Immunophilins Control T Lymphocyte Adhesion and Migration by Regulating CrkII Binding to C3G

Members of the Crk family of adaptor proteins play important roles in signal transduction from numerous cell surface receptors (1). Adaptor proteins can orchestrate the assembly of multimolecular complexes at receptor sites, thereby linking receptors to their downstream signaling cascades (2). In T lymphocytes, Crk proteins were found to associate with the ZAP70 protein tyrosine kinase in a cell activation-dependent manner (3, 4). A ZAP70-associated CrkL was found to interact with the Wiskott–Aldrich syndrome protein (WASP)/WASP-interacting protein (WIP) complex and induce its recruitment to the immunological synapse (5). Concomitant recruitment of protein kinase Cθ to the immunological synapse (6, 7) results in phosphorylation of WIP and disengagement of WASP from the WIP–WASP complex, thereby releasing it from WIP inhibition and enabling the activation of cellular events leading to actin polymerization and cytoskeletal rearrangement (5).

All Crk adaptor proteins possess a single N-terminal Src homology 2 (SH2) domain, which binds phosphotyrosine-containing sequences, followed by one (CrkI) or two (CrkII and CrkL) Src homology 3 (SH3) domains, which bind proline-rich motifs (8, 9). The two SH3 domains are connected via a linker region containing a single, highly conserved, protein conformation-regulating tyrosine residue, plus several proline residues (10, 11). Phosphorylation of the tyrosine residue (Tyr²¹¹ and Tyr²⁰⁷ in human CrkII and CrkL, respectively) by Abl, or by other protein tyrosine kinases, promotes intramolecular binding of the linker region to the self SH2 domain, thereby sequestering the SH2 and N-terminal SH3 (SH3N) domains and preventing them from interaction with physiological binding partners (11, 12).

Recent in vitro studies suggested an additional regulatory role for the proline-rich motif within the CrkII linker region, which is under the control of peptidyl-prolyl cis-trans isomerases (PPIases) (13–16). In these studies, a truncated recombinant protein consisting of the SH3N-linker-SH3C of the chicken CrkII (CrkII^SLS) was found to be sensitive to PPIases, and undergo cis-trans isomerization at Pro²³⁸ located within the linker region. Combined chemical shift and relaxation rate analyses revealed that CrkII^SLS exists in solution in two different geometric conformations, as follows: a major one (90%) in which Pro²³⁸ adopts a cis conformation and CrkII^SLS is folded in a closed and inactive conformation, and a second minor conformation (10%) in which Pro²³⁸ acquires a trans conformation and CrkII^SLS adopts an open, active conformation, allowing the protein to interact with its binding partners.

Despite the in vitro responsiveness of CrkII^SLS to PPIase-induced conformational changes, the direct effect of PPIases on full-length CrkII has not yet been resolved, nor has the ability of CrkII to serve as a substrate for PPIases in vivo. In addition, the sequence around Pro²³⁸ in the chicken CrkII is only partially conserved in mammalians, raising questions about the ability of PPIases to isomerase CrkII and regulate its function in human cells.

The aim of the current study is to analyze the ability of immunophilins, predominantly cyclophilin A (CypA) and 12-kDa FK506-binding protein (FKBP12), to affect CrkII activity in human T cells, using the Jurkat leukemia T cell line as a model system.

Our findings demonstrate that the in vivo activity of human CrkII, but not of CrkI, is regulated by the immunophilins, CypA and FK506-binding protein. The results are further substantiated by experiments that use cyclosporine A (CsA) and FK506, two distinct immunophilin inhibitors that are known to inhibit the calcineurin–NFAT signaling pathway. The overall findings suggest that a mixture of immunophilin inhibitors, containing the CsA and FK506, can inhibit selected effector T cell functions via a CrkII-dependent, and NFAT-independent mechanism.

Sandimmune (CsA; 50 mg/ml) was from Novartis Pharma AG (Basel, Switzerland), and FK506 (Prograf or tacrolimus, 5 mg/ml) was from Astellas Pharma. Recombinant human CypA, 2-ME, and Triton X-100 were from Sigma-Aldrich. AEBSF, aprotinin, and leupeptin were from ICN Biomedicals (Aurora, OH). ECL, glutathione-coupled Sepharose beads, and protein A–Sepharose were from Amersham Pharmacia Biotech (Uppsala, Sweden). Recombinant stromal cell-derived factor 1α (SDF1α) protein and human fibronectin were from R&D Systems (Minneapolis, MN) and Sigma-Aldrich, respectively. Trypsin (0.25%-EDTA [1:200]) was from Biological Industries (Beit Haemek, Israel). Mouse anti-Crk (I/II) mAb was from BD Transduction Laboratories (Lexington, KY), and mouse anti-phosphotyrosine mAb (4G10) was from Upstate Biotechnology (Lake Placid, NY). Mouse anti-CD3ε mAb (OKT3) was prepared in nude mice by i.p. injection of the OKT3 hybridoma and collection of the ascites fluid. Rabbit polyclonal anti-C3G Ab and mouse mAb specific to CypA were from Santa Cruz Biotechnology (Santa Cruz, CA). Mouse anti–β-actin mAb (AC-15) was from Sigma-Aldrich. Mouse anti-human Myc mAb (clone 9E102)–producing hybridoma was obtained from the American Type Culture Collection (CLR 1729), and ascites was prepared in BALB/c mice. Cy3-conjugated goat Abs directed against mouse or rabbit IgG were from Jackson ImmunoResearch Laboratories, and HRP-conjugated goat Abs directed against mouse or rabbit IgG were from Amersham Pharmacia Biotech.

