Abstract
The role of phosphoinositide 3 kinases (PI 3-K) in chemokine-induced NK cell chemotaxis was investigated. Pretreatment of NK cells with wortmannin inhibits the in vitro chemotaxis of NK cells induced by lymphotactin, monocyte-chemoattractant protein-1, RANTES, IFN-inducible protein-10, or stromal-derived factor-1α. Introduction of inhibitory Abs to PI 3-Kγ but not to PI 3-Kα into streptolysin O-permeabilized NK cells also inhibits chemokine-induced NK cell chemotaxis. Biochemical analysis showed that within 2–3 min of activating NK cells, pleckstrin is recruited into NK cell membranes, whereas PI 3-Kγ associates with these membranes 5 min after stimulation with RANTES. Recruited PI 3-Kγ generates phosphatidylinositol 3,4,5 trisphosphate, an activity that is inhibited upon pretreatment of NK cells with wortmannin. Further analysis showed that a ternary complex containing the βγ dimer of G protein, pleckstrin, and PI 3-Kγ is formed in NK cell membranes after activation with RANTES. The recruitment of pleckstrin and PI 3-Kγ into NK cell membranes is only partially inhibited by pertussis toxin, suggesting that the majority of these molecules form a complex with pertussis toxin-insensitive G proteins. Our results may have application for the migration of NK cells toward the sites of inflammation.
Natural killer cells are antiviral and antitumor effector cells (reviewed in 1 . Chemokines are proinflammatory cytokines that are responsible for the accumulation of various cell types at the inflammatory sites 2 . They are divided into four subfamilies depending on the arrangement of the first cysteine residues in the N-terminal region. These are designated as CXC (α), CC (β), C (γ), and CX3C (δ) chemokines. The importance of chemokines and NK cells in the eradication of viral infection has been explored by Salazar-Mather et al. 3 , who observed that MIP-1α3 recruits NK cells into the sites of CMV foci, resulting in reduced infection with this virus. In addition, members of the CC chemokines induce the polarization of NK cells, and facilitate NK/target cell conjugate formation 4 . The intracellular signaling pathways induced by chemokines in NK cells in order for these cells to polarize and extravasate into various tissues are largely unknown.
Chemokine receptors are coupled to the heterotrimeric G proteins in various cell types 5 . The heterotrimeric G proteins are composed of three subunits, α, β, and γ. The chemotactic effect of chemokines is mediated through pertussis toxin (PT) substrates 2, 5 . Similarly, MIP-1α-, MCP-1-, RANTES-, SDF-1α-, IP-10-, and lymphotactin-induced NK cell chemotaxis is mediated through the heterotrimeric G proteins in NK cells (reviewed in 6 . We have suggested previously that the βγ subunit of G proteins is involved in IP-10-, or lymphotactin-induced NK cell chemotaxis 7 . Recently, the involvement of Gβγ subunits in the chemotaxis of transfected cell lines has been reported 8, 9 . However, the downstream signaling molecules important for this activity have not been examined.
In addition to G protein-coupled receptors (GPCR), receptor tyrosine kinases transmit intracellular signals resulting in the activation of various molecules, such as the phosphoinositide 3-kinase (PI 3-K), among others. PI 3-K is present in three different forms, I, II, and III. These enzymes phosphorylate the D3 hydroxy position of the inositol ring of phosphatidylinositol 10 . IA is composed of a catalytic subunit of p110 α, p110 β, or p110 δ. These catalytic subunits are associated with a regulatory p85 subunit. The latter contains SH2 domains, SH3 domain, and proline-rich segments, which facilitate its interaction with multiple sites present in various proteins. On the other hand, IB (p110 γ subunit) does not have a site that binds the regulatory p85, but instead has a motif, presumed to be pleckstrin homology (PH), that binds the βγ subunit of G proteins 11 . PH domains are present in more than 100 proteins, including GTP exchange factors such as SOS and ARNO; GTPase-activating proteins such as Ras-GAP; and in proteins such as dynamin, spectrin, general receptors for phosphoinositide, in kinases such as Bruton’s tyrosine kinase, and in phospholipases such as PLCβ, PLCγ, or PLCδ, among other molecules 12 . The PH domains associate with phospholipids to recruit the PH-containing proteins in the proximity of their substrates present in the cell membranes. For example, PH domain of PLCγ associates with phosphatidylinositol 3,4,5 trisphosphate (PI3, 4, 5 P3) and is recruited into the membrane by this phospholipid, where it hydrolyzes its substrate, the PIP2 13 . Also, the PH domain of SOS is recruited into the membranes, where it activates Ras 14 . In addition to binding phospholipids, PH domains are important for protein-protein interactions. For example, the PH domains of the β adrenergic receptor kinase (G protein-coupled receptor kinase 2) bind the βγ dimer of PT-sensitive G proteins 15 . Although Gβγ activates PI 3-Kγ, resulting in the activation of the mitogen-activated protein kinase pathway 16 , or Jun kinase 17 , the binding of PI 3-Kγ-PH to Gβγ has been disputed 18 . Pleckstrin, the major substrate for protein kinase C in platelets, is composed of two PH domains (N-terminal and C-terminal) separated by 150 amino acids. Activation of platelets with thrombin results in the phosphorylation of pleckstrin by protein kinase C. Phosphorylated pleckstrin in turn inhibits PI 3-Kγ activity as a result of activating GPCR 19 . This inhibition was reversed upon the addition of purified Gβγ, suggesting that phosphorylated pleckstrin may interact with the βγ dimer. However, a correlation of coupling and/or function between Gβγ, pleckstrin, and PI 3-Kγ has not been reported. In this study, we explored the possibility that there is an interaction among these components as a result of activating NK cells with chemokines.
