Rap1 Is a Potent Activation Signal for Leukocyte Function-Associated Antigen 1 Distinct from Protein Kinase C and Phosphatidylinositol-3-OH Kinase

Cells and cDNA transfection.

BAF cells, a pro-B-cell line (48), were suspended with RPMI 1640 medium (GibcoBRL, Grand Island, N.Y.) containing 10% fetal calf serum (FCS; Sigma Chemical Co., St. Louis, Mo.), 50 μM β-mercaptoethanol, and 10% WEHI-3 conditioned medium as a source of interleukin-3 (IL-3). For the stable expression of human LFA-1 in BAF cells, 10-μg aliquots of human αL and β2 integrin cDNAs in CDM8 (Invitrogen, Carlsbad, Calif.) were cotransfected with 2 μg of pTK-Hyg (Clontech Laboratories, Palo Alto, Calif.) by electroporation (Bio-Rad Laboratories, Hercules, Calif.) at 370 V and 960 μF. Transfected cells were selected with hygromycin at 0.5 mg/ml (Wako Pure Chemical Industries, Ltd., Osaka, Japan) and expanded for the analysis of LFA-1 expression with a FACScan (Becton Dickinson, Oxnard, Calif.). Several LFA-1-expressing BAF (BAF/hLFA-1) clones were isolated by a limiting dilution and subjected to adhesion assays. BAF/hLFA-1, which shows activation-dependent adhesion to ICAM-1, was further introduced by electroporation using the genes encoding signaling molecules as described below and selected using G418 (1 mg/ml; Gibco) by the procedure described above. G418-resistant clones were isolated and propagated in the presence of G418 and hygromycin.

The T-cell leukemia cell line Jurkat and promyelocyte-like leukemia cell line HL-60 were maintained in RPMI medium containing 10% FCS. They were introduced with the indicated genes by the same procedure as outlined above.

Mutant plasmids of PKCs and Ras/Rho family GTPases.

The constitutively active PKCα, -βI, and -βII were made by replacing glutamic acid with alanine at position 25 (PKCαE25, PKCβIE25, and PKCβIIE25) as described elsewhere (36, 41). PCR was used to insert a Myc epitope tag (EQKLISEEDL) at the carboxyl terminus of PKCαE25, PKCβIE25, or PKCβIIE25. The mutations were verified by sequencing both strands of the plasmids. Similar mutants of PKCδ and PKCζ that have a glutamic acid replacement in the pseudosubstrate sequences (PKCδE144/145 and PKCζE119) were obtained from S. Ohno (Yokohama City University) and P. J. Parker (Imperial Cancer Research Fund, United Kingdom). Constitutively active H-Ras, R-Ras, Rap1, Rac, and Rho mutants were produced from their cDNAs by a single point replacement of glycine with valine at positions 12 (H-RasV12, RacV12, and Rap1V12), 38 (R-RasV38), and 14 (RhoV14). Constitutively active RalA has a single point mutation created by replacement of glutamic acid with leucine at position 72 (RalL72) and was provided by H. Koide (Tokyo Institute of Technology University). A dominant negative form of Rap1 has a point mutation created by substitution of serine for asparagine at position 17 (Rap1N17). Epitope tags were introduced at the amino-terminal end of the mutant small GTPases (Myc epitope tag for RacV12, R-RasV38, and RhoV14; T7 epitope tag for Rap1). Membrane-targeted PI 3-kinase catalytic p110α subunit (p110-CAAX), PDK-1, and Akt/PKB were obtained from J. Downward (Ludwig Institute for Cancer Research, United Kingdom), K. E. Anderson (The Babraham Institute, United Kingdom), and P. N. Tsichlis (Fox Chase Cancer Center), respectively. All constructs were subcloned in pcDNA3 (Invitrogen, Carlsbad, Calif.) for transfection.

Antibodies.

