To determine whether expression of β1 integrin on the cell surface requires its intracellular interaction with the actin cytoskeleton, β1 integrin was immunoprecipitated with the anti-β1 integrin antibody or control IgG from HBE or MCF-7 cells which adhered to collagen IV. Western blot analysis revealed that actin coprecipitated with β1 integrin but not with the IgG precipitates from quiescent HBE cells (Fig. 1A). In contrast, actin did not coprecipitate with β1 integrin or the IgG precipitates from adherent MCF-7 cells (Fig. 1A). To determine whether the failure to coprecipitate actin with β1 integrin was due to the absence of actin expression, whole cell lysates were prepared from HBE or MCF-7 cells, both of which were quiescent and adherent to collagen IV. Immunoblot analysis indicated that total amounts of actin and β1 integrin were comparable between HBE and MCF-7 cells (Fig. 1B).

Interaction between β1 integrin and the actin cytoskeleton is mediated by the actin-binding proteins which are involved in the focal adhesion complex where integrin β1 through its cytoplasmic domain links with the actin cytoskeleton. We next determined which actin-binding proteins mediate this linkage between β1 integrin and actin. Immunoblot analysis of the β1 integrin or control IgG precipitates for α-actinin, talin and vinculin, revealed that only α-actinin among three proteins coprecipitated with β1 integrin but not with the IgG precipitates from quiescent HBE cells (Fig. 2A). Similar to HBE cells, α-actinin, but not talin or vinculin, coprecipitated with β1 integrin but not with the IgG precipitates from quiescent MCF-7 cells (Fig. 2A). Western blotting for the whole cell lysates revealed that total amounts of α-actinin and vinculin in HBE and MCF-7 cells were comparable and that the amount of talin was slightly smaller in MCF-7 cells than that in HBE cells (Fig. 2B). The results indicated that α-actinin coprecipitated with β1 integrin in both HBE and MCF-7 cells.

The cytoplasmic domain of β1 integrin is short and has the putative binding domains for α-actinin 18. To determine the difference in the cytoplasmic domain of β1 integrin between HBE and MCF-7 cells, total RNA was extracted from the cells and the cDNA that covers the full-length of the cytoplasmic domain of β1 integrin was amplified. Reverse transcription-polymerase chain reaction (RT-PCR) products from HBE or MCF-7 cells were identical in lengh with 462-bp on agarose gels (Fig. 3A). Sequence analysis of the cDNA fragments of HBE and MCF-7 cells revealed that both fragments were identical to the wild-type cytoplasmic domain of β1 integrin isoform A (Fig. 3B), indicating that there is no difference in the cytoplasmic domain of β1 integrin between HBE and MCF-7 cells.

The importance of β1 integrin phosphorylation to its functions which include cell adhesion and invasion has been suggested by many investigations 19202223. To determine this possibility, we examined the phosphorylation state of β1 integrin in HBE and MCF-7 cells. Metabolic labeling of cells with [³²P]orthophosphoric acid indicated that β1 integrin (110 kDa), and not α-actinin (100 kDa), was phosphorylated in either HBE and MCF-7 cells (Fig. 4A). When the β1 integrin immunoprecipitates were probed with anti-phosphotyrosine (PY) antibody, β1 integrin, and not α-actinin, reacted with the antibody in both HBE and MCF-7 cells (Fig. 4B). This indicates that β1 integrin is phosphorylated at least at tyrosine in MCF-7 cells as well as in HBE cells.

To determine the role of phosphorylation of β1 integrin in the link formation between β1 integrin and actin, the β1 integrin immunoprecipitates were dephosphorylated by PTP or PP. Treatment of the β1 integrin immunoprecipitates from HBE cells with PTP resulted in tyrosine dephosphorylation of β1 integrin as assessed by the anti-PY antibody and concomitant dissociation of actin from it (Fig. 5A). In contrast, tyrosine dephosphorylation of β1 integrin from MCF-7 cells by PTP did not affect the absence of coprecipitation of actin with β1 integrin but successive incubation of the PTP-treated β1 integrin with the supernatant of the β1 immunoprecipitates that contained unbound actin or with exogenous human platelet actin caused coprecipitation of actin with β1 integrin (Fig. 5A). The result indicated that β1 integrin in either cells was tyrosine phosphorylated. On the other hand, dephosphorylation of the β1 integrin immunoprecipitates from either cells by PP2A₁ did not result in any alteration in coprecipitation of two proteins or the reactivity of β1 integrin with the anti-PY antibody (Fig. 5B). Actin remained coprecipitated with β1 integrin from HBE cells, while it remained dissociated from β1 integrin from MCF-7 cells. However, incubation of the PP2A₁-treated β1 integrin from MCF-7 cells with the supernatant of the β1 immunoprecipitates or with exogenous human platelet actin caused coprecipitation of actin with β1 integrin (Fig. 5B). The results indicated that β1 integrin in HBE cells is not serine/threonine phosphorylated but in MCF-7 cells.

