Lu/BCAM (Lutheran/basal cell-adhesion molecule) is a laminin 511/521 receptor expressed in erythroid and endothelial cells, and in epithelial tissues. The RK573–574 (Arg573-Lys574) motif of the Lu/BCAM cytoplasmic domain interacts with αI-spectrin, the main component of the membrane skeleton in red blood cells. In the present paper we report that Lu/BCAM binds to the non-erythroid αII-spectrin via the RK573–574 motif. Alanine substitution of this motif abolished the Lu/BCAM–spectrin interaction, enhanced the half-life of Lu/BCAM at the MDCK (Madin–Darby canine kidney) cell surface, and increased Lu/BCAM-mediated cell adhesion and spreading on laminin 511/521. We have shown that the Lu/BCAM–spectrin interaction mediated actin reorganization during cell adhesion and spreading on laminin 511/521. This interaction was involved in a laminin 511/521-to-actin signalling pathway leading to stress fibre formation. This skeletal rearrangement was associated with an activation of the small GTP-binding protein RhoA, which depended on the integrity of the Lu/BCAM laminin 511/521-binding site. It also required a Lu/BCAM–αII-spectrin interaction, since its disruption decreased stress fibre formation and RhoA activation. We conclude that the Lu/BCAM–spectrin interaction is required for stress fibre formation during cell spreading on laminin 511/521, and that spectrin acts as a signal relay between laminin 511/521 and actin that is involved in actin dynamics.
- adhesion molecule
- cell adhesion
- Lutheran/basal cell-adhesion molecule (Lu/BCAM)
- Madin–Darby canine kidney (MDCK)
Lu (Lutheran)/BCAM (basal cell-adhesion molecule) glycoprotein is a member of the IgSF (Ig superfamily) that carries both the Lu blood group antigens and the BCAM tumoral antigen. It is represented by two isoforms, Lu and Lu(v13), that only differ by the length of their cytoplasmic domain (59 compared with 19 amino acids) . The specific 40 amino acids at the cytoplasmic end of Lu glycoprotein comprise phosphorylation sites  and an SH3 (Src homology 3)-binding motif, consistent with a receptor signalling function. Lu/BCAM represents a unique receptor for laminin 511/521 (comprising α5β1γ1 and α5β2γ1 chains respectively)  in normal and sickle RBCs (red blood cells) [4–6]. In human blood cells, Lu/BCAM is expressed only in erythrocytes. Lu/BCAM is also present at the basal layer of epithelia, on the surface of a subset of muscle cells and on blood vessel endothelium [7–9]. In these tissues, Lu/BCAM is recognized as a co-receptor with integrins for laminin 511/521 . Disruption of the Lu gene in mouse provided evidence that Lu/BCAM was involved in the maintenance of normal basement-membrane organization in kidney and intestine .
In RBCs, Lu/BCAM interacts directly with spectrin through the juxtamembrane RK573–574 (Arg573-Lys574) motif of its cytoplasmic tail [11–13]. Identified at the inner membrane of RBCs, spectrins are considered as the central components of a ubiquitous and complex spectrin–actin scaffold, called the spectrin-based skeleton . This network, attached to diverse cellular membranes, is involved in diverse functions, including the resilience and stability of membranes, the establishment of specialized membrane domains and in vesicle trafficking. Spectrins exist as elongated flexible heterotetramers of 200 nm made up of two α- and β-subunits that constitute the filaments of the network, the nodes of which are cross-linked by actin filaments. In mammals, two genes encode for α-chains (αI and αII) and five for β-chains. Although αI-spectrin is the only isoform expressed in mature RBCs, αII-spectrin is the most common form in nucleated cells. Moreover, αI-spectrin is not expressed in epithelial cells. Each spectrin subunit is organized as an alignment of spectrin repeats, made up of three α-helices each; α-spectrins contain 20 spectrin repeats. In RBCs, the Lu/BCAM-binding site in αI-spectrin has been delimited to the single α4 repeat . As Lu/BCAM represents a minor component of the RBC membrane (from 1.5 to 4 copies/RBC), we have speculated that the Lu/BCAM–spectrin interaction might be critical for cell signalling rather than for maintenance of the membrane mechanical properties .
The adhesion properties of Lu/BCAM depend on its cytoplasmic domain phosphorylation, indicating an inside–out activating signal for its laminin 511/521 receptor function . The physiological stress mediator adrenaline (epinephrine), acting through the β2-adrenergic receptor, increases Lu/BCAM-mediated adhesion of sickle RBCs to laminin 511/521 through a PKA (protein kinase A)-dependent pathway . Lu/BCAM phosphorylation by PKA occurs in sickle, but not in normal, RBCs, concomitantly with an enhanced cell adhesion to laminin 511/521 under physiological conditions . In polycythaemia vera patients, who present an increased risk of thrombosis, Lu/BCAM is constitutively phosphorylated, and this feature is associated with an increased RBC adhesion to endothelial laminin 511/521 .
