Immune Complex-induced Integrin Activation and L-plastin Phosphorylation Require Protein Kinase A*

Integrins in resting leukocytes are poorly adhesive, and cell activation is required to induce integrin-mediated adhesion. We recently demonstrated a close correlation between phosphorylation of Ser5 in L-plastin (LPL), a leukocyte-specific 67-kDa actin bundling protein, and activation of αMβ2-mediated adhesion in polymorphonuclear neutrophils (PMN) (Jones, S. L., Wang, J., Turck, C. W., and Brown, E. J. (1998) Proc. Natl. Acad. Sci. U. S. A.95, 9331–9336). However, the kinase that phosphorylates LPL Ser5 has not been identified. We found that cAMP-dependent protein kinase (PKA), but not a variety of other serine kinases, can specifically phosphorylate LPL and LPL-derived peptides on Ser5 in vitro. The cell-permeable cAMP analog 8-bromo-cAMP and the adenylate cyclase activator forskolin both induce LPL phosphorylation in cells. Two PKA inhibitors, H89 and KT5720, inhibited immune complex (IC)-stimulated LPL phosphorylation as well as IC-induced activation of αMβ2-mediated adhesion in PMN. The dose response of H89 inhibition of PMN adhesion correlated with its inhibition of LPL phosphorylation in response to IC. IC stimulation also transiently increased intracellular cAMP concentration in PMN. Thus, PKA functions in an integrin activation pathway initiated by IC binding to Fcγ receptors in addition to its better known role as a negative regulator of cell activation by G protein-coupled receptors. In contrast, LPL Ser5 phosphorylation and PMN adhesion induced by formylmethionyl-leucylphenylalanine or phorbol myristate acetate were not affected by PKA inhibitors, suggesting that a different kinase(s) is responsible for LPL phosphorylation in response to these agonists. Phosphoinositidyl 3-kinase also is required for FcγR but not formylmethionyl-leucylphenylalanine- or phorbol myristate acetate-induced LPL phosphorylation and activation of αMβ2. Two phosphoinositidyl 3-kinase inhibitors blocked FcγR-induced cAMP accumulation, demonstrating that this kinase acts upstream of PKA. These data demonstrate a necessary role for PKA in IC-induced integrin activation and LPL phosphorylation.

However, the kinase that phosphorylates LPL Ser 5 has not been identified. We found that cAMP-dependent protein kinase (PKA), but not a variety of other serine kinases, can specifically phosphorylate LPL and LPL-derived peptides on Ser 5 in vitro. The cell-permeable cAMP analog 8-bromo-cAMP and the adenylate cyclase activator forskolin both induce LPL phosphorylation in cells. Two PKA inhibitors, H89 and KT5720, inhibited immune complex (IC)-stimulated LPL phosphorylation as well as IC-induced activation of ␣ M ␤ 2 -mediated adhesion in PMN. The dose response of H89 inhibition of PMN adhesion correlated with its inhibition of LPL phosphorylation in response to IC. IC stimulation also transiently increased intracellular cAMP concentration in PMN. Thus, PKA functions in an integrin activation pathway initiated by IC binding to Fc␥ receptors in addition to its better known role as a negative regulator of cell activation by G proteincoupled receptors. In contrast, LPL Ser 5 phosphorylation and PMN adhesion induced by formylmethionylleucylphenylalanine or phorbol myristate acetate were not affected by PKA inhibitors, suggesting that a different kinase(s) is responsible for LPL phosphorylation in response to these agonists. Phosphoinositidyl 3-kinase also is required for Fc␥R but not formylmethionylleucylphenylalanine-or phorbol myristate acetate-induced LPL phosphorylation and activation of ␣ M ␤ 2 . Two phosphoinositidyl 3-kinase inhibitors blocked Fc␥R-induced cAMP accumulation, demonstrating that this kinase acts upstream of PKA. These data demonstrate a necessary role for PKA in IC-induced integrin activation and LPL phosphorylation.
