A Critical Role of Protein Kinase Cδ Activation Loop Phosphorylation in Formyl-Methionyl-Leucyl-Phenylalanine-Induced Phosphorylation of p47^phox and Rapid Activation of Nicotinamide Adenine Dinucleotide Phosphate Oxidase

Reagents

The N-formyl peptide fMLF, PMA, and isoluminol were purchased from Sigma-Aldrich. HRP and superoxide dismutase were obtained from Roche. The anti-hemagglutinin (HA) Ab and mouse mAb to β-actin were acquired from Santa Cruz Biotechnology. Rabbit polyclonal Abs against the nonphosphorylated PKCδ and phospho-PKCδ (Thr⁵⁰⁵) were obtained from Cell Signaling Technology.

Preparation of PKC and p47^phox expression constructs

The mouse cDNAs for PKCα, δ, ε, and ζ were provided by Dr. I. B. Weinstein (Columbia University, New York, NY). The mouse cDNA for PKCβII was a gift from Dr. C. L. Aschendel (Purdue University, West Lafayette, IN). The mouse PKCθ cDNA was a gift from Dr. Z. Sun (University of Illinois, Chicago, IL). An HA tag was placed in the C termini of these constructs. The cDNAs were subcloned into the pcDNA3 vector (Invitrogen Life Technologies). PKC chimeras and truncation mutants were prepared by PCR amplification of the selected PKCδ and PKCβII fragments using overlapping oligonucleotide primers with the following sequences: PKCδ-8 (forward) 5′-GCCCAATTGGCACCCTTCCTGCGCATCT-3′; PKCδ-2044 (reverse) 5′-GCGCAATTGAATGTCCAGGAATTGCTCAAACTT-3′; PKCβ-43 (forward) 5′-TAACAATTGGCTGACCCGGCTGCGGG-3′; PKCβ-2076 (reverse) 5′-GCGCAATTGGCTCTTGACTTCAGGTTTTAAAAATT-3′; PKC-C (forward) 5′-AAAGGCAGCTTTGGCAAGGT-3′; PKC-C (reverse) 5′-ACCTTGCCAAAGCTGCCTTT-3′; and PKCδ-995 (forward) 5′-GCGCAATTGAACAACGGGACCTATG GCAAG-3′.

PCR was performed with denaturing at 94°C (30 s), annealing at 55°C (30 s), and extension at 72°C (3 min), for a total of 25 cycles. The PCR products were subcloned into the pcDNA3 vector. Site-directed mutagenesis of the PKCδ and p47^phox genes was conducted with the QuikChange kit from Stratagene. Selected serine residues in p47^phox were mutated to alanine (see Fig. 1⇓B). The cloned PCR products and mutated cDNAs were sequenced to confirm accuracy.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1.

PKCδ-mediated p47^phox phosphorylation and reconstitution of fMLF-induced NADPH oxidase in COS-phox cells. A, COS-phox cells were transfected with expression constructs for FPR, G_αi2, and either one of the listed PKC isoforms or a control vector as indicated. Twenty-four hours after transfection, cells were assayed for O^⨪₂ production based on isoluminol-enhanced chemiluminescense in counts per second (cps). The fMLF-induced increase in O^⨪₂ generation was inhibited in the presence of superoxide dismutase (data not shown). A representative set of data, from one of five experiments that generated similar results, is shown. B, Amino acid sequence of the C-terminal region of p47^phox (residues 301–390) is shown. Potential phosphorylation sites to be mutated are underlined. C, Results from in vitro kinase assay using wild-type (WT) and mutant p47^phox proteins fused to GST as substrates. The substrates (10 μg each) were incubated with 0.2 μg of recombinant PKCδ and [γ-³²P]ATP at 30°C for 30 min. The samples were analyzed by SDS-PAGE, and the autoradiogram was quantified. Several experiments (n > 3) were performed, and duplicate data (mean ± SD) from a representative experiment are shown. D, The same set of p47^phox mutants were individually expressed in the COS^91/22 cells, cotransfected with expression vectors for FPR, PKCδ, and p67^phox, and assayed for O^⨪₂ production as described in A. The level of O^⨪₂ was based on integrated chemiluminescence (int. CL) collected during the first 20 min after fMLF (1 μM) stimulation. The expression levels of the p47^phox mutants were detected by Western blotting (bottom) using an anti-p47^phox Ab, showing equal expression and loading of all samples. A representative set of data is shown as mean ± SEM, and similar results were obtained in two other repeating experiments.