In resealed RBC ghosts, disruption of the Lu/BCAM–spectrin interaction by a spectrin peptide encompassing the α4 repeat resulted in a weakened linkage of Lu/BCAM to the spectrin-based skeleton, and induced cell adhesion to laminin 511/521 . Increased adhesion of RBCs from HS (hereditary spherocytosis) patients with a marked deficiency in spectrin, demonstrated the biological relevance of this interaction: intact HS RBCs exhibited reinforced adhesion to laminin 511/521 under physiological conditions that resulted, at least in part, from an impaired interaction between Lu/BCAM and the membrane skeleton . All of these observations indicate that the Lu/BCAM interaction with erythroid spectrin negatively regulates cell adhesion to laminin 511/521.
In the present study, we have shown that Lu/BCAM bound in vitro and ex vivo to non-erythroid αII-spectrin in kidney epithelial cells. This interaction modulated spreading and adhesion of MDCK (Madin–Darby canine kidney) cells to laminin 511/521. Most importantly, we have shown that the Lu/BCAM–spectrin interaction was involved in a laminin 511/521-to actin signalling pathway promoting actin skeleton reorganization.
The protease inhibitor cocktail, microspin GST (glutathione transferase) purification module and Protein A–Sepharose were purchased from Amersham Biosciences. Br-cAMP was from Calbiochem. Primers used in PCR and mutagenesis experiments were from MWG Biotech. The in vitro transcription and translation TNT® T7 quick kit for PCR-amplified DNA was obtained from Promega. Dako was from DakoCytomation. PIC (phosphatase inhibitor cocktail) for serine/threonine phosphatase, a purified human laminin 511/521 mixture, fibronectin and 30% (w/v) BSA were supplied by Sigma. Lab-Tek® II chamber slides™ were from Nalge Nunc. Sulfo-NHS-LC-biotin [sulfo-succinimidyl-6-(biotinamido)hexanoate] and immunopure immobilized streptavidin beads were obtained from Pierce. G-LISA™ RhoA, Rac and Cdc42 (cell division cycle 42) Activation Assay Biochem Kit™ (absorbance-based) and cell-permeant Rho inhibitor (exoenzyme C3 transferase, CT04) were from Cytoskeleton. Image-iTFX signal enhancer, calcein/AM (calcein acetoxymethyl ester), bis-benzimidazole Hoechst 33342, AlexaFluor® 488 Phalloidin and NuPAGE® Novex Bis-Tris gels were purchased from Invitrogen. Cell culture media and reagents were from Gibco BRL. Except when otherwise mentioned, reagents were purchased from Sigma–Aldrich. Transwell polycarbonate filters (12 mm diameter, 0.4 μm pore) were obtained from Costar. Millipore electrical resistance apparatus was from Bedford. The bio-imaging analyser was obtained from Fujifilm-Raytest. An evolution VF camera was obtained from Mediacybergetics. A 490 nm Multiskan RC was purchased from Lab Systems.
Antibodies directed against Lu/BCAM were used: mAb (monoclonal antibody) clone F241 and the rabbit pAb (polyclonal antibody) 602 (collaboration with Dr D. Blanchard, EFS, Nantes and INTS, Paris, France). The immunopurified pAb directed against the αII-spectrin SH3 domain was produced in our laboratory . Other antibodies used were: goat biotinylated anti-Lu/BCAM antibody (R&D systems), sheep anti-human Ubc9 (ubiquitin-conjugating enzyme 9; AG Scientific), and Alexa Fluor® 488-conjugated anti-mouse and Alexa Fluor® 568-conjugated anti-rabbit antibodies (Molecular Probes, Invitrogen).
Human kidney carcinoma epithelial cells A498 (A.T.C.C., HB44) were maintained in MEM (minimal essential medium) Glutamax I supplemented with 10% FBS (fetal bovine serum), 100 units/ml antibiotic/antimycotic, 1 mM sodium pyruvate and 0.1 mM non-essential amino acids. MDCK cells (A.T.C.C., CCL-34) were maintained in DMEM (Dulbecco's modified Eagle's medium) Glutamax I supplemented with 10% FBS, 100 units/ml antibiotic/antimycotic and 0.1 mM non-essential amino acids. Both cell lines were grown in a humidified atmosphere at 37 °C with 5% CO2.