Leukocyte integrins are able to modulate avidity for their ligands. While circulating in blood or lymph, leukocytes maintain their integrins in a low adhesive state. In contrast, when leukocytes migrate out of the vasculature, their integrins have higher avidity, allowing them to participate in the processes of extravasation and migration through extracellular matrix. This regulation of integrin adhesion is essential for appropriate leukocyte function, since integrin activation is required for leukocyte migration to sites of inflammation and for lymphocyte recirculation through lymph nodes, but inappropriate activation leads to significant injury of normal tissues (1,2). Many molecules specifically found at inflammatory sites or in lymph nodes can induce the transition of integrins from low to high avidity, and exposure to these molecules provides the stimulus both for transendothelial migration and for migration through the extracellular matrix to sites of infection and inflammation (3,4).
The molecular mechanisms that modulate integrin avidity are not well understood. There is evidence for activation-induced increases in integrin affinity (5-7), integrin clustering (8), and integrin diffusion (9). At a molecular level, phosphoinositidyl 3-kinase (PI 3-kinase) 1 (10), protein kinase C (11,12), and the Ca 2ϩ -dependent protease calpain (13) all can have a role in leukocyte integrin activation. In addition, there is evidence for involvement of the actin cytoskeleton in regulation of integrin avidity (9,14). We recently demonstrated that cellpermeant peptides from the leukocyte-specific actin bundling protein L-plastin (LPL) will activate myeloid ␤ 2 integrins (15). These studies showed that LPL phosphorylation is closely associated with integrin activation and led to the hypothesis that LPL phosphorylation is a necessary step for integrin activation by many proinflammatory agents. This hypothesis has focused attention on the unknown enzyme(s) that phosphorylate LPL.
A well studied mechanism for leukocyte integrin activation is exposure to immune complexes (IC). IC binding to polymorphonuclear leukocyte (PMN) IgG Fc receptors (Fc␥R), especially Fc␥RIIA, induces integrin activation via a PI 3-kinase and PKC-dependent pathway, which is closely associated with LPL phosphorylation (12,15). IC-induced activation of the leukocyte-specific integrin ␣ M ␤ 2 (Mac-1, CD11b/CD18) is required both for sustained adhesion to IC in vitro (12,16) and for a normal IC-induced inflammatory response in vivo (17). Thus, IC-induced ␣ M ␤ 2 activation is probably a critical event in an- 1 The abbreviations used are: PI 3-kinase, phosphatidylinositol 3-kinase; Fc␥R, receptor for the Fc piece of IgG; fMLP, formylmethionylleucylphenylalanine; IC, immune complex(es); LPL, L-plastin; PKA, cAMP-dependent protein kinase A; PKG, cGMP-dependent protein kinase G; PMA, phorbol 12-myristate 13-acetate; PMN, polymorphonuclear neutrophil(s); mPMN, murine PMN; S5A, synthetic peptide based on amino acids 2-19 of LPL in which Ser 5 has been changed to Ala; S7A, synthetic peptide based on amino acids 2-19 of LPL in which Ser 7 has been changed to Ala; SCR, synthetic peptide in which amino acids 2-19 of LPL have been scrambled; PKC, protein kinase C; 8-Br-cAMP and -cGMP, 8-bromo-cyclic AMP and GMP, respectively; HBSS, Hanks' balanced salt solution; BSA, bovine serum albumin. tibody-dependent host defense against infectious diseases as well as in idiopathic inflammatory diseases due to autoantibodies. In the current study, we have used this model for integrin activation to examine the hypothesis that LPL phosphorylation is an essential step in integrin activation. We show that the cAMP-dependent Ser/Thr kinase PKA phosphorylates LPL in vitro on Ser 5 , the site of in vivo phosphorylation, and that IC-induced LPL phosphorylation requires PKA. Inhibition of PKA blocks sustained PMN adhesion to IC because ␣ M ␤ 2 activation is prevented. These data not only support the hypothesis that LPL phosphorylation is a step in leukocyte integrin activation but also show a surprising and unexpected role in leukocyte activation for cAMP and PKA, generally considered to be potent endogenous inhibitors of activation. Finally, PKA is not required for LPL phosphorylation or integrin activation in response to formylmethionyl-leucylphenylalanine (fMLP) or phorbol 12-myristate 13-acetate (PMA). Thus, there are at least two distinct signaling pathways and kinases that lead to integrin activation through LPL Ser 5 phosphorylation.