Cell culture and transfection

The transgenic COS-phox cells were generated as previously described (18). The cells were maintained at 37°C with 5% CO₂ in DMEM supplemented with 10% heat-inactivated FBS, 2 mM l-glutamine, 100 IU/ml penicillin, 50 μg/ml streptomycin, 0.2 mg/ml hygromycin (Sigma-Aldrich), 0.8 mg/ml geneticin (Invitrogen Life Technologies), and 1 μg/ml puromycin (Calbiochem). LipofectAMINE 2000 reagent (Invitrogen Life Technologies) was used for transient transfection of 4 μg of DNA (2.5 μg of an expression vector for human formyl peptide receptor (FPR), 0.5 μg of a G_αi2 expression vector, and 1 μg of an expression vector for the PKC isoform of interest) into COS-phox cells in a 100-mm culture dish (0.5–1 × 10⁶ cells). A transfection efficiency of 45–50% was achieved in a typical experiment. Cells were analyzed 21–24 h after transfection. The PLB-985 human myelomonoblastic leukemia cell line (30) was maintained in RPMI 1640 medium supplemented with 2 mM l-glutamine, 10% FBS, 100 IU/ml penicillin, and 50 μg/ml streptomycin. Exponentially growing cells at a starting density of 2 × 10⁵/ml were differentiated for 6 days with 0.5% dimethylformamide in RPMI 1640 containing 1% Nutridoma-SP and 0.5% FCS, with a medium change on day 3. Approximately 95% of the cells underwent granulocytic differentiation by day 6, based on morphologic analysis after May-Grünwald-Giemsa staining. The differentiated cells were used in studies with fMLF and platelet-activating factor (PAF; Biomol).

Measurement of O^⨪₂ production

O^⨪₂ production in COS-phox cells was determined by an isoluminol-ECL assay (31), in 6-mm diameter wells of a 96-well, flat-bottom, white tissue culture plates (E & K Scientific), as detailed in a previous publication (17). The chemiluminescence count per second was continually recorded at 1-min intervals. Samples containing 250 U of superoxide dismutase, in addition to the stimuli, were run in parallel. The relative level of O^⨪₂ produced was calculated based on the integrated chemiluminescence. In some experiments, integrated chemiluminescence in a given period of time after fMLF stimulation was expressed as detailed in figure legends.

Analysis of protein expression

Whole cell extracts were generated by lysis of transfected COS-phox cells with 200–500 μl of PAGE buffer containing protease inhibitors (Protease Inhibitor Mixture Set I; Calbiochem). Samples were sonicated for 15 s on ice using a Model 60 Sonic Dismembrator (Fisher Scientific), and heated at 95°C for 5 min. Whole cell extracts were analyzed on a 10% denaturing gel. Resolved protein samples were transferred to nitrocellulose membranes (Hybond ECL; Amersham Biosciences) for Western blotting using ECL detection (Pierce).

In vitro kinase assay

The coding sequence of a full-length p47^phox cDNA was fused to a GST gene in the pGEX4T1 vector (GE Healthcare), generating a GST-p47^phox fusion construct. The fusion protein was expressed in E. coli (DH5α DE3), and purified on glutathione affinity column according to the manufacturer’s instructions. In vitro kinase assays were performed using immunoprecipitated, recombinant PKCδ constructs. Briefly, HEK 293T cells in 100-mm dishes were transfected with the HA-tagged expression constructs of PKCδ, or with the control vector (pCDNA3) (2 μg/dish). Cellular proteins were extracted with PKC extraction buffer (50 mM HEPES (pH 7.5), 150 mM NaCl, 0.1% Tween 20, 1 mM EDTA, 2.5 mM EGTA, 10% glycerol) containing protease inhibitors (10 μg/ml aprotinin, 10 μg/ml leupeptin, 0.1 mM PMSF) and phosphatase inhibitors (1 mM NaF, 0.1 mM Na₃VO₄, 10 mM β-glycerophosphate). HA-tagged PKC proteins were immunoprecipitated from 300 μg of cell extracts using 3 μg of an anti-HA Ab and 30 μl of protein A/G-Sepharose after 3 h of incubation at 4°C. The immunoprecipitates were washed twice with PKC extraction buffer and once with the PKC kinase buffer (50 mM HEPES (pH 7.5), 10 mM MgCl₂, 1 mM DTT, 2.5 mM EGTA, 1 mM NaF, 0.1 mM Na₃VO₄, 10 mM β-glycerophosphate), and resuspended in 20 μl of PKC kinase buffer. The kinase assay was initiated by adding 40 μl of the kinase buffer containing 10 μg of GST-p47^phox fusion protein or histone H1.1 (Santa Cruz Biotechnology) and 5 μCi of [γ-³²P] ATP. The reactions were performed at 30°C for 30 min and terminated by adding Laemmli sample buffer and boiled for 5 min. Samples were electrophoresed on 10% SDS-polyacrylamide gel, transferred to nitrocellulose membrane, and exposed to a PhosphorImager screen. The autoradiograph data were quantified using the ImageGauge software (v.3.12; Fuji Photo Film). The blots were then subjected to Western blotting with anti-HA Ab to verify equal loading of the proteins in each reaction.

PKCδ-catalyzed phosphorylation of p47^phox is necessary for fMLF-induced O^⨪₂ production

We recently found that reconstitution of fMLF-induced NADPH oxidase activation in the transgenic COS-phox cells requires not only the heterologous FPR but also signaling molecules, such as PKCδ, that are abundant in neutrophils but scanty in COS cells (17). This observation is consistent with results from the use of neutrophil cores, PKCδ-selective inhibitors and antisense depletion (14, 15, 16, 17), all of which suggest an important role of PKCδ in NADPH oxidase activation. The major isoforms of PKC in neutrophils are PKCα, βII, δ, and ζ, which belong to three different subclasses of the PKC family and are known to phosphorylate p47^phox, an essential component of the phagocyte NADPH oxidase (11, 32). In transfected COS-phox cells, however, only the δ isoform can reconstitute fMLF-induced NADPH oxidase activation (Fig. 1⇑A). This result suggests that PKCδ is important for FPR-mediated signaling leading to the catalysis of p47^phox phosphorylation and O^⨪₂ production.

