias inherent in the study; sequencing methods were disproportionately applied to peaks derived from assays with PKCδ plus Src (since PKCδ autophosphorylation was identified as a byproduct of studies designed to identify sites for Src-dependent PKCδ tyrosine phosphorylation).
As an initial approach to determine whether these novel phosphorylation sites are modified on PKCδ in vivo, mouse PKCδ was overexpressed in cardiomyocyte cultures metabolically labeled with 32P and then treated with H2O2 (to activate PKCδ and promote Src-dependent PKCδ tyrosine phosphorylation). PKCδ was then immunoprecipitated and subjected to phosphopeptide mapping analysis. A representative RP-HPLC chromatogram depicted in Fig 2B shows that PKCδ is phosphorylated at Thr141, Ser504, and Ser662 (the hydrophobic motif) in H2O2-treated cardiomyocytes. These studies provide important evidence that PKCδ-Thr141 phosphorylation can be detected in vivo in cardiomyocytes. The identification of Ser504 as a phosphoaccepter site was more surprising, since autophosphorylation at this site was not detected in the in vitro kinase assays. Ser504 is an evolutionarily conserved phosphorylation site adjacent to the threonine phosphoacceptor site in the activation loops of PKCδ and aPKCs [ζ and λ/ι]; cPKCs or other nPKCs do not contain a phosphorylatable residue at this position. While it is interesting to speculate that this serine may play a redundant role with Thr505 to structure the catalytic pocket of PKCδ for catalysis, our recent mutagenesis studies suggest otherwise, showing that a single T505A substitution is sufficient to prevent Src-dependent changes in PKCδ substrate specificity. Detailed studies that consider a possible role for Ser504 to substitute for Thr505 as a modification that regulates PKCδ’s substrate specificity are in progress.
Previous studies implicated autophosphorylations at residues N-terminal to the pseudosubstrate domain of PKCβII in the control of PKCβII activation. Since Thr141 maps to a similar position in PKCδ, we used a heterologous overexpression strategy, with GFP-tagged WT-PKCδ and PKCδ constructs harboring non-phosphorylatable alanine or phospho-mimetic aspartate substitutions at position 141 to examine whether Thr141 phosphorylation plays a similar role to influence PKCδ activation or downregulation. Fig 3A shows that WT-PKCδ, PKCδ-T141A, and PKCδ-T141D expression is similar in resting COS7 cells. WT-PKCδ and PKCδ-T141A are recovered primarily in the cytosolic fraction, but a smaller pool of WT-PKCδ and PKCδ-T141A constitutively localize to the particulate fraction. PKCδ-T141D also is primarily recovered in the cytosolic fraction. However, a 4.2±0.5-fold greater amount of PKCδ-T141D constitutively partitions to the particulate fraction, relative to WT-PKCδ (n=5, p<0.05). This increase in constitutive PKCδ-T141D partitioning to the particulate fraction is indicative of in vivo PKCδ activation. While a reciprocal decrease in PKCδ-T141D recovery in the soluble fraction is not obvious in Fig 3A, this is due to technical issues related to the detection of small differences in PKCδ immunoreactivity in soluble fractions that contain the bulk of the PKCδ immunoreactivity; the predicted differences in soluble PKCδ levels are resolved when protein loading for soluble fractions is decreased 5-fold (data not shown). Additional studies showed that lipid cofactors such as PMA or the cell-permeable DAG analog diC8 induce a similar complete translocation of all three enzymes from the soluble to the particulate fraction; no PKCδ, PKCδ-T141A, or PKCδ-T141D immunoreactivity remains in the soluble fraction of PMA- or diC8-treated COS7 cells.
