Cyclic AMP-increasing Agents Interfere with Chemoattractant-induced Respiratory Burst in Neutrophils as a Result of the Inhibition of Phosphatidylinositol 3-Kinase Rather than Receptor-operated Ca2+ Influx

Superoxide anion and arachidonic acid were produced in guinea pig neutrophils in response to a chemotactic peptide formyl-methionyl-leucyl-phenylalanine (fMLP). Both responses were markedly, but the former response to a phorbol ester was not at all, inhibited when the cellular cAMP level was raised by prostaglandin E1 combined with a cAMP phosphodiesterase inhibitor. Increasing cAMP was also inhibitory to fMLP-induced activation of phosphatidylinositol (PI) 3-kinase and Ca2+ influx without any effect on the cation mobilization from intracellular stores. The fMLP-induced respiratory burst was abolished when PI 3-kinase was inhibited by wortmannin or LY294002, but was not affected when Ca2+ influx was inhibited. On the contrary, fMLP released arachidonic acid from the cells treated with the PI 3-kinase inhibitors as well as from nontreated cells, but it did not so when cellular Ca2+ uptake was prevented. The chemotactic peptide activated PI 3-kinase even in cells in which the receptor-mediated intracellular Ca2+ mobilization and respiratory burst were both abolished by exposure of the cells to a permeable Ca2+-chelating agent. Thus, stimulation of fMLP receptors gave rise to dual effects, activation of PI 3-kinase and intracellular Ca2+ mobilization; both effects were necessary for the fMLP-induced respiratory burst. Increasing cellular cAMP inhibited the respiratory burst and arachidonic acid release as a result of the inhibitions of PI 3-kinase and Ca2+ influx, respectively, in fMLP-treated neutrophils.

The intracellular signaling pathways responsible for the formyl-methionyl-leucyl-phenylalanine (fMLP) 1 -induced respiratory burst in phagocytes remain to be fully understood as yet, although the activation mechanism of the respiratory burst oxidase is currently elucidated as an assembly of a number of membrane-bound and cytosolic components, including gp91 phox , p22 phox , p47 phox , p67 phox , and Rac (see Ref. 1 for review).
In general, one of the best strategies for identification of essential cellular signals involved in a cell response to a receptor stimulation is to search for the target with which an inhibitor interacts to abolish the response efficiently. Pertussis toxin was a good inhibitor and was successfully applied to the fMLPinduced respiratory burst. Prior exposure of guinea pig neutrophils to low concentrations of pertussis toxin for several hours prevented the cells from producing superoxide anion in response to the subsequent addition of fMLP (2). The prevention resulted from the toxin-induced ADP-ribosylation of G protein (Gi-2), the chemical modification by which the modified G protein is uncoupled from receptors. Evidence has been thus provided for involvement of the G protein in the fMLP-induced respiratory burst via phospholipase C activation (2)(3)(4).
An additional example of useful inhibitors of the phagocytic respiratory burst is wortmannin, a fungal metabolite with hydrophobic sterol structure. Baggiolini and his colleagues (5,6) first reported that incubation of human neutrophils with wortmannin or 17-hydroxywortmannin for 5 min inhibited fMLPinduced O 2 Ϫ generation and degranulation without affecting the chemoattractant-induced Ca 2ϩ mobilization. The inhibition induced by wortmannin was so selective that it did not inhibit similar responses of neutrophils to a phorbol ester, a direct inhibitor of protein kinase C. Wortmannin has lately been proven to abolish the cellular signaling as a result of its direct interaction with phosphatidylinositol (PI) 3-kinase, providing convincing evidence for essential roles of this enzyme or the products of the enzyme in signaling pathways connecting G protein-coupled receptors and the respiratory burst (7,8).
The purpose of the present paper is further application of this useful strategy to identification of important signaling pathways responsible for the fMLP-induced respiratory burst in guinea pig neutrophils. Increases in cAMP within cells have long been known to inhibit receptor-mediated respiratory bursts in reticulocytes and macrophages (e.g. Refs. 9 and 10 and references cited therein). We have identified intracellular signals that were impaired in parallel with the inhibition of fMLP-induced O 2 Ϫ generation upon the addition of cAMP-increasing agents to the cells.
