INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Integrins are a family of transmembrane receptors composed of at least 25 different heterodimers that mediate both cell-substrate and cell-cell adhesion (Hynes, 1992). Among their other functions, integrins play a central role in cell migration. Integrin-dependent migration is important in many biologic processes including embryonic development, wound healing, inflammation, and tumor metastasis.

Cell migration is a complex and poorly understood process that involves both actin reorganization and integrin-dependent focal adhesion remodeling. For cells to migrate on a substrate, they must adhere and de-adhere in a coordinated manner and generate tensile force in the direction of migration. The force required to promote migration is generated by the actin cytoskeleton and integrin-dependent protein complexes that anchor actin to specific sites on the cell membrane (Lauffenburger and Horwitz, 1996). The actin reorganization required to promote cell polarization and directional migration is regulated by specific intracellular signaling pathways (Lauffenburger and Horwitz, 1996; Horwitz and Parsons, 1999). Although most members of the integrin family are capable of mediating cell migration (Lauffenburger and Horwitz, 1996), experiments utilizing chimeric or truncated integrins indicate that the cytoplasmic domain of the 4 subunit preferentially enhances cell migration and inhibits cell spreading compared with other  subunits of the 1 subclass of integrins (Chan et al., 1992; Kassner et al., 1995). In addition, unlike most integrins, 41 is relatively excluded from mature focal adhesion complexes (Kassner et al., 1995). These findings support the hypothesis that 4-containing integrins, which are widely expressed on leukocytes, play a specialized role in promoting rapid cell migration. Further, these findings support a model that links inhibition of cell spreading (i.e., cell rounding) to enhanced migration. Recently, the integrin 91, an integrin that is structurally most closely related to 41, was shown to promote transendothelial neutrophil migration through its interactions with vascular cell adhesion molecule-1 on activated endothelial cells (Taooka et al., 1999). The cytoplasmic domain of the 9 subunit shares 52% homology with the 4 subunit but is divergent from all other integrin-cytoplasmic domains. We therefore questioned whether the 9 cytoplasmic domain would also specifically enhance cell migration and inhibit cell spreading.

Recently, 41-dependent cell migration was shown to be dependent on the specific interaction of the 4 cytoplasmic domain with the adapter protein paxillin (Liu et al., 1999). Paxillin participates in several intracellular signaling pathways that influence cell migration (Clark and Brugge, 1995; Turner, 2000). Site-directed mutagenesis of a tyrosine (Y) residue to an alanine (A) residue at position 991 (Y991A) in the 4 cytoplasmic domain that inhibited the 4-paxillin interaction also inhibited 41-dependent migration and cell shape changes (Liu et al., 1999; Liu and Ginsberg, 2000). In the current study, we utilized a series of chimeric and truncated versions of the 9 subunit to determine whether the 9 cytoplasmic domain preferentially enhances migration and inhibits cell spreading on an 91-specific ligand and what role, if any, paxillin plays in these processes.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Reagents and Antibodies

The 91-specific ligand used in this study was a recombinant form of the third fibronectin type III repeat of chicken tenascin-C (Prieto et al., 1993) containing alanine (A) substitutions for both glycine (G) and aspartate (D) residues within the arginine (R)-G-D site (TNfn3RAA; Yokosaki et al., 1998). The cDNA for TNfn3RAA was obtained from Anita Prieto and Kathryn Crossin (Scripps Research Institute, La Jolla, CA) and was prepared in Escherichia coli as previously described (Prieto et al., 1993). The mouse mAb, Y9A2, increased against human 91, was prepared as previously described (Wang et al., 1996). The following monoclonal antibodies were purchased commercially: monoclonal antibodies against paxillin (clone 349, BD-Biosciences-Transduction Laboratories, Lexington, KY) and against hemagglutinin (HA)-tag (12CA5, American Type Culture Collection [ATCC], Rockville, MD).

