INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

The functional activity of adhesion receptors expressed on T lymphocytes can be rapidly modulated by signals that T cells receive from the external environment. These adhesion regulatory signals can result in rapid changes in both adhesion receptor expression and function. One example of activation-induced changes in adhesion receptor expression is the rapid proteolytic cleavage of the L-selectin receptor upon activation of T cells and neutrophils (Kishimoto et al., 1989; Jung and Dailey, 1990). In contrast, adhesion mediated by integrin receptors expressed on T cells is regulated by rapid changes in the functional activity of integrins, rather than changes in levels of integrin expression on the cell surface. Thus, resting T lymphocytes express integrins that bind poorly to extracellular matrix (ECM) proteins or cell surface molecules. However, activation of T cells via the antigen-specific CD3/T cell receptor (CD3/TCR) complex, several different coreceptors, or inflammatory chemokines results in increased integrin-mediated adhesion within seconds to minutes (Dustin and Springer, 1989; van Kooyk et al., 1989; Shimizu et al., 1990; Diamond and Springer, 1994; Campbell et al., 1998). This rapid, but transient, change in integrin-mediated adhesiveness allows T cells to adhere to endothelial cells lining the blood vessel wall, interact with and respond to antigen-presenting cells and other cellular targets in tissue, and to navigate through, and respond to, the rich microenvironment in tissue sites. The significance of activation-dependent regulation of integrin function is further illustrated by recent reports of novel forms of leukocyte adhesion deficiency characterized by defective activation of 2 integrin-mediated adhesion (Kuijpers et al., 1997 and Harris, unpublished observations).

Integrin receptors consist of two noncovalently associated subunits, an  subunit and a  subunit. Integrin  subunit cytoplasmic domains are essential for activation-induced changes in cellular adhesion, as well as for transmitting signals back inside the cell upon interaction with ligand. Expression of  subunit cytoplasmic domains alone is often sufficient to transmit signals, and overexpression of integrin  subunit cytoplasmic domain sequences can inhibit endogenous integrin-mediated function (Akiyama et al., 1994; Chen et al., 1994; Finkelstein et al., 1997; Tahiliani et al., 1997). The use of integrin chimeras has been particularly fruitful, since the role of the integrin cytoplasmic domain can be analyzed in the context of a background of endogenous 1 integrin expression found in nucleated cells. A number of studies have utilized 1 chimeric proteins and nonhuman cells, such as CHO cells, to characterize effects of mutations in the 1 cytoplasmic domain on integrin function. These studies have demonstrated critical roles for specific regions of the 1 integrin cytoplasmic domain, particularly two well-conserved NPXY motifs, for localization of 1 integrins to focal contacts (Reszka et al., 1992; Vignoud et al., 1997). Antibodies that recognize activation-dependent epitopes on integrin  subunits have also been used to demonstrate a role for NPXY motifs in the 1 and 3 cytoplasmic domains in the regulation of integrin conformation (Bazzoni et al., 1995; Hughes et al., 1995; O'Toole et al., 1995; Luque et al., 1996; Puzon-McLaughlin et al., 1996). However, it remains unclear whether novel epitopes are precise predictors of the activation-induced changes in adhesion mediated by 1 integrins (Bazzoni and Hemler, 1998).

The expression of significant levels of endogenous 1 integrins on essentially all nucleated cells has complicated further attempts to study the role of the 1 cytoplasmic domain in 1 integrin function. Given the critical role that activation plays in regulating 1 integrin-mediated adhesion in T lymphocytes, we sought to develop an experimental system that would allow for rapid structure/function analysis of the 1 integrin subunit in T cells without the complications of endogenous 1 integrin subunit expression. Jurkat T cells express the 41 and 51 integrins and exhibit activation-dependent adhesion to the 1 integrin ECM ligand fibronectin (FN) (Mobley et al., 1994). We report in this study the isolation of a mutant of the Jurkat T cell line that does not express the 1 integrin subunit and does not adhere to FN. Transient expression of the integrin 1 subunit by DNA-mediated gene transfer yields transfectants that express the 1 integrin subunit and demonstrate activation-induced adhesion to FN. By transient expression of 1 integrin cytoplasmic mutants in this cell line, we demonstrate a critical role for the carboxy-terminal end of the 1 integrin cytoplasmic domain, as well as the NPXY motifs, in activation-dependent regulation of integrin-mediated adhesion in human T cells.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Cells

The Jurkat E6-1 cell line was obtained from the American Type Tissue Collection (ATCC, Rockville, MD). Jurkat 6A, a subclone of E6-1, and the A1 mutant were maintained in RPMI 1640 (Life Technologies, Grand Island, NY) supplemented with 10% FCS (Atlanta Biological, Norcross, GA), L-glutamine, and penicillin/streptomycin (complete medium).

Generation and Isolation of A1 Cells

Mutant A1 was isolated as previously described (Mobley et al., 1994, 1996). Briefly, Jurkat 6A cells were irradiated with 300 rads of -rays, and mutants were selected by collecting cells not adherent to FN upon stimulation with the phorbol ester phorbol dibutyrate (Sigma, St. Louis, MO). After several rounds of selection, cells were analyzed for expression of CD2, CD3, CD28, 4 integrin, 5 integrin, and 1 integrin. The A1 mutant lacked surface expression of the 1 integrin subunit, as assessed by flow cytometric analysis. This mutant was cloned several times by limiting dilution to obtain a clonal population of 1 integrin-deficient cells.