Cells-Neutrophils were isolated from peritoneal exudates of Hartley guinea pigs of 400 g body weight 18 h after intraperitoneal injection of sodium casein,and cultured for 4 h with or without 50 ng/ml of pertussis toxin at 37°C in the RPMI 1640 medium under 95% air, 5% CO 2 , as described previously (11). The cells were loaded with [ 3 H]arachidonic acid (carrier-free, 1 Ci/5 ϫ 10 7 cells/dish) or Fura2/AM (3 M) for 4 h or for a final 1 h, respectively, during the culture for the purpose of monitoring arachidonic acid release and [Ca 2ϩ ] i changes subsequently. Following the culture, the monolayer was washed with the Ca 2ϩ -free Hepes-buffered medium which had been prepared by replacing CaCl 2 by 1 mM EGTA in the regular Krebs-Ringer-Hepes medium (134 mM NaCl, 4.7 mM KCl, 1.2 mM KH 2 PO 4 , 1.2 mM MgSO 4 , 2.5 mM CaCl 2 , 5 mM glucose, 0.2% bovine serum albumin, 20 mM Hepes, pH 7.4). The cells were then scraped off with a rubber policeman, washed, and finally suspended in the regular Krebs-Ringer-Hepes medium to be subjected to further analyses. Freshly isolated (noncultured) cell suspensions were used in some experiments in which cells were not intended to be loaded with the toxin, the fluorescent dye, or the radioactive metabolite; essentially the same results were obtained between the fresh and cultured cells as to O 2 Ϫ generation, 32 P labeling ,and their susceptibility to stimulants or inhibitors studied.
Superoxide Generation-The suspension of neutrophils (10 6 cells/0.5 ml) was first incubated for 10 or 15 min at 37°C in the presence or absence of inhibitors or cAMP-generating agents (henceforth both referred to solely as inhibitors for brevity) in the regular Krebs-Ringer-Hepes medium before being then stimulated with fMLP or PMA in the presence of 80 M cytochrome c. The stimulation was stopped by adding ice-cold Ca 2ϩ -free Krebs-Ringer-Hepes medium (fortified with 10 mM EGTA), and the mixture was quickly centrifuged (1,500 ϫ g for 10 s) to afford the supernatant to be assayed for O 2 Ϫ generation by measurement of the reduction of cytochrome c based on the difference spectrum at 550 -540 nm.
Measurement of [Ca 2ϩ ] i -The Fura2-loaded neutrophils (5 ϫ 10 6 cells/0.5 ml), after being exposed to vehicles or inhibitors for 10 or 15 min at 37°C, were analyzed, with the additions shown in the figures, for Fura2 fluorescence at 510 nm with the excitation at 340 and 380 nm in the regular or Ca 2ϩ -free Krebs-Ringer-Hepes medium. Fluorescence was determined in a dual excitation wavelength fluorometer (Hitachi, F-2000) and expressed as a ratio of the fluorescence at the two excitation wavelengths to calculate [Ca 2ϩ ] according to the equation described previously (12).
Arachidonic Acid Release-Aliquots (4 ϫ 10 6 cells/0.5 ml) of [ 3 H]arachidonate-labeled neutrophil suspensions, prepared as described above, were first exposed to inhibitors for 10 min and then incubated for 15 min with stimulants at 37°C. The incubation was stopped by the addition of 0.5 ml of Ca 2ϩ -free Krebs-Ringer-Hepes medium fortified with 10 mM EGTA. This was followed by immediate centrifugation to give the supernatant, the radioactivity in which, was determined as arachidonate release from the cells.