The full-length wt (wild-type) cDNA sequence of Lu/BCAM (wt-Lu) and bearing RK573-574AA (Arg573→Ala/Lys574→Ala; mt-Lu) (+1 is taken as the methionine residue of the initiation codon) were cloned into the pcDNA3 expression vector as described previously [12,20]. A D343R mutation (+1 is taken as the methionine residue of the initiation codon; the mutated residue is the same as residue 312 in reference  where +1 was taken as the first residue of the mature protein) was obtained by in vitro mutagenesis using the sense primer 5′-AGAGTGGAGGATTACCGGGCGGCAGATGACGTG-3′, and the antisense primer 5′-CACGTCATCTGCCGCCCGGTAATCCTCCACTCT-3′. Stable MDCK cells expressing wt and mutant Lu/BCAM, were obtained and amplified as described previously . To obtain clones showing similar Lu antigen membrane expression [SABC (specific antibody-binding capacity) units], cell clones were selected by serial limit dilution . The SABC for wt-Lu and for mt-Lu were 456.337 and 401.887 respectively.
Protein extraction from MDCK cells
After trypsin treatment, the pellet corresponding to 2×106 MDCK cells was incubated for 1 h at 4 °C in 400 μl of lysis buffer [20 mM Tris/HCl (pH 8), 150 mM NaCl, 5 mM EDTA, 0.2% BSA and 0.010–0.030% Triton X-100 supplemented with protease inhibitor cocktail]. After centrifugation (15000 g for 15 min at 4 °C), the soluble cell extract (S) was loaded on to a 4–12% NuPAGE® Novex Bis-Tris gel under reducing conditions. Proteins were analysed by Western blot using a goat biotinylated anti-Lu/BCAM antibody (0.1 μg/ml) and sheep anti-human Ubc9 antibody (1:3000 dilution) as an extraction control.
Cells (1×107) were lysed for 30 min on ice in lysis buffer [20 mM Tris/HCl (pH 8), 500 mM NaCl, 2 mM MgCl2 and 1% Triton X-100 supplemented with protease inhibitor cocktail]. Cell lysates were then sonicated (30 s pulse ‘high’, 30 s rest) and centrifuged for 10 min at 11000 g.
Lu/BCAM was immunoprecipitated by incubating the supernatants with mouse anti-Lu mAb F241, an irrelevant antibody (mAb anti-IgG1) or only with Protein A–Sepharose as a negative control overnight at 4 °C. The beads were centrifuged at 11000 g for 20 min, and were washed five times with wash buffer 1 [20 mM Tris/HCl (pH 7.4), 600 mM NaCl, 2 mM MgCl2 and 1% Triton X-100 supplemented with protease inhibitor cocktail] and twice with buffer 2 [20 mM Tris/HCl (pH 7.4)] to remove traces of detergent. Immunoprecipitated proteins were first eluted with glycine/HCl buffer (pH 3.2) and then equilibrated in Tris/HCl buffer (pH 9) for incubation for 5 min. Samples were loaded on to 4–12% NuPAGE® Novex Bis-Tris gels under reducing conditions. Western blot analyses were performed using the anti-αII-spectrin SH3 domain pAb (1:100000 dilution) or goat biotinylated anti-Lu/BCAM antibody (at 0.1 μg/ml).
GST pull-down assays
GST peptides were produced as described previously  and immobilized on Sepharose-4B–glutathione bead microspin columns. Briefly, PCR-amplified cDNA fragments encoding the C-terminal ends of Lu/BCAM (residues 569–628), were fused with the GST protein in the pGEX-5X-3 plasmid. mt-Lu RK573-574AA was obtained by in vitro mutagenesis as described previously . Spectrin repeat units corresponding to αIR3 amino acids (Ala253-Gly384), αIIR3 (Gln255-Leu382),h αIR4 (Ser361-Glu486) and αIIR4 (Glu361-Glu486), αIR3-5 (Ala253-Leu594), αIIR3-5 (Gln255-Leu594), αIR5 (Asp466-Leu594) and αIIR5 (Leu467-Leu594) were obtained as a [35S]methionine-labelled protein with the in vitro transcription and translation TNT® T7 quick kit for PCR-amplified DNA.
The products (50 μl) were incubated overnight at 4 °C with GST–wt-Lu and GST–mt-Lu constructs (100 μg) immobilized on Sepharose 4B–glutathione beads equilibrated in PBS and antiproteases. After five washes with PBS/0.05% Tween 20, the bound proteins were solubilized in Laemmli sample buffer , loaded on to a 4–12% NuPAGE® Novex Bis-Tris gel, visualized and quantified by Kodak ID Image analysis software.