Loss of linkage between the cell-cell adhesion molecule E-cadherin and the actin cytoskeleton, that is mediated by a three-protein link of β-catenin, α-catenin and α-actinin, is induced by association of the EGF receptor having the protein tyrosine kinase activity with β-catenin that leads to tyrosine phosphorylation of β-catenin and concomitant dissociation of α-catenin from it 2425. By analogy with this phenomenon, the β1 integrin immunoprecipitates were probed with the antibodies to several serine/threonine protein kinases to determine whether serine/threonine protein kinase associates with β1 integrin. Among these, the anti-CaMKII antibody, which reacts broadly with all CaMKII isoforms, reacted with doublet of proteins having molecular mass of around 56 kDa in the β1 integrin immunoprecipitates but not in the control IgG precipitates from MCF-7 cells (Fig. 6A). These proteins which were reactable with the anti-CaMKII antibody were not detected in the β1 integrin or control IgG precipitates from HBE cells (Fig. 6A). To confirm the CaMKII kinase activity of the doublet of proteins, in vitro kinase assay was carried out using a specific substrate peptide Autocamtide II, in the presence of calmodulin and inhibitor peptides of cAMP-dependent protein kinase (PKA) and protein kinase C (PKC). Whereas the β1 integrin immunoprecipitates from HBE cells exhibited a very low level of the kinase activity that was slightly above that of the precipitates without anti-β1 integrin antibody, the relative kinase activity of the immunoprecipitates from MCF-7 cells was 17-fold (Student's t-test, p < 0.01) above that of the β1 integrin immunoprecipitates from HBE cells (Fig. 6B). When 10 μM KN-62, a specific inhibitor of CaMKII 262728, was present, the kinase activity of the β1 integrin immunoprecipitates from HBE cells was 7-fold above (p < 0.05) that in the absence of KN-62 (Fig. 6B). Contrary to this, 10 μM KN-62 significantly inhibited the kinase activity of the β1 integrin immunoprecipitates from MCF-7 cells by 92% (p < 0.01) (Fig. 6B).

HBE cells obtained form Clonetics were maintained in serum-free defined medium MCDB 170 30 supplemented with 10 ng/ml EGF, 1.4 μM hydrocortisone, 5 μg/ml insulin, 0.1 mM ethanolamine, 0.1 mM phosphoethanolamine, and 25 μM prostaglandin E2 24 at 37° C in a humidified atmosphere of 95% air and 5% CO₂. MCF-7 cells were maintained in 5% fetal bovine serum (FBS)-containing RPMI 1640 medium. Subconfluent culture of HBE or MCF-7 cells were harvested with 0.25% (w/v) trypsin-2.65 mM EDTA solution and the cells were replated on 60-mm collagen IV-coated dishes (Becton-Dickinson). HBE or MCF-7 cells were then growth arrested by incubation in an EGF-deprived medium for 3 days or in 0.1% FBS-containing medium for 2 days, respectively.