Generation of 9 Constructs

The previously described pBlueScript (BS)-SK9 cDNA plasmid was used as the template to generate all 9 constructs (Yokosaki et al., 1994). To generate the 9 chimeras containing the cytoplasmic domains of 2, 5, and 4, a mutation in the 9 sequence was generated at amino acid position 972 in the transmembrane domain near the start of cytoplasmic domain that changed a leucine residue to a valine residue and created an SpeI restriction site. The mutation was created by polymerase chain reaction (PCR) with the use of a 5' forward primer that was upstream of an EcoRI site in the extracellular domain, 5'-tttcctttcatgaggtca-3' and a 3' reverse primer that created the SpeI restriction site for subcloning into the multiple-cloning site of pBS-SK9, 5'-ttct ta ctag tacggccagcagcaggaagat-3'. The nucleotides in bold type represent the sites of mutagenesis. The PCR product was digested with EcoRI and SpeI, purified, and subcloned into pBS-SK9. The 2 and 5 cDNA used in the PCR reactions to generate the cytoplasmic domains was from a teratocarinoma-2 cell line (ATCC). To generate the 92 chimera, a 5' forward primer that was specific for the cytoplasmic domain of 2 and contained an SpeI restriction site, 5'-ggcttactagtctggaagctcggcttcttc-3', and a 3' reverse primer specific for the cytoplasmic domain of 2 that contained a NotI restriction site for subcloning into the multiple-cloning site of pBS-SK9, 5'-atcttgcggccgcaagaaatccatgcacgcaaa-3', were used. The PCR product was digested with SpeI and NotI, purified, and subcloned into pBS-SK9. To generate the 95 chimera, a 5' forward primer that was specific for the cytoplasmic domain of 5 and contained an SpeI restriction site, 5'-ttcttactagtctggaaacttggattcttcaaacgc-3', and a 3' reverse primer specific for the cytoplasmic domain of 5 that contained an XbaI restriction site for subcloning into the multiple-cloning site of pBS-SK9, 5'-atctttctagagtggggggactggttcttca-3', were used. To generate the 94 chimera, the 4 expression plasmid, pc4DM8 (generously provided by Dr. David Erle), was used as a template, and a 5' forward primer that was specific for the cytoplasmic domain of 4 and contained an SpeI site, 5'-gctccactagtctggaaggctggcttcttt-3', and a 3' reverse primer specific for the cytoplasmic domain of 4 that contained a NotI site, 5'-ctgctgctgctggcggccgcggtaccttattaatcatcatt-gcttttac-3', were used. To generate the 9 cytoplasmic deletion mutants (9DMs), the pBS-SK9 was used as the template in PCR reactions that used the 5' forward primer, 5'-tttcctttcatgaggtca-3', for all of the 9DMs (Figure 1), and the 3' reverse primers specific for the 9 cytoplasmic domain that contained a NotI restriction site, 5'-ttcttgcggccgcttacatcttccagagcagcacggc-3', 5'-ttcttgcggccgcttatcggcgaaagaagcccatctt-3', 5'-ctgctgctgctggcggccgctctagattattattctttgtacc-ttcggcg-3', 5'-ctgctgctgctggcggccgctctagattattacttctcagcttcgataat-3', 5'-ctgctgctgctggcggccgctctagattattattcattctctttccggtt-3', 5'-ctgctgctgctggcggccgctctagattattactggacccagtcccaact-3', for 9DM1-9DM6, respectively. All reverse primers were designed to introduce two translation stop codons in tandem at the end of the coding sequence before the restriction sites, and all constructs were confirmed by nucleotide sequencing. The 9 site-directed mutants containing a tryptophan (W) to A substitution at either position 999 (W999A) or 1001 (W1001A) in the 9 subunit were generated from pBS-SK9 with the use of a QuikChange Site-Directed Mutagenesis kit (Stratagene, La Jolla, CA) according to the manufacturer's protocol. The 5' forward and 3' reverse primers used to generate 9(W999A) were 5'-ggaaagagaatgaagacagt gcggactgggtccagaaaaacc-3' and 5'-ggtttttctggac ccagtccgcactgtcttcattctctttcc-3', and the primers used to generate 9(W1001A) were 5'-ggaaagagaatgaagacagttgggac gcggtccagaaaaacc-3' and 5'-ggtttttctggaccgcgtc ccaactgtcttcattctctttcc-3'. The nucleotides in bold type represent the sites of mutagenesis. All constructs were confirmed by nucleotide sequencing. All 9 constructs were subcloned into the previously described full-length 9 expression plasmid pcDNAIneo9 (Yokosaki et al., 1994) after excision of the pBS-SK9 constructs with HindIII and NotI and subcloning into pcDNAIneo9. For subcloning into pBABEpuro, the 9 constructs were excised from pBS-SK9.

View larger version (44K): [in this window] [in a new window]   Figure 1.   Schematic diagram of 9, 9 chimeras, and 9DMs. The cytoplasmic amino acid sequences of 9, 9 chimeras (A), and 9DMs (B) are shown. Amino acids indicated in bold type represent areas of homology between all constructs and the underlined amino acids represent areas of homology with the 4 cytoplasmic domain.

Generation of Stable Cell Lines

The CHO cells lines were generated by calcium phosphate precipitation with vectors made in pcDNAIneo (Invitrogen, San Diego, CA.) and were maintained in Dulbecco's minimal essential medium (DMEM) supplemented with 10% fetal calf serum (FCS) and the neomycin analogue G418 (1 mg/ml; Life Technologies, Rockville, MD). Transfected cells were analyzed for expression of 91 integrins by flow cytometry with the anti-91 antibody, Y9A2. Mouse embryonic fibroblasts (MEF; from ATCC) and MEF paxillin^/ cell lines were infected with the use of 9 constructs in the retroviral vector pBABEpuro (Morgenstern and Land, 1990). Retroviruses were generated by calcium phosphate-mediated transfection into the Phoenix-E replication-incompetent ecotropic virus packaging cell line (Kinoshita et al., 1998; Swift et al., 1999). Specifically, 8 µg of plasmid DNA were added to 70% confluent Phoenix-E cells growing in 60-mm tissue culture dishes at 37°C in 3 ml of 10% FCS DMEM for 16 h. The medium was removed, 3 ml of fresh 10% FCS DMEM were added, and the cells were cultured for 16 h. Virus-containing supernatants were harvested and filtered through a 0.22-µm filter and then added to 50% confluent cultures in the presence of 5 µg/ml polybrene and cultured for 18-20 h. The virus-containing medium was removed and the cells were cultured in 10% FCS DMEM supplemented with 10 µg/ml puromycin (Sigma, St. Louis, MO). MEF and paxillin^/ MEF cells expressing the 91 constructs were identified by flow cytometry with the anti-91 antibody, Y9A2. Fluorescence-activated cell sorting was performed to isolate heterogeneous populations of cells expressing high levels of 91 integrins on their cell surfaces (Yokosaki et al., 1998). All cell lines continuously expressed high surface levels of 91 as determined by flow cytometry with Y9A2.