Human leukemia cell lines used in this study include the Jurkat T cell, clone 6.1, and Jurkat TAg T cell, which stably expresses the SV40-derived large T Ag. In addition, Jurkat T cell subclones, JKCrkI and JKCrkII, were prepared by S. Gelkop from our laboratory by transfecting Jurkat TAg T cells with Myc-tag–containing expression vectors for CrkI and CrkII, respectively, and selection of clones that express ∼2-fold CrkI or CrkII, compared with the cell of origin. All cell lines were maintained at a logarithmic growth phase in complete RPMI (RPMI 1640 supplemented with 5% heat-inactivated FCS, 2 mM l-glutamine, 50 U/ml penicillin, 50 μg/ml streptomycin [all from Biological Industries], and 0.5 μM 2-ME (Sigma-Aldrich). Cells were grown in 75-cm² growth-area tissue culture flasks (Cellstar, Greiner, Germany) in an atmosphere of 5% CO₂, at 37°C. Stimulation by TCR triggering was carried out using anti-CD3ε mAbs (OKT3; 1:1000 dilution of ascites, 30-min incubation on ice) plus cross-linking with a secondary Ab (goat anti-mouse IgG, 1:200) for 1 min at 37°C.

pGEX-5X vectors were from Amersham Pharmacia Biotech, and pGEX plasmids encoding different GST-Crk fusion proteins were gifts of M. Matsuda (National Institutes of Health, Tokyo, Japan). The pGST-C3G and pGST-SH3b were gifts of C. Guerrero (Universidad de Salamanca-Consejo Superior de Investigaciones Cientificas, Salamanca, Spain) (17). GST-isomerase vectors were gifts of J. Bolstad and W. Chen (University of Calgary, Calgary, AB, Canada) (human [h]FKBP12 and hFKBP12.6), T. Ratajczak (University of Western Australia, Crawley, WA, Australia) (hFKBP51 and hCyp40), B. Chambraud (INSERM U488, Paris, France) (hFKBP52), M. Emerman (University of Washington, Seattle, WA) (hCypA), and J. Buchner (University of Technology, Munich, Germany) (hCyp40).

Bacterial expression vectors were transformed into Escherichia coli DH5α-competent cells, and GST-fusion proteins were prepared as described (4). Pull-down assays were performed by incubation of bead-adsorbed GST or GST fusion proteins (2–10 μg, as indicated) with cell lysates at 4°C on a rotator for 4 h. The beads were washed (three times) in lysis buffer, resuspended in 2× Laemmli sample buffer, and boiled for 5 min. The eluents were subjected to SDS-PAGE under reducing conditions, followed by immunoblotting.

Lysates of Jurkat T cells (∼10⁸ cells) were incubated with bead-immobilized GST or GST-CrkII (10 μg/sample) for 3 h at 4°C. The beads were extensively washed with lysis buffer, and bound proteins were eluted by boiling for 5 min in sample buffer and subjected to SDS-PAGE (on a 10% polyacrylamide gel) under reducing conditions. The gel was stained with GelCode Blue reagent, followed by extensive washes in double-distilled water. A gel slice including the 14 kDa up to 22-kDa protein bands was excised, reduced, alkylated, and trypsin digested. Mass spectrometry analysis of the peptides was performed on MALDI mass spectrometer (Biological Mass Spectrometry Unit, Weizmann Institute of Science, Rehovot, Israel), followed by peptide sequence comparison with protein database using the Sonar MS/MS search engine (ProteoMetrics; Genomic Solutions).

The phosphorylation indicator of Crk chimeric unit plasmid (PICCHUx, a gift of M. Matsuda, National Institutes of Health, Tokyo, Japan) consists of human CrkII (aa 1–236) sandwiched between cyan fluorescent protein (CFP)– and yellow fluorescent protein (YFP)–emitting variants of GFP, in addition to the ki-Ras–derived CAAX box, which anchors the protein predominantly to the plasma membrane (18). A cDNA plasmid encoding a constitutively active Lck was a gift of A. Altman (La Jolla Institute for Allergy and Immunology). The eukaryotic vectors pCEFL-GST-CrkII-wild-type (WT), pCEFL-GST-CrkII-R38L, pCEFL-GST-CrkII-W169L, and pCEFL-GST-CrkII-Y221F were a gift of K. Yamada (National Institute of Dental and Craniofacial Research, National Institutes of Health, Bethesda, MD) (19). For DNA transfer into cells, Jurkat TAg cells were maintained at a logarithmic growth phase, washed in serum-free RPMI 1640, resuspended at 7 × 10⁶ cells/ml in transfection buffer (1.8 g/ml sucrose, 2 nM DTT in RPMI 1640), and transferred into 0.4-cm–gap Gene Pulser cuvettes (Bio-Rad) (5 × 10⁶ cells/700 μl/cuvette). Plasmid DNA (10 μg/group, unless otherwise indicated) was added to each cuvette, and electroporation was performed using a Bio-Rad Gene Pulser (250 V, 950 μF). The cells were then cultured in 50 ml complete RPMI 1640 in 145-mm² tissue culture plates for 36 h.

Cell lysates were prepared by resuspension of cells in lysis buffer (25 mM Tris/HCl [pH 7.5], 150 mM NaCl, 5 mM EDTA, 1 mM Na₃VO₄, 50 mM NaF, 10 μg/ml each of leupeptin and aprotinin, 2 mM AEBSF, and 1% Triton X-100), followed by a 20-min incubation on ice. Lysates were centrifuged at 13,000 × g for 30 min at 4°C, and the nuclear-free supernatants were used for immunoprecipitation.

For immunoprecipitation, primary Abs were preadsorbed to protein A–Sepharose beads for 1 h at 4°C. Excess Abs were removed by three washes in cold PBS, and Ab-coated beads were incubated with cell lysates for 3 h at 4°C. Immune complexes were precipitated by centrifugation, followed by extensive washing in a lysis buffer. Equal volumes of 2× SDS sample buffer were added to immunoprecipitates or whole-cell lysates, which were vortexed, boiled for 5 min, and fractionated by SDS-PAGE.