Antibodies and Other Reagents

The anti-glycophorin mAb 10F7, the 1 integrin-specific mAb TS2/16, the 1 integrin-specific mAb TS2/7, the CD3-specific mAbs 38.1 and OKT3, and L-specific mAb TS1/22 were from ATCC (Rockville, MD). The 1 integrin-specific mAb B-D15 was purchased from Biosource International (Camarillo, CA). The 1 integrin-specific mAb P4C10, the 2 integrin-specific mAb P1E6, the 3 integrin-specific mAb P1B5, and the 5 integrin-specific mAb P1D6 were obtained from Life Technologies/BRL (Grand Island, NY). The phycoerythrin (PE)-conjugated 1 integrin-specific mAb 4B4-RD1 was purchased from Coulter Immunology (Hialeah, FL). The 4 integrin-specific mAb NIH49d-1 was provided by Dr. S. Shaw (National Institutes of Health, Bethesda, MD). The 7-specific mAb Fib504 was provided by Dr. G. van Seventer (University of Chicago, Chicago, IL). The 6-specific mAb GoH-3 was provided by Dr. A. Sonnenberg (Central Laboratory of the Netherlands Red Cross Blood Transfusion Service, Amsterdam). The CD2-specific mAb 95-5-49 was provided by Dr. R. Gress (National Institutes of Health, Bethesda, MD). The CD2-specific mAb 9-1 was provided by Dr. S.Y. Yang (Memorial-Sloan Kettering Cancer Center, New York, NY). The CD28-specific mAb 9.3 was a gift from Dr. J. Ledbetter (Bristol-Myers Squibb, Seattle, WA). Anti-HA mAbs were purchased from BAbCo (Richmond, CA) and Boehringer-Mannheim (Indianapolis, IN). Goat anti-mouse immunoglobulin G (IgG)-FITC and goat anti-rat IgG-FITC were obtained from Southern Biotechnology (Birmingham, AL). Goat anti-mouse IgG was purchased from Cappel/Organon-Teknika (Durham, NC). Phorbol 12-myristate 13-acetate (PMA) was purchased from Sigma Chemical, dissolved in DMSO (100 µg/ml), and stored at 70°C. Human FN was generously provided by Dr. J. McCarthy (University of Minnesota, Minneapolis, MN).

DNA Constructs

The expression vector pHSX was derived from pMH-Neo (provided by Dr. B. Bierer, Dana-Farber Cancer Institute, Boston, MA) as follows. A HindIII site at position 1028 of the multiple cloning site was deleted by restriction enzyme digest and converted to blunt ends. The SalI (position 1042) to XhoI (position 1063) sequence was deleted by restriction enzyme digest and ligation of cohesive ends. The 3.6-kilobase (kb) EcoRI fragment from pECE.1 encoding human 1 integrin (provided by Dr. E. Ruoslahti, Burnham Institute, La Jolla, CA) was cloned into the EcoRI site (at position 1070) of the multiple cloning site of the pHSX vector to create the pHSX-1 vector.

An XhoI site at position 3590 (position 2520 of 1) was introduced by site-directed mutagenesis (Transformer Site-directed mutagenesis kit, CLONTECH, Palo Alto, CA). A HindIII site at position 3844 (position 2774 of the 1 sequence) was also deleted by site-directed mutagenesis as described above. Sequences were verified by the dideoxy method using Sequenase T7 DNA Polymerase (United States Biochemical, Cleveland, OH) according to the manufacturer's instructions. The resulting vector, pHSX-1X, contains a single HindIII site at position 3427, which lies immediately 5' of the 1 cytoplasmic domain sequence, and an XhoI site at position 3590, just 3' of the cytoplasmic domain sequence.

Constructs encoding the mutants 1(793), 1(NPIY), and 1(NPKY) were generated by gene synthesis, as described previously (Zell et al., 1996), using the following oligonucleotides: 1(793) primer 3: (5'-TCA-TAT-CAC-GGA-TTG-ACC-ACA-GTT-GTT-ACGGCA-CTC-TTA-TAA-ATA-GGA-TTT-TCA-CCC-GTG-TCC-CAT-TTG GCA-3'); and 1(793) primer 4: (5'-TCG-CTC-GAG-TCA-TTT-TCC-CTCATA-TCA-CGG-ATT-GAC-CAC-3'). 1(NPIY) primer 3: TCA-TACTTC-G GA-TTG-ACC-ACA-GTT-GTT-ACG-GCA-CTC-TTT-TCA-CCCGTG-TCC-CAT-TTG-GCA (primers 1, 2 and 4 were as used for 1(793). 1(NPKY) primer 3:(5'-TCG-ACC-ACA-GTT-GTT-ACG-GCA-CTCTTA-TAA-ATA-GGA-TTT-TCA-CCC-GTG-TCC-CAT-TTG-GCA-3') and 1(NPKY) primer 4: (5'-TCG-CTC-GAG-TCA-TTT-TCC-CTC-GAC-CAC-AGT-TGT-TAC-GGC-AC-3'). Oligonucleotides were synthesized on a Perkin Elmer-Cetus-Applied Biosystems (Foster City, CA) 392 Oligonucleotide Synthesizer. Primer 4 was used directly from the crude eluate, whereas primer 3 was purified on OPC columns (Perkin Elmer-Cetus-Applied Biosystems). PCR reactions were performed using the method of Gibbs et al. (1989) as described by Zell et al. (1996). Reaction products were fractionated by agarose gel electrophoresis. The appropriate bands were excised from the gel, purified using Wizard PCR preps (Promega, Madison, WI), digested with HindIII and XhoI, and cloned into the HindIII/XhoI site of the pHSX-1X expression vector. All pHSX-1 constructs were confirmed by automated DNA sequencing.

The alanine substitutions at positions 794, 795, 796, 797, and 798 and the phenylalanine substitution at position 795 of the 1 cytoplasmic domain were made by site-directed mutagenesis (QuickChange Site-Directed Mutagenesis kit, Stratagene, La Jolla, CA) according to the manufacturer's instructions. Each mutation was confirmed by automated DNA sequencing of plasmid DNA. Primers for the point mutations K794A, Y795A, Y795F, G797A, and K798A were synthesized by Amitof Biotech (Boston, MA), and primers for the E796A mutation were synthesized by Life Technologies (Grand Island, NY). Primer sequences for the substitutions are listed below.