Anti-Myc epitope monoclonal antibody (MAb) 9E10, rabbit polyclonal anti-PKCδ and -ζ antibody (Santa Cruz Biotechnology, Inc., Santa Cruz, Calif.), anti-T7 epitope MAb (Novagen, Madison, Wis.), rabbit polyclonal anti-H-Ras and anti-RalA antibodies (Transduction Laboratory, Lexington, Ky.), antihemagglutinin (anti-HA) MAb 12CA5 (Boehringer Mannheim, Indianapolis, Ind.), anti-phospho-Akt antibody (New England Biolabs Inc., Beverly, Mass.), and peroxidase-linked anti-mouse or rabbit antibody (Amersham Co., Arlington Height, Ill.) were used for immunoprecipitation and immunoblotting as described below. To identify the integrins involved in cell aggregation, the inhibitory monoclonal anti-human LFA-1 (TS1/22), anti-mouse ICAM-1 (KAT-1) (63), anti-mouse LFA-1 (FD441.8 and KBA2), anti-mouse VLA-4 (PS/2) (37), and anti-mouse VCAM-1 (MVCAM-A,429) (Pharmingen, San Diego, Calif.) antibodies were used. Activating MAb against human CD3ɛ (OKT3) was kindly provided by T. Katagiri (National Institute of Neuroscience).

Immunoprecipitation.

BAF cells (5 × 10⁶) were collected by centrifugation and suspended at 0°C for 30 min with 1 ml of lysis buffer (1% Triton X-100, 10 mM Tris-HCl, 150 mM NaCl, 2 mM EDTA, 2 μg of aprotinin per ml, 2 μg of leupeptin per ml, 1 mM phenylmethylsulfonyl fluoride [pH 7.6]) (24). The supernatant was incubated with the indicated antibodies and protein G-Sepharose 4B (Pharmacia Biotech AB, Uppsala, Sweden). The immunoprecipitates were washed with lysis buffer three times and subjected to immunoblotting.

Immunoblotting.

Cell lysates (10 μl per 5 × 10⁴ cells) or immunoprecipitated proteins were subjected to electrophoresis on a sodium dodecyl sulfate 9 or 12% polyacrylamide gel (SDS-PAGE). Proteins were transferred to a polyvinylidene difluoride membrane (Pharmacia Biotech) and blotted with peroxidase-conjugated antibody as described elsewhere (24). The Amersham ECL (enhanced chemiluminescence) system was applied for detection.

Coating plates with ICAM-1.

ICAM-1 was purified by immunoaffinity chromatography using anti-human ICAM-1 antibody (RR1/1)-conjugated Sepharose 4B from cell lysates prepared from 10⁹ JY cells as described before (24). Each well of a polystyrene microtiter plate (96-well plate; Linbro-Flow, Chantilly, Va.) was incubated with 2 μg of the purified ICAM-1 for 90 min at room temperature and then further incubated in phosphate-buffered saline containing 1% bovine serum albumin (BSA) for 30 min at room temperature (24). The amount of ICAM-1 used for coating was chosen to give the maximum cell binding. In this coating condition, site density was about 1,200 sites/mm² quantified using ¹²⁵I-labeled RR1/1 (2 μCi/mg) at a final concentration of 20 μg/ml (24).

Cell adhesion assays.

Assays of adhesion using ICAM-1-coating plates were performed as described elsewhere (24). Cells were labeled with 2′,7′-bis-(2-carboxyethyl)-5 (and -6) carboxyfluorescein (Molecular Probes, Inc., Eugene, Oreg.) and suspended with RPMI 1640 containing 10 mM HEPES (pH 7.4) and 5% FCS. For inhibition, coated wells or labeled cells were incubated with 20 μg of anti-human ICAM-1 antibody (RR1/1) or anti-human LFA-1 antibody (TS1/22) per ml for 30 min at room temperature before the assay. Cells were transferred into coated wells at 5 × 10⁴/well and then incubated at 37°C for 30 min. Nonadherent cells were removed with four 21-gauge needle aspirations. Input and bound cells were quantitated in the 96-well plate using a fluorescence concentration analyzer (IDEXX Corp., Westbrook, Maine). For the experiment with wortmannin (Wako Pure Chemical Ltd., Tokyo, Japan), cells were treated with wortmannin at room temperature for 10 min as described (46).

Flow cytometric analysis.