P Labeling of Phospholipids as a Measure of PI 3-Kinase in Intact
Cells-Production of 32 P-labeled PIP 3 was estimated by the method described previously (13). Neutrophil suspensions (10 8 cells/ml) were labeled with 32 P i (150 Ci/ml) by 30-min incubation at 30°C in the medium consisting of 10 mM Hepes-NaOH, pH 7.4, 136 mM NaCl, 4.9 mM KCl, and 5.5 mM glucose. These labeled cells were washed twice, resuspended at the density of 10 7 cells/ml in the regular Krebs-Ringer-Hepes medium, and, as 250-l aliquots, incubated at 37°C for 10 or 15 min with or without inhibitors before the final stimulation with 0.1 M fMLP for 15 s or longer. After stimulation, cells were lysed by vigorous stirring in 1.55 ml of chloroform/methanol/8% HClO 4 (50:100:5), followed by addition with 1 ml of the chloroform/8% HClO 4 (1:1) mixture to separate the organic phase, which was washed with chloroformsaturated 1% HClO 4 and dried in vacuo. The lipids were dissolved in 20 l of chloroform/methanol (95:5) and spotted on a thin-layer plate (Silica Gel 60, Merck), which had been impregnated with potassium oxalate by the procedure of development in a solvent system of 1.2% potassium oxalate-containing methanol/water (2:3) and activated by heating at 110°C for 20 min before spotting. The plate was developed in chloroform/acetone/methanol/acetic acid/water (80:30:26:24:14), dried, and visualized for the radioactivities with a Fuji BAS2000 Bioimaging Analyzer.
Measurement of cAMP-Incubation of neutrophil suspensions with 0.3 mM IBMX was stopped by the addition of HCl to make the final concentration of 0.1 N, which was followed by maintaining the dishes in boiling water for 3 min and centrifugation. The resultant supernatant was submitted for radioimmunoassay of cAMP as described previously (11).

Selective Inhibition of Respiratory Bursts by cAMP-increasing Agents and Various Inhibitors-Fig. 1 shows typical O 2
Ϫ generating responses of guinea pig neutrophils to additions of increasing concentrations of fMLP and PMA. The concentration-dependent increases in O 2 Ϫ generation were observed for fMLP and PMA with the maximal responses being at 100 and 10 nM, respectively. The cell responses to fMLP were markedly reduced ( Fig. 1A), but those to PMA were not affected (Fig. 1B), when the cells had been exposed for 10 min to agents that can increase the cellular content of cAMP (an adenylate cyclase activator, PGE 1 , an inhibitor of cAMP-hydrolyzing phosphodiesterase, IBMX, and their combination) or a cell-permeable cAMP analogue (Bt 2 cAMP). Actually, the cAMP content in neutrophils was less than 1, 6.2 Ϯ 0.1, and 91 Ϯ 4.2 pmol/10 7 cells (with the number of observations of six) after 10-min incubation without any addition, with 0.3 mM IBMX alone and its combination with 10 M PGE 1 , respectively, under these conditions. Prior exposure of neutrophils to pertussis toxin also abolished the cell response to fMLP without affecting the response to PMA (data not shown).
The respiratory burst induced by fMLP was likewise inhibited by wortmannin ( Fig. 2A) and staurosporine (Fig. 2B) in concentration-dependent manners. LY294002, a wortmanninlike inhibitor of PI 3-kinase (14,15), gave rise to effects, just like wortmannin (8), very similar to those caused by cAMPincreasing agents; fMLP-induced O 2 Ϫ generation was, but the PMA-induced generation was not at all, inhibited by the inhibitor (Fig. 2C). It is conceivable that the increase in cAMP may interact with the step that functions, like PI 3-kinase, upstream of protein kinase C in G protein-initiated signaling pathways. Bisindolylmaleimide, an inhibitor of protein kinase C much more specific than staurosporine (16), inhibited, in concentration-dependent manners, the fMLP-induced respiratory burst as well as PMA-induced one (Fig. 2D). Thus, activation of protein kinase C, which was not susceptible to cAMP increases within cells or PI 3-kinase inhibitors such as wortmannin and LY294002, was essential for the G protein-initiated respiratory burst. In fact, staurosporine did, but wortmannin did not, inhibit PMA-induced O 2 Ϫ generation at all the concentrations of the phorbol ester employed (Fig. 2E).