K794A primer A: 5'-CTG-TGG-TCA-ATC-CGG-CGT-ATC-AGG-GAA-AAT-GAG-TAC-TGC-CC-3'; K794A primer B: 5-GAC-ACC-AGT-TAG-GCC-GCA-TAC-TCC-CTT-TTA-CTC-ATG-ACG-GG-3'. Y795A primer A: 5'-CTG-TGG-TCA-ATC-CGA-AGG-CTG-AGG-GAA-AAT-GAG-TAC-TGC-CCG-TGC-3'; Y795A primer B: 5'-GCA-CGG-GCA-GTA-CTC-ATT-TTC-CCT-CAG-CCT-TCG-GAT-TGA-CCA-CAG-3'. Y795F primer A: 5'-GCA-CGG-GCA-GTA-CTC-ATT-TTC-CCT-CAA-ACT-TCG-GAT-TGA-CCA-CAG-3'; Y795F primer B: 5'-CTG-TGG-TCA-ATC-CGA-AGT-TTG-AGG-GAA-AAT-GAG-TAC-TGC-CCG TGC-3'. E796A primer A: 5'-CTG-TGG-TCA-ATC-CGA-AGT-ATGCGG-GAA-AAT-GAG-TAC-TGC-CC-3'; E796A primer B: 5'-GGGCAG-TAC-TCA-TTT-TCC-CGC-ATA-CTT-CGG-ATT-GAC-CAC-AG 3'. G797A primer A: 5'-CTG-TGG-TCA-ATC-CGA-AGT-ATG-AGGCAA-AAT-GAG-TAC-TGC-CC-3'; G797A primer B: 5'-GGG-CAG-TACTCA-TTT-TGC-CTC-ATA-CTT-CGG-ATT-GAC-CAC-AG-3'. K798A primer A: 5'-CTG-TGG-TCA-ATC-CGA-AGT-ATG-AGG-GAG-CAT-GAG-TAC-TGC-CCG-TGC-3'; K798A primer B: 5'-GCA-CGG-GCA-GTA-CTC-ATG-CTC-CCT-CAT-ACT-TCG-GAT-TGA-CCA-CAG-3'.

Transient Transfections

A1 cells were transfected by electroporation using a BTX Electro Square Porator T820 (BTX, San Diego, CA). Ten million cells in log-phase growth were harvested, washed twice with Opti-MEM (Life Technologies), and suspended in 0.2 ml of Opti-MEM containing 30 µg of the indicated pHSX-1x construct and 20 µg of pHook2 (Invitrogen, San Diego, CA). Cells and DNA were incubated 10 min at room temperature, and then electroporated at 240 mV with a single 25-ms pulse. Cells were allowed to recover after transfection for 30 min at room temperature, and then seeded in complete medium and incubated 16-20 h at 37°C in 5% CO₂.

Selection of Transfectants

A1 cells transiently expressing pHSX-1 and hemagglutinin (HA)-tagged pHook2 were positively selected with detachable magnetic beads. Cells were harvested from culture 16-20 h posttransfection, and viable cells were recovered by Ficoll-Hypaque (Sigma Chemical) density gradient centrifugation, washed twice in PBS, and incubated with saturating amounts of mouse anti-HA mAb (clone 16B12) in PBS containing 0.1% BSA (PBS/0.1%BSA), for 30 min at 4°C. Cells were washed to remove unbound antibody and suspended in PBS/0.1% BSA (4 × 10⁶ cells/ml). Magnetic beads conjugated with rat anti-mouse IgG1 via a DNA linker (rat anti-mouse IgG1 Cellection kit, Dynal, Lake Success, NY) were used to isolate the transfectants according to the manufacturer's instructions. Briefly, rat anti-mouse beads were added to antibody-coated cells at 10 beads per target cell in the above cell suspension. Cells and beads were incubated for 1 h at 4°C on a rotator. The bead-rosetted cells were captured with a magnet, washed in 1 ml RPMI/1% FCS, and collected as per the manufacturer's instructions. The antibody-coated cells were detached from the magnetic beads by DNase treatment, and the cells were collected and counted.

Flow Cytometry

Single-color flow cytometric analysis was performed as described previously (Mobley et al., 1996) by staining cells with saturating amounts of mouse or rat anti-human antibodies, followed by appropriate FITC-conjugated secondary antibody for detection. In some cases PE-conjugated anti-1 antibody (4B4-RD1) was used. Stained cells were analyzed on a FACScan (Becton Dickinson, San Jose, CA).

Adhesion Experiments

Adhesion assays were performed as described by Mobley et al. (1995). Briefly, 96-well microtiter plates (Costar, Cambridge, MA) were incubated with the indicated concentrations of FN overnight at 4°C. Unbound binding sites were blocked with PBS/2.5% BSA. Cells were labeled with 2 µg/ml calcein-AM (Molecular Probes, Eugene OR) for 20 min at 37°C, washed, and added to wells containing the appropriate stimuli. PMA was used at 10 ng/ml; CD2 was stimulated with a 1:10 dilution of 95-5-49 hybridoma culture supernatant and a 1:2000 dilution of mAb 9-1 ascites fluid. Direct 1 stimulation was with a 1:10 dilution of TS2/16 hybridoma culture supernatant. For CD3 stimulation, wells contained 3 µg/ml mAb 38.1. For CD28 stimulation, cells were incubated for 30 min on ice with saturating amounts of 9.3, washed, and added to wells containing 1 µg/well goat anti-mouse IgG. The cells were allowed to settle in the plates for 60 min at 4°C, and then warmed rapidly for the indicated timepoints. Nonadherent cells were washed off, and adherent cells were quantitated using a fluorescence plate reader (Biotek). Percent adhesion was assessed as: All data are averages of triplicate wells for each condition.

Northern Blotting Analysis

Poly-A RNA was isolated using the FastTrack 2.0 mRNA isolation system (Invitrogen). Poly-A RNA (2 µg) was separated on a formaldehyde gel and transferred to nylon membrane (Hybond-N, Amersham, Arlington Heights, IL). Probes used were a 1.3-kb BglII human 1 fragment from pECE.1 (provided by Dr. E. Ruoslahti, Burnham Institute, La Jolla, CA) and the 1.0-kb BamHI cyclophilin fragment from pGEM4Z (provided by Dr. V. Dixit, University of Michigan, Ann Arbor, MI).

Biotinylation and Immunoprecipitation

Jurkat and A1 cells were biotin labeled, and immunoprecipitations were performed as previously described (Finkelstein et al., 1997). Precleared lysates were incubated with goat anti-mouse IgG-coupled Sepharose beads (Zymed, South San Francisco, CA) precoated with either the L-specific mAb TS1/22 or the 1-specific mAb P4C10. Immunoprecipitates were washed, boiled 5 min, and then separated on a 5% SDS-polyacrylamide gel. Proteins were transferred to polyvinylidene difluoride membrane (Millipore, Bedford, MA). The membrane was blocked with PBS/4% BSA, and incubated with a 1:1000 dilution of horseradish peroxidase-conjugated streptavidin (Life Technologies), and protein was detected by enhanced chemiluminescence (Pierce, Rockford, IL).