Monoclonal anti-human LFA-1 (TS1/22), anti-activation-dependent epitope of LFA-1 (NKI-L16), anti-human CD18 (TS1/18), anti-mouse ICAM-1 (KAT-1), anti-mouse LFA-1 (FD441.8), and anti-CD3 (OKT3) were used for flow cytometric analysis. Hanks' balanced salt solution containing 3% FBS, 0.1% sodium azide, and 10 mM HEPES (pH 7.4) was used as the staining medium. Cells (10⁶) were incubated with staining buffer containing 10 μg of antibodies per ml on ice for 30 min. They were then washed twice with staining buffer, further incubated with 1 μg/ml of fluorescein isothiocyanate (FITC)-conjugated goat anti-mouse immunoglobulin G (IgG) F(ab′)₂ fragments (Cappel, Durham, N.C.), and subjected to flow cytometric analysis with a FACScan (Becton Dickinson, San Jose, Calif.).

Measurement of soluble ICAM-1 binding.

A fusion protein (D1D2-IgG) consisting of the first and second domains of human ICAM-1 fused to the Fc fragment of human IgG1 was kindly provided by T. Takashi (Daiichi Pharmaceutical, Tokyo, Japan) (65, 66). Measurement of the binding of D1D2-IgG to BAF cells was performed as described elsewhere (61). BAF cells were suspended in 50 μl of RPMI 1640 containing 10 mM HEPES (pH 7.4) and 5% FCS and incubated at 2 × 10⁵ cells per 50 μl with D1D2-IgG (1 mg/ml). After 30 min of incubation at 37°C, cells were washed twice and then incubated with FITC-conjugated goat anti-human IgG Fc-specific antibody (10 μg/ml; Cappel, Durham, N.C.) for 20 min on ice. Unbound secondary antibody was removed by washing twice, and then fluorescence of live cells detected using a FACScan flow cytometer.

Establishment of activation-dependent adhesion to ICAM-1 through LFA-1.

To analyze the avidity modulation of LFA-1, we tried to establish a system in which cells can adhere to ICAM-1 via LFA-1 in an activation-dependent manner, by introducing human αL and β2 in BA/F3, a nonadherent IL-3-dependent mouse pro-B-cell line. We used this cell line because it was nontransformed with a high transfection efficiency. We isolated the stable transfectants with various expression levels of LFA-1 and examined ICAM-1 binding with or without PMA stimulation. The results from the representative clones are shown (Fig. 1). The clone expressing relatively low levels of LFA-1 (clone 17) did not adhere to ICAM-1 even when it was stimulated with PMA for 30 min (Fig. 1). The clone expressing intermediate levels of LFA-1 (clone 14) did not adhere to ICAM-1 by itself but became adherent upon stimulation with PMA (Fig. 1). PMA stimulation did not change the expression levels of LFA-1 (data not shown). The clone expressing relatively high levels of LFA-1 (clone 9) constitutively adhered to ICAM-1 (Fig. 1). Adhesion to ICAM-1 was inhibited completely by anti-human LFA-1 or anti-human ICAM-1 MAbs, confirming a specific LFA-1/ICAM-1-mediated adhesion in the reconstituted system (Fig. 1B). These results suggest that expression levels of LFA-1 in leukocytes are strictly regulated within narrow limits to allow ICAM-1 binding in an activation-dependent manner. To investigate LFA-1 activation signals, we chose clone 14 (BAF/hLFA-1), which showed activation-dependent adhesion, and transfected it with mutationally active signaling molecules.

Active cPKC and nPKC, but not aPKC, increased LFA-1 avidity for ICAM-1.