LPL amino-terminal peptide and other mutant peptides were synthesized in the Protein Chemistry Laboratory (Washington University, St. Louis, MO). The following previously characterized peptides (15) were used in this study: amino acids 2-19 of human LPL (ARGSVS-DEEMMELREAFA), a mutant peptide in which Ser 5 has been changed to Ala (S5A) (ARGAVSDEEMMELREAFA), a mutant peptide in which Ser 7 has been changed to Ala (S7A) (ARGSVADEEMMELREAFA), and a peptide in which the 19 amino acids have been scrambled (SCR) (AGDESEMEFVMASALRRE).
Generation of Anti-phospho-LPL Antibody-A polyclonal anti-phospho-LPL antibody was generated against a peptide encoding LPL amino acids 2-11 (ARGSVSDEEM) in which Ser 5 was phosphorylated. The serum from immunized rabbits was collected and purified first by "negative" affinity on unphosphorylated peptide and then "positive" affinity on the phosphorylated peptide by Quality Controlled Biochemicals Inc. (Hopkinton, MA). Western blotting with the affinity-purified antiserum demonstrated that the antibody specifically recognized phosphorylated LPL from Jurkat cells as well as PMN but not the unphosphorylated protein and that the selectivity to phosphorylated LPL was similar to using autoradiography of LPL from 32 P-loaded cells (Fig. 1). This confirmed that in leukocytes Ser 5 is a major LPL phosphorylation site, as previously demonstrated by transfection of HeLa cells (15).
Adhesion Assay-Tissue culture plates were coated with BSA, BSAanti-BSA IC, or fetal calf serum, and PMN were loaded with calcein as described previously (12). PMN were treated with control buffer or inhibitors for 30 min at 37°C, following which 1 ϫ 10 5 cells/well were added to BSA-or IC-coated wells and incubated at 37°C for the indicated times. For PMA-or fMLP-stimulated adhesion, PMA (50 ng/ml final concentration), fMLP (0.5 M final concentration), or Me 2 SO control was added after the addition of cells to BSA-coated wells. After the incubation period, the fluorescence (485-nm excitation and 530-nm emission wavelengths) was measured using a fMax fluorescence plate reader (Molecular Devices, Sunnyvale, CA) before and after washing four times with 180 l of phosphate-buffered saline. The percentage of adhesion was calculated by dividing the fluorescence after washing by the florescence before washing. In preliminary experiments, fluorescence was shown to be linearly related to cell number.
In Vitro Phosphorylation Assay-To assess peptide phosphorylation, purified kinases were mixed with 100 M LPL, S5A, S7A, or SCR peptides in a reaction mixture containing 20 mM Tris, pH 7.4, 20 mM MgCl 2 , 10 mM dithiothreitol, 100 M ATP, and [␥-32 P]ATP, incubated at 30°C for 20 min. The reaction mixtures were spotted onto P81 phosphocellulose paper and washed four times in 75 mM phosphoric acid and once in 95% ethanol, following which retained radioactivity was determined. For in vitro LPL phosphorylation, 2.5 g of purified recombinant LPL protein was mixed with purified kinases in the reaction mixture described above. After a 20-min incubation at 30°C, the reaction was stopped by the addition of 100°C SDS-polyacrylamide gel electrophoresis sample buffer, and protein components were resolved on 10% SDSpolyacrylamide gel electrophoresis, transferred to polyvinylidene difluoride membrane, and stained with Coomassie Blue to assess the protein loading. The membrane was then subjected to autoradiography to assess protein phosphorylation. Positive controls for the activity of each kinase (substrate peptides for PKA, PKG, PKC, and casein kinase II; myelin basic protein for Pak-1 and PKB; and autophosphorylation for M2K, M3K, and p38-regulated/activated protein kinase) were included in each assay to ensure that the kinase was active against known substrates. In Vivo LPL Phosphorylation-Purified PMN or Jurkat cells were loaded with 2 mCi/ml [ 32 