Down-regulation of Rac Activity during β2 Integrin-mediated Adhesion of Human Neutrophils*

In human neutrophils, β2 integrin engagement mediated a decrease in GTP-bound Rac1 and Rac2. Pretreatment of neutrophils with LY294002 or PP1 (inhibiting phosphatidylinositol 3-kinase (PI 3-kinase) and Src kinases, respectively) partly reversed the β2 integrin-induced down-regulation of Rac activities. In contrast, β2 integrins induced stimulation of Cdc42 that was independent of Src family members. The PI 3-kinase dependence of the β2 integrin-mediated decrease in GTP-bound Rac could be explained by an enhanced Rac-GAP activity, since this activity was blocked by LY204002, whereas PP1 only had a minor effect. The fact that only Rac1 but not Rac2 (the dominating Rac) redistributed to the detergent-insoluble fraction and that it was independent of GTP loading excludes the possibility that down-regulation of Rac activities was due to depletion of GTP-bound Rac from the detergent-soluble fraction. The β2 integrin-triggered relocalization of Rac1 to the cytoskeleton was enabled by a PI 3-kinase-induced dissociation of Rac1 from LyGDI. The dissociations of Rac1 and Rac2 from LyGDI also explained the PI 3-kinase-dependent translocations of Rac GTPases to the plasma membrane. However, these accumulations of Rac in the membrane, as well as that of p47phox and p67phox, were also regulated by Src tyrosine kinases. Inasmuch as Rac GTPases are part of the NADPH oxidase and the respiratory burst is elicited in neutrophils adherent by β2 integrins, our results indicate that activation of the NADPH oxidase does not depend on the levels of Rac-GTP but instead requires a β2 integrin-induced targeting of the Rac GTPases as well as p47phox and p67phox to the plasma membrane.

In response to inflammatory signals such as tumor necrosis factor-␣ (TNF), 1 polymorphonuclear neutrophils (PMNs) ad-here to the surface of endothelium and then crawl forward (diapedesis) and pass between neighboring endothelial cells (transmigration) to reach infected tissues. This is followed by ingestion of the invading microbes, resulting in their dissolution, largely through the release of granule contents into the phagolysosome and generation of oxygen radicals by the membrane-bound NADPH oxidase (1).
GTP-binding proteins of the Rho subfamily, which belongs to the Ras superfamily of small GTPases, are necessary for regulation of PMN functions. For example, the motile potential of PMNs is largely due to the formation of membrane protrusions, a process that requires relocalization and activation of Rac and Cdc42 (as well as other signaling molecules such as PI 3-kinase) at the leading edge of these motile cells (10). The requirement for Rho GTPases in these functional events can readily be ascribed to their dynamic regulation of the actin-based cytoskeleton, similar to that described in many other cell models (11,12). In PMNs, Rac and Cdc42 are involved in actin nucleation (13,14); Rac promotes dissociation of gelsolin from actin filaments (13), and Cdc42 activates WASP proteins and the ARP2/3 complex (15). In addition to participating in regulation of cytoskeletal dynamics, Rac1 and Rac2 are part of the multicomponent, plasma membrane-bound enzyme NADPH oxidase; however, the exact role that Rac GTPases play in the regulation of the respiratory burst in phagocytic cells has not yet been completely elucidated (16,17).
These small GTP-binding proteins cycle between a GDPbound inactive form to a GTP-bound active form (18). In their inactive form, Rho GTPases are bound to proteins called GDIs * This work was supported by grants from the Swedish Cancer Foundation, the King Gustaf V Memorial Foundation, the Network for Inflammation Research (funded by the Swedish Foundation for Strategic Research) (to T. A.), the U-MAS Research Foundations, the Crafoord Foundation, the Ö sterlund Foundation (to T. A. and K. D), the Royal Physiographic Society in Lund, and the Koch Foundation (to K. D.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
(RhoGDI and LyGDI), which compete more efficiently in vivo for GDP-bound than for GTP-bound Rho GTPases (19). GDIbound Rho GTPases are found in the cytosol, because their C-terminal, lipid-modified end is inserted into a hydrophobic pocket of the immunoglobulin-like domain of the GDI molecule, which prevents the Rho GTPases from interacting with the membrane (20). Guanine nucleotide exchange factors, activated by extracellular stimuli, are responsible for the GDP-GTP switch. In their GTP-bound state, these proteins interact with specific effectors to initiate downstream signals and functions. The subsequent hydrolysis of bound GTP to GDP is catalyzed by the family of GTPase-activating proteins (GAPs).
We have recently shown that ligation of the ␤ 2 integrins on PMNs resulted in activation of Ras (21) and RhoA (22), but it is not yet known whether other Rho GTPases are also regulated in the course of this event. In the present study, we examined ␤ 2 integrin-dependent regulation of Rac and Cdc42 GTPases in adherent PMNs.
Chemicals-Protein G-Sepharose was from Oncogene TM (Germany); protein A-Sepharose, Dextran, and Ficoll-Hypaque were purchased from Amersham Biosciences; and the protease inhibitors pefabloc, pepstatin, leupeptin, aprotinin, and antipain were from Roche Applied Science. Benzamidine, LY294002, and wortmanin were obtained from Sigma, the tyrosine kinase inhibitor PP1 was from Alexis Biochemicals, and all electrophoresis reagents were obtained from Bio-Rad. All other chemicals were of analytical grade and were purchased from Sigma.