Materials and Methods
Culture medium
Culture medium consisted of RPMI 1640 supplemented with 10% human AB serum (Ulleval Hospital, Oslo, Norway), 10 U/ml penicillin, 100 μg/ml streptomycin, 1 mM L-glutamine, 1% nonessential amino acids (all from Life Technologies, Paisley, U.K.), and 5 × 10−5 M 2-ME (Sigma, St. Louis, MO). AIM-V medium was from Life Technologies.
Preparation of NK cell membranes
IL-2-activated NK cells were prepared by adherence to plastic flasks, as previously described 20, 21 . The majority of these cells (more than 85%) show the CD16+CD56+CD3− phenotype, as determined by flow-cytometric analysis. Before performing the biochemical assays described in this work, these cells were incubated overnight in a serum-free (AIM-V) medium. This was necessary since serum has been reported to induce the recruitment of PH-containing proteins into the plasma membranes of various cell types 14 . NK cell membranes were prepared by incubating the cells in a lysis buffer containing 10 mM HEPES, pH 7.5, 3 mM MgCl2, 40 μg/ml PMSF, 10 μg/ml leupeptin, 2 μg/ml pepstatin A, and 2 μg/ml aprotinin. After homogenization and sonication, the mixtures were centrifuged at 1000 × g for 10 min. The supernatants were transferred into Beckman tubes; ultracentrifuged in a buffer containing 10 mM HEPES, 3 mM MgCl2, and 2 mM EDTA; and snap frozen at −70°C.
Chemotaxis assay
Nucleopore blind well chemotaxis chambers with a lower well volume of 200 μl were used. A maximum volume of 200 μl medium containing RPMI plus 1% BSA was placed in the lower wells in the presence or absence of various agents. Cells (1 × 105) were placed in the upper compartments of Boyden chambers above the filters. The chambers were incubated for 2 h at 37°C in a 5% CO2 incubator. The filters were then removed, dehydrated, and stained with 15% Giemsa stain for 7 min and then mounted on glass slides using a drop of immersion oil between the filters and the slides. Cells in 10 high power fields from two filters were counted and averaged for each sample. Migration index was calculated as the number of cells migrating toward the concentration gradients of chemokines, divided by the number of cells migrating toward medium only.
Permeabilization with streptolysin O (SLO)
The methods of permeabilization and introduction of Abs have been previously described 7, 21 . In brief, human NK cells (5 × 106/100 μl) were incubated with 200 ng/100 μl of activated SLO at 4°C for 15 min in a buffer containing 150 mM K+-glutamate, 5 mM nitrilotriacetic acid, 0.5 mM EGTA, 0.2% BSA, and 10 mM PIPES, pH 7.2. The cells were washed three times in the cold, and then warmed to 37°C, incubated for an additional 15 min, and then washed with a buffer containing 150 mM KCl and 20 mM PIPES, pH 7.2. Goat affinity-purified anti-PI 3-Kγ, rabbit affinity-purified anti-PI 3-Kα (Santa Cruz Laboratories, San Diego, CA), or goat IgG were incubated with SLO-permeabilized NK cells for 1 h at 4°C. The cells were extensively washed, and then incubated in culture medium for 3 h at 37°C to allow the cells to rest before examining their ability to migrate in the microchemotaxis chambers.
