The Small GTPase Rap1 Is Required for Mn2+- and Antibody-induced LFA-1- and VLA-4-mediated Cell Adhesion*

In T-lymphocytes the Ras-like small GTPase Rap1 plays an essential role in stimulus-induced inside-out activation of integrin LFA-1 (αLβ2) and VLA-4 (α4β1). Here we show that Rap1 is also involved in the direct activation of these integrins by divalent cations or activating antibodies. Inhibition of Rap1 either by Rap GTPase-activating protein (RapGAP) or the Rap1 binding domain of RalGDS abolished both Mn2+- and KIM185 (anti-LFA-1)-induced LFA-1-mediated cell adhesion to intercellular adhesion molecule 1. Mn2+- and TS2/16 (anti-VLA-4)-induced VLA-4-mediated adhesion were inhibited as well. Interestingly, both Mn2+, KIM185 and TS2/16 failed to induce elevated levels of Rap1GTP. These findings indicate that available levels of GTP-bound Rap1 are required for the direct activation of LFA-1 and VLA-4. Pharmacological inhibition studies demonstrated that both Mn2+- and KIM185-induced adhesion as well as Rap1-induced adhesion require intracellular calcium but not signaling activity of the MEK-ERK pathway. Moreover, functional calmodulin signaling was shown to be a prerequisite for Rap1-induced adhesion. From these results we conclude that in addition to stimulus-induced inside-out activation of integrins, active Rap1 is required for cell adhesion induced by direct activation of integrins LFA-1 and VLA-4. We suggest that Rap1 determines the functional availability of integrins for productive binding to integrin ligands.

Circulating lymphocytes require a dynamic and flexible regulation of their integrin-dependent adhesive properties. Although these cells circulate in blood and lymph or migrate through tissues, rapid transitions between adherent and nonadherent states are required. The leukocyte function-associated antigen 1 (LFA-1) 1 (integrin ␣ L ␤ 2 ) is a transmembrane heterodimer composed of a unique ␣ subunit (␣ L or CD11a) and a ␤ 2 subunit (CD18) that is common to a subset of leukocyte integrins. Expressed at the cell surface of resting leukocytes, LFA-1 is inactive. However, rapid conversion into an active state allows the integrin to bind to its ligands, intercellular adhesion molecules 1, 2, and 3 (ICAM-1, -2, and -3) (1)(2)(3). Relevant examples of the LFA-1/ICAM-1-mediated adhesion include leukocyte-endothelial cell interaction to direct lymphocyte homing or extravasation (4) and the establishment and strengthening of contacts between T cells and antigen-presenting cells (APC) (5,6). In addition, lymphocytes express VLA-4 (integrin ␣ 4 ␤ 1 ), which binds to vascular cell adhesion molecule 1 (VCAM-1) (7).
The binding of ligand by LFA-1 and VLA-4 is activated through inside-out signaling (8,9). Cytokines, chemokines, or other T cell surface receptors such as the antigen-specific T cell receptor or CD2 lead to the production of second messengers and subsequent regulation of these integrins. For LFA-1 these signals are also generated by stimuli that increase the intracellular Ca 2ϩ concentration or by protein kinase C upon phorbol ester treatment. Models explaining the regulation of LFA-1 activity incorporate two distinct mechanisms. The first postulates modulation of the intrinsic affinity of the integrin for its ligand, whereas the second suggests a critical role for increased integrin receptor clustering or integrin redistribution at the cell surface (avidity regulation) (8). In addition to inside-out signaling, integrins can be activated directly from the outside by divalent cations like Mn 2ϩ or Mg 2ϩ (in the presence of EGTA), which bind to the ectodomain (10,11), or after stimulation with activating monoclonal antibodies, like the ␤ 2 -specific antibody KIM185 (12) and the ␤ 1 -specific antibody TS2/ 16 (13).