Samples of whole-cell lysates, immunoprecipitates, or pulled-down proteins were resolved by electrophoresis on 10% polyacrylamide gels using Bio-Rad Mini-PROTEAN II cells. Proteins from the gel were electroblotted onto a nitrocellulose membrane (Schleicher & Schuell) at 100 V for 1 h in BioRad Mini Trans-Blot transfer cells. After 1-h blocking with 3% BSA in PBS at 37°C, the nitrocellulose membranes were incubated with the indicated primary Abs, followed by incubation with HRP-conjugated secondary Abs or protein A. Immunoreactive protein bands were visualized using an ECL reagent and autoradiography. Whenever required, nitrocellulose membranes were stripped by incubation in stripping buffer (100 mM 2-ME, 2% SDS, and 62.5 mM Tris/HCl [pH 6.8]) for 30 min at 50°C, followed by 1-h incubation with blocking buffer (3% BSA in PBS).

Treatment with CypA was performed according to Kofron et al. (20), with some modifications. For a pull-down assay, bead-immobilized GST-CrkII fusion proteins (10 μg) were incubated with recombinant CypA (6 mM, final concentration) in assay buffer (50 mM HEPES, 100 mM NaCl [pH 8]) for 1 h at room temperature. The beads were then washed and incubated with 280 μl Jurkat T cell lysate (equivalent to 2 × 10⁷ cells) for an additional 3 h at 4°C. After washings, bound proteins were eluted from the beads by 5-min boiling in sample buffer, followed by SDS-PAGE. For coimmunoprecipitation studies, Jurkat T cells (8 × 10⁷) were lysed in 100 μl lysis buffer diluted in assay buffer in a final volume of 250 μl. CypA (6 mM, final concentration) was added to one of the two test tubes for 1-h incubation at room temperature. Protein A beads (10 μl/group) that were preincubated with anti-Crk mAbs (1 μg/group) for 1 h at room temperature were added to each test tube for 3 h of incubation at 4°C. The beads were extensively washed, and bound proteins were eluted by 5-min boiling in sample buffer and subjected to SDS-PAGE. Proteins that were pulled down by CypA-treated GST-CrkII, or coimmunoprecipitated with Crk from lysates of CypA-treated Jurkat T cells, were electroblotted onto nitrocellulose membranes that were immunoblotted with anti-Crk and anti-C3G Abs.

CsA (Sandimmun, 50 mg/ml in oil solution) and FK506 (Prograf, 5 mg/ml in ethanol) were diluted in RPMI 1640 culture medium before each experiment. Jurkat TAg, JKCrkI, and JKCrkII cells were cultured in 75-cm² growth-area tissue culture flasks at a concentration of 50 × 10⁶ cells/group in the presence or absence of CsA (5 μg/ml) and/or FK506 (5 ng/ml), or as indicated in the figure legend, and tested on the next day. In longer treatments, the cells were incubated with the same drug concentrations, and half of the volume of the culture medium was replaced once every 2 d with fresh culture medium containing the same drug concentration.

PICCHUx-transfected Jurkat T cells were grown on coverslips preincubated with poly-l-lysine in a 24-well culture dish. Cells were fixed with 4% paraformaldehyde (Sigma-Aldrich)/PBS for 15 min at room temperature and permeabilized with 0.2% Triton X-100/PBS for 5 min. Cells were then blocked in 1% BSA/PBS for 1 h, followed by staining with rabbit anti-C3G (Santa Cruz Biotechnology) for 1 h. Subsequently, the cells were washed three times in PBS and incubated with Cy3-conjugated goat anti-rabbit Abs. The coverslips were mounted on slides using DAKO mounting medium and imaged by Olympus FluoView FV1000 laser-scanning confocal microscope.

Fluorescence resonance energy transfer (FRET) analysis of living cells by flow cytometry and confocal microscopy was performed according to He et al. (21). Briefly, cells were transfected with the PICCHUx plasmid constructed to include the human CrkII protein (aa 1–236) and YFP and CFP on both ends of CrkII, plus an N-terminal CAAX motif (18). Cells were cultured for 48 h and then split into two identical groups that were cultured for an additional 24 h in the presence or absence of CsA (5 mM) plus FK506 (5 nM). Cells were then resuspended in FACS buffer (PBS plus 0.5% BSA and 0.1% NaN₃) for analysis on a FACSCanto II device (BD Biosciences). Excitation of CFP occurred at 405 nm, whereupon emission was detected in the CFP emission window and simultaneously in the YFP emission window. If FRET occurs, CFP emission decreases, while simultaneously YFP emission increases. This can be visualized by a shift in ratio of YFP/CFP emission intensities. The data were analyzed using Flowjo7.6.3 software.

Alternatively, another group of transfected cells was similarly grown on Fluorodish for imaging by a confocal microscope (Olympus FluoView FV1000), and FRET was measured by the donor-sensitized acceptor fluorescence technique, as previously described (22, 23). Briefly, three images were acquired for each set of measurements, as follows: YFP excitation/YFP emission image (YFP channel); CFP excitation/CFP emission image (CFP channel); and CFP excitation/YFP emission image (FRET channel). Single-labeled (CFP or YFP) cells were used to calculate the non-FRET fluorescence bleed-through produced by the fluorophores into the FRET channel, and the non-FRET fluorescent intensity values were subtracted from the apparent FRET intensities obtained from the double-labeled cells under the same conditions. A set of reference images was acquired from single-labeled CFP- or YFP-expressing cells for each set of acquisition parameters, and a calibration curve was derived to allow elimination of the non-FRET components from the FRET channel. The FRET efficiency was calculated on a pixel-by-pixel basis using the following equation:

where FRET_corr is the pixel intensity in the corrected FRET image, and CFP is the intensity of the corresponding pixel in the CFP channel image.