Coimmunoprecipitation and immunoblotting
Coimmunoprecipitation assay was done as described 22 . Protein samples were incubated with 1/100 dilution of the proper Ab (mouse anti-pleckstrin-PH from Transduction Laboratories (Lexington, KY), rabbit anti-Gβ from NEN-DuPont (Boston, MA), and goat anti-PI 3-Kγ from Santa Cruz) in the immunoprecipitation buffer (10 mM Tris, pH 7.4, 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 0.2 mM sodium vanadate, 0.2 mM PMSF, 1% Triton-X, and 0.5% Nonidet P-40) overnight at 4°C with gentle shaking. About 20 μl of protein A/G agarose (Santa Cruz) was added to each sample and incubated for another 4 h, centrifuged, and washed three times, and the pellets were suspended in electrophoresis sample buffers before running on SDS-PAGE. Immunoblotting was performed by running the samples on SDS-PAGE and then electrotransferred into polyvinylidene difluoride (PVDF) membranes, blocked with 5% skim milk in TBS buffer for 2 h, washed, and incubated with 1/1000 of the primary Abs, and 1/500 dilution of the secondary Abs. Development was done by using either horseradish peroxidase-color development reagents (Bio-Rad) or enhanced chemoluminescence (ECL) reagents (Amersham, Arlington Heights, IL) according to the manufacturer’s specifications.
Pretreatment with PT or wortmannin
Cells were incubated overnight in AIM-V medium. They were either left intact or were treated overnight with 100 ng/ml activated PT (Sigma). These cells were harvested and membranes were prepared as described above. Pretreatment with wortmannin (Sigma) was done by incubating NK cells (1 × 106) with various concentrations of this metabolite at 37°C for 30 min, or with the appropriate concentration of DMSO, washed, and examined as above.
Phosphatidylinositol kinase assay
NK cells were either left intact or pretreated with 100 nM of wortmannin. These cells were incubated with 100 pg/ml RANTES for 5 min, and membranes were prepared from these cells. The membranes were immunoprecipitated with anti-PI 3-Kγ overnight at 4°C in the presence of protein A/G agarose. The immunocomplex was washed once with ice-cold PBS and once with a buffer containing 0.5 M LiCl, 100 mM Tris-HCl, pH 7.5, and 1 mM sodium orthovanadate. It was then washed once with distilled water, and once with a washing buffer containing 20 mM Tris-HCl, pH 7.5, 0.5 mM EGTA and 100 mM NaCl. The immunocomplex was suspended in 50 μl of washing buffer, and was added to 0.5 μl of PI4, 5 P2 (20 mg/ml; Sigma). This mixture was incubated for 10 min at 25°C with 10 μCi/sample of [γ-32P]ATP (Amersham). MgCl2 (20 mM) was then added to this mixture, which was incubated for additional 15 min. The reaction was stopped by the addition of 150 μl of chloroform/methanol/11.6 N HCl, with the volume ratio of 100:200:2. The organic phase was separated by the addition of 100 μl chloroform. The lipid in the organic phase was separated on TLC aluminum silica gel 60 precoated sheets (Merck, Darmstadt, Germany). TLC plates were developed and resolved in chloroform/methanol/ammonium hydroxide/distilled water, with the volume ratio of 124:114:15:21. Radioactive PIP products were visualized by autoradiography.
Statistical analysis
Significant values were determined by using the two-tailed Student’s t test.