The observation that a T141D substitution leads to the constitutive recruitment of greater amounts of PKCδ to the particulate fraction suggests that a phospho-mimetic substitution at Thr141 favors PKCδ activation. This conclusion gains support from studies that examine the kinetics of PMA-dependent PKCδ downregulation. While WT-PKCδ, PKCδ-T141A, and PKCδ-T141D were consistently recovered at similar levels under resting culture conditions (arguing that the Thr141 substitution does not leads to a gross change in the stability of the enzyme), Fig 3B exposes differences in the kinetics of PKCδ downregulation. PMA treatment leads to a time-dependent decrease in WT-PKCδ abundance; WT-PKCδ abundance falls over the first 4 hr of PMA treatment and remains at this reduced level for the next 20 hr. While these results are at odds with early studies showing that chronic PMA treatment leads to the complete downregulation of PKCδ, more recent studies with newer lots of anti-PKCδ antibodies (that are considerably more sensitive than the reagents previously available) indicate that substantial amounts of residual PKCδ immunoreactivity can be detected in various cells types treated with PMA for 24 hr. PMA treatment also leads to a time-dependent decline in PKCδ-T141A abundance. However, PMA-dependent PKCδ-T141A downregulation is delayed compared to PMA-dependent downregulation of WT-PKCδ. Of note, residual levels of WT-PKCδ and PKCδ-T141A are similar in COS7 cells treated with PMA for 24 hr. In contrast, PMA treatment leads to a considerably more rapid and complete downregulation of PKCδ-T141D. PKCδ-T141D abundance falls dramatically following a 2 hr incubation with PMA, and only trace amounts of PKCδ-T141D immunoreactivity are detected in COS7 cells treated with PMA for 4 hr; PKCδ-T141D immunoreactivity is not detected in COS7 cells treated with PMA for 24 hr. These results are consistent with the notion that a T141A substitution stabilizes the closed conformation of the enzyme and prevents activation/downregulation and that an autophosphorylation (or a negative charge) in the vicinity of the PKCδ pseudosubstrate domain lowers pseudosubstrate domain binding affinity for the catalytic pocket and favors activation.
GFP-tagged WT-PKCδ and PKCδ-T295A, and PKCδ-T141A were overexpressed in COS7 cells and immunoprecipitated for use in kinase assays to determine whether Thr295 and Thr141 contribute to the control of PKCδ catalytic function. Kinase assays were performed with cTnI as substrate - without and with Src - to determine whether Thr295 and Thr141 influence PKCδ-Thr505 phosphorylation, PKCδ tyrosine phosphorylation by Src, and/or PKCδ phosphorylation of cTnI. Assays also were performed over a range of ATP concentrations to examine whether these phosphorylation sites influence the ATP requirements for PKCδ autophosphorylation or PKCδ phosphorylation of cTnI. Fig 4A shows that WT-PKCδ, PKCδ-T295A, and PKCδ-T141A are recovered as constitutively Thr505-phosphorylated enzymes, indicating that T295A and T141A substitutions do not disrupt in vivo PKCδ-Thr505 phosphorylation. While we have not yet generated phospho-site specific antibodies (PSSAs) to track PKCδ phosphorylation at Thr295 or Thr141, we noted that Thr295 is flanked by a +2 position Arg (i.e., T295QR); the Cell Signaling Technologies anti-pTXR PSSA specifically recognizes this phosphorylation motif. Fig 4 shows that WT-PKCδ and PKCδ-T141A are recovered from COS7 cells without any appreciable anti-pTXR immunoreactivity co-migrating with PKCδ. The anti-pTXR PSSA detects in vitro WT-PKCδ and PKCδ-T141A autophosphorylation; this autophosphorylation reaction is detected at 1 μM ATP and is maximal at 6 μM ATP. The anti-pTXR PSSA does not detect autophosphorylation of the PKCδ-T295A mutant. The observation that WT-PKCδ and PKCδ-T141A undergo similar in vitro autophosphorylation reactions at a site recognized by the anti-pTXR PSSA, whereas the PKCδ-T295A mutant does not, provides strong evidence that the anti-pTXR PSSA specifically recognizes in vitro PKCδ autophosphorylation at PKCδ-Thr295. While an effect of the T295A substitution to influence anti-TXR immunoreactivity indirectly by controlling PKCδ autophosphorylation reaction a different TXR motif in PKCδ can not formally be excluded, other TXR motifs at T38ER and T588KR in human PKCδ have not been identified as autophosphorylation sites; this alternative interpretation of the results is considered substantially less likely. Since the anti-TXR PSSA appears to track PKCδ-Thr295 autophosphorylation, it was used to examine whether PKCδ-Thr295 autophosphorylation is dynamically regulated in vivo in COS7 cells. Fig 4B shows that WT-PKCδ and PKCδ-T295A constructs are recovered in similar amounts, without any basal anti-TXR immunoreactivity, from vehicle-treated COS7 cells. The anti-TXR antibody detects a PMA-dependent increase in WT-PKCδ phosphorylation, whereas no PMA-dependent increase in PKCδ-T295A phosphorylation is detected. These results suggest that Thr295 phosphorylation also accompanies PKCδ activation in vivo in a cellular context.
