The responses of human neutrophils (PMN) involve reorganization and phosphorylation of cytoskeletal components. We investigated the translocation of protein kinase C (PKC) isoforms to PMN cytoskeletal (Triton-insoluble) fractions, in conjunction with activation of the respiratory burst enzyme NADPH oxidase. In resting PMN, PKC-δ (29%) and small amounts of PKC-α (0.6%), but not PKC-βII, were present in cytoskeletal fractions. Upon stimulation with the PKC agonist PMA, the levels of PKC-α, PKC-βII, and PKC-δ increased in the cytoskeletal fraction, concomitant with a decrease in the noncytoskeletal (Triton-soluble) fractions. PKC-δ maximally associated with cytoskeletal fractions at 160 nM PMA and then declined, while PKC-α and PKC-βII plateaued at 300 nM PMA. Translocation of PKC-δ was maximal by 2 min and sustained for at least 10 min. Translocation of PKC-α and PKC-βII was biphasic, plateauing at 2–3 min and then increasing up to 10 min. Under maximal stimulation conditions, PKC isoforms were entirely cytoskeletal associated. Translocation of the NADPH oxidase component p47phox to the cytoskeletal fraction correlated with translocation of PKC-α and PKC-βII, but not with translocation of PKC-δ. Oxidase activity in cytoskeletal fractions paralleled translocation of PKC-α, PKC-βII, and p47phox. Stimulation with 1,2-dioctanoylglycerol resulted in little translocation of PKC isoforms or p47phox, and in minimal oxidase activity. We conclude that conventional PKC isoforms (PKC-α and/or PKC-βII) may regulate PMA-stimulated cytoskeletal association and activation of NADPH oxidase. PKC-δ may modulate other PMN responses that involve cytoskeletal components.
Neutrophils (PMN)3 are the main cellular defense of higher organisms against invading bacteria. Major bactericidal mechanisms of PMN involve chemotaxis to sites of infection or inflammation, phagocytosis, the generation of superoxide via NADPH oxidase, and degranulation (1, 2, 3). Stimulation of these host defense responses involves receptor-mediated activation of phospholipases and protein kinases (4, 5), as well as reorganization and phosphorylation of cytoskeletal components (6).
Our primary goal is to understand the roles that both the cytoskeleton and protein kinases play in signaling events associated with the functional responses of PMN. In particular, our focus is the activation of the respiratory burst enzyme, NADPH oxidase. This enzyme is a multicomponent complex capable of converting O2 to O2−, which can be further metabolized to other toxic oxygen species to aid in the destruction of host pathogens (7). NADPH oxidase consists of a flavocytochrome b558 heterodimer (p22phox and gp91phox) and cytosolic components (p47phox, p67phox, p40phox, and Rac 1 or 2) (7, 8, 9). The cytosolic proteins translocate to form a complex with the membrane-associated cytochrome during enzyme activation (10). The signaling mechanisms that regulate the assembly and activation of NADPH oxidase are unclear.
Recent findings have implicated a role for cytoskeletal elements in the regulation of NADPH oxidase activation. In a cell-free system that examines oxidase activation, the addition of G-actin and its subsequent polymerization to F-actin enhances oxidase activity (11). This suggests that polymerized actin may provide a scaffolding to align oxidase components, which allows increased oxidase activity. How this observation relates to in vivo NADPH oxidase activation is unclear, but whole cell studies of the oxidase have demonstrated relationships with the cytoskeleton. Quinn et al. (12) and Mukherjee et al. (13) have demonstrated cosedimentation of NADPH oxidase components and activity with actin- and fodrin-rich membrane subdomains from PMA- and FMLP-stimulated PMN, respectively. Therefore, oxidase components may be localizing to the submembranous cytoskeleton during PMN stimulation. Grogan et al. (14) demonstrated the existence of a p40phox/p67phox/coronin complex in PMN and implicated a role for the oxidase in regulating changes in the cytoskeleton. Coronin is an actin-binding protein identified in Dictyostelium discoideum thought to be involved in cell motility (15). Each of the NADPH oxidase components has been shown to be associated with cytoskeletal (Triton-insoluble) fractions in PMA-stimulated PMN (16, 17, 18, 19). Woodman et al. (18) reported that approximately >95% of measurable oxidase activity was present in cytoskeletal fractions of PMA-stimulated cells. These observations strongly suggest that cytoskeletal elements play a role in NADPH oxidase activation. However, the mechanisms involved are unclear. It is important to identify the interaction between signaling pathways, cytoskeletal elements, and NADPH oxidase components to fully understand the complicated mechanism of oxidase activation and its regulation.