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Generation and Characterization of A1, a 1 Integrin-deficient T Cell Line

We have previously used Jurkat T cells to analyze activation-dependent regulation of 1 integrin-mediated adhesion (Mobley et al., 1994, 1996). One approach that we have employed with this cell line is the use of low-dose -irradiation and selection by panning on FN-coated dishes to isolate mutants with defects in 1 integrin expression and/or function. Three classes of mutants were isolated using this technique. The first class of mutants expressed normal levels of 1 integrin receptors, but had lost expression of specific integrin-regulatory receptors, such as CD3/TCR or CD2 (Mobley et al., 1994). The second class of mutants expressed normal levels of cell surface 1 integrins and integrin regulators, such as CD2, CD3, and CD28, but had defects in integrin activation (Mobley et al., 1996). Finally, we isolated a class of mutant that lacks expression of the 1 integrin subunit. This mutant is described in this report. The Jurkat mutant, A1, was isolated due to its inability to bind to FN after stimulation with the phorbol ester phorbol dibutyrate. Flow cytometric analysis shows that the A1 mutant does not express the integrin 1 subunit, as assessed with three different 1 integrin-specific mAbs (Figure 1, A and B). Consequently, A1 cells also lack expression of the 1, 2, 3, 5, and 6 subunits, which normally associate with the 1 subunit. Cell surface expression of the 4 subunit was also dramatically reduced in comparison to wild-type Jurkat cells but was still present at low levels. This is likely due to pairing of the 4 subunit with the 7 subunit, since Jurkat cells express the 7 subunit, as assessed by staining with the 7 integrin-specific mAb Fib504 (Figure 1B). The expression of the 7 subunit on wild-type Jurkat appears to be broader, containing a higher expressing population that is not seen in the A1 cells. Expression of the 2 integrin LFA-1 was not affected by the loss of 1 integrin cell surface expression in the A1 mutant (Figure 1A). Furthermore, receptors that regulate integrin activity, including CD3/TCR, CD2, and CD28, were expressed on A1 cells at levels comparable to wild-type Jurkat cells.

View larger version (29K): [in this window] [in a new window]   Figure 1.   Flow cytometric analysis of wild-type Jurkat cells and 1 integrin-deficient A1 Jurkat mutant. (A) Negative control (open profiles) staining with the anti-glycophorin mAb 10F7 is shown in each panel for ease of comparison. Solid profiles indicate staining with the following indicated mAbs: 1-specific mAbs B-D15 (left 1 panel) and P4C10 (right 1 panel), the 1-specific mAb TS2/7, the 2-specific mAb P1E6, the 3-specific mAb P1B5, the 4-specific mAb NIH49d-1, the 5-specific mAb P1D6, the 6-specific mAb GoH-3, the CD2-specific mAb 95-5-49, the CD3-specific mAb OKT3, the CD28-specific mAb 9.3, and the L (LFA-1)-specific mAb TS1/22. (B) Negative control (open profiles) staining with secondary goat anti-mouse IgG alone is shown for comparison. Solid profiles represent staining with either a 7-specific mAb FIB504 or 1-specific mAb TS2/16 followed by appropriate secondary antibody.

The lack of cell surface expression of the 1 subunit in A1 cells was also assessed by immunoprecipitation from biotin-labeled cell lysates. Figure 2 shows that while the 1 subunit was efficiently immunoprecipitated from wild-type Jurkat cells, 1 could not be precipitated from A1 cell lysates. In contrast, the 2 integrin subunit could be immunoprecipitated from both wild-type Jurkat and A1 cell lysates (Figure 2). Northern blotting analysis was also performed to determine whether the loss of 1 integrin protein expression in A1 cells was due to loss of 1 integrin mRNA. Using a probe specific for human 1 integrin, we determined that A1 cells expressed less than 2% of the 1 mRNA detected in wild-type Jurkat cells, as assessed by densitometric analysis of Northern blots (our unpublished results). This minimal amount of 1 mRNA could only be detected after prolonged exposure of the autoradiograph. We conclude from these results that the A1 mutant does not express the integrin 1 subunit on the cell surface and expresses minimal amounts of 1 mRNA.

View larger version (63K): [in this window] [in a new window]   Figure 2.   The A1 mutant does not express 1 integrin protein. Jurkat cells and the A1 mutant were biotinylated and lysed in 1% Triton detergent, and lysates were subjected to immunoprecipitation (IP) with goat anti-mouse Sepharose beads conjugated with the 1-specific mAb P4C10 or the 2-specific mAb TS1/22. IPs were washed, run on 5% acrylamide gel, and transferred to nitrocellulose. Protein was detected by incubating the membrane with Streptavidin-HRP and visualized by chemiluminescence.

Functional Analysis of A1 Cells

We next compared the adhesion of the 1 integrin-deficient A1 cells and wild-type Jurkat cells to FN. As previously described, the basal level of wild-type Jurkat cell adhesion to FN can be enhanced within minutes by treatment with the phorbol ester PMA, mitogenic pairs of CD2-specific mAbs, or antibody cross-linking of the CD3/TCR or CD28. Direct stimulation of 1 integrins with the activating 1 integrin-specific mAb TS2/16 also resulted in increased adhesion of wild-type Jurkat cells (Figure 3). In contrast, A1 cells were unable to adhere to FN, even in the presence of these various integrin-activating signals. Although FN has been reported to be a ligand for the 47 integrin (Postigo et al., 1993), the low level of 47 expressed on A1 cells was not able to mediate the adhesion of A1 cells to FN. Therefore, the loss of 1 integrin expression on A1 cells is associated with an inability to adhere to the 1 integrin ligand FN.

View larger version (28K): [in this window] [in a new window]   Figure 3.   A1 cells do not adhere to the 1 integrin ligand FN. Jurkat and A1 cells were analyzed for adhesion to the indicated concentrations of FN for 20 min at 37°C without stimulation (), or after stimulation with PMA (), the activating 1-specific mAb TS2/16 (), the CD3-specific mAb 38.1 (), mitogenic pairs of CD2-specific mAbs (), or the CD28-specific mAb 9.3 (). The data are represented as the mean percent adhesion of triplicate wells ± SD. Results shown are one of at least three representative experiments.