Isolation of Human PMNs-Blood was collected from healthy donors, and PMNs were isolated under endotoxin-free conditions as previously described (24). In short, the blood was subjected to dextran sedimentation followed by a brief hypotonic lysis of erythrocytes. The lysis was stopped by adding 3 ml of buffer A (565 mM NaCl, 2.7 mM KCl, 6 Ligation of ␤ 2 Integrins-For adhesion of PMNs, Petri dishes (Easy Grip TM ) containing 20 g/ml fibrinogen in phosphate-buffered saline (PBS) were incubated either overnight at 4°C or for 2 h at room temperature and then washed twice with PBS and once with calciumcontaining medium. PMNs (20 ϫ 10 6 ) were subsequently incubated in the fibrinogen-coated dishes at 37°C in the presence of TNF (20 ng/ml) for different periods of time. For treatment of suspended PMNs, polypropylene tubes (15 ml) were blocked for 2 h with 10% fetal calf serum and then rinsed extensively with PBS and once with calciumcontaining medium, after which the cells (10 6 /ml) were incubated in the tubes under gentle rotation at 37°C in the absence (control cells) or presence of TNF (20 ng/ml) for the indicated periods of time. The reactions were terminated by putting the Petri dishes or tubes on ice.
GST Pull-down Assays and Western Blotting-The cDNA of the Rac and Cdc42 binding domain from PAK1B (PAKcrib; amino acids 56 -267) was cloned into the bacterial expression vector pGEX-2T and was expressed in Escherichia coli as a fusion protein with glutathione Stransferase (25). PMNs were lysed in a buffer composed of 50 mM Tris-HCl, pH 7.5, 1% Triton X-100, 100 mM NaCl, 10 mM MgCl 2 , 5% glycerol, 1 mM Na 3 VO 4 , and protease inhibitors (20 g/ml aprotinin; pepstatin, leupeptin, and antipain (1 g/ml each); 2.5 mM benzamidine; 2 mM pefabloc). Lysates were centrifuged at 15,000 ϫ g for 10 min, and the Triton X-100-soluble fraction was recovered. A bacterial lysate containing the GST-PAKcrib fusion protein was added to PMNs lysates together with glutathione-Sepharose beads. After 1 h, the beads were collected by centrifugation and washed three times in 25 mM Tris-HCl, pH 7.5, 1% Triton X-100, 1 mM dithiothreitol (DTT), 100 mM NaCl, and 30 mM MgCl 2 . The beads were then resuspended in Laemmli sample buffer and boiled under reducing conditions for 5 min. The precipitated proteins were subjected to 12% SDS-PAGE and transferred to polyscreen PVDF membranes. The membranes were blocked in PBS supplemented with 0.2% Tween 20 and 3% milk and then incubated for 1 h with a primary antibody (1 g/ml dilution of anti-Rac2, anti-Cdc42, or anti-Rac Abs or 0.15 g/ml anti-Rac1 mAb) and thereafter washed three times for 5 min in PBS supplemented with 0.2% Tween 20. The membranes were subsequently incubated for 1 h with peroxidase-conjugated anti-mouse IgGs (1:10,000) in PBS supplemented with 0.2% Tween 20 and 3% milk. The blots were extensively washed, and antibody binding was visualized by enhanced chemiluminescence (ECL).
Measurement of NADPH Oxidase Activity-Nitro blue tetrazolium (NBT) is an electron acceptor used to indirectly detect the production of superoxide by PMNs (26). Upon electron acceptance, soluble and yellow NBT is converted to blue-black formazan that can be quantitated spectrophotometrically after extraction from cells with N,N-dimethylformamide. Briefly, PMNs (2.5 ϫ 10 6 /ml) in calcium-containing medium were preincubated with 0.2% NBT for 10 min, after which PMNs were plated on fibrinogen-coated plates in the presence of TNF for the indicated time periods. Suspended PMNs were taken as control cells (zero time point). Thereafter, adherent cells (5 ϫ 10 6 ) were scraped off, and the total cell suspension was transferred to Eppendorf tubes. The tubes were spun for 3 min, and the resulting pellets were dissolved in 1 ml of N,N-dimethylformamide and left at 56°C for 1 h. The tubes were then again spun for 3 min, after which the optical densities (515 nm) of cell extracts were determined in a spectrophotometer. The OD values obtained from resting cells ranged between 0.07 and 0.1.