Recently, the small GTPase Rap1 has been demonstrated to play a role in T cell receptor-, CD31-, and cytokine-induced adhesion mediated by LFA-1, VLA-4, and VLA-5 (14 -16). Overexpression of the active mutant RapV12 induced adhesion to immobilized ICAM-1 and VCAM-1. T cell receptor-and CD31-induced activation of integrins was abolished by the introduction of constructs that inhibit Rap1 signaling, i.e. Rap GTPase-activating protein (RapGAP), which lowers the level of Rap1GTP, the Rap1 binding domain (RBD) of RalGDS, which presumably inactivates Rap1 by binding to it, and the putative dominant negative mutant RapN17 (15). Interestingly, transgenic mice constitutively expressing Rap1aV12 in their T cell lineage have been generated (17). Expression of active Rap1A in primary T cells from these mice is sufficient to induce insideout signaling leading to ␤ 1 and ␤ 2 integrin-mediated adhesion. * 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  In these cells, active Rap1A did not modulate integrin affinity, as was measured by soluble ICAM-1 binding. However, thymocytes expressing Rap1AV12 showed integrin clustering at the surface, pointing to a role for Rap1 in integrin avidity modulation.
In the present study we show that overexpression of Rap1inhibitory signaling proteins, RapGAPs and RBD, interferes with the ability of Mn 2ϩ and KIM185 to induce LFA-1-mediated adhesion to ICAM-1 and with the ability of Mn 2ϩ and TS2/16 to induce VLA-4-mediated adhesion to VCAM-1. Both divalent cations and the activating antibodies do not induce accumulation of active Rap1GTP, consistent with the notion that they are thought to bypass the requirement for intracellular signaling events (18,19). This indicates that direct activation of integrins does not depend on further Rap1 activation. Although TPA-induced adhesion (inside-out integrin activation) may similarly not require activation of Rap1 (14), we demonstrate that it is blocked by inhibition of Rap1 signaling. Apparently, both stimulus-induced inside-out regulation of integrins and activation of integrin-mediated cell adhesion by cations and integrin-activating antibodies require Rap1. Use of pharmacological inhibitors reveals that Mn 2ϩ -, KIM185-, but also RapV12-induced LFA-1-mediated adhesion, is critically dependent upon intracellular Ca 2ϩ levels but independent of protein kinase C or MEK-ERK signaling. Furthermore, RapV12-induced LFA-1-mediated adhesion requires calmodulin. We propose that Rap1 activity, either basal or induced, regulates the functional availability of integrins for the adhesion process by or in combination with a calcium/calmodulindependent mechanism.
Cell Culture, Cell Line, and Transfection-The Jurkat T cell line JHM1 2.2 was provided by Dr. D. Cantrell (Imperial Cancer Research Fund, London, UK) with the kind permission of Dr. A. Weiss (University of California at San Francisco). Jurkat T cells were grown at 37°C in RPMI 1640 (Invitrogen) supplemented with 10% heat-inactivated (30 min at 56°C) fetal bovine serum and 0.05% glutamine in the presence of penicillin and streptomycin. The erythroleukemic K562 cells, either wild type or stably expressing LFA-1, were grown in 75% RPMI 1640 and 25% Iscove's modified Dulbecco's medium (Invitrogen) supplemented with 10% heat-inactivated fetal bovine serum and 0.05% glutamine in the presence of penicillin and streptomycin. The stable transfectants were cultured in the presence of 2 mg/ml Geneticin as has been described (22). Cells were transiently transfected by electroporation using 35 g of plasmid DNA. Cells (1.2 ϫ 10 7 cells/ml in 0.4 ml of complete medium) were pulsed at 250 V and 960 microfarads with 5 g of TK-luciferase plasmid DNA, construct plasmid as indicated in the figure legends, and added vector plasmid to keep DNA amounts constant. Subsequently, 24 h after transfection, cells were transferred to serum-free medium and used 42-48 h after transfection.