We first investigated the adhesion-stimulatory activities of PKCs which were representative of each isotype, including PKCα, -βI, and -βII for cPKC, PKCδ for nPKC, and PKCζ for aPKC, because they are major isotypes of PKC expressed in leukocytes (20). A mutation in the pseudosubstrate region of PKCs can transform the inactive PKC to the constitutively active form (41). The constitutively active mutants of PKCα (PKCαE25), PKCβI (PKCβIE25), PKCβII (PKCβIIE25), nPKCδ (PKCδE144/45), and PKCζ (PKCζE119) were introduced into BAF/hLFA-1. Several stable clones expressing different amounts of each active PKC isotype were isolated and subjected to adhesion assays. The BAF/hLFA-1 clones expressing either active PKCα, PKCβI, PKCβII, or PKCδ constitutively adhered to ICAM-1; the adhesion levels were comparable to those for the ICAM-1 binding of parental cells stimulated with PMA (Fig. 2). The levels of binding to ICAM-1 correlated with the levels of expression of introduced PKCs (data not shown). The expression levels of LFA-1 in these transfectants were identical to that of parental BAF/hLFA-1 (see Fig. 6), and PMA stimulation did not alter the LFA-1 expression (data not shown). PMA stimulation of the transfectants further enhanced adhesion levels (Fig. 2). This could be due to further activation of the transfected and endogenous PKCs by PMA. In contrast, the enforced expression of constitutively active PKCζE119 did not increase the adhesiveness of LFA-1 to ICAM-1 in the absence of PMA (Fig. 2). As previously reported (72), forced expression of activated PKCζ caused expression of proteins with molecular masses of 60, 52, and 30 kDa in addition to intact 80 kDa, which were the carboxyl-terminal fragments generated by proteolytic degradation (Fig. 2B). These results indicated that the conventional PKCα, -βI, and -βII and novel PKCδ, but not atypical PKCζ itself, were all able to activate LFA-1 binding to ICAM-1.

Membrane-targeted PI 3-kinase activated LFA-1, which was not dependent on PDK-1 and Akt/PBK.

Next, to examine the modulation of LFA-1 avidity by PI 3-kinase, we introduced a constitutively active p110 subunit from PI 3-kinase (p110-CAAX). p110-CAAX alone was found to induce LFA-1/ICAM-1-mediated adhesion in an expression level-dependent fashion (Fig. 3A).

Recently, it was demonstrated that the sequential activation of PDK-1 and Akt-1 by PI 3-kinase was critical for the antiapoptotic effect of growth factors (11). Membrane localization of PDK-1 was reported to be sufficient to activate downstream targets of PI 3-kinase such as Akt (1, 8, 32). To investigate whether PDK-1 can substitute PI 3-kinase with regard to activation of LFA-1, a membrane-targeted PDK-1 was introduced into BAF/hLFA-1. As shown in Fig. 3B, active PDK-1 did not induce the adhesion to ICAM-1. Constitutively active Akt, Myr-Akt, also failed to induce the adhesion to ICAM-1 (Fig. 3C).

It was reported that the PI 3-kinase/Akt pathway promoted cell survival of BA/F3 cells in the absence of IL-3 (42). To confirm whether activities of PI 3-kinase, PDK-1, and Akt were elevated in these transfectants, we examined cell survival upon withdrawal of IL-3. As shown in Fig. 3D, the viable cell number of parental cells rapidly decreased in IL-3-free medium, but the survival of all these transfectants was prolonged.

These results suggest that activation of LFA-1 is mediated by an effector molecule distinct from PDK-1 and Akt-1.

Activation of LFA-1 by small GTPases of the Ras/Rho family.

The effects of small GTPases of the Ras/Rho family on the adhesiveness of LFA-1 were examined by introducing the active forms of H-Ras, R-Ras, Rap1, RalA, Rac, and Rho into BAF/hLFA-1. Several clones expressing different amounts of active small GTPases were isolated to confirm expression dependency. The BAF/hLFA-1 clones expressing RacV12, H-RasV12, and Rap1V12 constitutively adhered to ICAM-1 without stimulation (Fig. 4A). The expression levels of RacV12, H-RasV12 and Rap1V12 correlated with the adhesion levels (data not shown). In contrast, RhoV14, R-RasV38, and RalL72 failed to stimulate adhesion to ICAM-1 (Fig. 4A), although the expression levels of R-RasV38 and RhoV14 were higher than those of RacV12 (Fig. 4B).

PI 3-kinase dependency of the activation of LFA-1 by Ras/Rho family members.

Since H-Ras controls the cytoskeletal organization through Rac (52), we introduced the dominant negative RacN17 together with H-RasV12. As shown in Fig. 5A, the expression of RacN17 did not reduce the induction of LFA-1/ICAM-1-mediated adhesion by H-RasV12, suggesting that the effect of H-RasV12 is not mediated by Rac.