P]phosphoric acid for 2 h at 37°C, washed once with HBSS, and resuspended in HBSS ϩϩ with 0.1% human serum albumin. Cells were treated with various inhibitors or buffer control and added to wells coated with BSA or IC. Alternatively, PMA, fMLP, 8-Br-cAMP, 8-Br-cGMP, or forskolin was added to cells in suspension. The cells were incubated at 37°C for 30 min and lysed by adding 2 ϫ lysis buffer (50 mM Tris 8.0, 1% deoxycholate, 1% Triton X-100, 150 mM NaCl, 100 mM NaF, 10 mM disodium pyrophosphate, with the protease inhibitors 5 mM diisofluorophosphate, 1 mM phenylmethylsulfonyl fluoride, 5 g/ml leupeptin, 5 g/ml aprotinin) for 20 min on ice. The lysates were centrifuged in a microcentrifuge at 14,000 rpm for 15 min at 4°C. LPL was immunoprecipitated from the lysates with 2.5 g of anti-LPL mAb LPL 4A.1 and 50 l of a 1:1 slurry of goat anti-mouse Sepharose for 2 h at 4°C. The immunoprecipitates were washed twice with 1ϫ lysis buffer and twice with phosphate-buffered saline and 0.05% Tween 20 and then boiled in SDS sample buffer for 2 min. The proteins were separated by SDS-polyacrylamide gel electrophoresis, transferred to polyvinylidene difluoride membrane, and stained with Coomassie Blue to detect the total immunoprecipitated LPL protein. LPL phosphorylation was detected by autoradiography. Alternatively, phosphorylated LPL in nonradioactive cells was detected by Western blotting immunoprecipitates with the anti-phospho-LPL antibody. In some experiments, membranes were blotted with LPL 4A.1 (26) to detect total phosphorylated and unphosphorylated LPL protein.
Determination of Intracellular cAMP-1 ϫ 10 6 PMN (in the presence of 100 M isobutylmethylxanthine) were treated with control buffer or inhibitors and then added to IC-or BSA-coated 12-well TC plates and incubated for the indicated times in a 37°C water bath. After the incubation, the cells were immediately centrifuged at 1200 rpm for 3 min at 4°C. Cell pellets were extracted with 65% (v/v) ice-cold ethanol. Cell extracts were collected and cleared by centrifugation at 3000 rpm for 15 min at 4°C. The supernatants were collected and dried in a vacuum oven. The dried extract material was dissolved in assay buffer, and the content of cAMP was determined as described using the acetylation procedure with the cAMP EIA kit (Linco Research Inc., St. Charles, MO) according to the manufacturer's directions.

cAMP-dependent Protein Kinase Specifically
Phosphorylates L-plastin on Ser 5 in Vitro-We have determined that Ser 5 is the predominant LPL phosphorylation site (15). Since this serine is within a consensus sequence for phosphorylation by the cAMP dependent protein kinase PKA (27), we tested whether PKA could phosphorylate LPL. We found that the purified catalytic domain of PKA phosphorylated an amino-terminal LPL peptide (containing LPL amino acids 2-19) but not the peptide in which Ser 5 was mutated to Ala or a peptide in which LPL sequence 2-19 was scrambled ( Fig. 2A). Mutation of Ser 7 to Ala as a control did not block PKA phosphorylation ( Fig. 2A). PKA also potently phosphorylated whole recombinant LPL protein in vitro (Fig. 2B). In contrast, PKG did not phosphorylate recombinant LPL (Fig. 2B) or LPL peptide (data not shown). While Ser 5 also is within consensus sites for phosphorylation by PKC and casein kinase II, neither the catalytic domain of PKC nor casein kinase II phosphorylated LPL peptides or recombinant LPL protein in vitro (data not shown). In addition, we have tested Pak1, protein kinase B, and some downstream kinases in the mitogen-activated protein kinase pathway, such as M2K (MAPKAPK2), M3K (MAPKAPK3), and p38-regulated/ activated protein kinase (22)(23)(24); all failed to phosphorylate LPL or its amino-terminal peptide in vitro (Ref. 15 and data not shown). Thus, in contrast to other candidate kinases, PKA can specifically phosphorylate LPL on Ser 5 in vitro.