Activation of NADPH oxidase is mediated, at least in part, by phosphorylation (20). PMA directly binds to and activates protein kinase C (PKC) isoforms, implying that the cytoskeletal-associated activation of NADPH oxidase by PMA could be mediated by PKC. The PKC family of serine/threonine protein kinases is divided into three classes, based on differential dependencies on Ca2+ and lipids for activation (21, 22, 23). The conventional class (cPKC, PKC-α, PKC-βI, PKC-βII, and PKC-γ) requires Ca2+, phosphatidylserine (PS), and diacylglycerol (DAG). The novel class (nPKC, PKC-δ, PKC-ε, PKC-η, PKC-θ, and PKC-μ), like the cPKC class, requires PS and DAG, but is Ca2+ independent. The atypical class (atypical PKC, PKC-ζ, and PKC-ι) is Ca2+ and DAG independent, but requires PS for activation. PMA, a DAG analogue, is a potent activator of conventional and novel PKC isoforms. Five PKC isoforms have been identified in human PMN. These include: PKC-α, PKC-βI, PKC-βII (cPKCs); PKC-δ (a nPKC); and PKC-ζ (an atypical PKC) (24, 25, 26, 27). Thus, PKC-α, PKC-β, or PKC-δ are prime candidates for mediating PMA-induced responses, as well as responses to receptors that trigger the liberation of second messengers by phospholipases.
Activation of PKC isoforms is generally associated with translocation from one site in the cell to another (21, 23). Associations between PKC isoforms and cytoskeletal proteins have been demonstrated in some cell types. PKC-α colocalizes with MARCKS (myristoylated alanine-rich C kinase substrate) in macrophages (28). Blobe et al. (29) demonstrated a specific association of PKC-βII, but not PKC-βI, with actin, which altered the substrate preferences of the βII isoform. PKC-δ colocalizes with, and is thought to phosphorylate, vimentin, an intermediate filament, in differentiated HL-60 cells (30).
Very little is known about the propensity of human neutrophil PKC isoforms to localize to Triton-insoluble fractions or the relationship this localization has to NADPH oxidase activation. Redistribution of p47phox to Triton-insoluble fractions in response to PMA stimulation was diminished when PMN were pretreated with the serine/threonine protein kinase inhibitor H-7 (17), implying that the assembly of NADPH oxidase at the cytoskeleton is mediated by phosphorylation. Curnutte et al. (17) demonstrated translocation of PKC-β (using a general PKC-β Ab that recognizes both PKC-βI and PKC-βII) to Triton-insoluble fractions. However, other isoforms have not been studied. In this study, we examined the correlation between the translocation of three PMA-responsive PKC isoforms (PKC-α, PKC-βII, and PKC-δ) in human PMN, the translocation of p47phox, and the appearance of NADPH oxidase activity in Triton-insoluble fractions. We used two known PKC activators, PMA and 1,2-dioctanoyl-sn-glycerol (diC8). Results indicate that the translocation of cPKC isoforms PKC-α and PKC-βII correlated most closely with the appearance of p47phox and NADPH oxidase activity in cytoskeletal fractions.
Materials and Methods
PMA, Triton X-100, PMSF, aprotinin, leupeptin, pepstatin, n-tosyl phenylalanine chloromethylketone, benzamidine, iodoacetamide, ferricytochrome c, and diisopropylfluorophosphate (DFP) were obtained from Sigma (St. Louis, MO). HBSS and PBS were from Life Technologies (Gaithersburg, MD). Dextran T-500 was purchased from Pharmacia Biotech (Piscataway, NJ). Nitrocellulose was obtained from Schleicher & Schuell (Keene, NH). Superoxide dismutase (SOD) was purchased from DDI Pharmaceuticals (Mountain View, CA). NADPH was from Boehringer Mannheim (Indianapolis, IN). The sources of Abs were as follows: anti-PKC-α, a mAb directed against the catalytic domain, was from Upstate Biotechnology (Lake Placid, NY) (31); anti-PKC-βII (human sequence, amino acids 661–673) (32) was from Oxford Biomedical Research (Oxford, MI); anti-PKC-δ (human sequence, amino acids 658–676) was a gift from David J. Burns, formerly of Sphinx Pharmaceuticals (Durham, NC) (26, 33); polyclonal Abs raised in goats, anti-p47phox, and anti-p67phox were gifts from Thomas L. Leto of the National Institutes of Health (34). HRP-conjugated secondary Abs were obtained from Transduction Laboratories (Lexington, KY; goat anti-mouse IgG and goat anti-rabbit IgG) and Organon Technika (Durham, NC; rabbit anti-goat IgG). A PKC-δ antigenic peptide was purchased from Oxford Biomedical Research. PKC-α and PKC-δ standards were gifts from David J. Burns, formerly of Sphinx Pharmaceuticals. Rabbit brain cytosol was used as the PKC-βII standard and was generated as described under Materials and Methods for PMN cytosol (35). diC8 was from Serdary Research Laboratories (London, Ontario, Canada).