Measurement of Rac-GAP Activity in PMN Lysates-PMNs (30 ϫ 10 6 ) were lysed in 100 l of ice-cold lysis buffer (PBS containing 1% Triton X-100, 1 mM EGTA, 5% glycerol, 10 mM benzamidine, 10 g/ml leupeptin, 10 g/ml aprotinin, 1 mM pefabloc, 1 g/ml pepstatin and antipain), and the lysates were clarified by centrifugation (15,000 ϫ g, 10 min). The Rac-GAP assay was performed essentially as described elsewhere (27), with minor modifications. Briefly, 10 l of precleared lysates (containing about 50 g of proteins) were incubated in 30 l of Rac-GAP buffer (16 mM Tris-HCl, pH 7.5, 0.1 mM DTT, 1 mg/ml bovine serum albumin, 1 mM GTP) for 5 min at room temperature. Thereafter, the reaction was initiated by adding 4 l of [␥-32 P]GTP-loaded Rac1, and the sample was incubated for 5 min at room temperature under shaking. The reaction was stopped by adding 0.2 ml of ice-cold Rac-GAP buffer and placing the samples on ice for 2 min. Aliquots (50 l) were filtered through nitrocellulose filters (0.45-mm pore size) under vacuum, and the filters were washed three times with 1 ml of wash buffer (50 mM Tris-HCl, pH 7.7, 5 mM MgCl 2 ). The filters were then air-dried and placed in plastic vials. 5 ml of scintillation mixture (Ready Gel; Beckman) was added, and radioactivity bound to the filters was measured using a scintillation counter. To achieve GTP loading of Rac1, 1-3 g of affinity-purified histidine-tagged Rac1 (expressed in Sf9 cells) was incubated at room temperature for 5 min in loading buffer (16 mM Tris-HCl, pH 7.5, 20 mM NaCl, 0.1 mM DTT, 5 mM EDTA, 100 nM GTP, and 5 Ci of [␥-32 P]GTP with a specific activity of 5000 Ci/mmol). Thereafter, MgCl 2 (20 mM) was added to block further nucleotide exchange activity, and the tubes were placed on ice. An aliquot was filtered through a nitrocellulose filter, and the filter was washed three times with wash buffer (see above). The radioactivity remaining on the filters was counted and considered as total bound Rac1 (100%).
Pretreatment of PMNs with Anti-␤ 2 Integrin Antibodies-PMNs (10 ϫ 10 6 /ml) were resuspended in calcium-containing medium in polypropylene tubes and then incubated for 30 min at 37°C with 15 g/ml IB4 antibody or isotype-matched IgG 2a monoclonal antibody. Thereafter, the cells were placed on dishes coated with fibrinogen and stimulated with TNF (20 ng/ml). The reactions were terminated by placing the dishes on ice, after which GST-PAKcrib pull-down assays were performed (see above).
Immunoprecipitation-PMNs were lysed by adding the following lysis buffer: 100 mM Tris-HCl, pH 7.5, 1% Triton X-100, 5 mM EDTA, 5 mM EGTA, 50 mM NaCl, 5 mM NaF, 1 mM Na 3 VO 4 , and protease inhibitors (20 g/ml aprotinin; 1 g/ml each pepstatin, leupeptin, and antipain; 2.5 mM benzamidine; 2 mM pefabloc). Cell lysates were clarified by centrifugation (10 min at 15,000 ϫ g), and LyGDI in the supernatants was immunoprecipitated by exposure to the anti-LyGDI antiserum (3 g/ml) for 2 h and then to 40 l of a 50% slurry of protein G-Sepharose for 45 min. The beads were subsequently collected by centrifugation and washed three times in a wash buffer (50 mM Hepes, pH 7.4, 1% Triton X-100, 0.1%, SDS, 150 mM NaCl, 1 mM Na 3 VO 4 ). The beads were then resuspended in Laemmli sample buffer and boiled under reducing conditions for 5 min. The immunoprecipitated proteins were subjected to electrophoresis on 12% SDS-PAGE and transferred to polyscreen PVDF transfer membranes. To detect Rac in the anti-LyGDI immunoprecipitates, the membranes were incubated with either anti-Rac1 mAb or anti-Rac2 antiserum and subsequently with peroxidaseconjugated anti-mouse or anti-rabbit IgGs (1:10,000), as described under "GST Pull-down Assays and Western Blotting." Determination of the Translocation of Cytosolic Components of the NADPH Oxidase-PMNs were scraped off from the Petri dishes, suspended in disruption buffer (100 mM Tris-HCl, pH 7.5, 5 mM EDTA, 5 mM EGTA, 50 mM NaCl, 5 mM NaF, 1 mM Na 3 VO 4 , and protease inhibitors (20 g/ml aprotinin; 1 g/ml each pepstatin, leupeptin, and antipain; 2.5 mM benzamidine; 2 mM pefabloc)), and then placed in a cell disruption bomb at 4°C (28). The bomb was equilibrated at 1000 p.s.i. for 10 min, after which the pressure was quickly released. The cell suspension was subsequently centrifuged at 10,000 ϫ g for 10 min at 4°C to pellet nuclei, heavy membrane fractions, and undisrupted cells. The supernatant was further centrifuged at 100,000 ϫ g for 1 h, and the resulting pellet was resuspended in disruption buffer. The protein content was determined (29), and aliquots were mixed with 2ϫ Laemmli buffer supplemented with 50 mM DTT and boiled. The proteins (5-10 g) were separated on 12% SDS-PAGE, and immunoblot analysis was performed as described above, using anti-Rac1 mAb, anti-Rac2 antiserum, anti-p47 phox , or anti-p67 phox mAbs.
To measure translocation of Rac proteins to the detergent-insoluble fraction, the Triton X-100-insoluble fraction was resuspended in Laemmli buffer containing 50 mM DTT, and the samples were subjected to sonication. Aliquots were taken to estimate the protein content of each sample. Aliquots of proteins (20 -40 g) were separated by 12% SDS-PAGE and transferred to polyscreen PVDF transfer membranes. The membranes were incubated with either anti-Rac1 mAb or anti-Rac2 antiserum and then with peroxidase-conjugated anti-mouse or anti-rabbit IgGs (1:10,000), as described above.