Flat-bottom 96-well microtiter plates (Costar, Cambridge, MA) were coated overnight at 4°C with 100 μg/ml soluble human fibronectin in PBS, followed by blocking with 2% BSA/PBS for 1 h at 37°C, and then washed with medium (RPMI 1640 without additives). Cells (2 × 10⁵/100 μl) were allowed to attach for 2 h at 37°C, and nonadherent cells were then removed by gentle aspiration and rinsing with prewarmed RPMI 1640. Adherent cells were quantified, and data are expressed as the mean of the percentage of adherent cells relative to total cell input, in three replicate wells.

Cell migration was assessed in disposable ThinCert cell culture inserts (Greiner) placed in a 24-well multiwell cell culture plate in which the two compartments in each migration chamber are separated by a porous positron-emission tomography membrane with 5.0 μm pore size. One hundred microliters of cell suspension (5 × 10⁶ cells/ml) in migration buffer (RPMI 1640 containing 50 U/ml penicillin, 50 μg/ml streptomycin, 10 mM HEPES, and 5% FCS) was seeded into the upper compartment of each chamber, and 600 μl migration buffer was placed in the lower chambers. Active cell migration into the lower compartment was induced by the addition of recombinant SDF1α (100 ng/ml) chemoattractant into the lower chamber. After 3 h of incubation at 37°C and 5% CO₂, cells that migrated through the membrane to the lower chamber were counted using a hemocytometer.

Statistical analyses were performed by one-way ANOVA using SIGMA STAT 3.11 (Systat Software 2004, San Jose, CA) or t test. Values are presented as average ± SD. Data were considered significantly different at p < 0.05.

Previous studies demonstrated the ability of CypA to catalyze in vitro the conformational change of a recombinant molecule possessing the SH3N-linker-SH3C region of the chicken CrkII adaptor protein (13). In addition, pull-down and coimmunoprecipitation studies demonstrated the physical interaction of CypA with some of its substrate protein (24–26).

To test whether CrkII can serve as an in vivo substrate for CypA, we first analyzed the ability of CrkII to associate with CypA in a pull-down assay. Bead-immobilized GST-CrkII fusion proteins, or GST as a negative control, were used to pull down binding proteins from a human Jurkat T cell lysate. Proteins were then fractionated by SDS-PAGE, and the gel was stained with GelCode Blue reagent. Because human CypA possesses 165 aa and a calculated molecular mass of 18,012 Da, a gel slice that includes proteins with a molecular mass of 14–22 kDa was excised, reduced, alkylated, and trypsin digested. Proteins were then analyzed on a MALDI mass spectrometer, followed by a peptide sequence comparison with a protein database using the Sonar MS/MS search engine. Results revealed that CrkII associates with a number of proteins, including CypA, and a closely related protein termed pepityl-prolyl cis-trans isomerase-like-1 (Fig. 1).

MALDI mass spectrometry is an extremely sensitive method that can hardly distinguish between low amounts of high-affinity CrkII-binding proteins versus large amounts of low-affinity CrkII-binding proteins in a tested sample. To demonstrate CrkII–CypA interaction using a less sensitive method, we performed a pull-down assay in which bead-adsorbed CypA was used to pull down CrkII from a lysate of Jurkat T cells. Gel filtration and Western blot analysis revealed that CrkII, but not CrkI, associates with CypA (Fig. 1B, 1C). Because a single substrate might be subjected to isomerization by different isomerases, we tested whether CypA interaction with CrkII is novel, or whether other immunophilins exhibit similar capabilities. A pull-down assay using different bead-immobilized GST immunophilins revealed that, in addition to CypA, FKBP12 and Cyp40, but not FKBP12.6, FKBP51, or FKBP52, pulled down CrkII from a Jurkat T cell lysate (Fig. 1B, 1C). CrkI, which lacks the C-terminal SH3 domain, was observed in whole-cell lysates of Jurkat T cells, although in a significantly lower amount compared with CrkII. However, CrkI was undetectable in all pull-down groups, suggesting no binding to any of the immunophilins tested (compare lane 1 with all other lanes in Fig. 1B, 1C).

Structure–function studies demonstrated that the CrkII C-terminal SH3 domain (SH3C) possesses an amino acid sequence and an overall conformation of a classical SH3 domain (27). However, the canonical-binding surface of the CrkII-SH3C domain is unusually polar, suggesting that its interaction with typical proline-rich ligands is not possible or occurs at a very low affinity (27). Other studies showed a negative regulatory role for the CrkII-SH3C domain, which affects ligand binding to the CrkII-SH2 or -SH3N domains (13, 28–31). The negative regulation by the CrkII-SH3C domain may reflect an isomerase-mediated trans-to-cis conversion of CrkII at the linker region, resulting in steric hindrance that prevents ligand binding by the CrkII-SH2 and/or -SH3N domains.

To test whether binding to immunophilin alters the ability of CrkII to interact with known CrkII-SH3N binding partners in vivo, we selected the Crk SH3 domain–binding guanine-nucleotide releasing factor (C3G), a guanine nucleotide exchange factor for Rap1 (32), which is activated following binding to Crk adaptor proteins (33–35), and tested its ability to bind CrkII-SH3N under different experimental conditions. As a first step, we verified that C3G binds the CrkII–SH3N domain and not other regions within the molecule. Jurkat T cells were transfected with eukaryotic vectors encoding GST-tagged WT CrkII, CrkII mutants that abolish SH2 (R38L) or SH3N (W169L) binding to their ligands, or a CrkII mutant (Y221F) that is unable to undergo activation-dependent tyrosine phosphorylation in vivo (19). The cells were then lysed and subjected to immunoprecipitation with anti-GST Abs. Protein fractionation by SDS-PAGE and immunoblot analysis demonstrated binding of C3G to WT and mutant CrkII, but not to CrkII (W169L) that possesses an inactive mutant SH3N domain (Fig. 2).