PKA Activators Cause LPL Phosphorylation in Vivo-To examine whether PKA is involved in LPL phosphorylation in vivo, we tested the effects on LPL phosphorylation of the PKA activators 8-Br-cAMP, a cell-permeable cAMP analog (Fig. 3A), and forskolin, an adenylate cyclase activator (Fig. 3C). Both induced LPL phosphorylation in a dose-dependent fashion in Jurkat cells. In contrast, the cell-permeant cGMP analog 8-Br-cGMP did not induce LPL phosphorylation even at 5 mM (data not shown). The increased LPL phosphorylation induced by 8-Br-cAMP or forskolin was readily inhibited by the PKAspecific inhibitor H89 (Fig. 3, B and C). 8-Br-cAMP also induced LPL phosphorylation in PMN, but only at a concentration of 5 mM (data not shown). Thus, PKA can directly phosphorylate LPL in vitro, and pharmacologic PKA activation can induce LPL phosphorylation in intact cells.
PKA Inhibitors Prevent IC Induction of LPL Phosphorylation-cAMP has most frequently been found to inhibit PMN activation (28 -30), while LPL phosphorylation has been associated with PMN integrin activation (15). To investigate the potential significance of PKA in agonist-activated LPL phosphorylation, the effects of two PKA inhibitors, H89 (31) and KT5720 (32), were tested on LPL phosphorylation induced by IC, fMLP, or PMA. Both H89 and KT5720 potently inhibited LPL phosphorylation initiated by IC ligation of Fc␥ receptors in PMN (Fig. 4, A and B). Neither the casein kinase II inhibitor 5,6-dichloro-1-␤-D-ribofuranosylbenzimidazole (33) nor the CaM kinase II inhibitor KN-62 (34) inhibited IC-induced LPL phosphorylation, demonstrating specificity of the effects of these two structurally unrelated PKA inhibitors. Moreover, neither PMA nor fMLP activation of LPL phosphorylation was inhibited by H89 or KT5720 (Fig. 4C and data not shown). These results suggest that PKA is responsible for IC-induced LPL phosphorylation in PMN but is probably not the kinase that phosphorylates LPL Ser 5 in response to fMLP or PMA.
PKA Is Required for Sustained PMN Adhesion to an ICcoated Surface, but Not for PMA-or fMLP-activated Adhesion-

FIG. 2. cAMP-dependent protein kinase specifically phosphorylates L-plastin on Ser 5 in vitro.
A, PKA catalytic domain phosphorylation of LPL peptide 2-19 (LPL), the same peptide in which Ser 5 was replaced by Ala (S5A), the LPL peptide in which Ser 7 was changed to Ala (S7A), and a scrambled LPL peptide (SCR) was examined as described under "Experimental Procedures." As negative controls, reaction mixtures were prepared without peptide substrate (no pep) or without kinase (no kinase). B, purified recombinant LPL was incubated with [␥-32 P]ATP and PKA or PKG as described under "Experimental Procedures." LPL phosphorylation was analyzed by autoradiography ( 32 P) as described in Fig. 1. Total loading of LPL was assessed by Coomassie Blue staining. The identity of the 70-kDa band in the PKG lanes is unknown, but it was in the recombinant PKG preparation.