Confocal fluorescence microscopy and immunofluorescence of transfected cells
The transepithelial resistance was measured daily on stably transfected MDCK cells grown on polycarbonate filters (5×105 cells/filter). At day 7, polarized cells were fixed for 20 min with 4% paraformaldehyde, treated with 50 mM ammonium chloride in PBS, permeabilized for 10 min with 0.5% Triton X-100 (in PBS) and incubated with the anti-Lu mAb F241 (1:10 dilution) and the anti-αII-spectrin SH3 domain pAb (1:500 dilution) in Dako for 1 h at room temperature (22 °C). Filters were washed with PBS/0.5% BSA and incubated with AlexaFluor® 488- and 568-conjugated anti-mouse and anti-rabbit antibodies (1:200) for 1 h at room temperature in PBS/0.5% BSA. Samples were examined by confocal microscopy using a Nikon EC-1 system equipped with 60×NA (numerical aperture) 1.4 and 100×1.30 objectives.
Lab-Tek® II Chamber Slides™ were coated with 2 μg/cm2 of either laminin 511/521 or fibronectin overnight at 4 °C (in 400 μl). Wells were then washed twice with PBS and were subsequently coated with 1% BSA at 37 °C for 1 h before two additional washes. Mock-transfected MDCK cells (104) or transfected cells expressing wt-Lu, mt-Lu, the D343R Lu mutant or Lu(v13) were added to the wells (in 400 μl). Cells were incubated in serum-free medium for 90 min at 37 °C in the presence or absence of 1 μg/ml per well of the cell-penetrating form of the Clostridium botulinum C3 toxin, an ADP ribosyltransferase that selectively ribosylates Rho proteins, rendering them inactive. After this incubation time, cells were fixed for 20 min with 4% paraformaldehyde, treated with 50 mM ammonium chloride in PBS, permeabilized for 10 min with 0.5% Triton X-100 (in PBS) and saturated for 30 min in Image-iTFX signal enhancer solution. Cells were washed with PBS/0.5% BSA and incubated with Phalloidin AlexaFluor® 488 (1:50 dilution) for 1 h at room temperature. Samples were examined by confocal microscopy.
Cell-surface biotinylation assay
Polarized MDCK monolayers expressing either wt-Lu or mt-Lu were cultured on filters as described above. Newly synthesized proteins were labelled by adding 150 μCi of [35S]methionine/[35S]cysteine in the cell culture medium for 20 min at 37 °C. After washing with complete medium, cells were incubated in non-radioactive medium for 60, 90, 120, 150 and 180 min at 37 °C. After each incubation time, cells were washed twice with ice-cold PBS and incubated on the basolateral side with 0.5 mg/ml Sulfo-NHS-LC-biotin. After immunoprecipitation of total Lu/BCAM, half of the eluted proteins was diluted with 3× Laemmli buffer (total Lu/BCAM), and the other half was incubated in 500 μl of lysis buffer for 3 h with immunopure immobilized streptavidin beads to isolate membrane Lu/BCAM. Beads were washed three times with lysis buffer and Lu/BCAM was eluted in 20 μl of Laemmli buffer for 5 min at 100 °C. Eluates from both steps (total and membrane Lu/BCAM) were analysed by SDS/PAGE (8% gels) under reducing conditions, followed by Western blotting with the rabbit anti-Lu pAb 602 (1:5000 dilution). Samples of biotinylated Lu proteins were analysed using a bio-imaging analyser to quantify the newly delivered membrane proteins as a function of time.
The 12-wells plates were coated with either 2 μg/cm2 of laminin 511/521 or 1% BSA overnight at 4 °C. Wells were washed twice with PBS and were subsequently coated with 1% BSA at 37 °C for 1 h before two additional washes. MDCK cells were washed twice and added to the wells (105 cells/well) in serum-free medium. After different incubation times (30, 60 and 90 min) at 37 °C, spread and round cells were quantified in four representative areas by microscopy (Leitz, ×100) using a computerized image analysis system (Biocom VisioL@b 2000). The counted cells were then averaged and presented as the mean percentage of spread cells. A P value of less than 0.05 was considered statistically significant.
Mock-transfected MDCK cells or transfected wt-Lu or mt-Lu cells (2×106 cells) were labelled with 5 μM calcein/AM and 2×106 MDCK-mt-Lu control cells were labelled with 1 μg/ml Hoechst 33342 for 30 min at 37 °C/5% CO2. Cells were then washed twice with prewarmed serum-free medium. MDCK, wt-Lu or mt-Lu cells were then mixed with control cells as an internal standard at a 1:1 ratio. Mixed cells (1 ml) were then put in laminin 511/521-pre-coated 12-well plates. After 90 min of incubation at 37 °C/5% CO2, the cells were washed twice with prewarmed serum-free medium to remove unattached cells and visualized by fluorescence using an Evolution VF camera. Ten images were acquired for each sample, and adherent cells were counted using Image-Pro® Plus software. Results are expressed as the mean percentage of adherent transfected cells compared with control adherent cells.