We examined whether there is a causal relationship between PKCδ-catalyzed p47^phox phosphorylation and fMLF-induced O^⨪₂ production using COS-phox cells as a functionally coupled system for study of the PKCδ activation mechanism. The C-terminal autoinhibitory region of p47^phox contains multiple serine residues that can be phosphorylated by PKC isoforms and other serine/threonine kinases (9, 10, 12, 13). Based on a previous report (33), we mutated seven serine residues (underlined in Fig. 1⇑B) to alanine, either individually or in combinations. The resulting p47^phox mutants were fused to GST and used as substrates for in vitro kinase assay with recombinant PKCδ. Among the residues selected for mutagenesis, Ser³⁰³, Ser³⁰⁴, and Ser³²⁸ represent consensus phosphorylation sites for PKC (34). Mutations of Ser³⁰³ and Ser³⁰⁴ to alanine reduced phosphorylation of p47^phox by ∼40% (Fig. 1⇑C). Mutation of other selected serine residues reduced p47^phox phosphorylation by 25–30%. Additional reduction in p47^phox phosphorylation was observed with the triple mutations (S303/304/345A and S303/304/379A) and the quadruple mutations (S303/304/345/379A). Interestingly, mutation of Ser³⁴⁵, a consensus phosphorylation site for MAPKs (35), also impaired p47^phox phosphorylation by PKCδ.

The effect of p47^phox mutations on fMLF-induced O^⨪₂ production was determined in COS-7 cells that express gp91^phox and p22^phox but not p67^phox and p47^phox (COS^91/22). Heterologous expression of FPR, p67^phox, p47^phox, and PKCδ reconstituted fMLF-induced O^⨪₂ production (Fig. 1⇑D, first group). All p47^phox mutants described were expressed equally well in the transfected cells (Fig. 1⇑D, Western blot). There was a positive correlation between reduced phosphorylation of the p47^phox mutants (Fig. 1⇑C) and compromised O^⨪₂ production (Fig. 1⇑D). Moreover, expression of several p47^phox mutants, including S328A, S345A and the double mutants S359/370A and S345/S379A, lowered the basal level of O^⨪₂ production. The presence of residual activities in the O^⨪₂ production assay and in vitro kinase assay suggests the presence of additional phosphorylation sites in p47^phox that contribute to NADPH oxidase activation. Together, these results demonstrate that fMLF-induced PKCδ activation is functionally coupled to O^⨪₂ production through p47^phox phosphorylation.

The catalytic domain of PKCδ contains a structural determinant regulated by fMLF for NADPH oxidase reconstitution

PKC isoforms differ in their regulatory domains between the subfamilies. These differences underline the distinct roles of diacylglycerol and Ca²⁺ in regulating the catalytic activity of classic vs novel PKC (20). In comparison, significant structural homology exists between the catalytic domains of different PKC isoforms. To understand whether the catalytic or regulatory domain of PKCδ determines the ability of this kinase to reconstitute fMLF-induced O^⨪₂ production in reconstituted COS-phox cells, we prepared chimeric PKC constructs by domain swap between PKCδ and PKCβII, the latter being unable to reconstitute NADPH oxidase under the same experimental conditions (Fig. 1⇑A). Two chimeras, one with the regulatory domain of PKCβII fused to the catalytic domain of PKCδ (PKCβII/δ), and the other with the catalytic domain of PKCβII fused to the regulatory domain of PKCδ (PKCδ/βII), were generated (Fig. 2⇓A). These chimeras were examined individually in transfected COS-phox cells. Although both constructs were properly expressed after transfection, only the PKCβII/δ construct was able to restore fMLF-stimulated O^⨪₂ production (Fig. 2⇓B).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2.

Characterization of PKCδ/PKCβII chimeras and the catalytic and regulatory domains of PKCδ. A, Representation of the chimeric PKC constructs, showing the swapped catalytic and regulatory domains of PKCβΙΙ and PKCδ. B, Expression of the PKC chimeras in COS-phox cells was determined by Western blotting (top), using an anti-HA Ab against the tagged PKC isoforms and another Ab against GAPDH for equal loading control. The fMLF-induced changes in O^⨪₂ production are shown as a function of time vs counts per second (cps) in the transfected COS-phox cells expressing the chimeras (bottom). At least three independent experiments were performed, and similar results were obtained. C, Deletion mutants were generated to test the roles of the regulatory PKCδ-R and catalytic PKCδ-C domains of PKCδ in O^⨪₂ production. The catalytic and regulatory domains of PKCδ constructs were schematically shown (top), and their expression in the transfected COS-phox cells was confirmed by Western blotting using an anti-HA Ab against the tagged protein constructs (bottom). D, Twenty-four hours after transfection, the catalytic and regulatory domains of PKCδ constructs were analyzed in O^⨪₂ production assays in the absence or presence of fMLF (1 μM) stimulation. The changes in O^⨪₂ production were quantified based on integrated chemiluminescence (Int. CL) collected during the first 20 min after agonist stimulation. Data shown are mean ± SEM from three experiments, each performed in duplicate.