The ability of CrkII to associate with both C3G and CypA, and serve as a putative substrate for CypA, raises the possibility that CypA-mediated isomerization of CrkII may affect CrkII binding to C3G. To test this hypothesis, we incubated bead-immobilized GST-CrkII in the presence or absence of CypA for 1 h at room temperature, and then added a Jurkat T cell lysate for additional 3 h of incubation. GST–CrkII–binding proteins were pulled down and analyzed by Western blot for the presence of C3G. The results (Fig. 3A) showed that GST–CrkII preincubation with CypA led to a ∼5-fold increase of the amount of C3G that interacted with CrkII.

We further tested whether Jurkat T cell lysate preincubation with CypA would affect the ability of the endogenous CrkII to associate with C3G. Thus, we incubated Jurkat T cell lysates with or without recombinant CypA for 1 h at room temperature, followed by incubation with anti-Crk mAb-coated protein A beads. Immunoprecipitation and Western blot analysis demonstrated that the addition of CypA to the Jurkat T cell lysate resulted in ∼3-fold increase in the amount of C3G that coimmunoprecipitates with CrkII (Fig. 3D).

The results in Fig. 3 demonstrated that the presence of CypA in the assay system increases C3G association with CrkII. This may reflect an increase in the affinity of CrkII to C3G due to physical interaction of immunophilins with CrkII. Alternatively, increased binding of C3G to CrkII may occur by immunophilin-dependent cis-to-trans isomerization of CrkII. The latter possibility is likely to be dependent on the immunophilin catalytic activity that promotes a conformational change in CrkII, thereby increasing accessibility of the CrkII-SH3N to the C3G poly-proline region.

To distinguish between these two potential mechanisms, we precultured Jurkat T cell lysates for 24 h in the presence or absence of immunophilin inhibitors, CsA and/or FK506, and pulled down binding proteins using bead-immobilized GST-C3G or GST-SH3b [a truncated C3G corresponding to aa 208–662, which possesses multiple copies of proline-rich regions (17, 33)] fusion proteins. Immunoblot analysis revealed that each of the two fusion proteins, GST-C3G and -SH3b, pulled down both CrkI and CrkII (Fig. 4A). Cell pretreatment with either CsA or FK506 had a negligible effect on the expression levels of Crk I and CrkII (see Fig. 5C). However, the amount of CrkII pulled down by each of the fusion proteins was significantly reduced following drug pretreatment (Fig. 4A). In contrast, the drugs had no effect on the amount of CrkI pulled down by GST-C3G and -SH3b. The results indicate that the effect of CsA and FK506 on CrkII binding to C3G is not due to some unknown effects of the drugs on protein synthesis, content, or activity, because the drugs had no effect on the extent of CrkI binding to C3G.

To substantiate these findings in a more physiological system, we analyzed the effect of immunophilin inhibitors on Crk-C3G interaction using coimmunoprecipitation studies. In this study, Jurkat T cells were pretreated as previously with a mixture of CsA and FK506 for 16 h, and C3G binding to Crk adaptor proteins was tested by immunoprecipitation of Crk and immunoblot with anti-C3G Abs. Because Abs that distinguish between CrkI and CrkII do not exist, we used Jurkat T cell lines that constitutively express either Myc-CrkI (JKCrkI) or Myc-CrkII (JKCrkII) at levels that resemble those of the endogenous Crk proteins (see Fig. 5D).

We found that anti-Myc mAbs coimmunoprecipitated C3G from JKCrkII cells and that cell pretreatment with CsA and FK506 reduced the amount of coimmunoprecipitating C3G by ∼2-fold. In contrast, no effect of the drugs was observed on the amount of C3G coimmunoprecipitated with Myc-CrkI from JKCrkI cells (Fig. 5A). This is despite the fact that similar levels of C3G were observed in both cell lines, and that C3G levels were not affected by the drug treatment (Fig. 5B).

Anti-Crk immunoblot demonstrated that anti-Myc mAbs immunoprecipitated only Myc-CrkI from JKCrkI cells, and only Myc-CrkII from JKCrkII cells (Fig. 5C). Furthermore, drug treatment of both cell lines did not alter the amount of Myc-CrkI or Myc-CrkII expressed in the two cell lines, respectively (Fig. 5C).

To analyze whether immunophilin inhibitors affect the association of CrkII with C3G in vivo, we used drug-treated or untreated Jurkat T cells and performed a fluorescence colocalization assay. Staining of Jurkat T cells confirmed that both CrkII (19, 36) and C3G (37, 38) localize predominantly at the cytoplasm of resting cells (data not shown). However, due to the wide distribution of CrklI and C3G all over the cytoplasm, we were unable to distinguish between CrkII-associated C3G or free C3G, as both would have shown a very similar staining pattern.

Trying a different approach, we transiently transfected Jurkat T cells with a CrkII-encoding plasmid, termed PICCHUx (18), which possesses a C-terminal CAAX box (in which C is cysteine, A is an aliphatic residue, and X is a variety of residues) of ki-Ras that allows posttranslational addition of a lipid group and thereby increases the protein’s affinity to the cell membrane. In addition, the PICCHUx chimeric construct encodes the human CrkII_1–236 sandwiched between cyan- and yellow-emitting variants of GFP (CFP and YFP, respectively). Expression of this fluorescently labeled CrkII at the plasma membrane allows the distinction between CrkII-bound membrane-associated C3G versus free C3G that is distributed in the cell’s cytoplasm.