Flow Cytometry

Cultured cells were harvested by trypsinization and rinsed with phosphate-buffered saline (PBS). Nonspecific binding was blocked with normal goat serum at 4°C for 10 min. Cells were then incubated with primary antibody for 20 min at 4°C, followed by a secondary goat anti-mouse antibody conjugated with phycoerythrin (Chemicon, Temecula, CA). Between incubations cells were washed twice with PBS. The stained cells were resuspended in 100 µl of PBS, and fluorescence was quantified on 5000 cells with a FACScan (Becton Dickinson, Rutherford, NJ) flow cytometer.

Cell Adhesion Assays

The wells of nontissue culture 96-well microtiter plates (Nunc, Naperville, IL) were coated by incubation with 100 µl of TNfn3RAA for 1 h at 37°C. After incubation, wells were washed with PBS and then blocked with 1% bovine serum albumin (BSA) in DMEM at 37°C for 30 min. Control wells were filled with 1% BSA in DMEM. The cells were detached with the use of 2.5 ml of trypsin solution (Sigma), followed by 2.5 ml of trypsin-neutralizing solution (Sigma), washed once in DMEM, and resuspended in DMEM at 5 × 10⁵ cells/ml. The cells were incubated with or without 50 µg/ml Y9A2 for 20 min at 4°C before plating. Plates were centrifuged (top side up) at 10 × g for 5 min before starting the incubation for 1 h at 37°C in humidified 5% CO₂. Nonadherent cells were then removed by centrifugation (top side down) at 48 × g for 5 min. Attached cells were fixed and stained in 40 µl of a 1% formaldehyde, 0.5% crystal violet, 20% methanol solution for 30 min, after which the wells were washed three times with PBS. The relative number of cells in each well was evaluated after solubilization in 40 µl of 2% Triton X-100 by measuring the absorbance at 595 nm in a microplate reader (Bio-Rad, San Francisco, CA). All determinations were carried out in triplicate, and the data represent the means ± SEM for a minimum of three experiments.

Cell Migration Assays

For chemotactic migration assays, 24-well Transwell plates (Costar, Cambridge, MA) were used. The lower side of the Transwell filters (6.5-mm diameter, pore size 8.0 µm) were coated with TNfn3RAA dissolved in 250 µl of DMEM for 60 min at 37°C. After incubation with TNfn3RAA, filters were washed by adding 100 µl of PBS to the top well and 500 µl of PBS to the bottom well. After washing twice, filters were blocked with 1% BSA in DMEM for 30 min and again washed once in PBS. Cells were detached as described above and resuspended in DMEM at 5 × 10⁵ cells/ml. Migration and adhesion assays were performed at the same time, and the cells from the same dishes were used for both assays. Cells were incubated for 20 min on ice with or without the anti-91 antibody, Y9A2 (50 µg/ml), and then 100 µl were loaded (50,000 cells/chamber) in each chamber. Each chamber was inserted into a well containing 600 µl of DMEM supplemented with 1% FCS to serve as a chemoattractant and incubated at 37°C in humidified 5% CO₂ for 2 h for the CHO cells or 3 h for the MEF cells. Medium was then aspirated and the filters washed once with PBS. Cells on the bottom of the filters were fixed for 20 min in 500 µl of DifQuik fixative (Fisher, Springfield, NJ), and the nonmigrated cells on the top of the filter were gently removed with a Q-tip. Filters were allowed to completely dry, stained by DifQuik, washed in running distilled H₂0 and allowed to destain in distilled H₂O for 1 h. Filters were air-dried (3 h), removed from the chamber with a scalpel, and mounted onto glass slides with the use of a Permamount/xylene solution, and the migrated cells were counted. Migrated cells were counted under a 25× objective with the use of a gridded eyepiece (reticule). Ten high-powered fields (HPF) per slide were counted, the average was taken, and the number of migrated cells was expressed as migrated cells per 10 HPF. The data represent means ± SEM from a minimum of three experiments

Cell-spreading Assays

Glass coverslips (12-mm circle, Fisher) sterilized in 100% ethyl alcohol were placed into the wells of 24-well plates, washed twice with 500 µl of PBS, and coated with TNfn3RAA in 400 µl of serum-free DMEM for 60 min at 37°C. After incubation with TNfn3RAA, coverslips were washed twice with 500 µl of PBS and then blocked with 1% BSA in DMEM for 30 min at 37°C. Cells were detached (as described above) and resuspended in DMEM at 5 × 10⁵ cells/ml with 100 µl (50,000 cells), loaded onto coated coverslips, and incubated for either 3 h (MEF cells) or 6 h (CHO cells) at 37°C in humidified 5% CO₂. Medium was then removed, and the cells were washed in PBS once and fixed in 2% (freshly made) paraformaldehyde for 20 min at room temperature. Coverslips were mounted and analyzed under a 25× objective with the use of a reticule. Ten HPF per slide were observed, and the number of spread cells and the number of total cells were counted. Time-course experiments were first performed to determine the time required to display the greatest difference in cell spreading for each cell type studied. The difference in the rate of cell spreading for a given time was expressed as the number of spread cells/number of total cells in 10 HPF × 100. Data represent the mean percentages of spread cells ± SEM for a minimum of three experiments.