PMN were isolated from heparinized venous blood obtained from healthy human adult donors by dextran sedimentation, Isolymph (Gallard-Schlesinger Industries, Carle Place, NY) centrifugation, and hypotonic lysis of RBC (36). PMN were resuspended at 5 × 107 cells/ml in HBSS supplemented with 4.2 mM sodium bicarbonate and 10 mM HEPES, pH 7.4.
Preparation of Triton-soluble and Triton-insoluble fractions
PMN at 5 × 107 cells/ml were treated with DFP for 5 min (37, 38) and were prewarmed for 5 min at 37°C. At 90 s into the prewarm, NaN3 was added to a final concentration of 1 mM (36). DMSO or ethanol was added as a vehicle control to unstimulated cells. Stimulated cells were treated for varying times at varying concentrations of PMA or diC8. Stimulation reactions were quenched with 45 ml of ice-cold HBSS, followed by centrifugation (250 × g, 10 min, 4°C). Fractions were obtained according to the method of Curnutte et al. (17) with modifications. Briefly, PMN were resuspended in 0.75% Triton X-100, 60 mM PIPES, 25 mM HEPES, pH 7.4, 10 mM EGTA, 2 mM MgCl2, 8 μg/ml aprotinin, 156 μg/ml benzamidine, 1 mM iodoacetamide, 20 μg/ml leupeptin, 100 ng/ml pepstatin, 100 μM n-tosyl phenylalanine chloromethylketone, and 1 mM PMSF, and incubated on ice for 20 min. The cell lysate was layered on the same buffer containing 6% sucrose (1:3, v/v, cell lysate:6% sucrose buffer) and centrifuged at 150,000 × g for 30 min (SW55 rotor; Beckman Instruments, Palo Alto, CA). The top fraction was collected as the noncytoskeletal (Triton-soluble, TS) fraction. The pellet was resuspended by sonication (2 × 10 s) in the same buffer without Triton X-100 and iodoacetamide, and termed the cytoskeletal (Triton-insoluble, TI) fraction. All fractions were stored in 10% glycerol at −70°C. Protein concentrations were determined using the BCA protein assay (Pierce, Rockford, IL) (39) and BSA as standard.
Preparation of cytosol and membrane fractions
PMN at 5 × 107 cells/ml were treated for varying times at varying concentrations of diC8, with ethanol added as a vehicle control to unstimulated cells. Reactions were quenched with a 10-fold excess of cold modified HBSS. Cells were treated with DFP for 5 min (37, 38) and resuspended at 1 × 108 cells/ml in 50 mM Tris, pH 7.5, 2 mM EGTA, 10 μM benzamidine, 1 μg/ml leupeptin, 10 μM pepstatin, 0.2 μg/ml aprotinin, 50 mM 2-ME, and 1 mM PMSF. PMN were sonicated to ∼90% breakage on ice and centrifuged at 800 × g for 10 min to remove nuclei and unbroken cells. The postnuclear supernatant was layered on a 15%/40% discontinuous sucrose gradient (35) and centrifuged at 150,000 × g for 30 min at 4°C (SW55 rotor; Beckman Instruments). The top fraction was collected as the cytosolic fraction and centrifuged again at 150,000 × g for 60 min (Ti50 rotor; Beckman Instruments, Palo Alto, CA). The 15%/40% interface was collected as the membrane fraction. Protein concentrations were determined using the Pierce Coomassie Blue Plus protein assay and BSA as a standard.
Lipid phosphorus assay
Lipids were extracted from PMN fractions by modified Bligh and Dyer (40). Lipid phosphorus was determined as described by Rouser et al. (41) with modifications. Briefly, unknown samples were dried under N2 to remove organic solvents, and were oxidized at 180–190°C for 1 h after addition of 150 μl 70% perchloric acid. After cooling, the tubes were rinsed with 900 μl dH2O. Then 167 μl each of 2.5% ammonium molybdate (w/v) and of 10% ascorbic acid (w/v) were added and the tubes were incubated in a 50°C water bath for 15 min. Absorbance was read at 820 nm using Na2HPO4 dilutions as standards.
SDS-PAGE and Western blotting
Samples were prepared for SDS-PAGE according to total protein by addition of Laemmli sample buffer, boiled for 5 min, and loaded on 7% (PKC) or 9% (p47 and p67) SDS polyacrylamide gels (42). After separation, the proteins were transferred (43) to nitrocellulose overnight. Blots were stained with Ponceau-S (0.2% Ponceau-S, 3% TCA, and 3% sulfosalicylic acid) (44) to visualize m.w. markers and were destained with deionized water, followed by Tris-buffered saline with 0.1% Tween-20 (TBS-T). The blots were blocked with TBS-T containing 5% nonfat dry milk for 60 min, washed, and incubated with the primary Ab diluted in 3% BSA/0.02% NaN3 in PBS, pH 7.3, for 2 h (anti-PKC-α, 1:500; anti-PKC-βII, 1:100; anti-PKC-δ, 1:1500; anti-p47phox, 1:1000; anti-p67phox, 1:1000). Then the blots were washed and incubated with HRP-conjugated secondary Ab for 60 min. The blots were washed again, then incubated with enhanced chemiluminescence (ECL) reagent (Pierce, Rockford, IL) for 1 min, and exposed to film for visualization. Autoradiograms were analyzed and quantified by densitometry (PDI, Huntington Station, NY).