Two days after transfection of Jurkat T cells with PICCHUx, the cells were incubated for an additional 24 h in the presence or absence of CsA plus FK506 (CsA/FK506), and then stained with C3G-specific Cy3-labeled Abs. The results demonstrated that the majority of the overexpressed CrkII protein was membrane bound (Fig. 6). A fraction of the C3G proteins in untreated cells was associated with the cell membrane in addition to a smaller amount of C3G that distributed unevenly in the cytoplasm. Cell treatment with CsA/FK506 did not affect CrkII distribution within the cell’s outer membrane, but resulted in a significant redistribution of the membrane-bound C3G back into the cell’s cytoplasm. Quantification of CFP and YFP colocalization using the ImageJ colocalization plugin JACoP (39) demonstrated that the fraction of red overlapping with green in untreated cells (Mander’s M2 coefficient of 0.394) was reduced upon cell treatment with CsA/FK506 (M2 = 0.095).

The PICCHUx plasmid was designed to serve as a phosphorylation indicator for active Abl, because Abl-mediated phosphorylation of PICCHUx (on Tyr²²¹) leads to intramolecular interaction of phosphotyrosine²²¹ with the internal SH2 domain, which results in folding of the protein and increased FRET emission (18).

We tested whether FRET activity of PICCHUx is also regulated by immunophilins, and therefore sensitive to inhibition by immunophilin inhibitors. In this experiment, PICCHUx-transfected Jurkat T cells were treated or untreated with CsA/FK506, and their FRET activity, in which CFP and YFP fluorophores served as a donor and acceptor for FRET, was monitored. The results demonstrated a significantly higher FRET efficiency in drug-treated cells compared with the control cells (Fig. 7A, 7B, control = 29.2 ± 1.8, drug treated = 45.9 ± 1.9, n = 10).

We also established a FACS-based FRET analysis method of PICCHUx-expressing Jurkat T cells, in which the transfected and drug-treated cells were analyzed using FACSCanto II. Live cells were gated according to forward and sideward scatter and adjusted photomultiplier tube voltages and compensation for CFP and YFP to specifically assess FRET in double-positive cells. Five-minute kinetics of FRET emission ratios were analyzed using FlowJo Ver. 7.6.3 software. An increase in FRET emission ratio was found in all time points of the measured kinetics in drug-treated group compared with the control (Fig. 7C, 7D). The results correlate with the confocal microscopy FRET data, supporting the assumption that CrkII is a target for immunophilins, and that inhibitors of immunophilin catalytic activity increase the proportions of cis conformers of CrkII.

Integrin-mediated cell adhesion is a critical process required for the migration of normal cells during biological processes, including embryonic development, wound healing, and immune responses (40, 41). In addition, cell migration plays an important role in various pathological conditions, such as tumor metastasis to remote tissues and organs (42, 43).

Previous studies demonstrated that independent overexpression of Crk or C3G, which form constitutive functional complexes in vivo, augments integrin-mediated adhesion of hematopoietic cells to fibronectin (44, 45). The Crk-SH3N, which mediates C3G binding, plays an important role in this process, because expression of a SH3N-defective mutant inhibited cell adhesion to fibronectin (44–46). In addition, both Crk and C3G function as focal adhesion proteins and Crk-C3G complexes recruit to the leading edge of polarized cells, where they activate additional effector molecules and provide spatially instructive cues required for cell adhesion and migration (47–50).

To test the effect of CsA/FK506 on C3G-CrkII– versus C3G-CrkI–dependent cell adhesion, we used Jurkat T cells that overexpress either CrkII or CrkI and analyzed their ability to adhere to fibronectin-coated microtiter wells before or after treatment with immunophilin inhibitors. CrkI or CrkII overexpression in Jurkat T cells increased their adherence by ∼2-fold (Fig. 8A). Pretreatment with CsA/FK506 inhibited the adhesion of Jurkat, JKCrkI, and JKCrkII T cells by 44.9, 20.1, and 55.4%, respectively (Fig. 8A). Because all three cell lines express similar levels of endogenous CrkI and CrkII, we used the adhesion values of Jurkat T cells as basal levels and subtracted them from the values obtained by JKCrkI and JKCrkII cells to calculate the increment in adhesion reflecting the effect of the overexpressed genes (Fig. 8B). We found that CsA/FK506 had almost no effect on Myc-CrkI–dependent adhesion of JKCrkI cells, but downregulated the Myc-CrkII–dependent adhesion of JKCrkII cells by ∼58% (Fig. 8B).

Adhesion of cells in different biological systems precedes their migration, as observed during embryogenesis, wound healing, and immune responses. Both Crk and C3G are important components of both processes, and overexpression of genetically altered Crk or C3G impairs both cell adhesion and migration (46, 50–52).

To test the effect of immunophilin inhibitors on CrkI- versus CrkII-dependent cell migration, we used the same Jurkat sublines and tested their ability to migrate in vitro in response to the chemoattractant, SDF1α. The cells were introduced into chambers that are divided into two compartments by a microporous membrane. The assay ascertains efficiency of cell migration by determining the percentage of cells that migrate from one chamber, across a microporous membrane, into the second chamber in response to SDF1α, which binds to CXCR4 that signals the cells to respond by chemotaxis (53) migration (54, 55).

Overexpression of either CrkI or CrkII significantly increased the migration rate of the Jurkat T cells (Fig. 8C), and cell treatment with CsA/FK506 inhibited the migration of all three cell types (Fig. 8C). We used the migration values of Jurkat T cells as basal levels and subtracted them from the values obtained by JKCrkI and JKCrkII cells to calculate the increment in migration reflecting the effect of the overexpressed genes (Fig. 8D). We found that CsA/FK506 did not inhibit Myc-CrkI–dependent migration of JKCrkI cells, while inhibiting Myc-CrkII–dependent migration of JKCrkII cells by ∼46% (Fig. 8D).