Phosphorylation assays in transfected cells
MDCK-wt-Lu and -mt-Lu (5×106 cells) grown in six-well plates were washed three times in DMEM without phosphate and incubated in the same medium for 2 h at 37 °C. Cells were then incubated with 300 μCi of [32P]Pi (25 mM) for 90 min at 37 °C and one well with wt-Lu or mt-Lu cells was supplemented with 1×PIC (90 min) or 1 mM Br-cAMP (30 min). Cells were washed twice with ice-cold PBS and incubated in lysis buffer [20 mM Tris/HCl (pH 8), 150 mM NaCl, 5 mM EDTA, 1% Triton X-100 and 0.2% BSA, supplemented with protease inhibitor cocktail and 1×PIC] for 1 h at 4 °C. Immunoprecipitation of Lu/BCAM was carried out as described above. The beads were washed five times in lysis buffer. Immunoprecipitated proteins were first eluted for 5 min at 100 °C with 3×Laemmli buffer supplemented with 5% 2-mercaptoethanol and then loaded on to a 4–12% NuPAGE® Novex Bis-Tris gel under reducing conditions. Analysis was performed by Western blotting using a goat biotinylated anti-Lu/BCAM antibody (0.1 μg/ml) or using the bio-imaging analyser.
Rho family proteins activation and inhibition assay
The six-well plates were coated with laminin 511/521 or fibronectin as described above. Mock-transfected MDCK cells, and wt-Lu or mt-Lu cells (25×104 cells) were added to the wells for 90 min. Active Rac1, 2, 3 and RhoA were measured using the G-LISA™ RhoA, Rac or Cdc42 Activation Assay Biochem Kit™ according to the manufacturer's instructions. GTP Rac, RhoA and Cdc42 were used as positive controls. The absorbance signals were detected at 490 nm.
Results are expressed as means and were analysed using the Mann–Whitney non-parametric test to determine statistical significance between mt-Lu and wt-Lu. Statistical significance is defined as P< 0.05.
Lu/BCAM interacts with αII-spectrin in non-erythroid cells via its RK573-574 motif
We investigated the potential interaction of Lu/BCAM with the αII-spectrin isoform expressed in non-erythroid cells. Immunoprecipitation assays were performed using two cell lines expressing endogenous Lu/BCAM: the human kidney epithelial cells A498 and the human microvascular endothelial cells HMEC-1. After Lu/BCAM immunoprecipitation, Western blot analysis revealed the presence of αII-spectrin, indicating that it was co-immunoprecipitated with Lu/BCAM from A498 cells (Figure 1A, left-hand panel) and HMEC-1 cells (results not shown), and suggesting an interaction between these two proteins. Neither spectrin (270 kDa) nor Lu/BCAM (90 kDa) were detected when an irrelevant antibody (IgG1) or Protein A–Sepharose alone were used for immunoprecipitation. The RK573-574 motif of the cytoplasmic tail of Lu/BCAM is required for binding to αI-spectrin in RBCs . The involvement of these residues in the interaction with αII-spectrin was also investigated by immunoprecipitation using transfected MDCK cells expressing either wt-Lu or mt-Lu. As expected, αII-spectrin was co-immunoprecipitated with wt-Lu (Figure 1A, right-hand panel). Faint quantities of αII-spectrin, similar to those obtained with the IgG1 negative control, were detected from cells expressing mt-Lu (Figure 1A, right-hand panel), indicating that the RK573-574AA mutation abolished the Lu/BCAM–αII-spectrin interaction. These results indicated that endogenous and recombinant Lu/BCAM interacted with αII-spectrin in both epithelial and endothelial cells; this interaction involved the Lu/BCAM cytoplasmic RK573-574 motif.
Lu/BCAM binds to the α4 repeat of αII-spectrin
The Lu/BCAM cytoplasmic domain binds to the αI-spectrin α4 repeat (termed α4R) in RBCs . Sequence analyses revealed that the αII-spectrin α4R (residues Glu361 to Glu486) shares 57% identity with its αI-spectrin corresponding repeat. We tested the involvement of the αII-spectrin α4R in the interaction with Lu/BCAM using GST pull-down assays. α3R, α4R, α5R and α3→5R were produced as [35S]methionine-labelled peptides by in vitro transcription and translation (TNT products) (Figure 1B), and their binding activity to the cytoplasmic domain of Lu/BCAM, produced as a GST-fusion protein, was analysed. As shown in Figure 1(B), the spectrin peptides encompassing either the α3→5R or only the α4R were pulled down with the wt-Lu cytoplasmic tail (GST–wt-Lu), giving a signal 3–5-fold higher than those obtained with either GST alone or the mt-Lu cytoplasmic domain (GST–mt-Lu). In contrast, the pull-down signals obtained with α3R and α5R were similar for GST, GST–wt-Lu and GST–mt-Lu. These results indicated that αII-spectrin α4R was involved in the interaction with the Lu/BCAM RK573-574 motif.