␤ 2 Integrin-mediated Regulation of Rac and Cdc42
GTPases in PMNs-To engage ␤ 2 integrins on PMNs, the cells were incubated on a surface coated with fibrinogen (a ligand for ␤ 2 integrins) (30) and exposed to TNF. Stimulation of PMNs with TNF or other cytokines such as leukotriene B4 increases the content of the CD11b/CD18 ␤ 2 integrin on the surface of PMNs (31), and it is presumed that these ␤ 2 integrins exhibit augmented avidity for their ligands (32). In the present experiments, we measured the relative activities of Rac and Cdc42 in PMNs upon engagement of ␤ 2 integrins by using the GST-PAKcrib binding assay, which is based on the knowledge that the GST-PAKcrib fusion protein binds the GTP-bound forms of Cdc42 and Rac but does not bind GTP-bound RhoA (33). We found that the amounts of Rac-GTP decreased in a time-dependent manner in PMNs adhering to immobilized fibrinogen in the presence of TNF, and we observed the lowest levels (2.5-fold decrease over controls) after incubation for 20 -40 min on a fibrinogen-coated surface (Fig. 1A). PMNs express both Rac1 and Rac2, and the latter represents 95% of total Rac (34); thus, the results presented in Fig. 1A definitely reflect decreased levels of Rac2-GTP. To corroborate this finding, and to ascertain whether regulation of Rac1 occurs in adherent PMNs, we conducted GST-PAKcrib pull-down assays followed by Western blot analysis using specific anti-Rac1 or anti-Rac2 antibodies. As expected, ligation of the ␤ 2 integrins on PMNs induced by exposure to immobilized fibrinogen caused a timedependent decrease in the level of Rac2-GTP (Fig. 1C) and, with a similar time course, also lowered the amount of Rac1-GTP (Fig. 1B). In parallel experiments, we confirmed that engagement of ␤ 2 integrins on PMNs by plating them on fibrinogen in the presence of TNF led to a time-dependent activation of the respiratory burst as assessed by measuring the production of superoxide-induced formazan (Fig. 1D).
In contrast to Rac, ligation of the ␤ 2 integrins led to a timedependent activation of Cdc42, and we detected the maximum levels of Cdc42-GTP (2.5-fold increase over controls) in PMNs incubated on fibrinogen for 20 -40 min (Fig. 2). However, stimulating suspended PMNs with TNF did not modify the basal activities of Cdc42, Rac1, or Rac2 (Fig. 3A), which confirms that TNF regulates only the activities of Cdc42 and Rac in adherent PMNs. This was further illustrated in our next set of experiments, in which PMNs were preincubated for 30 min at 37°C with anti-CD18 mAb (IB4) or an isotype-matched control antibody and then placed on a surface coated with fibrinogen and exposed to TNF. The relative activities of Cdc42 and Rac GTPases were subsequently measured using the GST-PAKcrib binding assay (see above). We found that preincubation with anti-CD18 antibody almost totally abolished adhesion-dependent regulation of Cdc42 and Rac activities (Fig. 3B), whereas pretreatment with an isotype-matched control antibody had no significant effect. These results agree well with a study by Nathan (2) showing that TNF-mediated activation of the NADPH oxidase depended on adhesion to a surface and activation of the ␤ 2 integrins (2).
To further assess the validity of our GST-PAKcrib pull-down assay, we treated PMNs with fMLP (10 Ϫ7 M, 1 min) and found that this substance increased the activity of Cdc42 (196 Ϯ 22% over controls; n ϭ 4, p Ͻ 0.01), Rac1 (160 Ϯ 21% over controls; n ϭ 4, p Ͻ 0.05) and Rac2 (400 Ϯ 106% over controls; n ϭ 4, p Ͻ 0.05), which agrees with the results reported by other investigators (33,35,36). Thus, these findings show that the fMLP receptor and the ␤ 2 integrins collaborate in regulation of Cdc42 but oppose each other in the regulation of Rac GTPases. In control experiments, we observed that GST-PAKcrib, but not GST, could pull down Cdc42 and Rac GTPases (Fig. 3A), which rules out any nonspecific precipitation of the GTPases by the GST part of the fusion protein.
Identification of the Signaling Pathways Involved in ␤ 2 Integrin-induced Regulation of Rac and Cdc42 GTPases-To explore the difference in adhesion-dependent regulation of Rac and Cdc42, we pretreated PMNs with a variety of pharmacological agents and then incubated them on plates coated with fibrinogen in the presence of TNF. Thereafter, we used the GST-PAKcrib binding assay (described above) to measure the relative amounts of the active forms of Rac1, Rac2, and Cdc42. We pretreated the PMNs with the following agents: wortmanin or LY294002, both of which inhibit PI 3-kinase by distinct mechanisms and are known to block fMLP-induced generation of phosphatidylinositol 3,4,5-trisphosphate and activation of the respiratory burst in PMNs (37), or PP1, which is a potent and selective inhibitor of Src family tyrosine kinases (38). We have previously shown that PP1 (3 M) blocks basal and ␤ 2 integrin-induced overall tyrosine phosphorylation of proteins in PMNs (22). In our present study, pretreatment of PMNs with wortmanin or LY294002 significantly reduced, but did not totally block, adhesion-induced up-regulation of Cdc42 activity, and it partly reversed adhesion-elicited down-regulation of the activities of Rac1 and Rac2 (Fig. 4). Interestingly, pre-exposure to PP1 had no impact on activation of Cdc42 caused by ligation of the ␤ 2 integrins (Fig. 4, left panel), whereas it partly reverted adhesion-induced down-regulation of Rac1 (Fig. 4, middle panel) and Rac2 activities (Fig. 4, right panel).