Adhesion Assay-For adhesion assays, transiently transfected Jurkat cells were harvested, washed, and resuspended in TSM buffer (20 mM Tris-HCl, pH 8.0, 150 mM NaCl, 1 mM CaCl 2 , 2 mM MgCl 2 ) at a concentration of 5 ϫ 10 5 cells/ml. 96-Well Nunc Maxisorp plates were coated overnight at 4°C with goat anti-human IgG antibodies (Jackson ImmunoResearch, West Grove, PA; 4 g/ml) in sodium bicarbonate buffer (Sigma), washed, and blocked for 30 min at 37°C with 1% bovine serum albumin, TSM followed by incubation for 1 h at 37°C with 50 ng/ml recombinant ICAM-1 or 2 g/ml VCAM-1 human IgG Fc fusion proteins, depending on the integrin studied. For studies on cell adhesion to fibronectin (Sigma), 96-well Nunc Maxisorp plates were coated overnight at 4°C with 1-5 g/ml fibronectin in sodium bicarbonate buffer, washed with TSM, and blocked for 30 min at 37°C with 2% bovine serum albumin, TSM. Poly-L-Lysine (Sigma) was coated on plates for 1 h at room temperature as a 0.1% w/v solution in water. After washing, 50 l of TSM was added per well with or without the indicated stimuli. Subsequently, 50 l cell suspension was added per well, after which the cells were spun down for 1 min at 200 rpm in a Heraeus Sepatech Megafuge 1.0. Cells were allowed to adhere for 30 min at 37°C, and non-adherent cells were removed with warmed 0.5% bovine serum albumin, TSM. Adherent cells were lysed and subjected to a luciferase assay as described previously (23). Expression of transfected constructs was confirmed by immunoblotting of total cell lysates. Cells bound were calculated, and numbers were corrected for transfection efficiency and nonspecific effects of constructs by measuring luciferase activity of total input cells ((counts in cells bound/counts in total input cells) ϫ 100%). Activating anti-integrin ␤ 1 and ␤ 2 antibodies TS2/16 and KIM185, respectively, have been described previously (12,13) and were used at 10 g/ml. Mn 2ϩ was used at 4 mM, and TPA (Sigma) was used at 100 ng/ml. Preincubations with the following pharmacological inhibitors were performed for 30 min at 37°C: Roche 31-8220 (Calbiochem; Analysis of Rap1 Activation in Vivo-Jurkat cells were serumstarved overnight and resuspended at 25 ϫ 10 6 cells/ml in RPMI without serum. 200 l of this suspension was used per sample. After transfer to Eppendorf tubes, cells were left untreated for 15 min at 37°C. Next Mn 2ϩ (4 mM), KIM185 (10 g/ml), TS2/16 (10 g/ml), or TPA (100 ng/ml) was added for indicated periods of time. Subsequently cells were lysed for 15 min at 4°C by the addition of ice-cold lysis buffer (10% glycerol, 1% Nonidet P40, 50 mM Tris-HCl, pH 7.4, 200 mM NaCl, 2.5 mM MgCl 2 , 1 M leupeptin, 0.1 M aprotinin), and lysates were cleared by centrifugation at maximal speed in an Eppendorf centrifuge for 15 min at 4°C. The GTP-bound form of Rap1 was isolated using RalGDS-RBD as an activation-specific probe and subsequently quantified by Western blotting using anti-Rap1 antibody, as has been previously described (24,25).