Cells adherent to collagen type IV were lysed in RIPA buffer (10 mM Tris-HCl, pH 7.4, 0.15 M NaCl, 1% sodium deoxycholate, 1% Triton X-100, 0.1% SDS, 1 mM phenylmethylsulfonyl fluoride (PMSF), and 1 mM sodium orthovanadate) 25 and sonicated briefly using an ultrasonic cell disruptor (Misonix Inc.). The cell lysates were incubated with the anti-integrin β1 monoclonal antibody (Life Technologies) or control IgG followed by Protein A-Sepharose (Amersham Pharmacia). The immunoprecipitates were washed three times with PB buffer (20 mM phosphate buffer, pH 8.6, 0.5% Nonidet P-40, and 0.1% SDS) and boiled for 2 min in lysis buffer (0.0625 M Tris-HCl, pH 6.8, 2% SDS, 10% glycerol, and 5% 2-mercaptoethanol). Whole cell lysates were prepared by homogenizing cells in 1% SDS in 20 mM Tris-HCl, pH 7.4, 1 mM PMSF and 1 mM sodium orthovanadate, and then boiling for 2 min, before sonicating briefly. The protein content was determined using a protein reagent kit (Pierce) with bovine serum albumin (BSA) as the standards. The immunoprecipitates or whole cell lysates were dissolved in SDS-polyacrylamide gel electrophoresis (PAGE) and after electrophoresis the samples were transferred onto membranes (Millipore). The blots were probed with anti-actin (ICN), anti-β1 integrin, anti-α-actinin (Upstate Biotechnology), anti-talin (Chemicon), anti-vinculin (Sigma), anti-phosphotyrosine (PY) (Upstate Biotechnology), or anti-CaMKII antibody (Santa Cruz Biotechnology). After incubation of the membrane with horseradish peroxidase-conjugated secondary antibodies (Amersham Pharmacia), the reactivity was visualized using an ECL kit (Amersham Pharmacia) and X-ray films (Fuji-film).

Total RNA was extracted from unstimulated cells which adhered to collagen IV, using TRIzol reagent and then digested with DNase I (Life Technologies). A one-step RT-PCR was carried out using an RT-PCR kit (CLONTECH) at the anealing temperature of 55°C. The primer set for the cytoplasmic domain of β1A integrin was 5'-TCCTATTTTAACATTACCAA-3' and 5'-ACTGTGACTATGGAAATTGC-3' 16. The PCR products were electrophoresed on 1.6% (w/v) agarose gels and the cDNA was extracted from the gel bands. The DNA sequence was confirmed for both forward and reverse strands of cDNA fragments using the set of PCR primers on an automated DNA sequencer (ABI PRISM 3100, Applied Biosystems).

Adherent cells to collagen IV-coated dish were labeled with 100 μCi/ml of [³²P]orthophosphoric acid (Amersham Pharmacia) in phosphate-free medium for the last 24 h during incubation to arrest the cells. Immunoprecipitation of β1 integrin and SDS-PAGE of the immunoprecipitates were carried out as described above. After electrophoresis, gels were dried under vacuum and autoradiographed using X-ray films at -70°C for a weak.

Immunoprecipitation of β1 integrin was carried out as described above and the supernatant obtained by centrifugation of the immunoprecipitates with the Protein A-Sepharose beads was retained at room temperature. The β1 integrin immunoprecipitates were washed twice with PB buffer and once with phosphatase buffer (50 mM Tris-HCl, pH 7.0, 150 mM NaCl, and 0.1 mM EDTA). Then the immunoprecipitates were treated with or without 1.0 unit of PTP (LAR, Calbiochem) 313233 for 10 min or 0.1 units of serine/threonine-specific PP (PP2A₁, Calbiochem) 343536 for 15 min in 100 μl of phosphatase buffer at 30°C. In some experiments, the phosphatase-treated immunoprecipitates were washed with RIPA buffer and incubated with the retained supernatant of the β1 immunoprecipitates that contained unbound actin or with 5 μg/ml human platelet actin (Cytoskeleton, Inc.) in RIPA buffer containing 0.2 mM CaCl₂ for 30 min at room temperature.

Immunoprecipitation of β1 integrin was carried out as described above, and the β1 integrin immunoprecipitates were washed twice with PB buffer and once with 20 mM phosphate-buffered saline, pH 7.2. In vitro kinase assay for β1 integrin-associated CaMKII was carried out at 30°C for 10 min using [γ-³²P]ATP (NEN Life Science Products) and a CaMKII assay kit (Upstate Biotechnology) that included specific substrate peptide Autocamtide II, calmodulin, and inhibitor peptides of PKA and PKC. In some experiments, the kinase assay was done in the presence of KN-62 (1-[N,O-Bis(5-isoquinolinesulfonyl)]-N-methyl-L-tyrosyl-4-phenylpiperazine) (Sigma), a specific inhibitor of CaMKII 262728 at the indicated concentrations. The background kinase assay was carried out using the precipitates with the Protein A-Sepharose beads but without the antibody to β1 integrin. Radioactivity incorporated into the substrate peptide was determined with a liquid schintillation counter. After subtraction of the background, the activity was normalized to the amount of β1 integrin that was quantified by measurement of the band intensity of β1 integrin on SDS-PAGE gels using an image analyzer equipped with a digital camera (EDAS 290, Kodak).