Results
PI 3-Kγ controls chemokine-induced NK cell chemotaxis
PI 3-K has been shown to play an important role in NK cell activity upon perturbation of FcR 23, 24 . In addition, wortmannin, a fungal inhibitor of PI 3-K activity, inhibits RANTES-induced T cell chemotaxis 25 . To investigate the role of PI 3-K in chemokine-induced NK cell chemotaxis, we pretreated NK cells with different concentrations of wortmannin. Fig. 1⇓ shows that the C chemokine lymphotactin (Ltn), the CC chemokines MCP-1 and RANTES, or the CXC chemokines IP-10 and SDF-1α induced the in vitro chemotaxis of NK cells (p < 0.005, as compared with cells migrating toward culture medium only). Low concentration of wortmannin (10 pM) did not inhibit chemokine-induced NK cell chemotaxis. However, 1 nM of wortmannin inhibited the chemotactic activity induced by Ltn, MCP-1, or IP-10 (p < 0.02, as compared with the migration of cells not treated with wortmannin), whereas higher concentrations (10, 100, and 1000 nM) inhibited RANTES and SDF-1α (p < 0.05), as well as Ltn-, MCP-1-, and IP-10-induced NK cell chemotaxis. None of the concentrations of wortmannin used affected the viability of NK cells, as determined by trypan blue exclusion test (not shown). To establish which PI 3-K isotype is involved in the chemokine activity, we introduced Abs to PI 3-K into SLO-permeabilized cells. Results in the upper panel of Fig. 2⇓ show that affinity-purified goat anti-PI 3-Kγ inhibited Ltn, MCP-1, RANTES, IP-10, or SDF-1α induction of NK cell chemotaxis (p < 0.001, p < 0.003, p < 0.004, p < 0.02, and p < 0.001, respectively, as compared with the migration of cells treated with goat IgG). Neither goat IgG, nor Ab to PI 3-Kα inhibited the chemotactic activity of the five chemokines examined (Fig. 2⇓, upper panel). Because all five chemokines examined exerted similar effects, we have utilized the CC chemokine RANTES for further studies. Dose-response curve shows that 0.01 and 0.1 μg of anti-PI 3-Kγ/1 × 106 permeabilized NK cells are not inhibitory, whereas 1 and 12.5 μg doses are inhibitory (p < 0.005) for RANTES-induced NK cell chemotaxis. As a control, different concentrations of this Ab were incubated with unpermeabilized cells. None of these concentrations inhibited RANTES-induced intact NK cell chemotaxis, indicating that the Ab must enter permeabilized cells before it exerts its inhibitory effect.
PI 3-K is important for chemokine-induced NK cell motility. NK cells were either untreated (solid columns) or treated (1 × 106/ml) with 0.01–1000 nM of wortmannin for 30 min at 37°C, washed, and then examined. These cells (1 × 105/100 μl) were incubated in the upper wells of Boyden chambers. In the lower wells, 1 ng/ml Ltn, 1 ng/ml MCP-1, 100 pg/ml RANTES, 1 ng/ml IP-10, or 2 ng/ml SDF-1α was incubated. Two hours later, the filters were collected and examined. Migration Index was calculated as the number of migrating cells in the presence of chemokines divided by the number of migrating cells in the presence of culture medium only (white columns = control). About 10–15% of the cells migrated within the 2-h incubation period. Mean ± SD of three to six experiments.
Effect of anti-PI 3-K on chemokine-induced NK cell chemotaxis. In the upper panel, NK cells were permeabilized with SLO, and were either left untreated (solid columns), or were incubated (1 × 106/100 μl) with Abs to nonimmune goat IgG (12.5 μg/100 μl), PI 3-Kγ (12.5 μg/100 μl), or PI 3-Kα (12.5 μg/100 μl). NK cells (1 × 105/100 μl) were incubated in the upper wells of Boyden chambers, whereas 1 ng/ml Ltn, 1 ng/ml MCP-1, 100 pg/ml RANTES, 1 ng/ml IP-10, or 2 ng/ml SDF-1α was incubated in the lower wells. In the lower panel, different concentrations (ranging between 0.01–12.5 μg/100 μl) of anti-PI 3-Kγ were incubated with 1 × 106/100 μl of either SLO-permeabilized, or unpermeabilized NK cells. These cells were washed and incubated (1 × 105/100 μl) in the upper wells of Boyden chambers. In the lower wells, 100 pg/ml of RANTES was incubated. Two hours later, the filters were collected and examined. Migration Index was calculated as described in Fig. 1⇑.
These results suggest that PI 3-Kγ plays an important role in the chemotaxis of NK cells. To biochemically address the nature of PI 3-Kγ effect, we investigated its recruitment to the membranes after stimulating NK cells with RANTES. Fig. 3⇓A shows that PI 3-Kγ is not present in NK cell membranes. However, it is recruited into these membranes upon stimulation of these cells with RANTES, as determined by immunoblotting NK cell membranes with a specific Ab to the catalytic p110γ subunit (lane 2). No such recruitment was observed in the absence of stimulation with RANTES (lane 1 in Fig. 3⇓A). To examine the association of PI 3-Kγ with Gβγ, NK cells were activated with culture medium or with RANTES for 5 min. Membranes were prepared from these cells, immunoprecipitated overnight with anti-PI 3-Kγ, and then immunoblotted with Ab to the common β subunit of G protein. The results in lane 1 of Fig. 3⇓B show that in the absence of stimulation with RANTES, anti-Gβγ picked up a 37-kDa band representing Gβγ in NK cell membranes. Importantly, an association between PI 3-Kγ and βγ subunit of G proteins occurred after stimulation with RANTES (lane 2 in Fig. 3⇓B). Taken together, the results clearly show that chemokine activation of GPCR results in the recruitment of PI 3-Kγ into NK cell membranes, and its association with the Gβγ dimer.