Our results suggest that the catalytic domain of PKCδ contains structural features required for the reconstitution of NADPH oxidase upon fMLF stimulation. This possibility was further examined through individual expression of the catalytic and regulatory domains of PKCδ in COS-phox cells (Fig. 2⇑C). Both PKCδ domains were adequately expressed with the expected sizes (Fig. 2⇑C). When examined for O^⨪₂ production, the COS-phox cells transfected with the regulatory domain (PKCδ-R) failed to respond to fMLF, whereas cells expressing the catalytic domain (PKCδ-C) produced a significant amount of O^⨪₂ without agonist stimulation (Fig. 2⇑D). Of interest was the observation that fMLF stimulation resulted in an additional (∼53%) increase in O^⨪₂ production compared with the unstimulated cells that express the catalytic domain of PKCδ.

fMLF induces phosphorylation of Thr⁵⁰⁵ in PKCδ-expressing COS-phox cells

Based on our findings, we examined the possibility that the fMLF-induced signaling pathway acts on the catalytic domain and modulates its activity. It is known that several serine and threonine residues in the catalytic domain are phosphorylated before a newly synthesized PKC acquires its catalytic potency, a process sometimes referred to as maturation of PKC (reviewed in Ref. 21). Activation loop phosphorylation may be catalyzed by PDK-1 (22, 23). However, whether activation loop phosphorylation is necessary for the catalytic potency of PKCδ remains controversial as conflicting results were obtained from several studies (23, 26, 27, 28). Using an anti-phospho-PKCδ Ab recognizing a phosphorylated Thr⁵⁰⁵, we determined the level of activation loop phosphorylation as a function of time. In COS-phox cells expressing the catalytic domain of PKCδ (PKCδ-C), fMLF stimulation caused increases in Thr⁵⁰⁵ phosphorylation that peaked at 30 and 180 s (Fig. 3⇓A). A similar experiment was conducted with transfected COS-phox cells expressing the full-length PKCδ (Fig. 3⇓B). A transient increase in the level of Thr⁵⁰⁵ phosphorylation was also observed 30 s after fMLF stimulation, and was followed by a decrease at 60 and 120 s. A prominent increase in Thr⁵⁰⁵ phosphorylation was again detected 180 s after fMLF stimulation (Fig. 3⇓B). A comparison with the kinetics of fMLF-induced O^⨪₂ production (Figs. 1⇑A and 2B) suggested to us that the first increase in Thr⁵⁰⁵ phosphorylation corresponds to the initial rise of O^⨪₂ production in reconstituted COS-phox cells. Basal phosphorylation at Thr⁵⁰⁵ was detected in cells expressing the catalytic domain of PKCδ as well as the full-length PKCδ. Densitometry analysis showed a basal phosphorylation level of 0.66 ± 0.03 for PKCδ-C and 0.48 ± 0.02 for the full-length PKCδ, based on data from three independent experiments. Although cells expressing the catalytic domain of PKCδ displayed higher basal level phosphorylation, induced phosphorylation of Thr⁵⁰⁵ was more evident in transfected COS-phox cells expressing the full-length PKCδ. Pertussis toxin, which ADP-ribosylates the α subunit of G_i/o proteins on a C-terminal cysteine and prevents its interaction with the receptor (36), reduced fMLF-induced phosphorylation of Thr⁵⁰⁵ (Fig. 3⇓C), confirming that the fMLF-induced PKCδ activation loop phosphorylation is a G protein-mediated event. It was notable that pertussis toxin also reduced the basal phosphorylation level. Basal phosphorylation at Thr⁵⁰⁵ may result from the constitutive (ligand-independent) activity of FPR, which became more evident in cells overexpressing FPR (37). Because both constitutive and agonist-induced activation of FPR involves G_i coupling, they are subjected to inhibition by pertussis toxin.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 3.

Agonist-induced phosphorylation of PKCδ at Thr⁵⁰⁵. A, A representative Western blot (n = 3) showing fMLF-induced (1 μM) Thr⁵⁰⁵ phosphorylation of the truncated PKCδ catalytic domain (PKCδ-C) as a function of time, using an anti-phospho-Thr⁵⁰⁵ Ab. The assay was performed 24 h after transfection of COS-phox cells. The expression level of PKCδ catalytic domain was determined with an anti-PKCδ Ab. Quantification of the Western blot shown is presented below. B and C, Western blot showing fMLF-induced (1 μM) Thr⁵⁰⁵ phosphorylation of the full-length PKCδ. C, Transfected COS-phox cells were treated with pertussis toxin (500 ng/ml, 4 h) before fMLF stimulation. The samples were analyzed on the same gel with the untreated samples (B). The extent of Thr⁵⁰⁵ phosphorylation was expressed as the ratio of the signal detected by the anti-phospho-Thr⁵⁰⁵ Ab to the signal detected by the anti-PKCδ Ab, which was used for normalization of sample loading. Data shown are representative of at least three experiments, with similar patterns of Thr⁵⁰⁵ phosphorylation.