Transfection of Mutant A1 with Human 1 Integrin cDNA Restores Activation-dependent 1 Integrin-mediated Adhesion to FN

We transiently expressed cDNA encoding the human 1 integrin subunit in A1 cells by electroporation. Expression of the 1 subunit was generally observed in 10-30% of the electroporated cells (Figure 4D). Two-color flow cytometry was used to analyze the expression of the 4 and 5 integrin subunits, as well as CD3, CD2, and CD28, on the 1-negative and 1-positive subpopulations of A1 cells after electroporation (Figure 4). Expression of both the 4 and 5 subunits increased upon reexpression of the 1 subunit, but the level of expression of CD3, CD2, and CD28 was not affected.

View larger version (46K): [in this window] [in a new window]   Figure 4.   The 1 integrin subunit is transiently expressed in A1 cells without altered expression of CD2, CD3, or CD28. A1 cells were transiently transfected with pHSX-1x DNA. Sixteen hours posttransfection, viable cells were isolated and stained for two-color flow cytometry using PE-conjugated 1-specific mAb 4B4 (x-axis) and mAbs against the indicated surface molecules followed by goat anti-mouse-FITC (y-axis). Quadrant boundaries were set from cells treated with nonspecific secondary antibody only. Panel D represents cells stained with anti-1 mAb 4B4-PE alone. The 1+ population in this experiment was 30.2% of total, and is shown in the lower right quadrant of panel D. The expression of CD2, CD3, CD28, 4 integrin, or 5 integrin on 1 or 1+ cells is shown in panels A, B, C, E, and F, respectively.

We developed an enrichment strategy to determine the adhesive capabilities of A1 transfectants expressing the 1 subunit. A1 cells were cotransfected with the 1 expression vector and pHook2, a plasmid that encodes for a hapten-specific single chain antibody molecule that is expressed on the cell surface as an integral membrane protein (Chesnut et al., 1996). The antibody region is expressed extracellularly and also contains a hemagglutinin (HA) tag. Two-color flow cytometry with anti-HA and anti-1 mAbs was used to demonstrate that A1 cells transfected with pHook2 only expressed the HA-tagged pHook2 molecule but remained 1-negative. Upon transfection with both pHook2 and 1, the subpopulation of A1 cells expressing pHook2 also expressed the 1 subunit (Figure 5). The subpopulation of A1 transfectants expressing pHook2 was then isolated with an anti-HA mAb and magnetic beads. As shown in Figure 5, this selection procedure enriched for a population of A1 cells expressing the transfected 1 cDNA. Expression of 1 in this selected population was typically broad and did not approach the high level of endogenous 1 integrin expressed on the wild-type parental Jurkat cells. Thus, this method allowed us to rapidly select and isolate 1⁺ transfectants without using anti-1 mAbs that might alter integrin function.

View larger version (32K): [in this window] [in a new window]   Figure 5.   Coexpression of 1 and pHook2 in A1 cells. A1 cells were transiently transfected with pHook2 alone (left column), or with pHook2 and pHSX-1x together (right column). Cells were collected 16 h after transfection and analyzed by two-color flow cytometry using an anti-HA mAb and GAR-FITC (y-axis) and PE-conjugated 1-specific mAb 4B4 (x-axis). Transfectants were analyzed before (top and middle panels) or after (bottom panels) magnetic bead selection. Coexpression of 1 and the HA tag is represented as percentage of total cells in the upper right quadrant (top panels).

To measure the ability of 1⁺ A1 cells to adhere to FN, we isolated A1 cells transiently transfected with pHook2 alone or cotransfected with 1 cDNA (Figure 6A). Similar to what was observed with untransfected A1 cells, A1 cells transfected with pHook2 alone did not adhere to FN, even after stimulation. Thus, expression of the pHook2 molecule did not alter the adhesion of A1 cells to FN. A1 cells expressing wild-type human 1 exhibited increased basal adhesion to FN when compared with control A1 cells expressing pHook2 only. Furthermore, stimulation of 1⁺ A1 cells with PMA, CD2 ligation, or cross-linking of CD3/TCR or CD28 resulted in a rapid, dose-dependent increase in adhesion to FN (Figure 6A). Like the adhesion response observed with wild-type Jurkat cells (Mobley et al., 1994), the adhesion of 1⁺ A1 transfectants to FN was rapid but transient: adhesion peaked within 10-20 min of stimulation and began to decrease within 40 min after stimulation (our unpublished results). The adhesion of 1⁺ A1 transfectants to FN was inhibited by the 1 integrin-specific mAb P5D2, or by a combination of 4- and 5-specific mAbs, indicating that the increased adhesion of these cells was mediated by endogenous 4 and 5 pairing with the 1 subunit encoded by the transfected cDNA (Figure 6B and our unpublished results). Together, these results show that A1 cells are capable of transiently expressing functional 1 integrins, and that 1⁺ transfectants exhibit activation-induced cell adhesion to a 1 integrin ligand.

View larger version (25K): [in this window] [in a new window]   Figure 6.   Adhesion to FN of A1 cells transfected with pHook2 alone or with 1. A1 cells were transiently transfected with pHook2 alone, or together with pHSX-1x DNA. Cells were recovered after transfection, and 1-expressing transfectants were isolated as described in MATERIALS AND METHODS. (A) A1 cells expressing pHook2 alone or with 1 were analyzed for adhesion to the indicated concentrations of FN for 20 min at 37°C without stimulation (- - - - -), or after stimulation with PMA (), the activating 1-specific mAb TS2/16 (*), the CD3-specific mAb 38.1 (), mitogenic pairs of CD2-specific mAbs (), or the CD28-specific mAb 9.3 (). (B) A1 cells expressing pHook2 alone or with 1 were analyzed for adhesion as in panel A to 1 µg/well of FN in the absence or presence of a 1:200 dilution of ascites of the anti-1 mAb P5D2. Cells were either unstimulated (left bar in each set of bars) or stimulated with PMA, mitogenic pairs of CD2-specific mAbs, the activating 1-specific mAb TS2/16, or the CD3-specific mAb 38.1 (left to right in each set of bars). The data are represented as the mean percent adhesion of triplicate wells ± SD. Results shown are one of at least three representative experiments.