PI 3-kinase was reported to be a downstream effector of both Rac and H-ras (26, 54). To examine the involvement of PI 3-kinase in LFA-1 activation by H-Ras and Rac, the cells were treated with inhibitors for PI 3-kinase before the adhesion assay. LFA-1/ICAM-1-mediated adhesion in both RacV12- and H-RasV12-expressing cells was reduced to background levels following pretreatment with low doses of wortmannin or LY294002 for 10 min before the assay (Fig. 5B and data not shown). To further confirm the involvement of PI 3-kinase, we examined the phosphorylation status of Akt in these transfectants as Akt is phosphorylated upon activation of PI 3-kinase. Consistent with the above result, Akt was phosphorylated in RacV12 and H-RasV12-expressing cells as well as cells expressing p110-CAAX. Wortmannin inhibited the phosphorylation of Akt, in parallel with the LFA-1-mediated adhesion (Fig. 5C).

In contrast, the LFA-1/ICAM-1 adhesion induced by Rap1V12, and also PMA, was not affected by pretreatment with wortmannin or LY294002 (Fig. 5B and data not shown). Thus, H-RasV12 and RacV12 activated LFA-1 through PI 3-kinase, while the Rap1V12-induced activation of LFA-1 was independent of PI 3-kinase.

The above results showed that PKC, PI 3-kinase, and Rap1 upregulate the adhesiveness of LFA-1 to ICAM-1. To obtain further insights into the characteristics of the avidity modulation by these molecules, we examined whether they differ in the ability to cause a conformational change of LFA-1 and morphological changes in cells upon binding to ICAM-1.

Active PI 3-kinase and Rap1 but not PKC induced conformational changes of LFA-1.

The NKI-L16 epitope is a unique Ca²⁺-dependent activation epitope of LFA-1, which is absent on resting T cells but appears upon in vitro culturing, whose expression is correlated with the capacity of cells to aggregate through LFA-1 and ICAM-1 (71). The expression of NKI-L16 epitope is independent of ligand binding and reflects the conformational change in αL before ligand binding (4). BAF/hLFA-1 expresses low levels of the NKI-L16 epitope, and these expression levels were unchanged by the introduction of active PKC isotypes which increased the adhesiveness of LFA-1 (Fig. 6A and data not shown). In contrast, the levels of NKI-L16 epitope in BAF/hLFA-1 expressing p110-CAAX and Rap1V12 were four- to fivefold higher than in parental cells, while the expression levels of LFA-1 detected by TS1/22 MAb were comparable in all transfectants (Fig. 6A).

Next, we examined whether these signaling molecules can increase the ligand binding affinity of LFA-1 by using soluble ICAM-1 (sICAM-1) (D1D2-IgG). Mn²⁺-stimulated BAF/hLFA-1 cells were used as a positive control and increased sICAM-1 binding 10-fold (data not shown) compared to nontreated cells. As shown in Fig. 6B, introduction of the active PKC isotype failed to increase the affinity of LFA-1 binding to sICAM-1. However, p110-CAAX and Rap1V12 transfectants increased the binding to D1D2-IgG between two- and threefold compared to parental cells (Fig. 6B). These bindings were inhibited by anti-ICAM-1 antibody (data not shown). Examination of membrane distribution of LFA-1 by immunofluorescence microscopy revealed no significant difference between the parental cells and these transfectants (data not shown).

Taken together, these results demonstrated that the activation of LFA-1 by PI 3-kinase and Rap1, but not PKC, accompanied changes in the conformation and ligand binding affinity of the LFA-1 molecule, suggesting that the activation mechanisms of PI 3-kinase and Rap1 are different from those by PKC.

Rap1V12, but not active PI 3-kinase and PKC, induced cytochalasin D-sensitive cell aggregation.

Rap1V12-expressing cells were found to strongly aggregate during culture (Fig. 7A). Cell aggregation was consistently observed in all independent clones expressing Rap1V12. Cell aggregation was not found in BAF/hLFA-1 or cells expressing active PKC isotypes and p110-CAAX (Fig. 7A). The cell aggregation was completely inhibited either by anti-human LFA-1 MAb (TS1/22) or anti-mouse ICAM-1 antibody MAb (KAT-1) but not by MAbs against mouse LFA-1, VLA-4, and VCAM-1 (Fig. 7A). This indicates that cell aggregation was mediated by human LFA-1 and mouse ICAM-1. Cellular aggregation was not observed in cells expressing RacV12 and H-RasV12 (Fig. 7A). Rap1V12 did not change the expression levels of introduced human LFA-1 or endogenous mouse LFA-1 and ICAM-1 (Fig. 7B).