Induction of activation loop phosphorylation correlates with fMLF stimulation of O^⨪₂ production

The relationship between PKCδ activation loop phosphorylation and NADPH oxidase activation was investigated using PLB-985, a human myelomonoblastic leukemia cell line (30). After 6 days of differentiation with 0.5% dimethylformamide, PLB-985 acquired the ability to respond to fMLF with O^⨪₂ production (Fig. 4⇓A). In comparison, PAF, a priming agent that does not directly activate NADPH oxidase in neutrophils, did not stimulate a significant production of O^⨪₂ in differentiated PLB-985 cells (Fig. 4⇓B). Both fMLF and PAF induced a rapid Ca²⁺ flux, suggesting the presence of functional receptors on the cell surface that could mediate proximal signaling (Fig. 4⇓, C and D). However, only fMLF induced biphasic phosphorylation of human PKCδ at Thr⁵⁰⁷ (equivalent to Thr⁵⁰⁵ in mouse PKCδ). These results demonstrate a correlation between induction of PKCδ activation loop phosphorylation and the ability of the agonist to stimulate O^⨪₂ production.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 4.

A comparison of fMLF and PAF for their induction of O^⨪₂ production, Ca²⁺ flux, and PKCδ phosphorylation in PLB-985 cells. PLB-985 or dimethylformamide-differentiated PLB-985 (6 days after addition of 0.5% dimethylformamide; Diff.) were stimulated with fMLF or PAF (100 nM) for their responses in O^⨪₂ production (A and B), Ca²⁺ flux (C and D) and activation loop phosphorylation at Thr⁵⁰⁷ (equivalent to Thr⁵⁰⁵ in mouse PKCδ) (E and F). The O^⨪₂ production assays were conducted with real-time measurement of isoluminol-ECL as described in Materials and Methods. Each sample contains ∼1 × 10⁶ cells. Ca²⁺ flux was measured with 5 × 10⁵ cells/sample using Indo-1/AM as an indicator, and expressed as the ratio of emission fluorescence measured at 405 nm and 485 nm (relative fluorescence). Phosphorylation assays were performed after a 24-h serum starvation of PLB-985 and differentiated PLB-985 (n = 3). Graphs represent the level of Thr⁵⁰⁷ phosphorylation, determined as the ratio of the band density detected by the anti-phospho-Thr⁵⁰⁷ Ab vs the band density detected by the anti-PKCδ Ab. The latter serves as a loading control.

Mutation of Thr⁵⁰⁵ in the activation loop impairs fMLF-induced NADPH oxidase activation

Phosphorylation of a PKC in the catalytic loop, turn motif and hydrophobic motif is thought to be a prerequisite for its catalytic potency (21, 38). Previously published reports indicate that PKCδ is an exception in that its activation does not require activation loop phosphorylation (26, 27, 28). These studies, however, used pseudosubstrate peptide or histone rather than a physiologically relevant protein as substrates in the in vitro phosphorylation assay. Because PKCδ is essential for the reconstitution of NADPH oxidase in COS-phox cells, and p47^phox is a substrate for PKC-catalyzed phosphorylation, we examined the effect of PKCδ phosphorylation in its catalytic domain on fMLF-induced O^⨪₂ production. Phosphorylation of Ser⁶⁴³ in the turn motif (39) was determined using an Ab against the phospho-Ser⁶⁴³ of PKCδ. Persistent phosphorylation of Ser⁶⁴³ was observed in the absence of fMLF and over the entire course of fMLF stimulation in PKCδ-transfected COS-phox cells (data not shown). Therefore, it is unlikely that Ser⁶⁴³ phosphorylation plays a critical role in the initiation of fMLF-induced NADPH oxidase activation in the transfected COS-phox cells.

The effects of serine/threonine phosphorylation in the catalytic domain of PKCδ were further examined using site-directed mutagenesis. Point mutations were individually introduced into the full-length cDNA of PKCδ, producing alanine substitutions at Glu⁵⁰⁰, Thr⁵⁰⁵ (activation loop), Ser⁶⁴³ (turn motif) and Ser⁶⁶² (hydrophobic motif) in the catalytic domain as illustrated in Fig. 5⇓A. In addition, Lys³⁷⁶, an ATP binding site (40), was replaced with an arginine, and the resulting construct (PKCδ^K376R) was used as a negative control for the reconstitution assay.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 5.

Different effects of PKCδ catalytic domain mutations in fMLF-induced O^⨪₂ production. Site-directed mutagenesis was performed in the catalytic domain of PKCδ, as schematically shown in A and detailed in Materials and Methods. Each mutant of PKCδ was analyzed along with PKCδ^K376R (deficient in ATP binding) in transfected COS-phox cells. B, Characterization of the activation loop substitutions, E500A and T505A. C, Characterization of the turn motif mutant S643A, and the hydrophobic motif mutant S662A. The expression level of the wild-type (WT) and mutant PKCδ was examined with anti-HA for detection of the tagged PKCδ constructs. The fMLF-induced changes in O^⨪₂ production was shown based on integrated chemiluminescence (Int. CL) determined in the first 20 min after agonist (1 μM fMLF) stimulation. Data shown are mean ± SEM from three experiments for each mutant.