Previous data suggest that LPL phosphorylation precedes ␣ M ␤ 2 activation, since IC, PMA, and fMLP all induce phosphorylation normally in the absence of expression of ␣ M ␤ 2 (25). Cell-permeant peptides containing the Ser 5 phosphorylation site or constitutively phosphorylated Ser 5 activate ␣ M ␤ 2 , suggesting a role for LPL phosphorylation in integrin activation. To assess the relevance of LPL phosphorylation by PKA for integrin activation, we investigated the effects of PKA inhibitors on activation of ␣ M ␤ 2 -mediated adhesion in PMN. While initial adhesion of PMN to an IC-coated surface requires only Fc␥ receptors, sustained adhesion requires activation of ␣ M ␤ 2 (12,16). H89 inhibited IC-induced adhesion significantly at a late time point (30 min) but had no effect on early PMN attachment (Fig. 5A), which is identical to the defect of ␤ 2 -deficient PMN (16). The kinetics of attachment to IC by H89-treated PMN parallel those of wortmannin-treated PMN (Fig. 5A), which fail to activate ␣ M ␤ 2 in response to FcgR ligation (12). Thus, like wortmannin (12), H89 specifically blocked ␣ M ␤ 2 activation by IC. Both H89 and wortmannin blocked IC-induced LPL phosphorylation (Fig. 5A, inset, Fig. 4, and Ref. 15). Furthermore, the dose response of H89 inhibition of LPL phosphorylation correlated closely to the inhibition of sustained IC-induced adhesion (Fig. 5B), consistent with the hypothesis that PKA phosphorylation of LPL is a required step in ICinduced activation of ␣ M ␤ 2 . In contrast, neither H89 nor KT5720 inhibited fMLP or PMA activation of PMN ␣ M ␤ 2 -mediated adhesion (Fig. 5, C and D). In fact, H89 or KT5720 moderately increased both control and fMLP-induced adhesion to BSA or fetal calf serum-coated surfaces (Fig. 5, C and D, and data not shown), in agreement with previous studies demonstrating that cAMP inhibits fMLP-induced PMN activation  (35)(36)(37)(38)(39). Thus, H89 and KT5720 inhibition of sustained PMN adhesion to IC was not due to nonspecific toxic effects of these pharmacologic agents but rather to inhibition of Fc␥R initiated signal transduction. These results indicate that PKA plays a positive role in IC-initiated PMN adhesion, perhaps via phosphorylation of LPL.
PKA Is Required for Fc␥ Receptor-induced ␣ M ␤ 2 Integrin Activation-The time course of H89 inhibition of IC-induced adhesion (Fig. 5A) showed that H89 did not inhibit early adhesion but completely inhibited adhesion after 30 min, indicating that PKA probably has no role in Fc␥R-mediated initial adhesion but is involved in activation of ␣ M ␤ 2 by Fc␥R ligation. To further test this hypothesis, we examined the effects of PKA inhibitor H89 on IC-induced adhesion in PMN from mice with ␤ 2 integrin deficiency due to gene disruption by a neomycin cassette (20). ␤ 2 integrin-deficient PMN can only adhere to IC for a short period of time and then fall off because ␣ M ␤ 2 is required for sustained adhesion to IC (16,17). In murine PMN (mPMN), even early Fc␥R-mediated adhesion is dependent on ␣ M ␤ 2 . Therefore, ␤ 2 integrin-deficient mPMN adhere to IC less well than wild type even in early time points. Nonetheless, ␤ 2 integrin-deficient mPMN adhered specifically to IC. Adhesion was transient and disappeared after 30 -40 min (Fig. 6A). H89 had no effect on adhesion of these mPMN to IC but inhibited sustained adhesion in wild type mPMN, as it had in human PMN (Fig. 6, A and B). Adhesion of H89-treated wild type mPMN was reduced to the level of ␤ 2 integrin-deficient mPMN (Fig. 6B). These results prove that PKA is not required for initial Fc␥R-mediated adhesion but is necessary for Fc␥R-induced ␣ M ␤ 2 integrin activation.

Ligation of Fc␥ Receptors Transiently Increases Intracellular cAMP Level in PMN-To determine whether IC-induced activation of PKA resulted from an increase in intracytoplasmic cAMP concentration ([cAMP]), cytoplasmic [cAMP] was measured at various times after IC activation.
[cAMP] was transiently increased when PMN adhered to IC, with a peak at 5 min, and then returned to its initial level after 15 min (Fig. 7A). The increase in cAMP was inhibited by anti-Fc␥RII antibody IV.3, confirming its initiation by IC ligation of Fc␥Rs. As expected, PKA inhibitor H89 or anti-␤ 2 antibody IB4 treatment did not inhibit this increased cAMP level (Fig. 7B), demonstrating that cAMP elevation requires Fc␥RIIA ligation but precedes PKA and integrin activation.