Ligation of ␤ 2 Integrins Enhances the Activity of Rac-GAP-To further elucidate the mechanism by which ligation of PMNs ␤ 2 integrins leads to down-regulation of Rac activities, we performed experiments to determine whether increased Rac-GAP activity could induce this phenomenon in adherent PMNs. We incubated PMNs lysates for 5 min in the presence of semipurified Rac1 loaded with [␥-32 P]GTP; in this assay, a decrease in ␥-32 P-labeled Rac1 indicates an increase in total Rac-GAP activity. As in a previous investigation (27), the intrinsic GTPase activity of Rac1 was high; i.e. 30% of GTPbound Rac was hydrolyzed during the 5 min that we performed the Rac-GAP assay in the absence of PMNs lysate (Fig. 5). Upon the addition of cell lysate, the hydrolysis of GTP-bound Rac1 increased significantly. Notably, we detected markedly higher Rac-GAP activity in lysates of PMNs with ligated ␤ 2 integrins than in lysates of control cells; this was indicated by the fact that 23 and 37% of total labeled Rac-GTP remained in the lysates of adherent (integrin-ligated) and control PMNs, respectively (i.e. activity in the adherent cells increased 1.6fold, p Ͻ 0.05). Furthermore, pretreatment of the cells with LY294002 totally reversed the adhesion-induced down-regulation of Rac-GAP activity (p Ͻ 0.05), whereas PP1 only modestly reversed the Rac-GAP activity (p Ͻ 0.05) (Fig. 5).
Ligation of ␤ 2 Integrins Induces Translocation of Rac1 but Not Rac2 to the Triton X-100-insoluble Fraction-Because the GST-PAKcrib pull-down assays were carried out on lysates in which the Triton X-100-insoluble fraction had been removed, the adhesion-induced decrease in Rac activities could actually reflect depletion of GTP-bound Rac in the Triton X-100-soluble fraction. To test that hypothesis, we performed Western blot analysis to measure amounts of both Rac1 and Rac2 in the Triton X-100-insoluble fraction. Interestingly, incubating PMNs on immobilized fibrinogen in the presence of TNF led to massive translocation of Rac1 (Fig. 6A) but not Rac2 (Fig. 6B) to the Triton X-100-insoluble fraction. No such translocation of Rac1 was observed in suspended cells stimulated with TNF (Fig. 6A). In addition, we found that pretreating cells with an anti-CD18 antibody, but not with an isotype-matched control antibody, blunted the adhesion-dependent redistribution of Rac1 to the detergent-insoluble cellular components (Fig. 6C). Furthermore, a pretreatment of PMNs with LY294002 or wortmanin, but not PP1, blocked adhesion-induced redistribution of Rac1 to the detergent-insoluble fraction (Fig. 6D).
Ligation of ␤ 2 Integrins Causes Dissociation of the LyGDI-Rac Complex-Structural data have shown that GDIs are negative regulators of Rho GTPases; thus, it is assumed that Rho GTPases must be dissociated from RhoGDIs in order to be activated and/or relocalized (19,20). However, the signaling proteins that regulate such dissociation of GDIs from the small GTPases have not yet been identified. Therefore, in our subsequent experiments, we examined the possibility that association of Rac with LyGDI is regulated during integrin-mediated cell adhesion. To this end, PMNs were incubated on plates coated with fibrinogen in the presence of TNF. Thereafter, the cells were lysed, and LyGDI was immunoprecipitated with a specific antiserum. To assess association between LyGDI and Rac GTPases, the immunoprecipitated fractions were separated by SDS-PAGE, transferred to PVDF membranes, and then blotted with antibodies directed against Rac1 or Rac2. We found that LyGDI co-immunoprecipitated with Rac1 and Rac2 in resting cells (Fig. 7). Incubation of PMNs on fibrinogen in the presence of TNF for 30 min resulted in dissociation of both the LyGDI-Rac1 and the LyGDI-Rac2 complex, and this was significantly reversed by pretreatment of the cells with LY294002 but was not influenced by pretreatment with PP1 (Fig. 7).