When individually examined in the transfected COS-phox cells, the above PKCδ mutants produced different effects in fMLF-induced O^⨪₂ production. The E500A substitution resulted in a significant (p ≤ 0.01) decrease in the basal level O^⨪₂ production (Fig. 5⇑B). However, fMLF was still able to stimulate a significant increase in O^⨪₂ production. In contrast, cells transfected with the T505A mutant responded poorly to fMLF in the reconstituted COS-phox cells, generating only a fraction of O^⨪₂ when compared with cells expressing the wild-type PKCδ. Reduced basal level O^⨪₂ production was also observed in cells expressing the T505A mutant. An analysis of data from the transfection experiments found a ∼35% decrease in O^⨪₂ production with the S643A mutation, indicating that Ser⁶⁴³ is required for maximal O^⨪₂ production. In comparison, no change in O^⨪₂ production was observed with the S662A mutation. Collectively, these results indicate a critical role of Thr⁵⁰⁵ phosphorylation in basal and induced O^⨪₂ production in the transfected COS-phox cells.

Mutation of Thr⁵⁰⁵ abrogates PKCδ autophosphorylation and its ability to phosphorylate p47^phox

A major function of PKC in NADPH oxidase activation is the catalysis of p47^phox phosphorylation (12, 41, 42). To determine whether one or more of the mutations alters the ability of PKCδ to phosphorylate p47^phox, we conducted in vitro kinase assay using as substrate a full-length p47^phox fused to GST, as described in the legend for Fig. 1⇑. The wild-type PKCδ, expressed in transfected HEK 293T cells and immunoaffinity purified from the cell lysate, strongly phosphorylated the GST-p47^phox substrate (lower band in the phosphorimages in Fig. 6⇓) as well as the kinase itself (autophosphorylation; upper band in the phosphorimages in Fig. 6⇓). Mutation of Glu⁵⁰⁰ to alanine resulted in a small and statistically insignificant decrease (p > 0.05) in p47^phox phosphorylation and PKCδ autophosphorylation (Fig. 6⇓A). In contrast, the T505A mutation caused a marked reduction in PKCδ autophosphorylation and a complete loss of p47^phox phosphorylation (Fig. 6⇓B). We also examined Ser to Ala mutations in the turn motif (PKCδ^S643A) and hydrophobic motif (PKCδ^S662A). Although no significant changes were observed with the alanine substitution at Ser⁶⁶² (Fig. 6⇓D), a small reduction in PKCδ autophosphorylation and p47^phox phosphorylation were observed with the S643A mutation (Fig. 6⇓C). The phosphorylation patterns of these PKCδ mutants are consistent with their abilities to reconstitute fMLF-induced NADPH oxidase in the transfected COS-phox cells, shown in Fig. 5⇑. The lack of autophosphorylation in PKCδ^T505A suggests that Thr⁵⁰⁵ phosphorylation is a prerequisite for autophosphorylation at Ser⁶⁴³ and Ser⁶⁶².

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 6.

The kinase activity of wild-type (WT) and mutant PKCδ constructs using GST-p47^phox as substrate. In vitro kinase assays were performed using immunoprecipitated PKCδ mutants from transfected HEK 293T cells and a recombinant GST-p47^phox fusion protein as substrate, as detailed in Materials and Methods. After incubation at 30°C for 30 min in the presence of [γ-³²P]ATP, the reaction was terminated. The samples were analyzed by SDS-PAGE and autoradiography. Two phosphorylated bands were identified: Top, representing autophosphorylated PKCδ (apparent molecular mass ≈78 kDa). Bottom, representing phosphorylated GST-p47^phox (apparent molecular mass ≈72 kDa). The activation loop substitutions E500A (A) and T505A (B), the turn motif mutant S643A (C), and the hydrophobic motif mutant S662A (D) were compared. The GST-p47^phox content in each kinase assay was determined by Western blotting using an anti-GST Ab (bottom). The K376R mutant was used as a negative control in the kinase assays. At least three independent kinase assays were performed, and similar results were obtained. A representative set of data is shown.

The catalytic activity of PKCδ^T505A is substrate-specific

As noted, different views exist with regard to a role of activation loop phosphorylation in the catalytic activity of PKCδ (26, 27, 28). To determine whether the use of different substrates has contributed to the discrepancy in these studies, we conducted in vitro kinase assay with histone H1.1, a substrate used in one of the previous studies mentioned. Under the same experimental conditions as in Fig. 6⇑, different results were obtained when histone H1.1 was used as a substrate (Fig. 7⇓). Although autophosphorylation of PKCδ^T505A remained minimal, a significant phosphorylation of histone H1.1 was observed with the T505A mutant of PKCδ (Fig. 7⇓B). This finding may help to reconcile the difference in PKCδ catalytic specificity and suggest substrate-specific phosphorylation by PKCδ^T505A.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 7.