Inhibition of Rap1 Abolishes Divalent Cation-induced Adhesion to ICAM-1 and VCAM-1-
To determine whether Rap1 is involved in Mn 2ϩ -dependent activation of LFA-1 we measured the effect of Rap1 inhibitory proteins on Mn 2ϩ -induced adhesion of Jurkat cells to ICAM-1, the counter receptor of LFA-1. Cells were transfected with RapGAP I, RapGAP II (which both lower the level of Rap1GTP), or RalGDS-RBD (which can form an inactive complex with Rap1GTP) together with a luciferase construct to detect transfected cells. These cells were subsequently plated on ICAM-1 in the presence or absence of Mn 2ϩ , and after 30 min nonadherent cells were washed away and the percentage of luciferase-positive cells attached was measured. In control cells transfected with empty vector, Mn 2ϩ treatment strongly induced adhesion to ICAM-1 (Fig. 1a). In contrast, adhesion induced by Mn 2ϩ was blocked in cells transfected with RapGAP I, RapGAP II (Fig. 1a), or RalGDS-RBD (Fig. 1b). The ability of Rap1 inhibitory proteins to block divalent cationinduced integrin adhesion was not limited to LFA-1, as Rap-GAP I and RalGDS-RBD also blocked Mn 2ϩ -induced adhesion of VLA-4 to VCAM-1 (Fig. 1c). Furthermore, Mg 2ϩ /EGTA-induced adhesion to ICAM-1 was also blocked by RapGAP and RalGDS-RBD (data not shown).

RapGAP Proteins Act on Active Rap1-induced Adhesion-To
confirm the specificity of RapGAP proteins in blocking Rap1dependent adhesion, Jurkat cells were transfected with active RapV12 or the Rap1 guanine-nucleotide exchange factors (GEFs) C3G or PDZ-GEF to activate endogenous Rap1 either alone or in combination with RapGAP II (Fig. 2a). As in our previous studies, RapV12 strongly induced adhesion to ICAM-1, whereas overexpression of C3G or PDZ-GEF resulted in a more modest but significant induction of adhesion. Cotransfection of cells with RapGAP II, however, completely blocked adhesion induced by RapV12, C3G, or PDZ-GEF, indicating that RapGAP II inhibited Rap1-dependent signaling. Precipitation of GTP-bound transfected Rap1 with GST-Ral-GDS-RBD fusion protein (Fig. 2b) confirmed that transfected RapGAP II catalyzed hydrolysis of GTP of both wild-type Rap1 and RapV12. Together these results show that RapGAP overexpression blocks Rap1-induced adhesion by decreasing the Rap1GTP amount. We conclude that Rap1GTP is required for Inhibition of Rap1 Blocks Adhesion Induced by ␤ 1 and ␤ 2 Integrin-activating Antibodies-Integrin-mediated cell adhesion induced by Mn 2ϩ is proposed to occur by direct interactions of these cations with the extracellular domains of integrins rather than via intracellular signaling pathways. However, Mn 2ϩ might also influence other cell surface or intracellular signaling proteins. Therefore, we tested the effect of Rap1 on cell adhesion induced by the monoclonal antibody KIM185. This antibody directly and specifically activates LFA-1 integrin in its ligand binding capacity (12). Treatment of Jurkat cells with KIM185 strongly induced adhesion to ICAM-1 (Fig. 3a). Similar to Mn 2ϩ , cell adhesion induced by KIM185 was completely blocked by transfection of either Rap-GAP I or RalGDS-RBD. VLA-4-mediated adhesion to VCAM-1 induced by the activating anti-␤ 1 integrin antibody TS2/16 was also completely blocked by transfection of cells with either of two RapGAPs, RapGAP I or SpaI (Fig. 3b). The inhibitory effect of RapGAP on Mn 2ϩ -and KIM185-induced adhesion to ICAM-1 was also observed in the erythroleukemic cell line K562 stably expressing LFA-1 (22) (Fig. 3c). In wild-type K562 cells, which only express VLA-5 (␣ 5 ␤ 1 ) as a ␤ 1 class of integrin (26), Rap-GAP II abolished TS2/16-induced adhesion to fibronectin (Fig.  3d). Our findings demonstrate that Rap1 is required for adhesion induced by integrin-activating antibodies.