Inhibition by cAMP-increasing Agents of Receptor-operated Ca 2ϩ Influx-Fura2-loaded neutrophils exhibited marked increases in [Ca 2ϩ ] i in response to the addition of fMLP (Fig. 3). The increase in [Ca 2ϩ ] i in response to fMLP was transient. The rapid increase was followed by a rapid decrease to form a sharp peak of [Ca 2ϩ ] i within 30 -60 s, after which [Ca 2ϩ ] i leveled off at a much higher level than the prestimulation value in Ca 2ϩcontaining medium (Fig. 3A). Such sustained increases in [Ca 2ϩ ] i were mostly due to continuous Ca 2ϩ influx following receptor stimulation; the rapid [Ca 2ϩ ] i peak became much smaller in magnitude in the incubation medium from which Ca 2ϩ was nominally omitted (Fig. 3B). Addition of Ca 2ϩ , within 3 min after receptor stimulation, into the Ca 2ϩ -free medium caused a marked increase in [Ca 2ϩ ] i (Fig. 3B). The Ca 2ϩ addition to the cell suspension in the Ca 2ϩ -free medium, without receptor stimulation, did not cause any significant increase in [Ca 2ϩ ] i (data not shown, but see a dotted trace in Fig. 4 for example). Thus, the first and the second peaks of [Ca 2ϩ ] i in Fig.  3B reflect mobilization of Ca 2ϩ from the intracellular stores and Ca 2ϩ influx from the exterior, respectively, in response of neutrophils to fMLP receptor stimulation.
When fMLP receptors were uncoupled from G proteins by prior exposure of cells to pertussis toxin, the agonist failed to increase [Ca 2ϩ ] i significantly in either the presence (Fig. 3C) or the absence (Fig. 3D) of extracellular Ca 2ϩ . No significant [Ca 2ϩ ] i increase occurred upon further addition of Ca 2ϩ into the Ca 2ϩ -free medium bathing pertussis toxin-and fMLPtreated neutrophils, providing evidence for involvement of the toxin-sensitive G protein in the Ca 2ϩ influx (Fig. 3D). In the cells pre-exposed to the cAMP-increasing agents (PGE 1 plus IBMX), the peak value of fMLP-induced [Ca 2ϩ ] i increase was slightly smaller than the non-exposed cells and was followed by the return to the level much lower than that observed for control ( Fig. 3E as compared with 3A), suggesting that receptor-coupled Ca 2ϩ influx was inhibited upon increases in the cellular cAMP concentration. In fact, addition of fMLP into Ca 2ϩ -free medium caused the same increase in [Ca 2ϩ ] i in the cells treated with cAMP-increasing agents as in control cells, whereas the subsequent addition of Ca 2ϩ elicited much smaller increase in [Ca 2ϩ ] i in the former cells than in the latter cells (Fig. 3F). Thus, elevation of intracellular cAMP attenuated fMLP receptor-operated Ca 2ϩ influx without affecting the receptor-mediated mobilization of Ca 2ϩ from the internal stores. There was only 20 -30% decrease in fMLP-induced generation of inositol 1,4,5-trisphosphate, when guinea pig neutrophils had been exposed to the cAMP-generating agents (data not shown). Conceivably, such slightly attenuated generation of inositol 1,4,5-trisphosphate is still enough for the maximal mobilization of Ca 2ϩ from the stores.
The inhibition of receptor-operated Ca 2ϩ influx was very unique to the action of cAMP-generating agents, in the sense that other inhibitors of fMLP-induced superoxide anion release, such as wortmannin and staurosporine, did not inhibit but rather enhanced the receptor-coupled Ca 2ϩ influx (Figs. 3,

G-J).
Staurosporine may enhance Ca 2ϩ entry by antagonizing protein kinase C-induced phosphorylation of proteins that is possibly involved in inhibition of capacitative Ca 2ϩ influx in this cell type (17), although the mechanism for wortmannininduced enhancement (Fig. 3H) remains to be clarified.
The addition of Ca 2ϩ following thapsigargin into neutrophil suspensions in the Ca 2ϩ -free medium gave rise to a marked increase in [Ca 2ϩ ] i , despite lack of the increase without preaddition of the endoplasmic Ca 2ϩ pump inhibitor, due to the capacitated Ca 2ϩ entry following emptying of the intracellular stores but not mediated by receptor stimulation (Fig. 4). This capacitated Ca 2ϩ entry was not affected significantly, or inhibited only slightly, by exposure of cells to cAMP-generating agents (Fig. 4). Emptying the Ca 2ϩ stores after the thapsigargin treatment was evidenced by the failure of the treated cells to respond to fMLP by increasing [Ca 2ϩ ] i (data not shown). Thus, the major target of increased cAMP or cAMP-dependent protein kinase appears to be the protein(s) involved in receptoroperated, rather than store-operated, Ca 2ϩ channels.