Disruption of the Lu/BCAM–spectrin interaction increases Lu/BCAM extractability in epithelial cells
To verify the involvement of the spectrin–Lu/BCAM interaction as a link between Lu/BCAM and the membrane cytoskeleton, we examined the detergent solubility of recombinant wt-Lu and mt-Lu in epithelial MDCK cells. Cells were treated with increasing concentrations of Triton X-100 (0.01–0.1%), and the amount of solubilized Lu/BCAM was analysed by Western blot using Ubc9 (20 kDa) as a cytoplasmic marker and loading control. As shown in Figure 1(C), wt-Lu was not detected in the soluble fraction below 0.025% Triton X-100. Conversely, mt-Lu was extracted starting from 0.01% Triton X-100, indicating that disruption of the Lu/BCAM–spectrin interaction increases the cytoskeleton-unbound Lu/BCAM fraction in non-erythroid cells.
Interaction with spectrin is not involved in the membrane targeting of Lu/BCAM
The basolateral expression of Lu/BCAM in polarized epithelial cells depends on the integrity of a di-leucine motif at position 608–609 of the cytoplasmic domain . To determine the potential role of αII-spectrin in Lu/BCAM membrane targeting, confluent monolayers of MDCK clones expressing similar amounts of either wt-Lu (400000 molecules/cell) or mt-Lu (450000 molecules/cell) were grown on filters and then labelled with both the anti-Lu/BCAM mAb and anti-αII-spectrin pAb. Confocal microscopy analysis showed that both cell types exhibited a polarized phenotype and form tight junctions at confluence, as indicated by daily transepithelial resistance measurements (900 ohms/well after 7 days of culture). As shown in Supplementary Figure S1 (at http://www.BiochemJ.org/bj/436/bj4360699add.htm), spectrin (red) and wt-Lu (green) were expressed at the cell membrane and displayed a lateral expression, as expected. The merge panels (yellow) indicated that wt-Lu and spectrin were in the same lateral membrane compartment. The same pictures were obtained with MDCK cells expressing mt-Lu, indicating that disruption of the Lu/BCAM–spectrin interaction did not modify the Lu/BCAM membrane-expression pattern. Therefore this interaction is not required for the correct lateral targeting of Lu/BCAM in polarized epithelial cells.
Interaction with αII-spectrin plays a role in the stability of Lu/BCAM at the epithelial cell membrane
As the Lu/BCAM–spectrin interaction was not involved in the cell-surface localization of Lu/BCAM, we investigated the putative role of this interaction in Lu/BCAM membrane delivery and turnover using pulse–chase experiments combined with surface protein biotinylation.
As the newly delivered Lu/BCAM at the membrane has been reported to reach a peak between 60 and 90 min of chase , the chase experiment on polarized MDCK cells was started at 60 min after the radiolabelling step and followed up to 180 min. Total cell-surface proteins were labelled with biotin before cell lysis and immunoprecipitation. Total Lu/BCAM was immunoprecipitated and the biotinylated fraction, corresponding to Lu/BCAM expressed at the cell membrane, was purified using streptavidin–agarose beads. The amount of radiolabelled Lu/BCAM present on the membrane after 60, 90, 120, 150 and 180 min was determined by SDS/PAGE of biotinylated proteins followed by autoradiography. Newly synthesized wt-Lu and mt-Lu were correctly addressed to the lateral membrane as both were present at the cell surface after 60 min of chase (Figure 2A). This result confirmed that the interaction with spectrin was not involved in Lu/BCAM membrane targeting. However, the turnover of mt-Lu and wt-Lu were significantly different, as shown by the extended mt-Lu half-life at the membrane. The newly delivered amounts of mt-Lu at 60 min remained unchanged up to 180 min, whereas wt-Lu decreased soon after 60 min. This difference in kinetics was not due to protein degradation as Western blot analysis showed that equivalent amounts of each protein were immunoprecipitated for all chase times (Figure 2B).
To quantify the turnover of both proteins, the intensity of all bands was determined and the ratio ‘radiolabelled/biotinylated’ was calculated. As shown in Figure 2(C), radiolabelled wt-Lu decreased at the membrane over time, with almost half of the proteins being internalized after 90 min as compared with 60 min. Conversely, mt-Lu showed an enhanced stable membrane expression, as newly delivered mt-Lu did not undergo significant internalization during up to 180 min of chase, indicating that the lack of interaction with spectrin decreases the membrane turnover of Lu/BCAM in polarized MDCK cells.