Blocking Rap1 Specifically Affects Integrin/Ligand-mediated Adhesion-To show that the Rap1 requirement is specific for integrin-ligand binding, we tested cell adhesion to poly-Llysine. No difference in cell attachment was observed upon expression of Rap1V12 or RapGAP either in the absence or presence of TS2/16 (Fig. 4a). We also used another approach to demonstrate that RapGAP inhibits integrin-ligand-based cell binding. Increasing the amount of immobilized fibronectin concomitantly increased basal cell binding, which could be abolished by expression of RapGAP (Fig. 4b). Interestingly, we observed that higher fibronectin densities rescued the inhibitory effect of RapGAP on TS2/16-induced adhesion both with K562 and Jurkat cells. Mn 2ϩ -induced adhesion could still be inhibited by RapGAP II at the highest fibronectin concentration used (5 g/ml; data not shown). From these results we conclude that blocking Rap1 specifically inhibits adhesion me- diated by integrin-ligand binding, but that integrins and processes required for adhesion are still functional.
Mn 2ϩ -/KIM185 Treatment Does Not Activate Rap1-The requirement for Rap1 in Mn 2ϩ and antibody-induced integrinmediated cell adhesion suggested the possibility that both stimuli mediated adhesion via activation of Rap1. We therefore tested whether Mn 2ϩ , KIM185, or TS2/16 was able to induce Rap1 activation (Fig. 5). Cells were stimulated for 30 min, corresponding to the duration of the adhesion assay, and Rap1GTP was analyzed using RalGDS-RBD as the activationspecific probe. No activation of Rap1 was observed with Mn 2ϩ (Fig. 5a). In contrast, strong activation of Rap1 was observed after treatment with TPA. Surprisingly, Mn 2ϩ treatment did induce a modest and sustained activation of the MEK/ERK signaling pathway, as indicated by phosphorylation of ERK1 (p44) and ERK2 (p42) (Fig. 5a). Treatment of cells with KIM185 (Fig. 5b) or TS2/16 (Fig. 5c) neither increased the Rap1GTP level nor induced ERK activation. Thus, although both divalent cations and activating anti-integrin antibodies stimulate adhesion in a Rap1-dependent manner, this is not a result of acute activation of Rap1.
Inhibition of TPA-induced Adhesion by Rap1-interfering Proteins-The phorbol ester TPA is generally used to induce integrin-mediated cell adhesion by a still elusive pathway. Therefore, we investigated the possible requirement for Rap1 in TPA-induced integrin-mediated cell adhesion as well. Treatment of Jurkat cells with TPA resulted in a rapid and prolonged activation of Rap1 as well as induction of ERK phosphorylation (Fig. 6a). Pretreatment of Jurkat cells with pharmacological inhibitors of protein kinase C (Roche 31-8220 and GF 109203X) inhibited TPA-induced activation of both Rap1 and ERK (Fig. 6b). Although previous studies provide evidence that excludes the requirement for Rap1 activation in TPA-induced adhesion (14), overexpression of RapGAP I or RalGDS-RBD blocked TPA-induced adhesion to ICAM-1 (Fig.  6c). Pharmacological inhibition of protein kinase C selectively blocked TPA-but not Mn 2ϩ -or KIM185-induced adhesion to ICAM-1 (Fig. 7a). These findings show that TPA-induced inside-out activation of adhesion requires Rap1 activity, either basal or induced.

Rap1-induced Adhesion Requires Intracellular Calcium and
Calmodulin-To study the mechanism by which Rap1 regulates integrins, we tried to identify common elements between RapV12-and Mn 2ϩ -or activating antibody-induced cell adhesion. Several reports provide strong evidence that agonist-induced adhesion in a number of hematopoietic cell types requires influx of calcium from extracellular sources (19,27,28). Indeed, the calcium chelator BAPTA-AM strongly inhibited basal and Mn 2ϩ -, TPA-, and Kim185-induced LFA-mediated adhesion (Fig. 7a). We therefore examined next whether RapV12-induced adhesion also displayed a similar requirement. Chelation of intracellular calcium with BAPTA-AM strongly inhibited both basal and RapV12-induced adhesion (Fig. 7b). Preincubation of RapV12-transfected Jurkat cells with the imidazole compound SK&F 96365, which inhibits calcium channels in leukocytes and, thus, lowers intracellular calcium, efficiently blocked basal and RapV12-induced adhesion in a dose-dependent manner (Fig. 7c). The calmodulin inhibitors W-7, E6 berbamine, and calmidazolium chloride each blocked RapV12-induced adhesion (Fig. 7d). From these results we conclude that RapV12-induced adhesion, like adhesion induced by other stimuli, requires calcium signaling.