The results indicate that CsA/FK506 significantly inhibited CrkII- but not CrkI-dependent adhesion and migration of the Jurkat T cells.

T lymphocytes express the three major types of Crk adaptor proteins, and all three are involved in constitutive and induced protein–protein interactions. The fact that Crk proteins interact with signaling molecules in T cells in a TCR activation-dependent manner (3–5, 56–60) supports a role for these proteins in signaling cascades operating downstream of the activated TCR and in regulation of T cell responsiveness. However, little is known about the biological function of the distinct Crk proteins in T cells or the mechanisms that control their expression levels and relative proportions. Structure–function studies revealed that CrkII and CrkL, but not CrkI, can undergo autoinhibition following phosphorylation of a critical tyrosine residue in the linker region connecting the two SH3 domains (Tyr²²¹ and Tyr²⁰⁷ in CrkII and CrkL, respectively) that can interact intramolecularly with the N-terminal SH2 domain.

Recent studies suggested that activity of the CrkII adaptor protein can be regulated by an additional mechanism, mediated by immunophilins that reversibly catalyze CrkII and determine the interconversion between cis (closed and inactive) and trans (open and active) conformers (13, 14). These in vitro studies were performed on a truncated recombinant protein consisting of the SH3N-linker-SH3C of the chicken CrkII (CrkII^SLS) and raise questions about the physiological relevance of these findings in mammals. For example, it is not clear whether immunophilins can catalyze the in vitro isomerization of full-length chicken CrkII, whether this isomerization can occur in vivo, and what mechanisms regulate the cis into trans isomerization of CrkII and vice versa. In addition, it is not clear whether immunophilins can catalyze the in vivo isomerization of human CrkII, which proline residue(s) may serve as target(s) for PPIases, and what is the physiological impact of this conformational change.

In this work, we used the human leukemia T cell line, Jurkat, to analyze the potential in vivo regulation of CrkII and CrkII-dependent functions by immunophilins and immunophilin-specific inhibitors.

Pull-down assays and Western blot analyses demonstrated the in vivo physical association between CrkII and immunophilins, including CypA, Cyp40, and FKBP12 in resting Jurkat T cells (Fig. 1). Although ∼90% of the CrkII proteins acquire a cis conformation in solution, in which the protein possesses a closed conformation that is not accessible to interaction with its SH3N-binding partners (14), it is yet unclear whether excess of immunophilins will change the balance between the cis and trans conformers of CrkII, and toward which direction.

One way to approach this question is to test the effect of excess of immunophilins on the ability of CrkII to associate with a binding partner that interacts with the CrkII-SH3N domain.

One such candidate is C3G, a guanine nucleotide exchange factor essential for Rap-GTPase activation (32), which interacts with the Crk-SH3N domain (35), and is critical for integrin-mediated cell adhesion and migration (50, 61). Our findings, showing that CypA increases the in vitro association of CrkII with C3G (Fig. 3), suggest that excess of immunophilins shifts the cis-trans balance of CrkII proteins toward the trans conformation. Increase in the proportions of trans CrkII conformers increases the number of open CrkII proteins that are available for interaction with the C3G molecule via their SH3N domain.

To find out whether the effect of immunophilins on CrkII binding to C3G is dependent on the isomerase catalytic activity, we used GST-C3G to pull down CrkII from lysates of Jurkat T cells that were pretreated with the most commonly used immunophilin inhibitors, CsA and FK506. We found that neither CsA nor FK506 had a dramatic effect on CrkII binding to GST-C3G, apparently because each drug by itself suppressed just one group of immunophilins, leaving the other group uninterrupted and free to induce a conformational change in C3G. However, the combined effect of CsA/FK506 significantly inhibited CrkII binding to GST-C3G (Fig. 4). The drug effect was CrkII specific because binding of CrkI, a SH3C-deficient alternatively spliced form, to GST-C3G was not altered under similar experimental conditions. Furthermore, a shorter GST fusion protein (GST-SH3b) possessing the C3G proline-rich regions (aa 208–662) (17, 33), which is sufficient for interaction with Crk, pulled down less CrkII proteins from a lysate of CsA/FK506-treated Jurkat T cells. In contrast, CrkI interaction with GST-SH3b was unaltered by cell pretreatment with CsA/FK506.

A support for the assumption that a relatively large fraction of the cellular CrkII acquires a cis (closed) conformation that is inaccessible to C3G (or other CrkII-SH3N– binding partners) comes from the observation in Jurkat T cells that, despite having excess of CrkII over CrkI (see Fig. 4A, whole-cell lysates), more CrkI molecules associate with GST-SH3b compared with CrkII in the pull-down assay (see Fig. 4A, two right lanes). These findings suggest that a fraction of the CrkII molecules that is detectable by Western blot in the whole-cell lysate is incapable of associating with GST-SH3b, apparently because of the cis conformation of the proteins.

To further verify that immunophilin inhibitors can alter the conformation of CrkII, but not CrkI, we first established two Jurkat sublines that constitutively express either Myc-CrkI or Myc-CrkII. Immunoprecipitation of Myc-tagged proteins from these cells coimmunoprecipitated C3G (Fig. 5A), and pretreatment of the cells with CsA/FK506 did not alter the amount of C3G that was expressed by the cells (Fig. 5B). However, cell pretreatment with immunophilin inhibitors decreased the amount of C3G that coimmunoprecipitated with Myc-CrkII by ∼3-fold, suggesting that the immunophilin inhibitors increased the proportions of cis conformers of CrkII without altering the conformation of CrkI.