Interaction of Lu/BCAM with spectrin modulates spreading and adhesion of epithelial cells to laminin 511/521
Alteration of the spectrin-based skeleton, as well as disruption of the Lu/BCAM–spectrin interaction, resulted in an enhanced Lu/BCAM-mediated adhesion of RBCs and K562 cells to laminin 511/521 [15,18]. To investigate whether the interaction with spectrin could similarly modulate Lu/BCAM-mediated adhesion of epithelial cells, we analysed the spreading and adhesion to laminin 511/521 of MDCK cells expressing either wt-Lu or mt-Lu as described in the Experimental section (Figure 3A).
We first compared the ability of MDCK cells to spread on laminin 511/521 and BSA (used as a negative control). At 90 min after plating, only 12% and 14% of mock-transfected MDCK cells could spread on BSA and laminin 511/521 respectively (Figure 3B). MDCK cells expressing either wt-Lu or mt-Lu showed a similar behaviour on BSA, but exhibited 38% and 60% of spread cells on laminin 511/521 at 90 min respectively (Figure 3B). Cell spreading on laminin 511/521 was higher for mt-Lu- than for wt-Lu-expressing cells, and was observed as soon as 30 min after plating (29% compared with 14%).
Cell adhesion to laminin 511/521 was also investigated. As expected, Lu/BCAM expression increased cell adhesion to laminin 511/521, as measured by the percentage of adherent cells: 65% for wt-Lu compared with 20% for mock-transfected MDCK cells (Figure 3C). Cell adhesion to laminin 511/521 was particularly reinforced for mt-Lu-expressing cells with 95% of adherent cells. These results indicated that disruption of the Lu/BCAM–spectrin interaction significantly increased Lu/BCAM-mediated epithelial cell adhesion to laminin 511/521.
The reinforced adhesion of mt-Lu-expressing MDCK cells is not related to the Lu/BCAM phosphorylation status
The increased adhesion of pathological RBCs to laminin 511/521 is associated with PKA-mediated Lu/BCAM phosphorylation [2,17]. As disruption of the Lu/BCAM–spectrin interaction reinforced the MDCK cell adhesion to laminin 511/521, we investigated whether this feature was related to a PKA-stimulated Lu/BCAM phosphorylation. Phosphorylation assays were performed using MDCK cells expressing either wt-Lu or mt-Lu. Lu/BCAM was immunoprecipitated from 32P-radiolabelled cells in the presence of a serine/threonine PIC or in the presence of Br-cAMP, a stable membrane-permeant cAMP analogue that activates PKA. In the absence of activators, and 90 min after cell plating on laminin 511/521, both wt-Lu and mt-Lu presented a similar basal phosphorylation level, as shown by autoradiography and Western blot analysis (Figure 4, compare lanes 1 and 4). The amount of immunoprecipitated Lu/BCAM was estimated by Western blot analysis, and Lu/BCAM phosphorylation was normalized by calculating the ratio of the phosphorylation signal and the intensity of each band in Western blot analysis. The phosphorylation level of both proteins was similarly induced in the presence of PIC and Br-cAMP (Figure 4), indicating that disruption of the Lu/BCAM–spectrin interaction did not alter the Lu/BCAM phosphorylation status.
Lu/BCAM–spectrin interaction is required for stress fibre formation on laminin 511/521
As cell adhesion and motility require actin rearrangements, we tested whether the Lu/BCAM–spectrin interaction could induce specific actin reorganization during MDCK cell adhesion and spreading. Low-density-plated cells were stained with Phalloidin to label F-actin (filamentous actin), 90 min after seeding on either laminin 511/521 (Figures 5a–5e) or fibronectin (Figures 5f–5j). Analyses were performed by confocal microscopy focusing on the cell basal membrane. In mock-transfected cells, actin was detected as peripheral spike-like protrusions and cortical rings, as well as cytoplasmic punctated forms on both laminin 511/521 and fibronectin substrates (Figures 5a and 5f). The same actin expression pattern was observed in all wt-Lu-expressing cells plated on fibronectin (Figure 5g). When these cells were plated on laminin 511/521, they exhibited a better spreading with extended lamellipodia. An increase in actin stress fibres was observed in 46% of Lu/BCAM-expressing cells (Figure 5b). To determine the putative involvement of the Lu/BCAM–laminin 511/521 interaction in stress fibre formation, MDCK cells expressing the D343R Lu/BCAM mutant, which is unable to bind to laminin 511/521 (the mutated residue is the same as residue 312 in reference ), were analysed. These cells did not form stress fibres on either laminin 511/521 or on fibronectin (Figures 5c and 5h). Absence of stress fibres was also observed for MDCK cells expressing mt-Lu plated on laminin 511/521 or fibronectin (Figures 5d and 5i). These results indicated that stress fibre formation in MDCK cells plated on laminin 511/521 was induced by the Lu/BCAM–laminin 511/521 interaction, and that the Lu/BCAM–spectrin interaction was involved in a laminin 511/521-to-actin signalling pathway leading to actin reorganization.