To determine whether the aggregation induced by Rap1V12 is a direct consequence of the marked increase of the adhesiveness of LFA-1 or caused by postadhesion events accompanying the actin-based cytoskeletal reorganization (62), we treated aggregated cells with cytochalasin D. The treatment completely inhibited the Rap1V12-induced LFA-1/ICAM-1-dependent cell aggregation, showing that the cell aggregation required polymerization of the actin cytoskeleton (Fig. 7C). In contrast, cell aggregation was not prevented with C3 exoenzyme, a specific inhibitor of Rho (Fig. 7C), suggesting that Rho is not involved in Rap1V12-induced cell aggregation. On the other hand, the treatment with cytochalasin D failed to inhibit ICAM-1 binding of the cells expressing Rap1V12 as well as active PKCδ or p110-CAAX (Fig. 7C), demonstrating that the increase in the avidity of LFA-1 does not require actin polymerization.

Taken together, these results indicate that Rap1 not only upregulates LFA-1 avidity but also promotes cytoskeletal rearrangement to induce stable cell-cell interactions and morphological changes.

Rap1 was involved in TCR-mediated LFA-1 activation and LFA-1/ICAM-1-dependent cell aggregation of PMA-stimulated HL-60 cells.

Since Rap1V12 has notable functions in stimulating cellular adhesion through LFA-1, we investigated whether Rap1 was physiologically involved in LFA-1/ICAM-1 adhesion using a dominant negative form of Rap1, Rap1N17. We examined LFA-1 activation by TCR cross-linking in Jurkat T cells and LFA-1/ICAM-1-mediated cell aggregation upon PMA-induced monocytic differentiation of HL-60 cells, because Rap1 was reported to be activated in both of these systems (references 53 and 73 and unpublished data).

As shown in Fig. 8A, the enforced expression of Rap1N17 in Jurkat T cells reduced TCR-induced LFA-1/ICAM-1 adhesion to less than half of that in TCR-stimulated parental cells. Cell surface expression levels of both the TCR-CD3 complex and LFA-1 were found to be similar in clones expressing the dominant negative Rap1 and in parental cells (Fig. 8A and data not shown). In contrast, the expression of Rap1N17 was little effect on PMA-induced LFA-1/ICAM-1 adhesion (Fig. 8A). We also examined whether the residual adhesion activity that was not inhibited by Rap1N17 was due to the insufficient expression of Rap1N17 or other signaling pathways. The high concentration of a PKC-specific inhibitor, calphostin (more than 500 nM) or wortmannin (more than 100 nM), had only marginal effects on TCR-induced LFA-1/ICAM-1 adhesion (data not shown). These results demonstrated that Rap1 played an essential role in regulating the adhesiveness of LFA-1 by TCR-mediated signaling in Jurkat T cells.

Our previous papers reported that the cell aggregation observed in PMA-stimulated HL-60 cells was completely inhibited by anti-αL or anti-ICAM-1 antibodies (23, 25). As shown in Fig. 8B, the expression of Rap1N17 in HL-60 cells prevented such LFA-1/ICAM-1-mediated cell aggregation in a dose-dependent manner. Cell surface expression of LFA-1 and ICAM-1 was almost identical in transfectants and parental cells (Fig. 8B). We also examined the effects of dominant negative forms of H-Ras, Rac, R-Ras, and Rho on the adhesion of PMA-stimulated HL-60 cells. Our previous report showed that H-RasN17 inhibited the induction of monocytic differentiation markers, including the expression of ICAM-1 (22). RacN17, RhoN19, and R-RasN43 did not affect LFA-1/ICAM-1-dependent cell aggregation in PMA-stimulated HL-60 cells (unpublished data). Thus, Rap1 was critically involved in stable LFA-1/ICAM-1-dependent cellular interactions for PMA-induced differentiation of HL-60 cells into macrophages.