Coimmunoprecipitation and Western Blot Analysis

CHO cell lines expressing different chimeric 91 integrins were surface labeled with sulfo-N-hydroxysuccinimide-biotin (Pierce, Rockford, IL) according to the manufacturer's protocol. Cells were then lysed on ice for 30 min in an immunoprecipitation buffer: 20 mM Tris-HCl (pH 7.4), 150 mM NaCl, 10 mM EDTA, 10 mM benzamidine HCl, 0.02% sodium azide, 1% Triton X-100, 0.05% Tween 20, 2 mM phenylmethylsulfonyl fluoride, 5 µg/ml aprotinin, and 5 µg/ml leupeptin, as previously described (Liu et al., 1999). Briefly, cell lysates were clarified by centrifuging at 16,000 × g for 20 min at 4°C and then incubated with protein G-Sepharose coated with the anti-91 antibody, Y9A2, or an irrelevant mouse immunoglobulin (Ig) G at 4°C overnight. The beads were washed with the same buffer four times, and precipitated polypeptides were extracted with SDS sample buffer. Precipitated cell surface biotin-labeled polypeptides were separated by SDS-PAGE under nonreducing conditions and detected with streptavidin peroxidase followed by ECL (Amersham Pharmacia Biotech, Piscataway, NJ). In parallel, lysates of unmodified cells were precipitated with the anti-91 antibody, Y9A2, and coimmunoprecipitated paxillin was detected with biotin-labeled anti-paxillin antibodies (BD-Biosciences-Transduction Laboratories) as previously described (Liu and Ginsberg, 2000).

Integrin Cytoplasmic Domain Model Proteins and Affinity Chromatography

The design and recombinant production of cytoplasmic domain model proteins was performed as previously described (Liu et al., 1999; Liu and Ginsberg, 2000). Briefly, PCR was used to generate a HindIII-BamHI fragment for each wild-type or mutant integrin cytoplasmic domain, and this fragment was subcloned into the modified pET15b vector as previously described (Liu and Ginsberg, 2000). Recombinant proteins were expressed in BL21 (DE3) pLysS cells (Novagen, Madison, WI), isolated by Ni²⁺-charged resins, and further purified to >90% homogeneity with the use of a reverse phase C₁₈ high-performance liquid chromatography column (Vydac, Hesperia, CA). Masses of all proteins were assessed by electrospray ionization mass spectrometry on an API-III quadrupole spectrometer (Sciex, Toronto, Canada) and varied by <0.1% from the predicted masses. Recombinant integrin cytoplasmic tails were bound to Ni²⁺-charged His-Bind resins (Novagen) and used for affinity chromatography as previously described (Liu and Ginsberg, 2000). Briefly, 1 mg of each recombinant integrin tail dissolved in 5 ml of 20 mM 1,4-piperazinediethanesulfonic acid, 50 mM NaCl, pH 6.8 (PN buffer), plus 1 ml of 100 mM sodium acetate (pH 3.5) was bound to 100 µl of Ni²⁺-charged His-Bind resins (Novagen) at 4°C overnight. Resins were then washed with PN buffer twice and stored in an equal volume of PN buffer plus 0.1% sodium azide. The expression and isolation of recombinant human paxillin was performed as previously described (Liu et al., 1999; Liu and Ginsberg, 2000). Aliquots of recombinant HA-tagged glutathione S-transferase (GST)-paxillin were mixed with 300 µl of buffer A plus 20 µg/ml aprotinin, 5 µg/ml leupeptin, 1 mM phenylmethylsulfonyl fluoride, 0.1% Triton X-100, 3 mM MgCl₂, and 1 mg/ml BSA, added to integrin tail-loaded resins, and incubated at room temperature with rotation for 2 h (Liu et al., 1999; Liu and Ginsberg, 2000). Resins were washed three times with the same buffer, and bound proteins were extracted with SDS sample buffer, separated on SDS-PAGE, and detected with antibody specific for HA-tag.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

The 9 Cytoplasmic Domain Specifically Enhances Cell Migration

To determine whether the cytoplasmic domain of the 9 subunit specifically enhances cell migration, chimeric  subunits composed of the extracellular and transmembrane domain of the 9 subunit fused to the cytoplasmic domains of 4, 2, or 5 were constructed (Figure 1A). A comparison of the cytoplasmic sequences of 9, 4, 2, and 5 reveals that all 4 subunits contain a similar membrane-proximal region but that otherwise only the 9 and 4 sequences are similar. Overall, the 9 cytoplasmic domain is 52% homologous to the 4 cytoplasmic domain.