NADPH oxidase assay
Cytoskeletal fractions from stimulated PMN were assayed for NADPH oxidase activity as the SOD-inhibitable reduction of ferricytochrome c, as described by Caldwell et al. (35) with modifications. Briefly, reaction tubes contained 48 mM KPO4 buffer (pH 7), 10 μM flavin adenine dinucleotide, 76 μM ferricytochrome c, 1 mM EGTA, 7.6 mM MgCl2, 1 μM GTPγS, 0.15 mM SDS, and 60–120 μg Triton-insoluble protein in a final volume of 237.5 μl. SOD (final concentration, 50 μg/ml) was added to one-half the reaction mixture. Addition of NADPH (final concentration, 0.2 mM) started the reaction, which was monitored at 550 nm on a UV-2401PC Shimadzu (Columbia, MD). Superoxide produced was calculated from the linear slopes, using an extinction coefficient of 21 mM−1cm−1 (45), and normalized to Triton-insoluble protein.
Characterization of the fractions
We first characterized the Triton-soluble and Triton-insoluble fractions for their protein and phospholipid content. Total protein was divided equally, 2.16 ± 0.06 and 2.17 ± 0.09 mg protein/108 cell equivalents (n = 11) in Triton-soluble and Triton-insoluble fractions, respectively. This protein distribution was comparable with detergent fractionations previously reported in PMN (46). Samples from several experiments were subjected to SDS-PAGE and stained with Coomassie blue. The pattern of proteins in each fraction was consistent between experiments and similar to that reported by Woodman et al. (18) (data not shown). Additionally, the prominent protein bands did not change with cell stimulation. The Triton-soluble fraction, as expected, contained the majority of the phospholipid (∼190 nmol phospholipid/mg protein), while the Triton-insoluble fraction contained ∼30 nmol phospholipid/mg protein. These values did not change with cell stimulation. This phospholipid distribution was similar to that for detergent fractionations previously reported in PMN (47) and other cells (48).
Translocation of PKC and p47phox and Triton-insoluble NADPH oxidase activation are dependent on the concentrations of PMA during stimulation
An assessment of PKC isoform translocation as a measure of activation in Triton-soluble and Triton-insoluble fractions was chosen for the following reasons. First, measurement of PKC activity in Triton-containing fractions is problematic, due to the inhibitor effect of the detergent (data not shown). Second, activity measurements do not differentiate between the activities of individual PKC isoforms within a class. Finally, PKC translocation generally is indicative of PKC activation (21, 23). Fig. 1 A shows the concentration dependence of PKC translocation from Triton-soluble to Triton-insoluble fractions. Isolated human PMN were stimulated with concentrations of PMA ranging from 0–1600 nM for 5 min and separated into Triton-soluble and Triton-insoluble fractions. Decreases in the level of each PKC isoform in the Triton-soluble fractions, concomitant with increases in the Triton-insoluble fractions, were observed with increasing PMA concentrations. There was an almost complete loss of PKC in the Triton-soluble fractions with cell stimulation, as only low percentages of the total present in unstimulated cells were present after treatment with 1600 nM PMA (PKC-α, 5.5%; PKC-βII, 2.7%; PKC-δ, 5.4%; based on densitometric analysis). PKC-βII (33% total recovery at 1600 nM PMA, compared with unstimulated cells) appears to be more sensitive to degradation than PKC-α or PKC-δ (67.4% and 39.3% recoveries, respectively).
We also assessed the distribution of NADPH oxidase components p47phox and p67phox under these conditions (Fig. 1 B). Translocation of p47phox to Triton-insoluble fractions occurred in response to PMA stimulation, in a concentration-dependent manner. Densitometric analysis revealed that 43% of p47phox remained Triton soluble at the highest concentration of PMA used, and the recovery of this protein was 99.3%. Conversely, p67phox was entirely Triton insoluble under all conditions, even in unstimulated PMN. These results confirm and extend previous observations (17, 18) of the localization of these proteins before and after PMA stimulation, supporting the validity of our method. As a result of its unchanging nature, p67phox served as a good marker for protein integrity in the Triton-insoluble fractions under the various conditions of cell stimulation.