Fluorescent analysis of Jurkat T cells, ectopically expressing membrane-bound Crk (PICCHUx) proteins, demonstrated a high degree of colocalization of Crk and C3G at one pole of the cell membrane. In contrast, this colocalization was partially disrupted by cell treatment with CsA/FK506 (Fig. 6). These findings are in agreement with the assumption that CsA/FK506 shifts the balance of CrkII proteins toward the cis conformation, which does not support CrkII-SH3N association with binding partners, such as C3G.

The assumption that CrkII is a substrate for catalytically active PPIases was further analyzed on PICCHUx-expressing Jurkat T cells. PICCHUx encodes a chimeric CrkII_1–236 that is targeted to the cell membrane via a CAAX domain, and is sandwiched between CFP and YFP (18). Confocal analysis demonstrated a basal level of FRET emission from these cells, which was significantly increased by cell pretreatment with CsA/FK506 (Fig. 7A, 7B). Similar results were obtained when cell analysis was performed on a FACS (Fig. 7C, 7D).

These results support the assumption that CrkII is a target for catalytically active immunophilins, and that inhibition of their isomerase activity (by CsA/FK506) promotes the accumulation of cis conformer of CrkII, as reflected by the increased FRET emission.

The accumulation of cis conformers of CrkII normally occurs following Abl kinase activation that phosphorylates CrkII and promotes the intramolecular phospho-Tyr²²¹ binding to the CrkII-SH2 domain. One possible explanation for the FRET data assumes that CsA/FK506 affects some unknown tyrosine kinase(s) that phosphorylates CrkII on a tyrosine residue and promotes the intramolecular folding, rather than an immunophilin-mediated trans-to-cis conformational change of CrkII. This assumption is likely to be incorrect, because tyrosine phosphorylation levels of CrkII in Jurkat T cell lysates were unaltered by cell treatment with CsA/FK506 (results not shown).

The CsA/FK506-induced increase in FRET emission is not as dramatic as in other systems, apparently reflecting the previously published observation that ∼90% of the CrkII^SLS in solution possesses the cis conformation (14). Furthermore, the ability of CsA/FK506 to modulate the conformation of the PICCHUx protein supports the hypothesis that PPIases act on a Pro residue, which is distinct from Pro²³⁸ that is predicted to be buried in the SH3N-SH3C binding interface in the CrkII^SLS cis isomer (16).

C3G is a guanine nucleotide exchange factor, which regulates the activity of the Rap1-GTPase (32), and is the first protein identified to physically associate with Crk, specifically with the Crk-SH3N domain (33, 35). Because Crk-C3G interaction plays a critical role in integrin-mediated cell adhesion and migration (49, 50, 61), we tested whether immunophilin-mediated regulation of Crk adaptor proteins may alter the Crk-C3G–dependent adhesion and migration of human T lymphocytes. Studies in this paper demonstrated that cell treatment with immunophilin inhibitors increased the proportions of cis conformers of CrkII, representing a closed conformation, which is not accessible for C3G binding. In agreement, immunophilin inhibitors suppressed the ability of Myc-CrkII–overexpressing JKCrkII cells to adhere to a fibronectin-coated surface (Fig. 8A, 8B) and migrate via a porous membrane toward a chamber rich in SDF1α chemokines (Fig. 8C, 8D). However, the same immunophilin inhibitors had a minimal effect on adhesion and migration of Myc-CrkI–overexpressing JKCrkI cells. This minimal inhibition may reflect the drug effects on the endogenous CrkII. Although the cell adhesion data support a role for immunophilin-mediated regulation of CrkII in integrin binding to fibronectin, results of the migration assay may also indicate the involvement of immunophilins and CrkII in the regulation of SDF1α-mediated signaling pathway.

Both CsA and FK506 are known to inhibit NFAT in activated T cells and induce immunosuppression by preventing NFAT-mediated activation of essential genes, including the IL2 (62). However, the effect of CsA and FK506 on CrkII appears to be independent of the drug effects on NFAT, because the expression levels of CrkI, CrkII, C3G, and other relevant proteins are not altered by T cell pretreatment with the drugs. Furthermore, CrkI and CrkII are alternatively spliced forms of a single gene, and, because CsA and FK506 do not change their expression levels and relative proportions in T cells, the results suggest that the drugs affect CrkII via a NFAT-independent mechanism.

The hypothesis that immunophilin inhibitors may negatively affect T cell activation by altering CrkII-dependent signaling pathways raises new questions relevant to the mechanism by which immunophilins regulate CrkII. For example, it is not clear which of the many types of immunophilins can interact with and regulate CrkII activity in T cells or in other cell types, and what are the mechanisms by which different immunophilins interact with CrkII. It is also unclear whether immunophilins differentially affect CrkII in resting versus activated cells, and what determines the conversion of CrkII into cis versus trans conformers. In analogy to protein kinase Cθ, which regulates the ZAP70-Crk-WASP-WIP signaling pathway in activated T cells (3, 5), but plays distinct roles in effector versus regulatory T cells (7, 63), it is also possible that regulation of CrkII and CrkII-dependent signal transduction by immunophilins is differentially regulated in distinct T cell subpopulations. Another question relates to the specific proline residues that serve as targets for immunophilins, which need to be mapped because immunophilins may use different proline residues and induce distinct conformational changes in CrkII, leading to dissimilar physiological outcomes.

Finally, the present studies support the idea that Crk-mediated signaling events in effector T cells can be manipulated exogenously by drugs, such as CsA and FK506, which target Crk-specific immunophilins and affect CrkII-dependent effector functions.

We thank Dr. S. Gelkop for preparation of the JKCrkI and JKCrkII cell lines and Drs. M. Matsuda, C. Guerrero, J. Bolstad, W. Chen, T. Ratajczak, B. Chambraud, M. Emerman, J. Buchner, K. Yamada, and A. Altman for the gifts of reagents.

The authors have no financial conflicts of interest.