Deletion of the Carboxy-terminal Five Amino Acids of the Cytoplasmic Domain Abolishes Adhesion of 1⁺ A1 Cells to FN

To study the structural requirements within the 1 integrin subunit cytoplasmic domain that are responsible for activation-dependent adhesion to FN, we generated a panel of cDNA clones encoding for 1 integrin subunits with specific cytoplasmic domain mutations (Figure 7). Each mutant 1 integrin subunit was then expressed transiently in A1 cells and analyzed for adhesion to FN (Figure 8 and 9). Although the level of 1 expression in transient transfectants varied from experiment to experiment, each 1 integrin cytoplasmic mutant tested was expressed on the cell surface at levels comparable to wild-type 1 in any given experiment (Figures 8A and 9). A deletion of the carboxy-terminal five amino acids (KYEGK) of the 1 cytoplasmic domain, designated 1(793), was unable to mediate adhesion to FN after stimulation (Figure 8B). Even at concentrations of immobilized FN that mediated maximal adhesion of Jurkat cells to FN, both basal and stimulated adhesion of A1 transfectants expressing the 1(793) was reduced when compared with the wild-type 1 subunit. Interestingly, adhesion induced by direct stimulation of the 1 activating mAb TS2/16 was also inhibited by deletion of the carboxy-terminal five amino acids. To establish whether the 1(793) transfectants exhibited delayed adhesion to FN, we tested the adhesion up to 40 min after stimulation. At all time points tested, the 1(793) transfectants were unable to adhere to FN (our unpublished results).

View larger version (36K): [in this window] [in a new window]   Figure 7.   Amino acid sequence of the wild-type 1 integrin cytoplasmic domain and the various cytoplasmic domain mutations analyzed in this study. The amino acid sequences of the wild-type 1 integrin cytoplasmic domain and each mutant construct are written in single letter code. Amino acid substitutions are boxed.

View larger version (36K): [in this window] [in a new window]   Figure 8.   Lack of adhesion to FN of A1 cells expressing the 1(793) mutant. A1 cells were transiently transfected with pHook2 alone, or together with pHSX-1x or pHSX-1(793) DNA. Cells were recovered after transfection, and 1-expressing transfectants were isolated as described in MATERIALS AND METHODS. (A) Cells were stained with PE-conjugated 1-specific mAb 4B4 as described in Figure 4 legend. (B) A1 cells expressing pHook2 alone or together with 1 or 1(793) were analyzed for adhesion to the indicated concentrations of FN for 20 min at 37°C without stimulation (- - - - -), or after stimulation with PMA (), the activating 1-specific mAb TS2/16 (), or the CD3-specific mAb 38.1 (). For each transfectant tested, adhesion after stimulation with CD2-specific mAbs or the CD28-specific mAb 9.3 was comparable to that seen for CD3 stimulation (our unpublished results). The data are represented as the mean percent adhesion of triplicate wells ± SD. Results shown are one of at least three representative experiments.

View larger version (37K): [in this window] [in a new window]   Figure 9.   Substitution of the tyrosine at position 795 with alanine, but not phenylalanine, reduces adhesion of A1 transfectants to FN. A1 cells were transfected with pHook2 alone, or together with the indicated 1 construct. Cells were recovered after transfection, and 1-expressing transfectants were isolated as described in MATERIALS AND METHODS. (A-E) A1 cells expressing pHook2 alone or together with the indicated 1 construct were analyzed for adhesion to FN (1 µg/well in panels A-D, 3 µg/well in panel E) for 20 min at 37°C without stimulation (), or after stimulation with PMA (), the activating 1-specific mAb TS2/16 (open bars), or the CD3-specific mAb 38.1 (hatched bars). For each transfectant tested, adhesion after stimulation with CD2-specific mAbs or the CD28-specific mAb 9.3 was comparable to that seen for CD3 stimulation (our unpublished results). Each panel represents a separate experiment. Transfectants were stained with PE-conjugated anti-1 antibody 4B4 followed by FACS analysis to determine the 1 expression on transfectants for that particular experiment. Histograms representing 1 expression are shown. The data are represented as the mean percent adhesion of triplicate wells ± SD. Results shown are one of at least three representative experiments with each mutant 1 cytoplasmic domain construct.

Mutation of the Tyrosine in the NPKY Motif to Alanine Inhibits Activation-dependent 1 Integrin-mediated Adhesion to FN

Since deletion of the carboxy-terminal five amino acids in the 1 cytoplasmic domain prevented 1 integrin function in A1 cells, we performed alanine scanning mutagenesis in this region of the 1 cytoplasmic tail (Figure 7). Each of the five amino acids in the carboxy-terminal end of the 1 cytoplasmic domain was mutated to an alanine, and the effect of these mutations on activation-dependent adhesion of A1 transfectants to FN was assessed. As shown in Figure 9, mutation of amino acids 794, 796, 797, and 798 to alanine did not affect either basal or stimulated adhesion to FN after expression in A1 cells. In contrast, mutation of the tyrosine residue at position 795 to an alanine reduced the ability of the 1 subunit to mediate adhesion to FN in A1 cells (Figure 9). The Y795A mutation reduced both the basal level of adhesion in the absence of stimulation, as well as the increased adhesion induced by stimulation of the integrin regulators CD3, CD2, or CD28 (Figure 9 and our unpublished results). Interestingly, adhesion induced by the activating 1 integrin-specific mAb TS2/16 was minimally affected by the Y795A mutation. When the tyrosine at position 795 was changed to a phenylalanine, both basal and stimulated adhesion were comparable to that observed with transfectants expressing the wild-type 1 integrin subunit. Thus, these results suggest a critical role for the tyrosine residue in the NPKY motif in the carboxy-terminal end of the 1 cytoplasmic domain in activation-dependent adhesion of Jurkat T cells to FN.

Deletion of the Conserved NPXY Motifs in the 1 Integrin Cytoplasmic Domain Inhibits Activation-dependent Adhesion of Jurkat T Cells to FN

To determine the role of the conserved NPXY motifs within the 1 cytoplasmic tail in mediating 1 integrin regulation in T cells, we generated 1 cytoplasmic mutants containing a deletion of either the amino-terminal NPIY sequence or the carboxy-terminal NPKY sequence (Figure 7). In each case, surface expression of the deletion mutants was similar to that seen for full-length 1 (Figure 10A). However, a deletion of either the NPIY sequence or the NPKY sequence resulted in the expression of a 1 integrin subunit that was unable to mediate adhesion to FN (Figure 10B). Both basal and stimulated adhesion of A1 cells to FN were reduced in transfectants expressing either the 1NPIY or 1NPKY mutant when compared with the wild-type 1 subunit. However, while CD3- and CD28-mediated activation of adhesion was inhibited completely by these two deletions, direct stimulation of 1 integrins by the 1 integrin-specific mAb TS2/16 was only partially reduced when compared with wild-type 1 (Figure 10B).