Fig. 2,A shows the percentage of the total of each PKC isoform and of p47phox in the Triton-insoluble fractions summarized from several experiments, as assessed by densitometry. Two patterns of PKC isoform translocation are observed. Translocation of PKC-δ was apparent at 30 nM PMA, reached a maximum at 160 nM PMA and then declined. In contrast, the concentration curve for translocation of cPKC-α and cPKC-βII was shifted to the right and did not decline. The concentration curves for translocation of p47phox (Fig. 2,A) and the appearance of NADPH oxidase activity in the Triton-insoluble fraction (Fig. 2 B) were similar to the curves for translocation of the cPKCs. However, the translocation of p47phox plateaued at 160 nM, instead of 300 nM PMA. The similarities in concentration dependence of translocation of cPKCs and p47phox and appearance of Triton-insoluble NADPH oxidase activity, coupled with the different concentration dependence of translocation of nPKC-δ, suggest the involvement of cPKC in NADPH oxidase regulation.
Translocation of PKC and p47phox and appearance of Triton-insoluble NADPH oxidase activation are dependent on time of PMA stimulation
We determined the time course for PMA-induced translocation of PKC isoforms and p47phox, shown in Fig. 3, A and B, respectively. Human PMN were stimulated with 300 nM PMA for 0–10 min, and were separated into Triton-soluble and Triton-insoluble fractions, as described in Materials and Methods. A control for cell exposure to the stimulus during preparation of the Triton-soluble and Triton-insoluble fractions was included, in which either the vehicle (DMSO, 0v) or the stimulus (PMA, 0p) was added at the time the stimulation reactions were terminated. Slight effects on the translocation of PKC and p47phox were apparent in the 0p control, indicating that PMA was not completely inert during the processing period. However, the effects of PMA were more dramatic during the time course conducted at 37°C. A rapid decrease in the level of each PKC isoform in Triton-soluble fractions was observed, with most of each isoform disappearing by 3 min (Fig. 3,A). Concomitant increases in the Triton-insoluble fraction occurred. In some experiments, PKC-βII appeared to return to the Triton-soluble fraction at the 10-min time point, but this finding was inconsistent. PMA also induced a time-dependent decrease in the level of p47phox in Triton-soluble fractions, with a concomitant increase in Triton-insoluble fractions (Fig. 3,B). However, in contrast to the PKC isoforms, 38% of p47phox remained in the Triton-soluble fraction at the 10-min time point. Similar to the findings in Fig. 1 B, p67phox was entirely associated with the Triton-insoluble fraction, and the level of this protein did not change with time of PMA stimulation.
In Fig. 3,A and less evident in Fig. 1,A, PKC-δ appears to be undergoing a gel shift as a result of PMA stimulation. Using the antigenic peptide for the PKC-δ Ab, we confirmed that both bands visualized on the blots were PKC-δ (data not shown). This shift in mobility may be the result of phosphorylation of PKC-δ (see Discussion). For purposes of quantitation by densitometry (see Figs. 2, 4, and 5), the two PKC-δ bands were added together for any given condition.
Fig. 4,A summarizes the appearance of PKC isoforms and p47phox in the Triton-insoluble fractions during the PMA time course, using densitometric analysis. As in the concentration curve experiments, two patterns were observed. The redistribution of cPKCs and p47phox to the Triton-insoluble fractions has a biphasic appearance, with an initial increase and a plateau at 2–3 min, then a further increase and a second plateau between 5 and 10 min. The curves for PKC-α and PKC-βII are similar, although the level of PKC-βII (as percentage of total PKC-βII in unstimulated PMN) was lower due to less recovery of this isoform. The appearance of p47phox in the Triton-insoluble fraction slightly preceded the appearance of PKC-α and PKC-βII. In contrast to the pattern with the cPKCs and p47phox, the level of PKC-δ in the Triton-insoluble fraction did not change after the maximum increase was reached at 2 min. Like PKC-α, PKC-βII, and p47phox, the time course for the appearance of Triton-insoluble NADPH oxidase activity has a biphasic appearance (Fig. 4 B). After an initial 30-s lag, oxidase activity dramatically increased up to 2 min. There was no change in activity from 2–5 min, but there was a further increase at the 10-min time point. Thus, appearance of oxidase activity in the Triton-insoluble fraction lags behind the appearance of p47phox and PKC isoforms and resembles the pattern of translocation of the cPKCs, but not that of nPKC-δ. This suggests that translocation and phosphorylation of oxidase components by a cPKC isoform must occur before superoxide production is observed.
Comparison between DAG and PMA for inducing translocation of PKC and p47phox and appearance of Triton-insoluble NADPH oxidase activity
The actions of PMA on neutrophils can be mimicked, to a limited extent, by more physiological PKC activators, such as cell-permeable DAGs. We examined whether diC8, a cell-permeable DAG, affected the distribution of PKC isoforms and p47phox in Triton-soluble and Triton-insoluble fractions, as well as appearance of NADPH oxidase activity in Triton-insoluble fractions. PMN were treated with either solvent (unstimulated), 70 μM diC8 for 3 min, or 300 nM PMA for either 3 or 10 min. Preliminary experiments determined that these conditions of diC8 stimulation resulted in maximum effects. Triton-soluble and Triton-insoluble fractions were prepared and the summarized levels of each protein and the level of NADPH oxidase activity in the Triton-insoluble fractions are shown in Fig. 5. In unstimulated PMN, 29% of the total PKC-δ was Triton insoluble. In contrast, only 0.1% of PKC-α and no PKC-βII were Triton insoluble in unstimulated PMN. Similarly, only 4.2% of p47phox and virtually no NADPH oxidase activity were present in Triton-insoluble fractions from unstimulated cells.