Recruitment of PI 3-Kγ and its association with Gβγ into NK cell membranes after stimulation with RANTES. A, NK cells were either stimulated with culture medium (lane 1) or with 100 pg/ml RANTES (lane 2) for 5 min at 37°C. Membranes were prepared from these cells, and then immunoblotted with an Ab to PI 3-Kγ. In the control (c), nonimmune isotype-matched Abs were used. B, NK cells were incubated overnight in the presence of serum-free medium, and were stimulated with either culture medium (lane 1) or 100 pg/ml RANTES (lane 2) for 5 min. Membranes were prepared from these cells, immunoprecipitated with anti-PI 3-Kγ overnight, and then immunoblotted with Ab to the common β of G proteins. In the control (c), nonimmune isotype-matched Abs were used. The molecular mass of PI 3-Kγ is 110 kDa.
Chemokines induce the recruitment of pleckstrin into NK cell membranes
After incubation in serum-free medium for 18 h, NK cells were activated for 5 min with culture medium, Ltn, MCP-1, RANTES, IP-10, or SDF-1α. Membranes were prepared from these cells, and were examined for the presence of pleckstrin. Fig. 4⇓A shows that all five chemokines induced the recruitment of pleckstrin into NK cell membranes (lanes 2–6). To investigate whether chemokines recruit pleckstrin from the cytosol, NK cells were treated with either culture medium or RANTES. The results show that pleckstrin is abundant in NK cell lysates, but not in NK cell membranes (lane 1 in the left and right panels of Fig. 4⇓B). However, it is recruited into the membranes after stimulating the cells for 5 min with RANTES (lane 2 in the left and right panels of Fig. 4⇓B), suggesting that stimulation with chemokines facilitates the distribution of this protein from the cytosol into the membranes. A physical association between pleckstrin and Gβγ dimer was then examined. NK cells were stimulated with culture medium or with RANTES, and membranes were prepared from these cells, immunoprecipitated with anti-pleckstrin-PH, and then immunoblotted with anti-Gβ Ab (left panel in Fig. 4⇓C). Reciprocally, the membranes were immunoprecipitated with anti-Gβ, and then immunoblotted with anti-pleckstrin-PH (right panel in Fig. 4⇓C). Only after stimulation with RANTES (lane 2 in both panels of Fig. 4⇓C) and not with culture medium (lane 1 in both panels), an association occurred between pleckstrin and Gβγ.
Recruitment of pleckstrin, and its association with Gβγ in NK cell membranes. A, NK cells (1 × 106/ml) were incubated for 5 min with culture medium (lane 1), 1 ng/ml Ltn (lane 2), 1 ng/ml MCP-1 (lane 3), 100 pg/ml RANTES (lane 4), 1 ng/ml IP-10 (lane 5), or 2 ng/ml SDF-1α (lane 6). Membranes were prepared from these cells, and were immunoblotted with anti-pleckstrin-PH. A 45–47-kDa representing pleckstrin is marked with an arrow. In the control (c), nonimmune isotype-matched Abs were used. B, NK cells were treated with either culture medium (lane 1 in the right and left panels) or 100 pg/ml RANTES for 5 min (lane 2 in the right and left panels). Extracts or membranes were prepared from these cells, and were immunoblotted with anti-pleckstrin-PH. C, NK cells were treated with either culture medium (lane 1 in the right and left panels) or RANTES (lane 2 in the right and left panels). Membranes were prepared from these cells, and were either immunoprecipitated with anti-Gβ, and then immunoblotted with anti-pleckstrin-PH (right panel), or were immunoprecipitated with anti-pleckstrin-PH, and then immunoblotted with anti-Gβ (left panel).
βγ dimer of PT-resistant G proteins forms a complex with pleckstrin and PI 3-Kγ in NK cell membranes
To demonstrate the nature of G proteins involved in the recruitment of pleckstrin and PI 3-Kγ, NK cells were either left untreated or pretreated with 100 ng/ml of PT. Results in Fig. 5⇓, A and B, show that RANTES induced the recruitment of pleckstrin and PI 3-Kγ, respectively, as compared with culture medium-treated cells (lane 3 versus lane 1). However, there was only a partial inhibition of the recruitment of pleckstrin (lane 2 in Fig. 5⇓A) or PI 3-Kγ (lane 2 in Fig. 5⇓B) upon pretreatment of NK cells with 100 ng/ml of PT. Higher concentrations of PT affected the viability of the cells (data not shown).