The full-length 9 subunit, the 94, 92, and 95 chimeras, and vector alone were stably expressed in CHO cells that do not express endogenous 9 and were examined for their ability to promote migration on the 91-specific ligand, TNfn3RAA. To determine whether the 9 constructs were expressed on the cell surface as 91 integrins and to sort cells expressing high levels of the 91 integrins, flow cytometry was performed with the anti-91 antibody, Y9A2 (Figure 2A). All stably transfected cells used in subsequent adhesion and migration assays expressed similar high surface levels of 91 integrins. Cell adhesion assays performed on TNfn3RAA (Figure 2, B and D) demonstrated that each chimera supports equivalent cell adhesion at both 3 µg/ml (Figure 2B) and 10 µg/ml (Figure 2D), and in each case adhesion was blocked by the anti-91 antibody, Y9A2. To determine whether the 9 cytoplasmic domain preferentially enhances migration compared with the 2 and 5 cytoplasmic domains, as had been previously demonstrated for the 4 cytoplasmic domain, chemotactic migration assays were performed on filters coated with either 3 µg/ml (Figure 2C) or 10 µg/ml (Figure 2E) of TNfn3RAA. As expected, wild-type 9 and each of the chimeras tested supported greater migration on TNfn3RAA than that seen in mock-transfected cells. However, both the 9 and the 4 cytoplasmic domains caused similar enhancement of cell migration compared with the cytoplasmic domains of 2 and 5. Migration of all 9-expressing cells was also inhibited by the anti-91 antibody, Y9A2, demonstrating that the enhanced migration was specific to the 91 integrins. These results indicate that the 9 cytoplasmic domain preferentially enhances cell migration and to the same level as the 4 cytoplasmic domain.

View larger version (45K): [in this window] [in a new window]   Figure 2.   Adhesion and migration of 9-expressing CHO cells. (A) Flow cytometric evaluation of cell surface expression of the 91 integrins on 9-, 9 chimera-, and mock-expressing CHO cells. Open peaks represent fluorescence (FL) of unstained CHO cells, and shaded peaks represent fluorescence of CHO cells stained with the anti-91 antibody, Y9A2. (B and D) 9-, 9 chimera-, or mock-expressing CHO cells were added to 96-well plates coated with either 3 µg/ml TNfn3RAA (B) or 10 µg/ml TNfn3RAA (D) after incubation with (below the dashed line) or without (above the dashed line) the anti-91 mAb, Y9A2. Cells were allowed to attach for 60 min, and nonadherent cells were removed by centrifugation. Adherent cells were stained with crystal violet and quantified by measurement of absorbance at 595 nm. (C and E) 9-, 9 chimera-, and mock-expressing CHO cells suspended in serum-free medium were seeded onto membranes coated with 3 µg/ml TNfn3RAA (C) or 10 µg/ml TNfn3RAA (E) in the upper well of 24-well plates after preincubation with (below the dashed line) or without (above the dashed line) the anti-91 mAb, Y9A2. After a 2-h incubation in the presence of 1% FCS in the bottom well, nonmigrated cells on the top side of the membrane were removed, and migrated cells on the bottom side of the membrane were fixed, stained, and counted with the use of a phase-contrast microscope in 10 HPF and expressed as number of migrated cells. Data (B-E) represent the means (±SEM) of triplicate experiments.

The Membrane-proximal 17 Amino Acids of the 9 Cytoplasmic Domain Are Sufficient to Mediate Enhanced Cell Migration

To identify the cytoplasmic sequences critical for mediating 9-dependent enhancement of cell migration, a series of 9DMs (Figure 1B) was stably expressed in CHO cells and examined for their ability to mediate 91-dependent adhesion and migration. All of the 9DMs were stably expressed on the cell surface at similarly high levels (Figure 3A). The cells expressing 9DM3-9DM6 all bound to TNfn3RAA-coated plates at similar levels and to the same level as cells expressing full-length 9 (Figure 3, B and D). The most severe truncations, 9DM1 and 9DM2, impaired 91-mediated adhesion, especially to 10 µg/ml TNfn3RAA. As expected from their lack of adhesion, 9DM1 and 9DM2 were unable to mediate 91-dependent migration to the same level as the full-length 9 subunit (Figure 3, C and E). In contrast, the 9DM4-9DM6 all mediated migration comparable to full-length 9. However, 9DM3 was unable to mediate 91-dependent migration, although it mediated adhesion to TNfn3RAA to the same level as the 9DM4-9DM6. These results indicate that 9-mediated enhancement of cell migration requires the 17 amino acids retained in the 9DM4 construct and is particularly sensitive to the loss of amino acids within the sequence IIEAEK that is present in 9DM4 but absent in 9DM3.

View larger version (50K): [in this window] [in a new window]   Figure 3.   Adhesion and migration of 9- and 9DM-expressing CHO cells. (A) Flow cytometric evaluation of cell surface expression of the 91 integrin from 9- and 9DM-expressing CHO cells as described in Figure 2. (B and D) Adhesion of 9- and 9DM-expressing CHO cells on 3 µg/ml TNfn3RAA (B) or 10 µg/ml TNfn3RAA (D) with or without preincubation with the anti-91 antibody, Y9A2, as described in Figure 2. (C and E) Migration of 9- and 9DM-expressing CHO cells on 3 µg/ml TNfn3RAA (C) or 10 µg/ml TNfn3RAA (E) with or without preincubation with the anti-91 antibody, Y9A2, as described in Figure 2. Data (B-E) represent the means (±SEM) of triplicate experiments.