DiC8 treatment induced only small changes in the distribution of PKC isoforms and p47phox and in the level of NADPH oxidase activity. Of the PKC isoforms, only PKC-α levels showed an increase (10% change in distribution) in Triton-insoluble fractions in response to diC8. This was accompanied by a similar small increase (10.8% change in distribution) in the level of p47phox, and a small increase in NADPH oxidase activity (unstimulated: 0.001 nmol O2−/min/mg; diC8-stimulated: 0.03 nmol O2−/min/mg). As already noted, PMA induced substantial translocation of PKC isoforms and p47phox, as well as in appearance of NADPH oxidase activity. The two patterns observed in Fig. 4 were also apparent in this experiment, in that there were increases between 3 and 10 min in Triton-insoluble levels of PKC-α, PKC-βII, and p47phox, and in oxidase activity, but not in the level of PKC-δ. Taken together, these data with different agonists add support to the concept of a closer relationship between cPKC isoforms and NADPH oxidase activation, than between nPKC-δ and the oxidase.
Although diC8 induced little (PKC-α) or no (PKC-βII, PKC-δ) measurable translocation of PKC isoforms to Triton-insoluble fractions, effects of this agonist on levels of PKC isoforms in Triton-soluble fractions were noted (data not shown). PKC-α levels were reduced in the Triton-soluble fraction from 99.9% in unstimulated PMN to 94.4% with diC8 treatment, consistent with the appearance of ∼10% of PKC-α in the Triton-insoluble fraction. The level of PKC-βII in Triton-soluble fractions declined from 100% in fractions from unstimulated PMN to 89.4% in fractions from diC8-treated PMN and that of PKC-δ declined from 70.8% to 58.4%. These decreases in the Triton-soluble fraction without corresponding increases in the Triton-insoluble fraction imply that diC8 does activate PKC-βII and PKC-δ and increases their sensitivity to proteolysis.
DiC8 induces translocation of PKC-βII and PKC-δ from cytosol to membrane fractions
To confirm the ability of diC8 to activate PKC-βII and PKC-δ in PMN, we examined whether this agonist stimulated translocation of these isoforms to membrane fractions. Previous studies using oleoyl-acetyl-glycerol had shown that this DAG could induce translocation of PKC activity from cytosol to membrane fractions (49). As shown in Fig. 6, treatment of PMN with diC8 for 30 s resulted in a concentration-dependent translocation of PKC-βII and PKC-δ from cytosol to membrane fractions, which began to plateau at 70 μM diC8. Higher concentrations of diC8 were not tested, because they appeared to partially solubilize the PMN. Virtually all of the cytosolic PKC-δ disappeared at 25 and 70 μM diC8, with recovery of ∼15–20% of the isoform in the membrane fraction. Thus, ∼80% of the PKC-δ present in membrane and cytosolic fractions from unstimulated cells was not recovered in either fraction after maximal diC8 treatment, suggesting compartmentalization and/or proteolysis of this isoform. Translocation of some PKC-δ to fractions other than the membrane also is implied by the observation that the total recovery of PKC-δ in Triton-soluble and Triton-insoluble fractions was higher (∼60%, see above). In contrast, the effects of diC8 on the translocation of PKC-βII were less dramatic. The level of PKC-βII in cytosolic fractions decreased by ∼20–25% in response to 25–70 μM diC8, accompanied by an increase in membrane fractions from 0% in unstimulated cells to 3–5% with diC8 stimulation. Thus, total recovery of PKC-βII after maximal diC8 treatment was ∼80–85%, suggesting less compartmentalization or proteolysis of this isoform compared with PKC-δ.
Translocation of PKC-βII and PKC-δ to membrane fractions was also dependent on the time of diC8 treatment (Fig. 7). Translocation in response to diC8 occurred rapidly, with close to maximal effects at 15 s and maximal effects at 30 s for both isoforms. There was a slight tendency for both isoforms to partially return from the membrane to the cytosol at 60 and 120 s, but the changes were small. Overall, the results in Figs. 5–7 demonstrate that stimulation of PMN with diC8 induces translocation of PKC-βII and PKC-δ to membrane fractions, but not to Triton-insoluble fractions.