Effect of PT or wortmannin on the recruitment of pleckstrin or PI 3-Kγ. A, NK cells were either untreated, or treated with 100 ng/ml PT overnight (lane 2), washed, and were incubated with either culture medium (lane 1) or 100 pg/ml RANTES (lanes 2 and 3) for 5 min. Membranes were prepared from these cells and were immunoblotted with anti-pleckstrin-PH. B, NK cells were treated as in A, except that the membranes were immunoblotted with anti-PI 3-Kγ. C, NK cells were treated with either culture medium (lanes 1 and 3), or 100 nM wortmannin for 30 min at 37°C (lane 2), washed extensively, and then incubated with culture medium (lane 1), or with RANTES (lanes 2 and 3) for 5 min. Membranes were prepared from these cells and were immunoblotted with anti-pleckstrin-PH. In the controls (c), nonimmune isotype-matched Abs were used.
Recruitment of pleckstrin or PI 3-Kγ into NK cell membranes is resistant to wortmannin treatment
To investigate whether PI 3-Kγ plays a role in the recruitment of pleckstrin into NK cell membranes, NK cells were pretreated with culture medium or wortmannin, washed, and then stimulated with RANTES. In lane 1 of Fig. 5⇑C, a faint band representing pleckstrin was present in NK cell membranes in the absence of stimulation. The existence of such a band was variable in cells generated from different donors. More important, treatment with RANTES resulted in the recruitment of pleckstrin into NK cell membrane (lane 3 in Fig. 5⇑C). Concentration of wortmannin (100 nM), which inhibited RANTES-induced NK cell chemotaxis (Fig. 1⇑), did not inhibit the recruitment of pleckstrin into NK cell membranes (lane 2 in Fig. 5⇑C).
Similarly, pretreatment of NK cells with wortmannin failed to inhibit the association of PI 3-Kγ with NK cell membranes upon stimulation with RANTES. Fig. 6⇓A shows that incubation of NK cells for 5 min with RANTES, and not with culture medium resulted in the association of PI 3-Kγ with NK cell membranes (left versus right lanes; similar to the results obtained in Fig. 2⇑A). Pretreatment of NK cells with 100 nM wortmannin before incubation with RANTES did not inhibit the association of PI 3-Kγ with NK cell membranes (Fig. 6⇓A, middle versus right lane). To correlate the inhibitory effect of wortmannin on chemokine-induced NK cell chemotaxis with the recruitment of PI 3-Kγ into NK cell membranes, we examined the ability of membrane-associated PI 3-Kγ to generate phospholipids. After stimulation with RANTES, NK cell membranes were immunoprecipitated with Ab to PI 3-Kγ overnight. This immune complex was mixed with PI4, 5 P2 in the presence of [32P]ATP. PI 3-Kγ activity and the generation of PI3, 4, 5 P3 were only seen after stimulation with RANTES and not with culture medium (Fig. 6⇓B, right and left lanes, respectively). The ability of PI 3-Kγ immunoprecipitated from NK cell membranes after stimulation with RANTES to generate PI3, 4, 5 P3 was inhibited upon prior pretreatment of NK cells with 100 nM of wortmannin (Fig. 6⇓B, middle versus right lane).
Wortmannin inhibits the activity of membrane-associated PI 3-Kγ. A, NK cells (1 × 106/ml) were either left intact (left and right lanes) or were pretreated with 100 nM of wortmannin for 30 min at 37°C (middle lane). These cells were extensively washed, and were incubated with either culture medium (left lane) or 100 pg/ml of RANTES for 5 min (middle and right lanes). Membranes were prepared from these cells, and were immunoblotted with anti-PI 3-Kγ. B, NK cells were treated similarly as in A. Their membranes were immunoprecipitated with anti-PI 3-Kγ, and then mixed with PI(4, 5)P2 in the presence of [γ-32P]ATP. The generation of PI(3,4,5)P3 was spotted on TLC silica gel plates (designated as PIP). Orig = the mixtures before migrating on TLC plates.