In vitro kinase assays were performed in the presence of Src, as a strategy to further interrogate the mechanisms that regulate PKCδ activity. Fig 4A shows that WT-PKCδ and PKCδ-T141A are phosphorylated by Src at Tyr311 in a similar manner; Src (and PKCδ-Tyr phosphorylation) does not lead to any obvious change in PKCδ autophosphorylation at Thr295. However, the PKCδ-T295A construct (harboring a T→A substitution in the vicinity of Tyr311 in the hinge region) shows a relative defect in Src-dependent Tyr311 phosphorylation. The functional consequences of altered Thr295 autophosphorylation and Src-dependent Tyr311 phosphorylation were examined in kinase assays with cTnI as a substrate. cTnI phosphorylation was tracked by immunoblot analysis with an anti-cTnI-pSer23/Ser24 PSSA that specifically recognizes the Ser23/Ser24-phosphorylated form of cTnI and the Cell Signaling anti-pTXR PSSA that selectively recognizes cTnI phosphorylation at KRPT144LR (as validated in (4)). Allosterically-activated forms of WT-PKCδ, PKCδ-T295A, and PKCδ-T141A (in assays with PS/PMA) phosphorylate cTnI at Ser23/Ser24 in a similar manner. In each case, cTnI-Ser23/Ser24 phosphorylation is detected at 1 μM ATP and maximal at 6 μM ATP. WT-PKCδ and PKCδ-T141A become cTnI-Thr144 kinases in assays with Src. Control experiments indicate that cTnI-Thr144 phosphorylation can not be attributed to a contaminant with serine/threonine kinase activity in the Src preparation, since in vitro kinase assays with Src and cTnI (without PKCδ) do not lead to any detectable 32P-incorporation into cTnI by PhosphoImager or evidence of cTnI phosphorylation by immunoblot analysis with the anti-cTnI-Thr144 PSSA. Of note, a T295A substitution decreases the cTnI-Thr144 kinase activity of the PKCδ enzyme. These results are consistent with previous studies showing that the effect of Src to convert PKCδ into a cTnI-Thr144 kinase can be abrogated by a Y311F substitution; the PKCδ-T295A mutant displays a defect in Src-dependent Tyr311 phosphorylation and decreased cTnI-Thr144 kinase activity. These results identify a novel PKCδ autophosphorylation reaction at Thr295 (a highly conserved residue in the hinge region) that is critical for Src-dependent PKCδ-Tyr311 phosphorylation and consequently Src-dependent changes in PKCδ substrate specificity.
We previously reported that PKCδ acts as a cTnI-Ser23/Ser24 kinase in assays with PS/PMA and as a dual Ser23/Ser24 and Thr144 kinase in assays with PS/PMA plus Src, and that a single T505A substitution prevents the Src-dependent acquisition of cTnI-Thr144 kinase activity. These studies implicated Thr505 autophosphorylation as a modification that critically regulates PKCδ phosphorylation of a heterologous substrate. The next set of experiments examined whether Thr505 phosphorylation regulates PKCδ autophosphorylation at Thr295. Fig 5 shows that WT-PKCδ is recovered as a constitutively Thr505-phosphorylated enzyme that undergoes an autophosphorylation reaction at Thr295 in assays performed with PS/PMA. WT-PKCδ phosphorylates cTnI at Ser23/Ser24 in assays with PS/PMA, and at Ser23/Ser24 and Thr144 in assays with PS/PMA and Src. The PKCδ-T505A construct also undergoes an autophosphorylation reaction in assays with PS/PMA (data not shown); PKCδ-T505A phosphorylates cTnI at Ser23/Ser24 in assays with PS/PMA (similar to WT-PKCδ). However, PKCδ-T505A autophosphorylation at Thr295 and Tyr311 phosphorylation by Src are defective, relative to WT-PKCδ; Src does not convert the PKCδ-T505A mutant into a cTnI-Thr144 kinase. Collectively, these results suggest a novel model for PKCδ activation. Our results indicate that PKCδ-Thr505 autophosphorylation has no effect on some aspects of catalytic activity (for example cTnI-Ser23/Ser24), but it is required for optimal PKCδ autophosphorylation at Thr295 and Src-dependent PKCδ-Tyr311 phosphorylation, a modification that converts PKCδ into a cTnI-Thr144 kinase.