The results of the present study demonstrate that the activation of PKC by PMA in human PMN induces the translocation of PKC-α, PKC-βII, and PKC-δ isoforms to Triton-insoluble fractions. Detergent-insoluble cellular fractions have been used by many investigators to represent the cytoskeleton (16, 17, 18, 19, 46, 47, 50, 51, 52, 53); thus, we speculate that these PKC isoforms translocate to cytoskeletal elements in response to PMA. This suggests that at least part of the function of these isoforms in PMN is conducted by regulation of cytoskeletal-mediated events. Furthermore, different patterns of translocation were observed for the cPKC isoforms, compared with nPKC-δ, suggesting that these classes of PKC have different translocation mechanisms, different localizations within the Triton-insoluble fractions, and/or different functional roles. Indeed, we found that the pattern of p47phox translocation and NADPH oxidase activation more closely resembled the pattern of cPKC translocation. Thus, our data suggest that the conventional class of PKC isoforms regulates the activation and/or cytoskeletal localization of NADPH oxidase.
Other reports have suggested that cPKCs participate in NADPH oxidase activation. Majumdar et al. (54) demonstrated that p47phox was a better substrate for partially purified neutrophil PKC-β than for an unidentified Ca2+-independent PKC. Additionally, a pseudosubstrate peptide that inhibited PKC-β, but not the Ca2+-independent PKC, reduced O2− release in electropermeabilized PMN. Duyster and coworkers (55) implicated a role for cPKC-β over nPKC-δ in O2− release by rat liver macrophages. Recently, Korchak et al. (56) used PKC-β antisense oligonucleotides to reduce PKC-β levels in HL-60 cells, differentiated to a PMN-like phenotype. The greatest effect on O2− release was seen in FMLP-stimulated PMN, with modest effects on O2− release induced by PMA or a phagocytic agonist. A similar study (57), performed in normal human monocytes, demonstrated marked inhibition of O2− release induced by opsonized zymosan in cells treated with PKC-α, but not PKC-β antisense. Our results support the possibility that cPKC-α participates in the regulation of NADPH oxidase, but we cannot exclude a role for cPKC-β.
The translocation of PKC-α and PKC-βII to the Triton-insoluble fraction preceded NADPH oxidase activation and occurred at similar concentrations of PMA. Thus, cPKC isoforms could either regulate the interaction of NADPH oxidase components with cytoskeletal elements and/or regulate NADPH oxidase activity after the enzyme complex becomes cytoskeletal associated. The inhibition of the translocation of p47phox to cytoskeletal fractions by nonselective protein kinase inhibitors (17) suggests that PKC and/or other protein kinases may regulate this process. However, comparison of the patterns of translocation of p47phox and the cPKCs shows that translocation of p47phox occurred at slightly earlier time points and at similar (cPKC-α) or lower (cPKC-βII) concentrations of PMA. This suggests that NADPH oxidase components, in particular p47phox, could regulate association of cPKC isoforms with the cytoskeleton. Curnutte et al. (17) reported marked reduction of PMA-stimulated translocation of cPKC-β in neutrophils from patients with a form of chronic granulomatous disease, who are missing p47phox, lending support to this possibility. Thus, the relationships between PKC, cytoskeletal elements, and the oxidase are likely to be complex.
Most of the components of NADPH oxidase are phosphorylated during stimulation of intact PMN (58, 59, 60, 61, 62), so the cPKCs may directly phosphorylate one or more components. Indeed, a mixture of the cPKCs purified from rat brain, as well as partially purified cPKC from human PMN, phosphorylate p47phoxin vitro (54, 63). In a cell-free system developed by Babior and colleagues (64), the cPKC-mediated phosphorylation of p47phox is a necessary step for NADPH oxidase activation to occur. However, other protein kinases can phosphorylate p47phoxin vitro (63, 65, 66), and other oxidase components are in vitro substrates for protein kinases besides PKC (62, 67). Additionally, the regulation by PKC could be at the level of intermediate cytoskeletal-associated proteins. Many cytoskeletal-associated proteins are substrates for PKC (28, 30, 68, 69, 70), and cytoskeletal reorganization is regulated by phosphorylation (6, 71). Thus, it is likely that multiple phosphorylation-dependent mechanisms contribute to the regulation of NADPH oxidase.
To further address the relationship between PKC and NADPH oxidase, it is important to identify the cytoskeletal elements with which PKC and oxidase components interact. Previous results implicate interactions between neutrophil NADPH oxidase components and the actin filament system (11, 12, 13, 14). Similarly, cPKC-α and cPKC-β are reported to bind, phosphorylate, or colocalize with actin or actin-binding proteins in macrophages and other cell types (28, 29, 69, 72, 73, 74, 75). Such studies have not been performed in PMN. PMN stimulated with PMA exhibit an increase in F-actin content (76) and membrane ruffling (77, 78). Thus, the actin-based cytoskeleton is likely to participate in the interactions between cPKC and NADPH oxidase components. However, actin-filament disrupting agents (e.g., cytochalasin B) do not affect superoxide release stimulated by PMA (36), possibly because the submembranous actin network is not disrupted (13). NADPH oxidase components and activity previously have been localized to the submembranous cytoskeleton (12, 13). Thus, it will be necessary to more precisely localize the cytoskeletal proteins interacting with the cPKCs and NADPH oxidase components, as well as to determine whether PKC isoforms and oxidase components colocalize.