The kinase activity of wild-type (WT) and mutant PKCδ constructs using histone H1.1 as substrate. In vitro kinase assays were performed using immunoprecipitated PKCδ mutants from transfected HEK 293T cells, as detailed in Materials and Methods. Recombinant histone H1.1 (Santa Cruz Biotechnology) was used as a substrate in place of GST-p47^phox. The activation loop substitutions E500A and T505A (A and B), the turn motif mutant S643A (C), and the hydrophobic motif mutant S662A (D) were compared. The histone protein content in each kinase assay was determined by SDS-PAGE using Coomassie brilliant blue staining (lower panels). The K376R mutant was used as a negative control in the kinase assays. At least three independent in vitro kinase assays were performed, and similar results were obtained from these experiments. A representative experiment is shown.

Agonist-induced phosphorylation of the activation loop in PKCδ contributes to the rapid production of O^⨪₂

The major PKC isoforms in neutrophils are PKCα, βII, δ, and ζ (9, 11, 32). Previous studies have shown that all these PKC isoforms are able to phosphorylate p47^phox when stimulated with PMA, although the kinetics of PKCδ binding and membrane translocation differ from those of the classic PKC isoforms (12, 32, 43). Two-dimensional phosphopeptide mapping indicates that, with the exception of PKCζ, all major PKC isoforms in neutrophils phosphorylate p47^phox with a similar phosphopeptide profile (32). Therefore, the catalytic potency of PKCδ is probably similar to that of PKCα and PKCβII with respect to p47^phox phosphorylation. The mechanism of activation, however, may be different between PKCδ and the two classic PKC isoforms. It has been well established that diacylglycerol binding to the regulatory domain of a PKC causes allosteric changes in its structure, allowing access of the substrate to the catalytic domain of the PKC. In phagocytes, PMA is a potent activator of NADPH oxidase due to its ability to mimic diacylglycerol in the activation of PKCs. PMA-induced O^⨪₂ production is sustained but lacks the rapid initial rise as seen in fMLF-stimulated cells. The discrepancy in the kinetics of NADPH oxidase activation may be attributed to the mechanism of PKC activation. Results derived from the current study indicate that fMLF-induced PKCδ phosphorylation in its activation loop may be such a mechanism. In differentiated PLB-985 human myelomonoblastic cells, fMLF-induced O^⨪₂ production was accompanied by PKCδ phosphorylation in its activation loop. In comparison, PAF failed to stimulate O^⨪₂ production and did not induce PKCδ phosphorylation in its activation loop. PAF is known as a priming agent and not a direct activator of NADPH oxidase. The priming effect results in part from PAF-induced phosphorylation of p67^phox but not p47^phox (44). Therefore, the ability to induce PKCδ phosphorylation in its activation loop may distinguish a full agonist for NADPH oxidase activation from a priming agent.

In neutrophils, O^⨪₂ production results from the activation of multiple signaling pathways that may be triggered simultaneously. In contrast, the approaches that we took focus on an individual kinase and the related signaling pathway to gain a better understanding of the underlying mechanism. Therefore, the functional impact of PKCδ and its activation loop phosphorylation in neutrophil NADPH oxidase activation may be broad or restricted to a particular pathway downstream of the activated FPR. fMLF stimulation results in guanine nucleotide exchange on the α subunit of the G_i proteins and the subsequent separation of the βγ subunits. These changes activate two pathways: the PLCβ pathway leads to hydrolysis of phosphatidylinositol 4,5-bisphosphate and the generation of the second messengers, diacylglycerol and inositol 1,4,5-trisphosphate, which are responsible for the activation of classic PKC isoforms and possibly novel PKC isoforms. The PI3K pathway causes phosphorylation of phosphatidylinositol 4,5-bisphosphate at the 3′ position of the inositol ring and the production of phosphatidylinositol 3,4,5-trisphosphate. The latter is responsible for a variety of cellular activation processes including the activation of kinases, such as PDK-1 and Akt, and guanine nucleotide exchange factors, such as P-Rex1 (45). PKCδ activation may result from PDK-1-catalyzed activation loop phosphorylation (23), thus representing a branch of the fMLF-induced signaling pathways. Indeed, both pertussis toxin and PI3K inhibitors are known to block fMLF-induced O^⨪₂ production and other neutrophil functions supporting the hypothesis that PKCδ involvement in oxidase activation is a terminal function of a more extensive fMLF signaling network. In COS cells, which are relatively inefficient in PLCβ activation unless additional G protein βγ subunits are expressed (46), the fMLF-induced O^⨪₂ generation may rely heavily on PKCδ activation loop phosphorylation that can be triggered by the PI3K pathway. In the study conducted by Yaffe and colleagues (14), it was found that phosphatidylinositol with 3′ position phosphorylation (PI(3,4)P₂, phosphatidylinositol 3,4,5-trisphosphate, and phosphatidylinositol (3)-phosphate) as well as PKCδ are essential for the reconstitution of NADPH oxidase in neutrophil cores. There are also published data that contradict the “restricted role” hypothesis and suggest a broader function of PKCδ in NADPH oxidase activation and other neutrophil activities. For example, studies showing PKCδ involvement in oxidant production were conducted not only with fMLF (17) but also with other agonists including leukotriene B₄ (15), PMA (14), and opsonized zymosan (16). In the current study, we observed that the first wave of PKCδ activation loop phosphorylation is followed by a rapid decline in its phosphorylation, suggesting possible activation of a protein phosphatase in response to PKCδ activation. Moreover, knockout of PKCδ gene in mice leads to compromised neutrophil responses to fMLF, TNF-α, and IL-8 in O^⨪₂ production as well as other activities including adhesion, chemotaxis, and lactoferrin release (47). Therefore, it is possible that PKCδ phosphorylates not only p47^phox but also other proteins that are involved in neutrophil activation.