Ca 2ϩ Influx as a Mediator of Arachidonic Acid Release Rather than O 2
Ϫ Generation-The chemoattractant-induced superoxide anion release was associated with an increase in Ca 2ϩ influx (Fig. 3). Both effects of fMLP were markedly attenuated upon increases in intracellular cAMP. The increase in fMLP receptor-coupled entry of extracellular Ca 2ϩ , however, never played any essential role in O 2 Ϫ generation, which occurred and was inhibited by cAMP-increasing agents in the same magnitude in either the presence or absence of Ca 2ϩ in the incubation medium (Fig. 5). In sharp contrast, arachidonic acid release provoked by fMLP receptor stimulation was highly dependent on the translocation of Ca 2ϩ from the extracellular fluids (Fig.  6). No significant release of arachidonate was observed if the medium was depleted of Ca 2ϩ (Fig. 6B).
Again, wortmannin and cAMP-increasing agents exerted different effects on the response of neutrophils. The fungal metabolite did not inhibit arachidonic acid release under any condition tested, whereas PGE 1 plus IBMX was an inhibitor of the fMLP-induced release as strong as Ni 2ϩ or Co 2ϩ , antagonists of transmembrane Ca 2ϩ inflow (Fig. 6A). Ni 2ϩ has been reported to be an inhibitor of receptor-operated Ca 2ϩ entry (18). Staurosporine, like wortmannin, did not inhibit fMLP-induced arachidonic acid release. Failure of cAMP-increasing agents to inhibit Ca 2ϩ ionophore-induced arachidonate production (Fig.  6A) is in agreement with the view that elevation of intracellu-lar cAMP inhibits the fatty acid production as a result of the inhibition of receptor-coupled Ca 2ϩ entry.
Prevention of fMLP-induced PI 3-Kinase Activation in Cells Whose cAMP Content Was Raised-32 P-Loaded neutrophils were incubated with fMLP for a short time of 15 s to detect fMLP-induced PIP 3 production as a result of PI 3-kinase activation (Fig. 7A). Progressively less activation of PI 3-kinase was evoked by fMLP as the concentration of PGE 1 was increased from 0.01 to 10 M in the IBMX-containing medium bathing the cells for 10 min prior to the addition of the chemoattractant. The range of the effective concentrations of PGE 1 was essentially the same when comparison was made between PI 3-kinase activation (Fig. 7A)  Ϫ generation depending on increasing concentrations of PGE 1 was inversely correlated with progressive increases of intracellular cAMP under the same conditions (Fig. 7B). Thus, the increase in intracellular cAMP antagonized both actions of fMLP to induce superoxide generation and to activate PI 3-kinase.
PI 3-Kinase Activation and Intracellular Ca 2ϩ Mobilization Are Differentially Responsible for Superoxide Anion Generation-No increase in [Ca 2ϩ ] i was observed, even in the presence of the cation in the incubation medium, following fMLP addition to neutrophils that had been exposed to BAPTA/AM, a cell-permeable Ca 2ϩ -chelating agent (Fig. 8, A and B). The intracellular Ca 2ϩ chelation abolished the superoxide-generating response of the cells to fMLP as well (Fig. 8D), but did not impair the chemoattractant-induced activation of PI 3-kinase (Fig. 8C). Such a manner in which BAPTA/AM inhibited fMLPinduced O 2 Ϫ generation formed a sharp contrast with the manner for the inhibition by cAMP-generating agents (Fig. 8D), which interfered with fMLP-induced PI 3-kinase activation under similar conditions (Fig. 8C) without affecting intracellular Ca 2ϩ mobilization (Fig. 3F). Thus, the mobilization of Ca 2ϩ from internal stores and the PI 3-kinase activation occurred in two separate signaling pathways emanating from fMLP receptor stimulation, eventually contributing to the same response of superoxide anion release from neutrophils. No arachidonate was released in response to fMLP from BAPTA-treated cells (data not shown), presumably because extracellular Ca 2ϩ did not enter the cells under these conditions (Fig. 8B).