Translocation of PKC isoforms to the Triton-insoluble fraction occurs at a slower rate than that observed for cytosol to membrane translocation, which is complete by less than 1 min (Ref. 26 , and data not shown). PKC may first associate with the membrane (where lipid activators are present) and then dock with cytoskeletal (Triton-insoluble) proteins that serve as substrates or anchor PKC in proximity with its substrate. In fact, many PKC-binding proteins bind to PKC in a PS-dependent manner (72, 79), suggesting that, in intact cells, membrane association of PKC may be required before binding of PKC to proteins can occur.
It is interesting that, unlike PMA, diC8 induced low levels of only cPKC-α translocation and NADPH oxidase activation in the Triton-insoluble fraction. DAGs are less potent than PMA for PKC activation and activation of NADPH oxidase in whole cells, although the optimal rates of O2− release are comparable (49, 80). However, effects of exogenously added DAGs on PKC and functional responses are transient, while those to PMA are more sustained (81, 82), most likely due to rapid metabolism of the DAG (81, 83). Therefore, the low levels of Triton-insoluble NADPH oxidase activity induced by diC8 may be caused by the rapidly reversible nature of the activation process, similar to results obtained with FMLP (36). The low ability of diC8 to induce translocation of PKC isoforms to Triton-insoluble fractions also may be related to the transient nature of PKC activation. Alternatively, recent evidence in living cells suggests that PKC may localize to different membrane domains, when PMA and diC8 are compared as agonists (84).
Although the translocation of the cPKCs to the Triton-insoluble fraction is linked to activation of the respiratory burst, the role of cytoskeletal-associated nPKC-δ in PMN is not known. About 30% of nPKC-δ was Triton insoluble in unstimulated PMN (Fig. 5), implying a constitutive cytoskeletal association, perhaps with the intermediate filament system (30). PMA induced rapid translocation of the remainder of the isoform from the Triton-soluble to the Triton-insoluble fraction. This response was clearly more sensitive than the response of the cPKCs, occurring at lower concentrations of PMA and at a faster rate. Translocation also was accompanied by a shift in mobility to slower migrating species, possibly as a result of tyrosine phosphorylation by Src family members (85, 86, 87). Members of the Src family of protein tyrosine kinases localize to Triton-insoluble fractions of human PMN upon PMA stimulation (50, 51) and, thus, could phosphorylate nPKC-δ. Further studies are needed to test this possibility and to determine the effects of these putative phosphorylation events on nPKC-δ function in PMN.
The association of PKC isoforms with the cytoskeleton in PMN has additional functional implications for the regulation of PKC directly. Triton-insoluble proteins may serve as an alternate mechanism for PKC activation. In MOLT-4 cells, Blobe et al. (29) demonstrated that PKC-βII was fully activated in the presence of Mg2+ when bound to actin in the absence of Ca2+ and lipids. Also, Triton-insoluble association may represent a means of PKC down-regulation, in which binding to proteins could make the protease-sensitive hinge region more accessible to proteolysis. This latter possibility is supported by our observation that PMA stimulation of PMN decreased the total recovery of each PKC isoform, in conjunction with translocation to the Triton-insoluble fraction.
In conclusion, we report the translocation of the PMA-responsive isoforms of PKC to the cytoskeleton (detergent-insoluble fractions) in PMN. The pattern of translocation was PKC class specific, and a close correlation between cytoskeletal association of cPKC isoforms and NADPH oxidase activity was established. The results strengthen the concept that PKC regulates functional responses in PMN through cytoskeletal-based interactions. To further understand the relationship between PKC, the cytoskeleton, and the activation of NADPH oxidase, studies are currently underway in our laboratory to determine whether oxidase components colocalize with PKC-α or PKC-βII in the Triton-insoluble fractions after PMN stimulation, and to identify other proteins in that fraction with which PKC isoforms associate.
We thank Dr. David Burns of Abbott Laboratories, and Dr. Thomas Leto of the National Institutes of Health for their kind and generous gifts of Abs.
This work was supported by Grant RO1-AI22564 from the National Institutes of Health.
Abbreviations used in this paper: PMN, polymorphonuclear leukocyte or neutrophil; PKC, protein kinase C; cPKC, conventional PKC; DAG, diacylglycerol; DFP, iisopropylfluorophosphate; diC8, 1,2-dioctanoyl-sn-glycerol; ECL, enhanced chemiluminescence; nPKC, novel PKC; phox, phagocyte oxidase; PS, phosphatidylserine; SOD, superoxide dismutase.