View larger version (36K): [in this window] [in a new window]   Figure 10.   A1 cells expressing NPXY deletions exhibit reduced adhesion to FN. A1 cells were transfected with pHook2 alone, or together with the indicated 1 construct. Cells were recovered after transfection, and 1-expressing transfectants were isolated as described in MATERIALS AND METHODS. (A) Transfectants were stained with PE-conjugated 4B4 antibody followed by FACS analysis to determine the 1 expression on transfectants for that particular experiment. Histograms representing 1 expression are shown. (B) Transfectants were analyzed for adhesion to FN (3 µg/well) for 20 min at 37°C without stimulation (), or after stimulation with the activating 1-specific mAb TS2/16 (), the CD3-specific mAb 38.1 (), or the CD28-specific mAb 9.3 (). The data are represented as the mean percent adhesion of triplicate wells ± SD. Results shown are one of at least three representative experiments with each mutant 1 cytoplasmic domain construct.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

Although 1 integrin-mediated adhesion of T cells has been shown to be dynamically regulated by activation signals that increase 1 integrin avidity, the structural basis for this regulation of integrin function remains incompletely defined. In this study, we sought to define the residues in the 1 cytoplasmic domain that are critical for activation-dependent regulation in T lymphocytes. The approach that we developed was made possible by the isolation of a 1 integrin-deficient variant of the Jurkat T cell line. This cell line, designated A1, does not express the 1 integrin subunit on the cell surface. Consequently,  subunits that associate with the 1 subunit, with the exception of low levels of the 4 subunit, are also not expressed on the surface of A1 cells. The A1 cell line has little detectable 1 mRNA, suggesting a defect in endogenous 1 integrin expression in A1 cells at the RNA level. The lack of 1 integrin expression on the cell surface results in loss of adhesion of A1 cells to FN when compared with wild-type 1+ Jurkat cells, even after stimulation with agonists that activate 1 integrin-mediated adhesion. Upon reexpression of the 1 integrin subunit in A1 cells, 1⁺ A1 cells adhered to immobilized FN in an activation- and 1-dependent manner, suggesting that the integrin-regulatory signaling cascades are still intact. Thus, the A1 cell line represents a valuable new cellular tool for the analysis of 1 integrin structure and function.

We used the 1 integrin-deficient A1 cell line to identify the regions of the 1 integrin cytoplasmic domain that are critical for activation-dependent regulation of 1 integrin-mediated adhesion. This question has been difficult to address, since all nucleated hematopoietic cells express the 1 integrin subunit. Although some prior studies have used integrin chimeric proteins expressed in nonhematopoietic cell lines (Vignoud et al., 1994; O'Toole et al., 1995), these cell lines do not exhibit the rapid changes in 1 integrin function that occur upon activation of T lymphocytes. Furthermore, these studies have, in general, involved the use of reporter antibodies that detect the expression of novel epitopes upon cell stimulation (Vignoud et al., 1994; O'Toole et al., 1995). Although these novel epitopes are clearly indicative of changes in 1 integrin conformation, the relationship between these conformational changes and adhesion has not been definitively established for many of these epitopes (Stewart et al., 1996; Yauch et al., 1997; Bazzoni and Hemler, 1998).

Since the A1 cell line does not express the 1 integrin subunit, we could directly determine whether specific mutations in the 1 integrin cytoplasmic domain affect activation-dependent adhesion of T cells to a 1 integrin ligand. We used a transient transfection approach to demonstrate that two highly conserved NPXY motifs in the human 1 integrin cytoplasmic domain are necessary for mediating both basal and stimulated adhesion of T cells to FN. Deletion of either the amino-terminal NPIY sequence or the carboxy-terminal NPKY sequence, or truncation of the carboxy-terminal five amino acids of the 1 integrin cytoplasmic domain, abolished adhesion to FN. These defects in adhesion were not due to effects of the 1 integrin cytoplasmic domain mutations on cell surface expression, since all of the mutant 1 integrin subunits were expressed on the cell surface at levels comparable to wild-type 1. Scanning alanine mutagenesis of the carboxy-terminal five amino acids revealed a critical role for the tyrosine residue at position 795 in the NPKY motif in 1 integrin function in T cells. While alanine substitutions at positions 794, 796, 797, and 798 did not affect 1 integrin-mediated adhesion, a change of the tyrosine at position 795 to alanine resulted in reduced basal and stimulated adhesion to FN compared with A1 cells expressing comparable levels of wild-type 1. However, a more conservative substitution of this tyrosine for phenylalanine did not affect 1 integrin-mediated adhesion in A1 cells, suggesting that phosphorylation of the tyrosine residue at position 795 is not required for adhesion of Jurkat T cells to FN. Since NPXY motifs have been shown to form tight turns in protein structure, and mutation of the tyrosine in the NPKY to alanine resulted in reduced 1 integrin-mediated adhesion in A1 cells, the NPKY motif may be a critical structural component of the 1 cytoplasmic domain.