The MEK-ERK Pathway Is Not Involved in Rap1-mediated Cell Adhesion-We were particularly interested in a possible role of the MEK-ERK pathway in Rap1-mediated cell adhesion, since positive and negative regulation of this pathway by Rap1 is a predominant effect of Rap1 reported in T lymphocytes and other cell types (29 -31). To measure whether the MEK-ERK pathway mediates Rap1-induced adhesion, we used the MEK inhibitors PD 98059 and U0126. Although both MEK inhibitors completely inhibited TPA-induced ERK activation (data not shown), they did not affect RapV12-or Mn 2ϩ -induced adhesion (Fig. 8). Alternatively, Rap1 may mediate cell adhesion by inhibiting ERK activation, suggesting that RapGAP might inhibit cell adhesion by preventing the inactivation of ERK. However, in the presence of the two MEK inhibitors, RapGAP II still inhibited cell adhesion. From these results we conclude that MEK and, therefore most likely ERK, is not involved in Rap1-mediated cell adhesion. This is confirmed by recent findings that active Rap1 did not interfere with the Ras signaling pathway (17). By utilizing the phosphatidylinositol 3-kinase inhibitors wortmannin and LY294002 and the phospholipase C inhibitor U73122, we did not find a role for phosphatidylinositol 3-kinase or phospholipase C in either RapV12-or Mn 2ϩinduced adhesion (data not shown). DISCUSSION Previously, we and others have shown that the small GTPase Rap1 is required for cytokine-induced inside-out activation of various integrins, including LFA-1, VLA-4, and ␣ M ␤ 2 (14,15,32,33). In this manuscript we now demonstrate that also direct activation of LFA-1-, VLA-4-, and VLA-5-mediated adhesion by Mn 2ϩ -or integrin-activating antibodies (monoclonal KIM185 for LFA-1 and TS2/16 for VLA-4 and VLA-5) is sensitive to inhibition of Rap1 signaling. This is shown by overexpression of either Rap-specific GAPs, which interfere with Rap1 signaling by reducing levels of endogenous GTP-bound Rap1, or overexpression of RalGDS-RBD, which inhibits Rap1 signaling by binding to GTP-bound Rap1 and sequestering it from potential effector proteins in Jurkat and K562 (Figs. 1 and 3). As an assay to monitor enhanced integrin function, we measured binding to immobilized integrin ligands, i.e. ICAM-1 for LFA-1, VCAM-1 for VLA-4 and fibronectin for VLA-4 and VLA-5. This unexpected finding raised the question of whether the inhibition is due to an integrin-specific effect or whether Rap1 affects a general requirement for integrin-mediated cell adhesion, like cytoskeletal rearrangements and cell spreading. However, inhibition of Rap1 signaling did not affect integrin-independent adhesion to poly-L-Lysine (Fig. 4a). In addition, inhibition of TS2/16-induced adhesion by blocking Rap1 signaling could be rescued by increasing the fibronectin concentration (Fig. 4b).
Basal Rap1 signaling may be sufficient to play a role in integrin-mediated adhesion, since we have not observed an increase in Rap1GTP upon treatment with either Mn 2ϩ or activating antibodies (Fig. 5). This may imply that Rap1 only provides a favorable setting for enhanced integrin function. In agreement with a facilitating function for Rap1 are the recent findings that overexpressing Rap1bV12 in megakaryocytes itself had no effect, but augmented fibrinogen binding to ␣ IIb ␤ 3 induced by a PAR4 thrombin receptor agonist (34). In addition, we have not observed increased adhesion of human megakaryoblasts overexpressing active Rap1 to immobilized fibrinogen, whereas we have found inhibition of TPA-induced adhesion by RapGAP. 2 However, RapV12 as well as Rap1-specific GEFs that activate endogenous Rap1 are able to induce increased integrin-mediated cell adhesion (Fig. 2), showing that activation of Rap1 is sufficient to enhance the function of LFA-1.