The 9 Cytoplasmic Domain Inhibits Cell Spreading

The cytoplasmic domain of 4 has previously been demonstrated to inhibit cell spreading compared with other  subunits of the 1 subclass of integrins (Kassner et al., 1995). To determine whether the 9 cytoplasmic domain could also inhibit cell spreading, spreading assays were performed with cells expressing each of the constructs described above on coverslips coated with 10 µg/ml TNfn3AA (Figure 4). After 6 h, the greatest difference in cell spreading was evident with cells expressing the full-length 91 and the 94 chimera being less spread (~10% spread) than cells expressing either the 92 or 95 chimeras (~35% spread). Similarly, cells expressing the 9DM3, which mediates adhesion but not migration on TNfn3RAA, were spread to the same level as cells expressing the 92 and 95 chimeras that do not mediate enhanced migration. The 9DM4-9DM6 that mediate enhanced migration also inhibited cell spreading to similar levels as that of 91- and 941-expressing cells. These results indicate that the membrane-proximal 17 amino acids of the 9 cytoplasmic domain are sufficient to mediate both enhanced migration and impaired spreading and that the 9 and 4 cytoplasmic domains share both functional properties.

View larger version (37K): [in this window] [in a new window]   Figure 4.   Spreading of 9-, 9 chimera-, and 9DM-expressing CHO cells. Cells were seeded onto sterile coverslips coated with TNfn3RAA (10 µg/ml) and allowed to spread for 6 h at 37°C. Percentage of spreading was determined by phase-contrast microscopy. Data represent the means (±SEM) of triplicate experiments.

The 9 Cytoplasmic Domain Associates with the Focal Adhesion Adapter Protein, Paxillin

Recently, the 4 cytoplasmic domain was demonstrated to associate with and directly bind to the adapter protein, paxillin, and this interaction was shown to be critical for 4-dependent enhanced migration and impaired cell spreading (Liu et al., 1999). Because the 9 cytoplasmic domain is highly homologous to the 4 cytoplasmic domain (52% homology), we predicted that the 9 cytoplasmic domain would also associate with paxillin. Indeed, bacterially expressed GST-paxillin directly and specifically bound to the recombinant 9 cytoplasmic domain immobilized on a Ni²⁺ resin-charged column but not the IIb cytoplasmic domain in vitro (Figure 5A). To determine whether 9 binds paxillin with a similar affinity as 4, recombinant protein-binding assays were again performed. Both the 9 and 4 cytoplasmic domains specifically bound paxillin in a concentration-dependent manner with very similar binding affinities, suggesting that the strength of the 9-paxillin and 4-paxillin interactions is quite similar (Figure 5B). To determine whether paxillin associates with 9 in vivo and to the same level as 4, cell lysates from CHO cells expressing 91, 941, and 921 were immunoprecipitated with the anti-91 antibody, Y9A2. The precipitates were resolved with the use of SDS-PAGE and immunoblotted with an anti-paxillin antibody. Paxillin was coimmunoprecipitated with full-length 9 and the 94 chimera to similar levels but not with the 92 chimera (Figure 5C). These combined results demonstrate that the 9 cytoplasmic domain, like the 4 cytoplasmic domain, directly interacts with paxillin both in vitro and in vivo.

View larger version (28K): [in this window] [in a new window]   Figure 5.   Direct association of paxillin with 9. (A) HA-tagged recombinant GST-paxillin was added to Ni2+-charged resins loaded with the 9 or IIb cytoplasmic domains. Bound fractions were collected and separated on 4-20% SDS-PAGE under reducing conditions, transferred to a nitrocellulose membrane, and stained with antibody specific for the HA-tag, 12CA5. S.M., starting material. Depicted are results of one of three experiments performed with similar results. (B) HA-tagged recombinant GST-paxillin was added to Ni2+-charged resins loaded with the 9, 4, or the IIb cytoplasmic domains as described above. Depicted are results of one of three experiments performed with similar results. , 9; , 4; , IIb.(C) CHO cells stably expressing 91, 921, or 941 integrins were surface labeled with biotin and subjected to immunoprecipitation with the anti-91 antibody, Y9A2, or an irrelevant mouse IgG. The precipitates were separated on 4-20% SDS-PAGE and transferred to a nitrocellulose membrane. Paxillin coimmunoprecipitation was detected with an anti-paxillin antibody (top), and precipitated surface proteins were detected with streptavidin peroxidase followed by ECL (bottom). Depicted are results of one of three experiments performed with similar results.

The Binding of Paxillin to 9DM4 and 9DM5 Correlates with Enhanced Migration and Inhibition of Cell Spreading

To determine whether paxillin binds to the 9DM3-9DM5, recombinant protein-binding assays were performed as described above. Both 9DM4 and 9DM5 bound paxillin to similar levels and to the same level as the 9 cytoplasmic domain (Figure 6). However, the IIb cytoplasmic domain, as expected, and 9DM3 did not bind paxillin. The inability of 9DM3 to bind paxillin and promote enhanced migration and inhibit cell spreading suggests that an 9-paxillin interaction may be required for 91-dependent migration and inhibition of cell spreading. The fact that both 9DM4 and 9DM5 bind paxillin and support enhanced migration and inhibition of cell spreading further suggests that 9, like 4, may require paxillin to mediate enhanced migration and inhibit cell spreading.