Since PI 3-kinase inhibitors blocked sustained adhesion to IC with similar kinetics as PKA inhibitors and also blocked LPL phosphorylation (Fig. 5A and Refs. 12 and 15), it was important to determine whether PI 3-kinase is necessary for PKA activation by IC. Both wortmannin and LY294002, two PI 3-kinase inhibitors with different modes of action, blocked IC-induced increase in cAMP (Fig. 7A), suggesting that PI 3-kinase functions upstream of PKA in the pathway for LPL phosphorylation and integrin activation, perhaps in regulation of adenylate cyclase. [cAMP] in each sample was determined after 5 min. The data are representative of three independent experiments.

DISCUSSION
Previous work using cell-permeant peptides from the amino terminus of LPL showed that these peptides were capable of inducing rapid integrin activation in PMN (15). Integrin activation by these peptides required Ser 5 and was prevented by PI 3-kinase inhibitors. Surprisingly, integrin activation by a cellpermeant peptide in which Ser 5 was phosphorylated was independent of PI 3-kinase, leading to the hypothesis that the well described role for PI 3-kinase in integrin activation (10) results from its involvement in a pathway resulting in phosphorylation of Ser 5 of LPL. These studies focused attention on the phosphorylation of LPL Ser 5 as a potentially critical regulatory step in activation of PMN integrins. In the present work, we present data indicating that PKA can phosphorylate LPL and does so when PMN Fc␥RIIA is ligated. First, PKA phosphorylates both LPL and an LPL amino-terminal peptide in vitro. PKA phosphorylation occurs on Ser 5 , as determined by failure of phosphorylation of the S5A mutant peptide and by recognition of PKA-phosphorylated LPL by a phospho-Ser 5 -specific antibody. Second, agents that increase intracytoplasmic [cAMP], which activates PKA, induce LPL phosphorylation. Finally, two pharmacologic PKA inhibitors block IC-induced LPL phosphorylation in PMN. While the data in whole cells demonstrate only that PKA activation can lead to LPL phosphorylation, together with the demonstration of direct phosphorylation in vitro, these data support the hypothesis that PKA can directly phosphorylate LPL in cells. It is likely that other kinase(s) can phosphorylate LPL as well, since fMLP and PMA-induced phosphorylation are not antagonized by the same PKA inhibitors that block IC-induced phosphorylation. While these kinases are not known, PKC is not a likely candidate, since neither recombinant LPL nor the LPL amino-terminal peptide was phosphorylated by the PKC catalytic domain in vitro. The existence of distinct kinases that can phosphorylate LPL is consistent with previous data suggesting two distinct signaling pathways for both LPL phosphorylation and integrin activation (12,15).
The hypothesis that LPL phosphorylation is a requisite step in PMN integrin activation suggests that interference with its phosphorylation should block activation-dependent integrinmediated adhesion. Since PKA inhibitors blocked LPL phosphorylation by IC, we tested the effects of these inhibitors on ␣ M ␤ 2 integrin activation by examining sustained adhesion to IC. These inhibitors blocked IC-induced, ␣ M ␤ 2 -dependent sustained adhesion of both human and murine PMN. The specificity of the effect was demonstrated by (i) failure of inhibitors of other serine kinases to affect ␣ M ␤ 2 activation; (ii) failure of the PKA inhibitors to affect the transient Fc␥R-mediated adhesion to IC that occurs in ␣ M ␤ 2 -deficient PMN; and (iii) failure of the PKA inhibitors to block ␣ M ␤ 2 -mediated adhesion induced by fMLP and PMA, which cause PKA-independent LPL phosphorylation. Thus, these data are further evidence for a close association between LPL phosphorylation and PMN integrin activation and support the hypothesis that LPL phosphorylation (which occurs normally in ␣ M ␤ 2 -deficient cells (25)) is a required step in activation of ␣ M ␤ 2 -mediated adhesion.