Ligation of ␤ 2 Integrins Causes Translocation of Rac1, Rac2, p47 phox , and p67 phox to a Membrane-enriched Fraction-We then wanted to find out whether a lack of Rac activation still would enable these cytosolic components of the NADPH oxidase, as well as p47 phox and p67 phox , to translocate to membrane-enriched fraction in PMNs undergoing integrin-mediated cell adhesion. To this end, we allowed the cells to adhere onto immobilized fibrinogen in the presence of TNF for various periods of time. The cells were then scraped off the plates and disrupted by nitrogen cavitation, and membrane-enriched fractions were prepared. The amounts of Rac1, Rac2, p47 phox , and p67 phox in these fractions were measured by Western blotting with specific antibodies. The results show that ligation of the ␤ 2 integrins caused time-dependent translocations of these cytosolic proteins to the plasma membrane (Fig. 8A). Furthermore, we found that the ␤ 2 integrin-induced accumulations of Rac1, Rac2, p47 phox , and p67 phox in the membrane fraction were significantly impaired in PMNs that had been pretreated with PP1. Similar results were found in PMNs pretreated with LY294002 with the exception of p67 phox that accumulated in the membrane fraction independently of PI 3-kinase activation (Fig. 8B). It is known that p47 phox but not p67 phox contains a Phox homology domain. The Phox homology domain of P47 phox binds phosphatidylinositol 3-phosphate, and such lipid modification of the protein triggers its attachment to the plasma membrane (39). Thus, it is logical that in adherent cells, the translocation of p47 phox but not that of p67 phox to the membrane does require PI 3-kinase activity. DISCUSSION In the present study, we found that ligation of ␤ 2 integrins on human PMNs differentially regulated the activities of Rac and Cdc42, as indicated by the observation that the amount of GTP-bound Cdc42 was increased, whereas the levels of GTPbound Rac1 and Rac2 were decreased.

FIG. 3. TNF-induced regulation of Cdc42 and Rac GTPases in PMNs requires adhesion and activation of the ␤ 2 integrins.
A, PMNs in suspension (10 6 /ml) were stimulated with TNF (20 ng/ml) for the indicated times, after which the cells (20 ϫ 10 6 ) were lysed. The blots show the amounts of active Cdc42, Rac1, and Rac2 determined by the GST-PAKcrib binding assay, as described in the legend to Fig. 1. The results of the GST pull-down assays are illustrated to the right. B, PMNs in suspension (10 ϫ 10 6 /ml) were preincubated with either an isotype-matched control antibody IgG 2a (20 g/ml) or the anti-␤ 2 integrin antibody (anti-CD18, IB4; 20 g/ml) for 30 min at 37°C. Thereafter, the cells were incubated on a fibrinogen-coated surface and stimulated with TNF (20 ng/ml) for 30 min. The PMNs (20 ϫ 10 6 ) were lysed, and the amounts of active GTP-bound Cdc42, Rac1, and Rac2 were determined using the GST pull-down assay as described in the legend to Fig. 1. WB, Western blot.

FIG. 4. The role of PI 3-kinase and Src family tyrosine kinases in ␤ 2 integrin-dependent regulation of Rho GTPases in PMNs.
PMNs in suspension (10 ϫ 10 6 /ml) were pretreated for 20 min at 37°C with the Src family tyrosine kinase inhibitor PP1 (3 M) or with a PI 3-kinase inhibitor, LY294002 (20 M, LY) or wortmanin (10 nM, W). Thereafter, the cells were incubated on a fibrinogen-coated surface and stimulated with TNF for 30 min. The cells (20 ϫ 10 6 ) were subsequently lysed, and amounts of active Cdc42, Rac1, and Rac2 in the lysates were measured using the GST-PAKcrib binding assay. The insets are representative Western blots of each Rho GTPase. The diagrams illustrate densitometric analysis of the relative activities of Cdc42, Rac1, and Rac2 in PMNs that had been pretreated with the indicated inhibitors (PP1, LY, or W) and then incubated on fibrinogen for 30 min. The data are expressed as percentage of unstimulated control cells and represent means Ϯ S.E. of 5-9 (for Cdc42), six (for Rac1), and three or four (for Rac2) separate experiments. Statistical significance versus cells plated for 30 min (unpaired Student's t test) was as follows: *, p Ͻ 0.05; **, p Ͻ 0.01. WB, Western blot.
Consequently, we then addressed the question of how the ligation of ␤ 2 integrins down-regulates Rac activities in PMNs. We believe that our findings of a significant adhesion-induced increase in Rac-GAP activity in whole-cell lysates could at least in part explain the simultaneous decrease in Rac activities. Several observations support our conclusion. First, we have previously observed that engagement of ␤ 2 integrins on PMNs increased the activity of p190RhoGAP (a GAP protein for RhoA and Rac) and its translocation to a crude membrane fraction (22). Second, other investigators have detected Rac-GAP activity in a membrane fraction of PMNs and suggested that its modulation could play a role in the regulation of the NADPH oxidase (40). Third, the present finding that PI 3-kinase is essential for ␤ 2 integrin-induced Rac-GAP activation in adherent PMNs can well explain the ability of LY294002 to partly reverse the ␤ 2 integrin-induced decrease in Rac-GTP levels. Despite this, we made the somewhat surprising finding that inhibition of Src family tyrosine kinases had only a very modest effect on the ␤ 2 integrin-mediated activation of Rac-GAP, notwithstanding the fact that they reversed the ␤ 2 integrin-induced decrease in Rac-GTP levels to a similar extent as did inhibition of PI 3-kinase. Consequently, the ␤ 2 integrin-induced decrease in Rac activities could be regulated by at least one mechanism additional to the modulation of Rac-GAP activity. A possible explanation would be that GTP-loaded Rac GT-Pases translocate to the cytoskeleton (detergent-insoluble) via a Src-dependent mechanism.