In the final set of experiments, GFP-tagged WT-PKCδ and PKCδ-T141D were overexpressed in COS7 cells and immunoprecipitated for use in kinase assays to determine whether a negative charge at Thr141 favors PKCδ activation. Fig 6A (top) shows that WT-PKCδ and PKCδ-T141D constructs are expressed at similar levels in COS7 cells and that an immunoprecipitation with an anti-GFP antibody (that recognizes the tag on the heterologously overexpressed PKCδ enzyme) completely clears these enzymes from COS7 cell lysates. Similar amounts of WT-PKCδ and PKCδ-T141D proteins are recovered in anti-GFP pulldowns (Fig 6A, bottom). WT-PKCδ and PKCδ-T141D enzymes were then used in kinase assays without and with PS/PMA (in the absence or presence of Src). Fig 6B shows that WT-PKCδ and PKCδ-T141D are recovered with similar levels of constitutive phosphorylation at Thr505, but not Thr295 (i.e., a T141D substitution that increases PKCδ partitioning to the particulate fraction does not lead to an increase in basal PKCδ-Thr295 autophosphorylation in COS7 cells). However, the PKCδ-T141D mutant displays a substantial amount of lipid-independent kinase activity that is not detected in assays with the WT-PKCδ enzyme. PKCδ-T141D executes a robust lipid-independent autophosphorylation reaction, detected as both an increase in Thr295 autophosphorylation by immunoblot analysis and an increase in 32P-incorporation into the mutant enzyme by PhosphorImager analysis. The PKCδ-T141D mutant also is a lipid-independent cTnI kinase. PKCδ-T141D activity is only slightly further increased when assays are performed in the presence of PS/PMA. The Src-dependent increase in Y311 phosphorylation is similar for WT-PKCδ and PKCδ-T141D enzymes. These results indicate that a negative charge at Thr141 (in the vicinity of the pseudosubstrate domain) disrupts autoinhibitory interactions that restrict basal PKCδ activity.
This study identifies seven novel PKCδ autophosphorylation sites in very distinct regions of PKCδ at Thr141 adjacent to the pseudosubstrate motif, Thr218 in the C1A–C1B interdomain region, Thr295, Ser302, and Ser304 in the hinge region, and two serine residues (S501RAS504T) N-terminal to the activation loop Thr505 phosphorylation site. While these novel autophosphorylation sites are evolutionarily conserved in PKCδ, they do not reside in specific phosphorylation motifs common to all PKC isoforms. Rather, these novel autophosphorylation sites map to functionally important regions of the protein that appear to be autophosphorylation ‘hot spots’ in other PKC isoforms. For example, PKCδ and PKCε contain sites for autophosphorylation in the C1A–C1B interdomain region, although the PKCδ C1A–C1B interdomain autophosphorylation site (NSRDT218IF) bears no resemblance to the C1A–C1B interdomain autophosphorylation site recently identified in PKCε (PDQVGS234QR). These results suggest that PKCs may have evolved to contain sites for autophosphorylation at key positions (rather than conserved phosphorylation motifs) in the enzyme. In fact, one could speculate that differences in the autophosphorylation motifs identified in individual nPKC isoforms may have evolved to accommodate the distinct catalytic requirements of these enzymes.