RANTES induces the formation of a ternary complex in NK cell membranes
To examine the association of these molecules in NK cell membranes, NK cells were stimulated with RANTES for 1–20 min, and membranes were prepared from these cells, immunoprecipitated with anti-PI 3-Kγ, and then immunoblotted with Abs to either PI 3K-γ (Fig. 7⇓A) or pleckstrin-PH (Fig. 7⇓B). An association between these two molecules occurred after 3 min (B). The same association was observed when NK cell membranes were immunoprecipitated with anti-pleckstrin first, and then immunoblotted with anti-PI 3-Kγ (C). However, the majority of the band detected after 3 min may be due to pleckstrin since there was only a low recruitment of PI 3-Kγ at this time (A). A strong association between pleckstrin and PI 3-Kγ occurred 5 min after stimulation with RANTES (B). This coincided with the robust recruitment of PI 3-Kγ into NK cell membranes at this time point (A). The association between these two molecules was also apparent 10 min after stimulation. Both pleckstrin and PI 3-Kγ almost disappeared from this complex 20 min after stimulation with RANTES (B and C), indicating a transient recruitment of these molecules after stimulation with RANTES. Time kinetic association between pleckstrin and Gβγ was also examined. After RANTES stimulation of NK cells for various times, membranes from these cells were immunoprecipitated with anti-pleckstrin-PH and then immunoblotted with anti-Gβ. Fig. 7⇓D shows that an association between these molecules occurred after 2 min and was strong after 3 and 5 min of stimulation. However, after 10 min, most of the Gβγ dimer dissociated from pleckstrin.
Formation of a ternary complex in NK cell membranes after stimulation with RANTES. NK cells were incubated with 100 pg/ml RANTES for 1–20 min. Membranes of these cells were immunoprecipitated with anti-PI 3-Kγ, and then immunoblotted with either anti-PI 3-Kγ (A) or anti-pleckstrin-PH (B). Reciprocally, NK cell membranes were immunoprecipitated with anti-pleckstrin-PH, and then immunoblotted with anti-PI 3-Kγ (C). In D, NK cells were treated as above, except that the membranes were immunoprecipitated with anti-pleckstrin-PH, and then immunoblotted with anti-Gβ.
Discussion
Chemotactic cytokines (chemokines) are proinflammatory mediators implicit for the recruitment of various cell types into the inflammatory sites (reviewed in Refs. 2 and 5). Depending on the presence and the arrangement of the first cysteine residues in the N-terminal region, chemokines are divided into four subfamilies: CXC (α), CC (β), C (γ), and CX3C (δ). The CC chemokine MIP-1α, MCP-1, RANTES, macrophage-derived chemokine, MIP-3α, or MIP-3β 20, 26, 27, 28, 29, 30 ; the C chemokine lymphotactin 7, 31, 32 ; the CXC chemokine IP-10, or SDF-1α 7, 27, 33 ; and the CX3C chemokine fractalkine 30, 34 have been shown to induce the chemotaxis of NK cells. Furthermore, MIP-1α recruits NK cells toward the livers of CMV-infected mice, which resulted in increased inflammation and decreased susceptibility to infection with this virus 3 .
Receptors for CC chemokines are coupled to the heterotrimeric G proteins, which are composed of three subunits, α, β, and γ. The presence and identity of the G proteins in human and rat NK cell membranes have been investigated. It was reported that these membranes expressed Gi, Go, Gs, Gq, Gz, and G13. These G proteins play vital roles as early transducers of various biological functions in NK cells, such as chemotaxis (reviewed in 6 . However, the downstream signaling molecules important for the chemotaxis of NK cells are not known. It has been observed that wortmannin, a fungal inhibitor of PI-3K, inhibits RANTES-induced T cell chemotaxis 25 . Because wortmannin has effects other than inhibition of PI 3-K, it is not clear whether this enzyme is involved in mediating chemokine-induced cellular chemotaxis. In this study, we showed that wortmannin as well as inhibitory Abs to PI 3-Kγ but not PI 3-Kα inhibit C, CC, and CXC chemokine-induced NK cell chemotaxis, suggesting that PI 3-K IB plays an important role in chemokine activation of NK cells.