We have also investigated the involvement of putative Rap1 effectors in the regulation of integrins. Although regulation of the MEK-ERK pathway is reported to be a predominant effect of Rap1 in T lymphocytes and other cell types (29 -31), MEK inhibitors neither blocked RapV12-nor Mn 2ϩ -induced adhesion (Fig. 8). Interestingly, this issue was recently analyzed in an in vivo model system (17). In agreement with our results, transgenic mice constitutively expressing Rap1aV12 in their T cell lineage did not demonstrate a modulating role for Rap1 in the MEK-ERK pathway. Furthermore, inhibitors of the proposed Rap1 targets phospholipase C and phosphatidylinositol 3-kinase did not interfere with RapV12-or Mn 2ϩ -induced adhesion. However, pretreatment of cells with the intracellular calcium chelator BAPTA-AM inhibited both basal adhesion and RapV12-, Mn 2ϩ -, or integrin-activating antibody-induced adhesion to a similar extent (Fig. 7). Utilization of the SK&F 96365 inhibitor, which blocks extracellular calcium channels required for leukocyte adhesion, also blocked RapV12-induced adhesion. This result is compatible with recent evidence that extracellular stimuli inducing T lymphocyte adhesion via LFA-1, such as T cell receptor stimulation and phorbol esters, do so by inducing an influx of extracellular calcium (28). Although in several studies Rap1 has been implicated in the regulation of cellular calcium levels (35)(36)(37), clear evidence is still lacking. A recent study demonstrated that Rap2b via interaction with phospholipase C⑀ is able to regulate intracellular calcium signaling in HEK293 cells (38). A large number of calcium-responsive signaling proteins have been shown to regulate integrin-dependent adhesion in various cell types, like calpain (28,39,40), calmodulin (41), and calcium/calmodulin-dependent kinase II (42)(43)(44). In the case of RapV12-induced adhesion, calmodulin is required (Fig. 7d). However, we could not demonstrate a role for calpain, calcium/calmodulin-dependent kinase II, or cal- cineurin in RapV12-induced integrin function (Fig. 7). The function of this calmodulin requirement is therefore still unclear.
Our results indicate that all stimuli that enhance the function of integrins tested, including inside-out signaling and direct activation, require Rap1. Previously, it was reported that the phorbol ester TPA may induce integrin activation independent of Rap1 activation (14). However, we have found that overexpression of RapGAPs or RalGDS-RBD efficiently blocked TPA-induced adhesion, indicating that Rap1-GTP is also required for TPA-induced integrin activation. How Rap1 enhances integrin function is still elusive. Cell surface expression of integrins is not affected by Rap1 signaling interference (14,15). This indicates that Rap1 is involved in the increase in integrin function either by inducing a conformational change or by inducing clustering of integrins, two nonexclusive modes of integrin activation (8,9,45). Importantly, the effect of Rap1 on integrin function occurs with both ␤ 1 and ␤ 2 integrins. The most plausible explanation for the role of Rap1 is regulation of auxiliary factors for integrin function. These factors may be signaling or structural proteins that associate with integrin cytoplasmic tails (46). Matsuda and co-workers utilize fluorescent resonance energy transfer analysis to visualize Rap1 activation in vivo and note that growth factor-induced Rap1 activation initiates in the perinuclear region (47). Perhaps Rap1 influences integrin processing, the repertoire of integrin-associated proteins that is delivered to surface, or modulates membrane microdomain organization (lipid rafts) (48,49) influencing integrin function (50).