The NPXY amino acid motifs are conserved among many integrin  chains. In the 1 cytoplasmic domain, the two NPXY motifs and a third region have been shown to be critical for localizing 1 integrins to focal contact sites (Reszka et al., 1992; LaFlamme et al., 1994; Vignoud et al., 1997). Studies of the NPXY motif in the 3 integrin subunit have also suggested a critical role for this motif in 3 integrin function (Ylänne et al., 1995; Filardo et al., 1995). Mutation of the amino-terminal NPXY motif in the 3 subunit cytoplasmic domain abolished the constitutive adhesion and spreading of v3-expressing melanoma cells to vitronectin. In contrast to our results, a truncation of the carboxy-terminal 11 amino acids in the 3 subunit, a region that includes the carboxy-terminal NPXY motif, did not affect adhesion to vitronectin (Filardo et al., 1995). Deletion of the amino-terminal NPXY motif in the 1 cytoplasmic domain, when expressed in the context of a chimeric integrin expressing the 3 extracellular domain, also abolished the binding of PAC1, a mAb that recognizes the activated form of the IIb3 integrin (O'Toole et al., 1995). Interestingly, substitutions of the asparagine or tyrosine residues in the carboxy-terminal NPXY motif also affected PAC-1 binding, although the level of inhibition was not as dramatic as that seen for mutations in the amino-terminal NPXY motif (O'Toole et al., 1995). Our results also suggest a role for the tyrosine residue in the carboxy-terminal NPXY motif in 1 integrin function, since the Y795A mutation in the 1 integrin cytoplasmic domain severely reduced adhesion of A1 cells to FN. A role for the carboxy-terminal end of the 1 integrin cytoplasmic domain in regulating 1 integrin-mediated cell adhesion is also suggested by the ability of a cell-permeable peptide representing the carboxy-terminal half of the 1 cytoplasmic tail to inhibit fibroblast adhesion (Liu et al., 1996).

Similar to 1-mediated adhesion, L2-mediated adhesion of lymphocytes requires activation for interaction with intercellular adhesion molecules (ICAMs) (Dustin and Springer, 1989; van Kooyk et al., 1989). When expressed in COS cells, deletion of the carboxy-terminal five amino acids of the 2 cytoplasmic domain, including a lysine and phenylalanine in a NPKF motif, abolished constitutive adhesion to purified ICAM-1 (Hibbs et al., 1991b). Furthermore, this study showed that expression of a 2 mutant in which the carboxy-terminal 5 amino acids were replaced with Glu-Val-Cys in a 2 integrin-deficient B lymphoblastoid cell line resulted in a loss of adhesion of transfectants to immobilized ICAM-1 upon phorbol ester stimulation. Additional studies have implicated the phenylalanine residue in the carboxy-terminal NPKF motif in the 2 cytoplasmic domain as being critical in LFA-1-mediated adhesion to purified ICAM-1 in COS cells (Hibbs et al., 1991a). Although mutation of this phenylalanine to tyrosine had no effect on adhesion, a change to alanine or leucine abolished adhesion. Thus, these results with a related integrin  subunit also support a central role for the carboxy-terminal end of integrin  subunits in integrin function.

Our studies with the A1 cell line to some extent parallel recent analyses of 1 integrin structure and function using the GD25 cell line, which was derived from 1-null embryonic stem cells (Wennerberg et al., 1996). There are several important differences between these two 1-negative cell lines that should be noted. Unlike A1 cells, GD25 cells express other FN-binding integrins, notably V3. This complicates the analysis of effects of 1 cytoplasmic domain mutations on 1-mediated interactions with FN using GD25 cells. In addition, the basal activity of 1 integrins on GD25 fibroblasts and A1 T cells differs, since 1 integrin-mediated adhesion of A1 cells is dependent to a large extent on activation signals that increase 1 integrin activity. Despite these differences, some common effects of 1 cytoplasmic domain mutations are observed. Most notably, the Y795F mutation, which did not affect A1 adhesion to FN, also did not affect 1 integrin-mediated attachment of GD25 cells to laminin-1 (Wennerberg et al., 1998) or to FN (Sakai et al., 1998). However, a dramatic effect of the Y795F mutation on directed migration of GD25 cells was observed, suggesting a critical role for this residue in movement rather than attachment. Mutation of the proline residues in either NPXY motifs, either singly or in combination, also impaired cell adhesion as well as expression of the 1 integrin epitope defined by mAb 9EG7 (Sakai et al., 1998). However, it is difficult to directly assess the effect of these mutations on cell adhesion relative to wild-type 1 in GD25 cells, since these mutations also affected cell surface expression of these mutant 1 subunits. Adhesion of GD25 cells to laminin-1 was profoundly affected by the mutation of two highly conserved threonines at positions 788 and 789 to alanine (Wennerberg et al., 1998). However, mutation of these two threonines did not abolish 1 integrin-dependent adhesion to FN, suggesting that the impact of mutating these threonine residues in GD25 cells is ligand-dependent. It remains to be determined whether these threonine substitutions would have a similar impact on activation-dependent adhesion to A1 Jurkat cells to FN or another 1 integrin ligand, such as VCAM-1.

Our studies also demonstrated a reduced ability of the 1 integrin-activating mAb TS2/16 to enhance the adhesion of cells expressing the 1(793), 1(NPIY), or 1(NPKY) mutant subunits. Although TS2/16 and other activating 1 integrin-specific antibodies can induce or stabilize the active form of 1 integrins in the absence of other modes of stimulation (Arroyo et al., 1992), our results suggest that optimal adhesion induced by direct activation of 1 integrins by TS2/16 does, in fact, require the integrity of the 1 integrin cytoplasmic tail. This is consistent with studies demonstrating that deletion of the carboxy-terminal 16 amino acids in the human 1 integrin cytoplasmic domain inhibited soluble FN binding by 51 expressed in CHO cells induced by the activating 1 integrin-specific mAb 8A2 (Puzon-McLaughlin et al., 1996). Our studies also demonstrate that the 1(793), 1(NPIY), and 1(NPKY) mutant subunits were unable to respond to all of the other stimuli that have been shown to activate 1 integrins in Jurkat cells, including phorbol ester stimulation and activation via CD3/TCR, CD2, and CD28. Furthermore, the 1(K794A), 1(Y795F), 1(E796A), 1(G797A), and 1(K798A) mutants were able to respond to all stimuli. This suggests that at the level of the 1 integrin cytoplasmic domain, all of these integrin-activating signals require common structural integrity of the 1 cytoplasmic domain.

In summary, we have produced and characterized a variant of the Jurkat T cell line that lacks expression of the 1 integrin subunit. This 1-deficient cell line was used to demonstrate a critical role for the carboxy-terminal end of the 1 cytoplasmic domain, and two conserved NPXY motifs, in the activation-dependent regulation of 1 integrin-mediated adhesion. Thus, this 1 integrin-deficient cell line provides a valuable new cellular reagent for the analysis of 1 integrin structure and function in the context of a cell that exhibits dynamic regulation of integrin adhesiveness without the complication of endogenous 1 integrin subunit expression.