In the course of these studies, two other laboratories identified seven novel sites for PKCδ autophosphorylation at Thr50, Thr141, Ser299, Ser302, Ser304, Thr451, and Ser506 (human numbering) (10;15). Studies from the Parker laboratory suggest that Ser299 autophosphorylation may provide a convenient marker for PKCδ activation, that could be used (in lieu of more cumbersome methods that assay for PKCδ translocation to membranes) to screen for PKCδ activation in tumor models. The notion that PKCδ autophosphorylation might play a functionally important role to regulate catalysis was not considered. Of the four autophosphorylation sites at Thr141, Ser302, Ser304 and Ser506 that were detected in both the previous publication and our study, Thr141 is particularly interesting. Thr141 is strategically positioned N-terminal to the pseudosubstrate domain, a region that maintains PKCδ in an inactive conformation through an intramolecular interaction with the catalytic pocket. Our mutagenesis studies show that PKCδ is phosphorylated at Thr141 in vivo in COS7 cells and that Thr141 autophosphorylation influences the kinetics of PKCδ activation/downregulation in cells. The PKCδ-T141D mutant is constitutively recovered (in greater amounts than WT-PKCδ) in the particulate fraction; a T141A substitution slows, and a T141D substitution accelerates, the tempo of PKCδ downregulation. In vitro kinase assays provide further evidence that a phosphorylation reaction at Thr141 relieves autoinhibitory constraints that limit PKCδ activity, showing that the PKCδ-T141D enzyme functions as a lipid-independent serine/threonine kinase. The conclusion that an autophosphorylation at Thr141 regulates PKCδ activity resonates with previous modeling studies of PKCβII, which proposed that negative charges due to autophosphorylation reactions at Ser16/Thr17 and an Arg residue at position 19 (a position that corresponds to Thr141 in PKCδ) participate in intramolecular binding interactions that influence enzyme activity.
Three autophosphorylation sites described in previous studies (at Thr50, Ser299, and Thr451) were not detected in our experiments (although one site was detected in the non-phosphorylated KPT50MY52PEWK peptide fragment, which contains putative phosphorylation sites at both Thr50 and Tyr52). Rather, this study identifies novel PKCδ autophosphorylation reactions at Thr218 and Thr295, sites not detected in previous studies. Thr218 resides in the C1A–C1B interdomain region, within a phosphorylation motif that is conserved in the C1A–C1B interdomain region of PKCθ. Studies of PKCθ implicate this C1A–C1B interdomain autophosphorylation reaction in mechanisms required for proper PKCθ targeting to lipid rafts and antigen receptor signaling responses in Jurkat and T cells (13). While the phosphorylation motif detected in PKCδ and PKCθ is not conserved in other PKC isoforms, regulatory autophosphorylation reactions have been identified in the C1A–C1B interdomain regions of PKCε and PKD1 (the founding member of a different family of PMA/DAG-sensitive kinases (16)). These results suggest that the C1A–C1B interdomain region may be an autophosphorylation ‘hot spot’ that is exposed on the surface and regulates the localization/actions of these enzymes in cells.
Finally, this study characterizes the variable V3 hinge region of PKCδ as a target for functionally important tyrosine and serine/threonine phosphorylations. We recently reported that Tyr311 and Tyr332 are the major sites for in vitro Src-dependent PKCδ phosphorylation. We also showed that PKCδ is dually phosphorylated at Tyr311 and Tyr332 in cardiomyocytes subjected to oxidative stress. Recent studies implicate Tyr311 phosphorylation as a modification that modulates PKCδ catalytic activity toward Thr144 on cTnI. Phosphorylation at Tyr332 does not contribute to the control of kinase activity, but rather generates a docking site for PKCδ binding partners such as the adapter protein Shc (17). Studies reported herein (along with recent results from the Parker laboratory (10)) extend the analysis by showing that the human PKCδ hinge region contains a cluster of autophosphorylation sites at Thr295, Ser299, Ser302 and Ser304. These sites, and their flanking sequences, are highly evolutionarily conserved in PKCδ; they are not conserved in other mammalian PKC isoforms (although the hinge regions of PKCβII and PKCε contain their own distinctive autophosphorylation motifs). The Parker laboratory detected PKCδ-Ser299 phosphorylation in a dually phosphorylated 