Indeed, it has been proposed that GTP loading promotes movement of Rho GTPases either to the plasma membrane (41) or to detergent-insoluble cytoskeletal fraction (42). Therefore, we investigated the possibility that the decrease in the amounts of Rac-GTP levels in adherent cells could actually be due to a translocation of GTP-bound Rac to the detergentinsoluble fraction. However, there are three major findings that argue against such an explanation. First, although we detected a low level of Rac2 in the detergent-insoluble fraction of PMNs, we found no differences between resting cells and cells that had undergone integrin-mediated cell adhesion. Second, although incubation of PMNs on fibrinogen induced a pronounced translocation of Rac1 to the cytoskeleton (detergent-insoluble fraction), the adhesion-induced redistribution of Rac1 to the cytoskeleton and down-regulation of Rac1 activity did not exhibit parallel time kinetics. Third, despite the fact that PI 3-kinase inhibition did partly reverse the adhesioninduced decrease of Rac1 and Rac2 activities in PMNs, it did not enhance, but instead completely blocked, the translocation of Rac1 to the cytoskeleton-enriched Triton X-100-insoluble fraction. Thus, the ␤ 2 integrin-induced transfer of Rac1 to the cytoskeleton does not depend on GTP loading, meaning that translocation of GTP-bound Rac to the cytoskeleton cannot explain the decrease in Rac activities. Furthermore, the fact that Rac1 was selectively redistributed to the cytoskeletal fraction suggests that Rac1 and Rac2 could have different functions in PMNs. In agreement with this, other investigators have proposed that Rac2 may be an important regulator of the NADPH oxidase, whereas Rac1 primarily would be involved in controlling the rearrangement of the actin-based cytoskeleton (43).
To better understand the above findings, we wanted to identify an alternative mechanism responsible for the ␤ 2 integrininduced translocation of Rac GTPases. It is known that in their inactive GDP-bound forms, Rho GTPases are located in the cytosol, where they are bound to RhoGDIs, and it is assumed that they must be released from the RhoGDIs to be relocalized (19,20). Consequently, we addressed the question of whether ␤ 2 integrin ligation mediates dissociation of Rac GTPases from RhoGDIs in PMNs, and if so, how this occurs. In resting PMNs, we found an association between Rac GTPases and LyGDI, which is a hematopoietic cell-specific RhoGDI. Furthermore, engagement of ␤ 2 integrins in adherent PMNs triggered a PI 3-kinase-dependent but Src tyrosine kinase-independent dissociation of Rac1 as well as Rac2 from LyGDI. In accordance, it has been reported (in vitro, however) that the binding of phosphoinositides to RhoGDI disrupts the RhoGDI-Rac complex (44). Thus, the PI 3-kinase-dependent relocalization of Rac1 to the cytoskeleton-enriched fraction could possibly be regulated by the ␤ 2 integrin-mediated dissociation of Rac1 from LyGDI. Nevertheless, since Rac2 was not redistributed to the detergent-insoluble components of the cells, although it was released from LyGDI to the same extent as was Rac1, dissociation of the LyGDI-Rac complex is definitely a prerequisite but is not the only event necessary for relocalization of Rac proteins to occur. The fact that only Rac1 contains a C-terminal polybasic sequence that is a binding site for phosphoinositides (45) could possibly explain why Rac1 but not Rac2 translocates to the cytoskeleton.
Yan and Berton (46) have shown that PMNs incubated on fibrinogen and exposed to TNF, identical to the experimental conditions used in the present study, produced reactive oxygen intermediates. By using NBT as an electron acceptor to detect the production of superoxide by PMNs, we have here further confirmed that latter finding and ascertain that the NADPH oxidase is also activated in the present experimental situation. Rac1 and Rac2 are known to be part of the multicomponent, plasma membrane-bound enzyme NADPH oxidase in PMNs (16,33,46), but extensive research has not yet revealed the mechanisms by which these GTPases regulate this oxidase in phagocytes. It has been found that GTP and GTP-loaded Rac are needed to activate NADPH oxidase in cell-free systems (16). However, it has also been proposed that the GDP-bound form of Suspended PMNs (10 7 /ml) were pretreated for 20 min at 37°C with either the Src family tyrosine kinase inhibitor PP1 (3 M) or the PI 3-kinase inhibitor LY294002 (20 M, LY) and were incubated on a fibrinogen-coated surface and stimulated with TNF (20 ng/ml) for 30 min. Thereafter, the cells were lysed, and aliquots (50 g) of the lysates were incubated for 5 min with histidine-tagged Rac1 that was preloaded with [␥-32 P]GTP as described under "Materials and Methods." The samples were then filtered, and the filters were subjected to liquid scintillation counting to determine the amount of radioactivity that remained bound to Rac1. Rac-GAP activity is shown as a decrease in the amount of [ 32 P]GTP bound to Rac1. To measure the intrinsic GTPase activity of Rac1 in such assays, lysis buffer was added to histidinetagged Rac1 that was preloaded with [␥-32 P]GTP (designated No lysate). The data are expressed as percentage of the maximum (time 0) for each lysate and represent the means Ϯ S.E. of five separate experiments (performed in triplicate). Statistical significance (paired Student's t test) was as follows: *, p Ͻ 0.05. Rac1 in association