In addition, we showed that PI 3-Kγ is recruited into NK cell membranes within 5 min after stimulation of these cells with the CC chemokine RANTES. At almost the same time, an association between PI 3-Kγ and the βγ dimer of G proteins occurred. In addition, we observed that pleckstrin is recruited from the cytosol into NK cell membranes within 2–3 min after stimulating NK cells with C, CC, or CXC chemokines. The recruitment of pleckstrin was insensitive to wortmannin pretreatment, suggesting that PI 3-Kγ products such as PI3, 4 P2 or PI3, 4, 5 P3 are not important for the recruitment of pleckstrin into NK cell membranes shortly after activating GPCR. These results contrast the recruitment of PLCγ-PH into the cell membranes upon stimulation with growth factors, which occurs as a result of binding the PLCγ-PH to phospholipids, which facilitates its association with the cell membrane. This activity was sensitive to wortmannin pretreatment 13 . Therefore, there may be multiple mechanisms by which PH domain-containing proteins are recruited into the membranes, depending on the pathway utilized by the ligands.
In addition, we observed that wortmannin did not inhibit the recruitment of PI 3-Kγ into NK cell membranes. To correlate the finding that wortmannin inhibits NK cell chemotaxis with the recruitment of this kinase into the membranes of these cells, we observed that wortmannin inhibited the activity of this kinase. Hence, the generation of PI3, 4, 5 P3 (Fig. 6⇑) or PI3, 4 P2 (data not shown) was abrogated upon pretreatment of NK cells with this metabolite. These results suggest that during activation with RANTES, PI 3-Kγ is recruited into NK cell membranes, placing it in close proximity to its phospholipid substrates, resulting in the phosphorylation of various phosphatidylinositol lipids.
Interestingly, pretreatment of NK cells with PT only partially inhibited the recruitment of either pleckstrin or PI 3-Kγ into their membranes, suggesting that the majority of pleckstrin or PI 3-Kγ that are recruited into NK cell membranes shortly after stimulation with RANTES are associated with the βγ subunit of PT-resistant G proteins. This is not surprising, considering that receptors for RANTES as well as other chemokines present on NK cells are coupled to the PT-insensitive Gq and Gz and the PT-sensitive Gi and Go 21 .
Taken together, it appears that shortly after activation of NK cells with chemokines, the βγ dimer of G proteins associates with pleckstrin, most likely through the PH domain since the Ab we have utilized is specific for PH domain of pleckstrin. PI 3-Kγ is also associated with the βγ/pleckstrin-PH complex forming a ternary complex in NK cell membranes as a result of activating NK cells with the CC chemokine RANTES. It is possible that the Gβγ dimer may act as a docking molecule for pleckstrin and PI 3-Kγ. This is similar to the function of Gβγ, which upon activation of the M2 and M3 muscarinic receptors, recruits G protein-coupled receptor kinase, resulting in the phosphorylation and desensitization of these receptors 35 . However, recruitment of PI 3-kinase, as described in this study, results in activation and chemotaxis. It appears that after forming this ternary complex, the βγ dimer dissociates, presumably reassociating with the α subunit in the membranes, whereas the PI 3-Kγ mediates NK cell chemotaxis by phosphorylating phospholipids such as phosphatidylinositol 4 phosphate (PI4 P) or phosphatidylinositol 4,5 bisphosphate (PI4, 5 P2), which are important for this process. These results may shed some light on the biochemical events utilized by chemokines to induce the polarization and the extravasation of NK cells into the sites of virus infection, or tumor growth 3, 4 .
Footnotes
↵1 This work was supported by grants from the Norwegian Cancer Society, the Norwegian Research Council, the Anders Jahres Fond, HFSP, and Hydro Norway. A.A.M. is a Senior Scientist of the Norwegian Cancer Society.
↵2 Address correspondence and reprint requests to Dr. A. A. Maghazachi, Department of Anatomy, University of Oslo, P.O. Box 1105 Blindern, Oslo, Norway N-0317. E-mail address: azzam.maghazachi{at}basalmed.uio.no
↵3 Abbreviations used in this paper: MIP-1α, macrophage-inflammatory protein-1α; GPCR, G protein-coupled receptor(s); IP-10, IFN-inducible protein-10; Ltn, lymphotactin; MCP-1, monocyte-chemoattractant protein-1; PH, pleckstrin homology; PI 3-K, phosphoinositide 3-kinase; PI(3,4)P2, phosphatidylinositol 3,4 bisphosphate; PI(3,4,5)P3, phosphatidylinositol 3,4,5 trisphosphate; PI(4,5)P2, phosphatidylinositol 4,5 bisphosphate; PLC, phospholipase C; PT, pertussis toxin; SDF-1, stromal-derived factor-1; SLO, streptolysin O.
- Received August 20, 1998.
- Accepted December 14, 1998.
- Copyright © 1999 by The American Association of Immunologists
References
In this issue