A role for cAMP in phagocyte activation is quite unexpected. In general, increases in [cAMP] have been found to inhibit fMLP-stimulated PMN activation, as measured by adhesion or respiratory burst activation (30,36,40,41) and have been found to inhibit integrin activation in several cell types (29,42,43). However, Fc␥R-mediated accumulation of cAMP and of PKA at phagosomes has been noted previously (44,45), although this cAMP has been variously interpreted as part of the activation process (44) and a component of the regulation of phagocytosis (45). A resolution of the apparent conflict between the well described negative regulatory role for cAMP in fMLP stimulation and its activating role in Fc␥R-mediated signaling may lie in the compartmentalization of cAMP in response to Fc␥R ligation. Localized increases in cAMP may be prevented from propagating through the cytoplasm by the colocalization of phosphodiesterase (45), and it is of note that in the present study increases in cytosolic [cAMP] on adhesion to IC could only be detected in the presence of a phosphodiesterase inhibitor. Pharmacologic increases in cAMP or increases that occur because of ligation of receptors (28) at a distance from the adhesion site may activate quite different processes than cAMP confined to the region of Fc␥R ligation, because of access to different downstream signaling pathways. In this regard, it is potentially interesting that LPL is phosphorylated by 8-bromo-cAMP addition in PMN only at very high concentrations, consistent with the possibility of divergent effects of this mediator on the different pathways leading to LPL phosphorylation. Nonetheless, LPL phosphorylation by high concentration 8-Br-cAMP in the absence of ␣ M ␤ 2 activation suggests that LPL phosphorylation is not sufficient for integrin activation.
Finally, our data suggest that PI 3-kinase is involved in an Fc␥R-initiated pathway to PKA activation, since wortmannin and LY294002 both block PKA-mediated LPL phosphorylation in PMN and Fc␥R-induced increase in [cAMP]. Activation of PI 3-kinase by PMN Fc␥R ligation is well described (46 -48), and inhibition of PI 3-kinase has been shown to block Fc␥R-mediated phagocytosis (47), LPL phosphorylation, and integrin activation (12,15). Our data demonstrate that PI 3-kinase inhibitors block Fc␥R-induced cAMP accumulation, which is consistent with the hypothesis that the role for PI 3-kinase is upstream of PKA activation in IC-induced integrin activation. While this could theoretically be either through regulation of adenylate cyclase or phosphodiesterase activity, the PI 3-kinase inhibitors blocked cAMP accumulation when phosphodiesterase activity was inhibited by isobutylmethylxanthine, strongly implicating PI 3-kinase in activation of adenylate cyclase. While the signaling pathway responsible for this regulation is unknown, there is evidence that the state of assembly of the actin cytoskeleton, which can be regulated by PI 3-kinase, affects adenylate cyclase activity in S49 cells (49) and in yeast (50,51). It is intriguing as well that an adenylate cyclase regulatory protein in Dictyostelium contains a pleckstrin homology domain (52), since these domains often bind inositol 1,4,5-trisphosphate, the product of PI 3-kinase activity. Whatever the precise mechanism, these studies have identified adenylate cyclase as a physiologically significant downstream target of PI 3-kinase after its activation by Fc␥R ligation.
In summary, this work has two significant implications for understanding the biochemical pathways that lead to PMN integrin activation. First, these studies identify a PKA-dependent activation pathway initiated by IC. Although PKA has generally been thought to negatively regulate PMN activation, most of these studies have been done primarily with fMLPinduced activation, and the current data are completely compatible with the previously observed divergence between ICand fMLP-induced PMN activation (12). Furthermore, IC-initiated PKA activation is dependent on PI 3-kinase via its regulation of adenylate cyclase, a novel pathway for regulation of cytoplasmic [cAMP].
Second, these data support the hypothesis that LPL phosphorylation is a requisite step in PMN integrin activation. PKA inhibitors block IC-induced integrin activation but not fMLPor PMA-induced activation, which is in perfect accord with their effects on LPL phosphorylation. Since PKA was identified as a candidate for involvement in integrin activation because of its ability to phosphorylate LPL in vitro, these data strongly support the involvement of LPL phosphorylation in integrin activation. While additional methods of approach will be required to define the mechanism by which LPL phosphorylation affects integrin function, it is highly likely that this is a significant regulatory step in the regulation of leukocyte integrin avidity.