View larger version (24K): [in this window] [in a new window]   Figure 6.   Association of paxillin with 9 deletion mutants. (A) Binding of recombinant paxillin to 9, 9DM3, 9DM4, 9DM5, or the IIb cytoplasmic domains as described in Figure 5. Depicted are results of one of two experiments performed with similar results. , 9DM3; , 9DM4; , 9DM5; , 9; , IIb.

The 9 Cytoplasmic Domain Mediates 91-dependent Migration in Paxillin Null Cells

To determine whether paxillin is required for 9-mediated enhanced migration, MEF cells null for paxillin (paxillin^/) were infected with retroviruses encoding 9, 94, 95, or vector alone and examined for their ability to promote migration on TNfn3RAA. As a control, wild-type MEF cells that contain endogenous paxillin were similarly infected. Neither the wild-type MEF nor the paxillin^/ MEF cells express endogenous 9. As in CHO cells, all constructs were surface expressed at similar high levels (Figure 7A), and all supported adhesion to TNfn3RAA (Figure 7, B and D). In the wild-type MEF cells, 9 and 94 mediated enhanced migration compared with 95 on both 3 µg/ml (Figure 7C) and 10 µg/ml (Figure 7E) TNfn3AA, as expected. In the paxillin^/ MEF cells, however, the 9 cytoplasmic domain mediated enhanced migration (Figure 7, C and E). At both 3 µg/ml (Figure 7C) and 10 µg/ml (Figure 7E) TNfn3RAA, the 91-expressing paxillin^/ MEF cells demonstrated enhanced migration, whereas the 94-expressing paxillin^/ MEF cells did not. Migration mediated by wild-type 9 and each 9 chimeric integrin was inhibited by the anti-91 antibody, Y9A2, in both the wild-type MEF and paxillin^/ MEF cells. Surprisingly, these findings suggest that paxillin is not required for 9-dependent enhancement of cell migration, as is the case for 4.

View larger version (49K): [in this window] [in a new window]   Figure 7.   Adhesion and migration of 9-expressing MEF and paxillin/ MEF cells. (A) Flow cytometric evaluation of cell surface expression of the 91 integrins from 9- and 9 chimera-expressing MEF cells (top row) and 9- and 9 chimera-expressing paxillin/ MEF cells (/; bottom row) as described in Figure 2. (B and D) Adhesion of 9- and 9 chimera-expressing MEF and paxillin/ MEF cells (/) on 3 µg/ml TNfn3RAA (B) or 10 µg/ml TNfn3RAA (D) with or without preincubation with the anti-91 antibody, Y9A2, as described in Figure 2. Note: twice the amount of the anti-91 antibody, Y9A2, was used to block adhesion of paxillin/ MEF cells (/) on 10 µg/ml TNfn3RAA. (C and E) Migration of 9- and 9 chimera-expressing MEF and paxillin/ MEF cells (/) on 3 µg/ml TNfn3RAA (C) or 10 µg/ml TNfn3RAA (E) for 3 h with or without preincubation with the anti-91 antibody, Y9A2, as described in Figure 2. Data (B-E) represent the means (±SEM) of triplicate experiments.

Paxillin Is Required for 91-dependent Inhibition of Cell Spreading

To determine whether paxillin is required for 9-mediated inhibition of cell spreading, MEF and paxillin^/ MEF cells stably infected with 9, 94, 95, or vector alone were analyzed in cell-spreading assays on 10 µg/ml TNfn3AA (Figure 8). After 3 h, the greatest difference in cell spreading was evident. The wild-type MEF cells expressing 95 were approximately twice as spread as cells expressing either 9 or 94, as expected. However, the paxillin^/ MEF cells expressing either 9 or 94 were at least as well spread as cells expressing 95. Thus, paxillin appears to be critical for both 9- and 4-dependent inhibition of cell spreading.

View larger version (35K): [in this window] [in a new window]   Figure 8.   Spreading of 9- and 9 chimera-expressing MEF and paxillin/ MEF cells. 9- and 9 chimera-expressing MEF cells and 9- and 9 chimera-expressing paxillin/ MEF cells (/) were seeded onto sterile coverslips coated with TNfn3RAA (10 µg/ml) and allowed to spread for 3 h at 37°C. Percentage of spreading was determined by phase-contrast microscopy. Data represent the means (±SEM) of triplicate experiments.

Point Mutations in the 9 Cytoplasmic Domain That Abolish Paxillin Binding Do Not Inhibit Cell Migration