with RhoGDI can potently stimulate NADPH oxidase (48,49). In addition, a pretreatment of PMNs with PP1 (50) or wortmanin (51) has been reported to block TNF-dependent ␤ 2 integrin-induced activation of the respiratory burst, treatments that we here have shown to reverse the down-regulation of Rac activities in adherent PMNs. However, we found, as have others, that stimulating PMNs with fMLP induced activation of both Rac (present results; see Refs. 33, 35, and 36) and the NADPH oxidase (37). Thus, together, the mentioned findings provide support for the notion that activation of the respiratory burst in intact PMNs does not depend on whether Rac is loaded with GDP or GTP. In favor of this assumption, Dang et al. (52) have reported that the function of Rac1 might be to promote interaction between cytochrome b 558 and p67 phox at the plasma membrane, and they found that such interplay did not depend on whether Rac1 was in its GDP-or GTP-bound form. In line with these findings, it is well established that Rac, along with p47 phox and p67 phox , must be translocated to the plasma membrane to ensure a proper chemotac-FIG. 6. ␤ 2 integrins induce translocation of Rac1, but not Rac2, to Triton X-100 insoluble fractions of PMNs. PMNs were either incubated on fibrinogen (adherent cells) or kept in suspension (suspended cells) and were stimulated with TNF (20 ng/ml) for the indicated times. Thereafter, the cells were lysed, the lysates were clarified by centrifugation, and the Triton X-100-insoluble fraction (i.e. the pellet) was resuspended in Laemmli buffer. Proteins (20 g) were separated by 12% SDS-PAGE, transferred to a PVDF membrane, and then immunoblotted with either an anti-Rac1 (A) or anti-Rac2 (B) antibody. PMNs in suspension were preincubated for 30 min at 37°C with either an isotype-matched control antibody IgG 2a (20 g/ml) or the anti-CD18 mAb (IB4) (20 g/ml) (C) or with either PP1 (3 M), LY294002 (20 M; LY), or wortmanin (10 nM; W) (D). Thereafter, the leukocytes were incubated on fibrinogen and stimulated with TNF (20 ng/ml) for 30 min. The cells were subsequently lysed, and the amount of Rac1 in the Triton X-100-insoluble fraction was determined by Western blot analysis, performed as described above for A and B. Each of the illustrated blots is representative of 4 -6 experiments. The arrows indicate the positions of Rac1 and Rac2.
FIG. 7. Rac GTPases dissociate from LyGDI during ␤ 2 integrinmediated adhesion of PMNs. PMNs in suspension (10 ϫ 10 6 /ml) were preincubated for 20 min at 37°C with either the Src family tyrosine kinase inhibitor PP1 (3 M) or the PI 3-kinase inhibitor LY294002 (20 M; LY). Thereafter, the cells were incubated on a fibrinogen-coated surface and stimulated with TNF for 30 min. The PMNs (20 ϫ 10 6 ) were then lysed, and clarified lysates were exposed to anti-LyGDI antiserum. The immunoprecipitated (IP) proteins were subjected to 12% SDS-PAGE, transferred to a PVDF membrane, and then immunoblotted with an anti-Rac1 antibody (upper panel). The membranes were subsequently stripped and reblotted with anti-Rac2 antiserum (middle panel) and an anti-LyGDI antiserum (lower panel), as described under "Materials and Methods." The arrows indicate the positions of Rac1, Rac2, and LyGDI. The illustrated blot is representative of three experiments. WB, Western blot.
FIG. 8. ␤ 2 integrins induce translocation of Rac1, Rac2, p47 phox , and p67 phox to a membrane-enriched fraction. A, PMNs (40 ϫ 10 6 ) were incubated on fibrinogen in the presence of TNF (20 ng/ml) for the indicated time periods and were subsequently disrupted by nitrogen cavitation, after which crude membrane fractions were prepared. The membrane proteins (5-10 g) were resolved by 12% SDS-PAGE, transferred to PVDF membranes, and immunoblotted with one of the following antibodies: anti-Rac1, anti-Rac2, anti-p47 phox , or anti-p67 phox . The arrows on the right indicate the position of Rac1, Rac2, p47 phox , or anti-p67 phox . A representative experiment (of three or four) is shown. B, PMNs in suspension (10 ϫ 10 6 /ml) were pretreated for 20 min at 37°C with the Src family tyrosine kinase inhibitor PP1 (3 M) or with the PI 3-kinase inhibitor, LY294002 (20 M; LY). Thereafter, the cells were incubated on a fibrinogen-coated surface in the presence of TNF for 30 min, and the amounts of Rac1, Rac2, p47 phox , and p67 phox in a membrane-enriched fraction were determined as described for A. A representative experiment (of three) is shown. tic peptide-induced activation of the respiratory burst in human PMNs (53). Here we found that ␤ 2 integrin induced translocations of Rac1, Rac2, p47 phox , and p67 phox to a membrane-enriched fraction. The fact that translocation of Rac and p47 phox /p67 phox to the membrane correlates with activation of the NADPH oxidase (53) and that inhibition of PI 3-kinase and/or Src family members significantly reduced the translocation of these proteins to the plasma membrane is enough to explain their effects on ␤ 2 integrin-induced activation of the respiratory burst (50,51).
In summary, we have made several novel and important observations regarding ␤ 2 integrin regulation of Rac1 and Rac2 in adherent PMNs. Based on these, we propose that ␤ 2 integrindependent activation of the respiratory burst in adherent PMNs is not dependent on the amounts of GTP-bound Rac but is instead triggered by relocalization of Rac GTPases as well as p47 phox and p67 phox to the plasma membrane.