The requirement of activation loop phosphorylation for PKCδ-catalyzed phosphorylation is substrate-specific

Studies conducted by Stempka and colleagues (26, 27) led to the conclusion that phosphorylation of Thr⁵⁰⁵ in the activation loop is not essential for mouse PKCδ to acquire its catalytic competence, although the same process is known to be important for the activation of other PKC isoforms (22, 23, 24, 48). This conclusion was based on two observations. First, recombinant PKCδ produced in bacteria, which lacked Thr⁵⁰⁵ phosphorylation due to the absence of a “PKC kinase” such as PDK-1, was nonetheless competent for allosteric modulation by activators such as PMA. Secondly, mutation of Thr⁵⁰⁵ to an alanine did not abolish its catalytic activity in in vitro kinase assay (26). It was proposed that the independence of PKCδ activity from its activation loop phosphorylation was due to the presence of Glu⁵⁰⁰, which carries the negative charge and helps to maintain the catalytic competence of PKCδ. Indeed, Stempka and colleagues found that mutation of Glu⁵⁰⁰ significantly weakens the catalytic activity of mouse PKCδ (27). A more detailed structure and function analysis of human PKCδ was conducted by Liu and colleagues recently (28). In this study, the authors observed that Phe⁵⁰⁰ and Phe⁵²⁷, along with the N-terminal extension of the kinase domain, play important roles in maintaining the kinase activity in the absence of activation loop phosphorylation at Thr⁵⁰⁷, which is equivalent to Thr⁵⁰⁵ in mouse PKCδ. Structural analysis has shown that interactions of Phe⁵⁰⁰ with Tyr³³⁴ and Phe⁵²⁷ with Ile⁴⁹⁹ are important for stabilization of the conformation of PKCδ^T507A. As a result, PKCδ^T507A is able to phosphorylate PKCα pseudosubstrate in vitro (28). However, our results indicate that mutation of Thr⁵⁰⁵ renders the mouse PKCδ incapable of phosphorylating p47^phox. The difference between our results and the findings made by Liu et al. (28) may be attributed to substrate specificity. In their study, Liu and colleagues found that the human PKCδ^T507A prefers certain peptide substrates, but displays poor catalytic activity on other substrates, especially when these substrates are used at higher concentrations. Their experimental data suggest that PKCδ^T507A-specific product inhibition is a plausible explanation (28). The possibility that PKCδ^T505A catalyzes phosphorylation in a substrate-specific manner is further suggested by our observation that mouse PKCδ^T505A could phosphorylate histone H1.1, despite minimal autophosphorylation of PKCδ^T505A (Fig. 7⇑B).

In most published studies, the catalytic activity of PKCδ was determined using synthetic peptides or histone as substrates. The current study explores the unique features of p47^phox in NADPH oxidase activation, and uses this physiologically relevant protein as substrate for the in vitro kinase assay. The same protein is also used in a whole cell-based reconstitution assay, thereby providing a functionally coupled approach for study of PKCδ activity in vitro and in cells. It is important to note that phosphorylation of substrate in intact cells may differ from phosphorylation assays conducted in test tubes. For instance, Liu and colleagues (28) observed that the human PKCδ^T507A mutant is fully capable of phosphorylating the PKCα pseudosubstrate peptide in vitro but fails to stimulate NF-κB and AP-1 reporter activity in cells. Moreover, this dephosphorylated mutant of human PKCδ is able to induce apoptosis (28). Therefore, the diversity of intracellular substrates for PKC isoforms is a factor to consider when evaluating the catalytic activity of PKCδ in the cells. As discussed in the study by Liu et al. (28), reduction in PKCδ catalytic activity at higher substrate concentration may result from either substrate inhibition or product inhibition. With respect to PKCδ-mediated p47^phox phosphorylation, the C-terminal autoinhibitory region contains multiple phosphorylation sites and is an excellent substrate for a number of PKC isoforms. A more detailed analysis will be necessary to delineate the structure and function relationship between PKCδ activation loop phosphorylation and p47^phox phosphorylation, and to determine whether failure to phosphorylate p47^phox by PKCδ^T505A is the result of substrate inhibition or production inhibition.

In summary, our results demonstrate that inducible PKCδ phosphorylation in its activation loop is required for fMLF-stimulated O^⨪₂ production in transfected COS-phox cells and in differentiated PLB-985 myelomonoblastic cells. These findings provide an explanation for the rapid production of O^⨪₂ following fMLF stimulation and suggest the possibility that different PKC isoforms may indeed have nonredundant functions in neutrophil activation. The conclusions drawn from this study also emphasize the importance of using physiologically relevant protein substrates in studies of PKC structure and function.