Different Signals Arising from fMLP Receptor Stimulation Are Responsible for Superoxide Generation and Arachidonate
Release-In the present paper, we measured two kinds of neutrophil responses, O 2 Ϫ generation and arachidonic acid release, to the addition of fMLP. Both responses were markedly attenuated by the agents (e.g. PGE 1 plus IBMX) causing accumulation of cAMP in cells. The other inhibitor of both responses was BAPTA/AM, a cell-permeable Ca 2ϩ -chelating agent, that abolished fMLP-induced intracellular Ca 2ϩ mobilization and the subsequent Ca 2ϩ influx (Fig. 8). It is also well known that prior exposure of neutrophils to pertussis toxin abolishes all the cell responses to fMLP including O 2 Ϫ generation and arachidonic acid release (see Ref. 19 for review). Other inhibitors (or maneuver of cells) tested here inhibited only either of these two responses. The chemoattractant-induced O 2 Ϫ generation was totally inhibited by protein kinase C inhibitors, staurosporine (and by bisindolylmaleimide also), and by direct and selective inhibitors of PI 3-kinase, wortmannin (and by LY294002 also), despite the failure of these inhibitors to affect fMLP-induced arachidonic acid release (Fig. 6). Vice versa, omission of Ca 2ϩ from the incubation medium, just like the addition into the medium of Ni 2ϩ or Co 2ϩ , direct inhibitors of transmembrane Ca 2ϩ permeation, abolished arachidonate-releasing responses of neutrophils to fMLP and A23187, a Ca 2ϩ -ionophore (Fig. 6) without any modification of fMLP-induced O 2 Ϫ generation and its susceptibility to cAMP-increasing agent-induced inhibition (Fig. 5).
It is very likely, therefore, that fMLP receptor stimulation triggered two signaling pathways each, with different susceptibility to certain inhibitors, separately leading to superoxidegenerating and arachidonate-producing responses of neutrophils. fMLP is one of the chemokine families whose receptors possessing seven membrane-spanning regions release, upon being occupied by the ligand, the ␤␥-subunits from coupled pertussis toxin-sensitive G proteins. The signaling pathways may bifurcate at a point distal to the G protein, as evidenced by the fact that both responses are totally inhibited by treatment of the cells with pertussis toxin.
Ca 2ϩ Influx as cAMP-susceptible Signals Essential for fMLPinduced Arachidonic Acid Release-The G␤␥-protein is a direct activator of phospholipase C␤ (20), a product of which, inositol 1,4,5-trisphosphate, causes mobilization of Ca 2ϩ from the internal stores in the endoplasmic reticulum that is immediately followed by external Ca 2ϩ influx across the plasma membrane through the receptor-operated and store-operated channels. The Ca 2ϩ influx was essential for the observed arachidonate release, which was abolished by omitting Ca 2ϩ from the medium or by inhibiting the flux by Co 2ϩ or Ni 2ϩ (Fig. 6). Stimulation of Ca 2ϩ influx by a Ca 2ϩ -ionophore gave rise to arachidonic acid release. Cytosolic phospholipase A 2 is likely to be activated as a result of Ca 2ϩ -dependent translocation into membranes where the substrate phospholipids are available (21)(22)(23)(24).
Thus, it is very reasonable to assume that cAMP-increasing agents inhibit fMLP-induced arachidonic acid release as a result of their unique action to suppress Ca 2ϩ influx following the mobilization of Ca 2ϩ from internal stores (Fig. 3F), in accordance with previous reports (9,10). It is tempting to speculate that receptor-operated Ca 2ϩ channel is one of the targets of cAMP-dependent protein kinase, since cAMP-increasing agents were incapable of inhibiting thapsigargin-induced Ca 2ϩ influx (Fig. 4) as well as Ca 2ϩ ionophore-induced arachidonic acid release (Fig. 6A). Cyclic AMP has recently been reported to inhibit thrombin receptor-operated Ca 2ϩ influx into platelets directly (25) or indirectly (26).