Point mutations in the 4 cytoplasmic domain have previously been shown to both inhibit paxillin binding and abolish 4-mediated enhancement of cell migration and inhibition of cell spreading (Liu et al., 1999; Liu and Ginsberg, 2000). To confirm our results that an 9-paxillin interaction is not required for 9-dependent enhancement of cell migration, two point mutations were made in the 9 cytoplasmic domain, 9(W999A) and 9(W1001A), based on mutations in the 4 cytoplasmic domain shown to inhibit the 4-paxillin interaction (Figure 9A). Although, as shown in Figure 6, the region of the 9 cytoplasmic domain containing these residues is not required for paxillin binding, recombinant protein-binding studies indicated that the 9(W999A) and 9(W1001A) mutations abolished the 9-paxillin interaction in vitro and the interaction of 9 with the paxillin family member, Hic-5 (Young, Taooka, Liu, Askins, Yokosaki, Thomas, and Sheppard, unpublished results). Thus, these mutations were used as additional tools to evaluate the in vivo significance of the 9-paxillin interaction. To evaluate the effects of the 9(W999A) and 9(W1001A) mutations in vivo, 9(W999A) and 9(W1001A) were stably expressed in CHO cells. Both mutants were similarly expressed at the same high surface levels as wild-type 9 (Figure 9B). Immunoprecipitation studies performed with the use of the anti-91 antibody, Y9A2, demonstrate that both the W999A and W1001A point mutations in 9 dramatically inhibit coimmunoprecipitation of paxillin with 9 (Figure 9C). In functional assays, cells expressing either 9(W999A) or 9(W1001A) mediated adhesion (Figure 9D) and migration (Figure 9E) to TNfn3AA (10 µg/ml), as well as cells expressing the wild-type 91 integrin. These results confirm our earlier findings and indicate that an 9-paxillin interaction is not required for 91-dependent enhancement of cell migration.

View larger version (36K): [in this window] [in a new window]   Figure 9.   Adhesion and migration of CHO cells expressing 9 cytoplasmic domain mutations that abolish paxillin binding. (A) The cytoplasmic amino acid sequences of 9, 9(W999A), and 9(W1001A) are shown. Amino acids depicted in bold type represent sites of mutagenesis. (B) Flow cytometric evaluation of cell surface expression of the 91 integrins from 9-, 9(W999A)-, and 9(W1001A)-expressing CHO as described in Figure 2. (C) CHO cells stably expressing 91 or the 9(W999A)1- or 9(W1001A)1-chimeric integrins were surface labeled with biotin and subjected to immunoprecipitation (IP) with the anti-91 antibody, Y9A2, or an irrelevant mouse IgG as described in Figure 5. Depicted are results of one of three experiments performed with similar results. (D) Adhesion of 9-, 9(W999A)-, and 9(W1001A)-expressing CHO cells on 10 µg/ml TNfn3RAA with or without preincubation with the anti-91 antibody, Y9A2, as described in Figure 2. (E) Migration of 9-, 9(W999A)-, and 9(W1001A)-expressing CHO cells on 10 µg/ml TNfn3RAA with or without preincubation with the anti-91 antibody, Y9A2, as described in Figure 2. Data (D-E) represent the means (±SEM) of triplicate experiments.

91-dependent Inhibition of Cell Spreading Is Dependent on an 9-Paxillin Interaction

To confirm that an 9-paxillin interaction is required for 91-dependent inhibition of cell spreading, the 9 mutants, 9(W999A) and 9(W1001A), that do not bind paxillin described above were analyzed in a cell-spreading assay on 10 µg/ml TNfn3AA (Figure 10). After 6 h, the greatest difference in cell spreading was evident. Cells expressing wild-type 9 were less spread (~12%) than cells expressing the 95 chimera (~30%), as expected. However, cells expressing the 9 mutants, 9(W999A) and 9(W1001A), were even more spread (~50%) than cells expressing the 95 chimera, indicating that an 9-paxillin interaction is required for 9 to inhibit cell spreading. These results confirm our earlier findings and suggest that an 9-paxillin interaction is required for 91-dependent inhibition of cell spreading. In addition, these combined results (Figures 7-10) suggest that 9 may use different intracellular signaling pathways to promote enhanced migration and inhibition of cell spreading.

View larger version (32K): [in this window] [in a new window]   Figure 10.   Spreading of 9-, 9(W999A)-, and 9(W1001A)-expressing CHO cells. 9-, 9(W999A)-, and 9(W1001A)-expressing CHO cells were seeded onto sterile coverslips coated with TNfn3RAA (10 µg/ml) and allowed to spread for 6 h at 37°C. Percentage of spreading was determined by phase-contrast microscopy. Data represent the means (±SEM) of triplicate experiments.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

In this study, we demonstrate that the 9 cytoplasmic domain specifically promotes cell migration and inhibits cell spreading. In these experiments, all performed on a ligand specific for the 91 extracellular domain, TNfn3RAA, the 9 cytoplasmic domain preferentially enhanced cell migration and inhibited cell spreading compared with the cytoplasmic domains of 2 or 5. These results were nearly identical to those reported previously for the 4 cytoplasmic domain utilizing similarly constructed chimeric integrin subunits containing the extracellular domain and transmembrane domains of either the 2 or 4 subunit (Chan et al., 1992; Kassner et al., 1995). As expected, based on these earlier studies, the 4 cytoplasmic domain also enhanced migration and inhibited spreading in our study. These results, together with recent studies demonstrating that 91 and 4 integrins are both critical to transendothelial leukocyte migration (Issekutz et al., 1996; Gao and Issekutz, 1997; Taooka et al., 1999), establish 9 and 4 integrins as members of a functional subfamily of integrins. This classification is further supported by the sequence similarity between 9 (Palmer et al., 1993) and 4 (Takada et al., 1989) and by several recent reports of overlapping ligand-binding specificity for the 9 and 4 integrins (Taooka et al., 1999; Eto et al., 2000