298ApS299RRpS302DS304ASSEPVGIY313QGFEK318 peptide liberated by a 4 hr trypsin digest. Ser299 phosphorylation was not detected in our experiments that used a lengthier trypsin digestion (16 hr), which cleaved the 298–318 peptide fragment into a singly phosphorylated 301RS302DS304ASSEPVGIY313QGFEK318 peptide (with phosphorylation at either Ser302, Ser304, or Tyr313) and a very small 298ASR300 fragment that was not captured in the experiments. The Parker laboratory identified a dynamic increase in PKCδ-Ser299 phosphorylation in the membrane compartment of PMA-treated COS7 or HeLa cells; a role for PKCδ-Ser299 phosphorylation to regulate PKCδ’s cellular or in vitro catalytic function was not considered. Studies reported herein focused on PKCδ autophosphorylation at Thr295, a different autophosphorylation reaction that was not detected in the previous study. In vitro kinase assays exposed a novel role for PKCδ-Thr295 autophosphorylation as a lipid-dependent modification that links PKCδ-Thr505 autophosphorylation to PKCδ regulation by Src. PKCδ-Thr295 autophosphorylation is reduced in the PKCδ-T505A mutant and a T295A substitution leads to a defect in Src-dependent PKCδ-Tyr311 phosphorylation. These results provide a molecular explanation for the previous observation that Src phosphorylates PKCδ at Tyr311 only when assays are performed in the presence of lipid cofactors. We had previously speculated that lipid cofactors induce a conformational change that renders PKCδ a better substrate for Src. However, studies reported herein expose an additional effect of PMA or DAG to trigger a series of ordered autophosphorylation reactions at Thr505 and Thr295 that are required for Src-dependent PKCδ-Tyr311 phosphorylation (schematized in Fig 7). While a mechanism whereby an autophosphorylation reaction at Thr295 would prime PKCδ for subsequent Tyr311 phosphorylation by Src is uncertain, these studies suggest that compounds targeted to the PKCδ-Thr295 autophosphorylation site might selectively interdict the catalytic function of Tyr311-phosphorylated PKCδ, without acting as general inhibitors of the enzyme. Since there is evidence that the hinge region of PKCβII is exposed (and becomes proteolytically labile) in the active conformation of PKCβII (8), the hinge region of PKCδ also is likely to provide an accessible surface for drug interactions. A compound that inhibits PKCδ-Thr295 phosphorylation would constitute a novel class of PKCδ inhibitors that would selectively prevent the cellular actions of tyrosine phosphorylated PKCδ (in the soluble fraction of cells exposed to oxidative stress), while preserving PKCδ actions/phosphorylations that result from GPCR activation and the generation of DAG in membranes.
PKC isoforms play key roles in mechanisms that regulate cell proliferation, survival, and migration. PKCs are dysregulated in many clinically important disorders and genetic studies in PKC knockout models in mice provide compelling evidence that PKC isoforms contribute to the pathogenesis of various immune disorders, atherosclerotic and diabetic cardiovascular diseases, and malignancies. Current methods to screen for PKC activation are quite limited. Most studies have relied on measurements of PKC protein expression, which typically correlates at best only weakly with PKC activation. While measurements of PKC translocation to membranes are more informative, translocation studies are too cumbersome to use as screens for PKC activation (and translocation provides an imperfect measure of enzyme activation under certain circumstance, since some PKCs are activated in the cytosol via non-traditional lipid-independent mechanisms during oxidative stress (18)). Studies to determine whether Thr295 in PKCδ’s hinge region is phosphorylated exclusively via an obligate intramolecular autocatalytic reaction will be critical to determine whether this modification can be exploited as a biomarker to screen for PKCδ activation in biologically relevant tissue samples.
This work was supported by USPHS NHLBI grant HL 77860 and TL1 RR024158 from the National Center for Research Resources (NCRR), a component of the National Institues of Health (NIH) and NIH Roadmap for Medical Research. It’s contents are solely the responsibility of the authors and do not necessarily represent the official view of the NCRR or NIH. Information on NCRR is available at http://www.ncrr.nih.gov/. Information on Re-engineering the Clinical Research Enterprise can be obtained from http://nihroadmap.nih.gov/clinicalresearch/overview-translational.asp.