Essential Role of cAMP-susceptible PI 3-Kinase in the fMLPinduced Respiratory Burst-The respiratory burst was inhibited by both inhibitors of PI 3-kinase and protein kinase C when it was induced by fMLP, but was not inhibited by the lipid kinase inhibitors when it was evoked by PMA (Fig. 2). Likewise, cAMP-increasing agents, which inhibit fMLP-induced PI 3-kinase and O 2 Ϫ generation, exerted no effect on the PMA-induced respiratory burst (Fig. 1). Phosphorylation of cytosolic components of the respiratory burst oxidase, particularly p47 phox , is one of the important activation mechanisms of the oxidase (1). At least 7 of the 11 serine residues on the p47 phox molecule responsible for the activation were phosphorylated in PMA-activated neutrophils (27), and the phosphorylation maintains the oxidase in the assembled/active state (28). Thus, PI 3-kinase, which is inhibited when cellular cAMP increased, is located upstream of protein kinase C in the signaling cascade arising from fMLP receptors and leading to the respiratory burst.
Priming Role of Ca 2ϩ Mobilization for the fMLP-induced Respiratory Burst-A number of recent reports have revealed that stimulation of fMLP receptors in neutrophils or neutrophil-type culture cell lines gives rise to phosphorylation of protein tyrosine residues (29 -31), PI 3-kinase activation (7,8,13,(32)(33)(34)(35)(36)(37), and Ras activation leading to MAP kinase cascades via activation of Raf-1 (30, 38 -42). It is unlikely that inositol 1,4,5-trisphosphate-induced Ca 2ϩ mobilization functions downstream, or bifurcates from any site, of these tyrosine kinaserelated signaling pathways, since IP 3 -generating phospholipase C␤ is directly activated by receptor-coupled G␤␥ (20). There are few papers reporting a decisive role of Ca 2ϩ mobilization to trigger any tyrosine kinase signaling system. For instance, concanavalin A can activate a tyrosine kinase (43) or MAP kinase (38,44) in Ca 2ϩ -depleted human neutrophils. Likewise, an increase in [Ca 2ϩ ] i was not essential for fMLPinduced MAP kinase cascade activation (39). In fact, fMLP activated PI 3-kinase even in cells in which the chemoattractant did not cause Ca 2ϩ mobilization, Ca 2ϩ influx, or O 2 Ϫ generation after the treatment of cells with BAPTA/AM (Fig. 8C).
Thus, a readily acceptable possibility is that internal Ca 2ϩ mobilization primes, permits, or supports the signaling pathways responsible for O 2 Ϫ generation. In other words, the Ca 2ϩ mobilization and the tyrosine kinase-related signalings including activation of PI 3-kinase are both indispensable for fMLP to provoke O 2 Ϫ generation. The small amount of Ca 2ϩ mobilized in the Ca 2ϩ -free medium (e.g. Fig. 3B) would be enough to support the respiratory burst.
The Site with Which an Increase in Cellular cAMP or cAMPdependent Protein Kinase Interacts-There is no evidence for direct phosphorylation by cAMP-dependent protein kinase (protein kinase A) of the p85 regulatory subunit or the p110 catalytic subunit of PI 3-kinase. Instead, a signaling pathway connecting the fMLP receptor stimulation and the PI 3-kinase activation appears to be a target of protein kinase A. Increasing cAMP or activation of cAMP-dependent protein kinase A has been reported to antagonize various receptor-initiated activation of MAP kinase cascades (42,(45)(46)(47)(48)(49)(50)(51)(52)(53). The target of protein kinase A was located upstream of MAP kinase kinase (45), between p21 ras and Raf-1 (47,51) or Raf-1 itself (42, 48 -50, 52). On the other hand, PI 3-kinase is directly activated by G␤␥ in neutrophils (54) and platelets (55) or by Ras (56), Rac (57), or Rho (58,59) proteins in various cell types. Alternatively, PI 3-kinase has been found to be located upstream of the Ras protein in certain growth factor-initiated signaling pathways (60 -62). Further studies are thus required before any decisive conclusion will be made as to how increasing cAMP inhibits fMLP-induced PI 3-kinase in neutrophils.