To determine the putative role of the 40 C-terminal amino acids of the Lu glycoprotein in regulating stress fibre formation, the same experiment was performed using MDCK cells expressing the short isoform Lu(v13). In this isoform the 40 C-terminal amino acids comprising the proline-rich SH3-binding domain, the di-leucine motif and the phosphorylation sites are truncated. Similarly to wt-Lu-expressing cells, 64% of cells expressing Lu(v13) were able to form stress fibres when plated on laminin 511/521 (Figure 5e), indicating that Lu/BCAM-mediated actin reorganization was independent of its proline-rich SH3-binding domain and of potential signalling events that could be mediated by its cytoplasmic tail.
Laminin 511/521 activates a RhoA-dependent signalling pathway via the Lu/BCAM–spectrin interaction
Rho GTPases are important regulators of the actin cytoskeleton and are involved in cell shape and motility. The activation of the small GTP-binding protein RhoA plays a role in stress fibre formation, whereas Rac and Cdc42 are responsible for actin polymerization in lamellipodia [25–27]. We compared RhoA, Rac and Cdc42 activities in mock-transfected MDCK cells and in cells expressing either wt-Lu or mt-Lu. Activities were evaluated from equal amounts of lysates obtained from cells plated for 90 min on either laminin 511/521 or fibronectin (Figure 6). No changes could be observed in Rac and Cdc42 activity for all conditions (Figures 6B and 6C). In contrast, cells expressing wt-Lu and plated on laminin 511/521 exhibited a significant increase of RhoA activity as compared with mock and mt-Lu cells (Figure 6A). These results suggested the involvement of RhoA activation in actin reorganization associated with the Lu/BCAM–spectrin interaction.
To investigate further the involvement of the RhoA pathway, we used the exoenzyme C3 transferase as a RhoA inhibitor. Pretreatment with the C3 transferase for 2 h is sufficient to inhibit stress fibre formation in wt-Lu and Lu(v13) MDCK cells plated on laminin 511/521, indicating that RhoA is necessary for the specific Lu/BCAM-induced actin reorganization (Figure 7).
Lu/BCAM interacts with spectrin in non-erythroid cells
The present study has identified the αII-spectrin isoform as a direct cytoplasmic partner of Lu/BCAM in vitro and ex vivo in different cellular contexts. The co-immunoprecipitation of endogenous spectrin with Lu/BCAM from kidney epithelial A498 cells provided physiological relevance of the GST pull-down interactions.
Positively charged juxtamembrane amino acids in adhesion molecules, such as CD44 and ICAM-2 (intercellular adhesion molecule 2), have been shown to interact with cytoskeletal protein 4.1 and ERM (ezrin/radixin/moesin) proteins . As in RBCs, the Lu/BCAM RK573-574 motif interacted directly with αII-spectrin α4R and it could be assumed that this interaction might occur in all cell types where these two proteins are expressed. In humans, αI and αII-spectrins share 55% identity in the amino acid sequence. α-Spectrins are composed of 20 repeats and αII-spectrin α4R exhibits 57% identity with its corresponding erythroid αI-spectrin repeat. Moreover, the Lu/BCAM RK573-574 motif and surrounding sequences are conserved in mammals.
Effects of the Lu/BCAM–spectrin interaction on Lu/BCAM membrane localization and stability
Lu/BCAM is endogenously expressed in epithelial kidney A498 cells at 70000 copies/cell . Its lateral expression in polarized epithelial cells depends on the di-leucine motif of its cytoplasmic domain .
Moreover, recruitment of Lu/BCAM at the lateral surface in epithelial cells, adjacent to laminin α5-containing extracellular matrix, suggests a role for laminin 511/521 in its membrane localization . Indeed, Lu/BCAM expression was dramatically reduced in various tissues of mouse embryos lacking laminin α5 chain, whereas it was significantly increased in the heart of transgenic mice overexpressing this chain [29,30]. The results of the present study indicated the predominance of the di-leucine motif for Lu/BCAM lateral targeting. Our immunostaining experiments indicated that interaction with spectrin was not required for the proper targeting of Lu/BCAM in polarized epithelial MDCK cells.
In conclusion, the RK motif does not appear to be involved in Lu/BCAM membrane targeting, but this interaction could play a role in Lu/BCAM membrane turnover. In the pulse–chase experiments, we observed that the Lu/BCAM RK573-574AA mutant had an extended half-life at the cell surface. Since αII-spectrin interacts with syntaxin family proteins, which play a key role in vesicle fusion during exocytosis and endocytosis , the results of the present study suggest that spectrin might be involved in Lu/BCAM turnover through the endocytic pathway.
Role of the Lu/BCAM–spectrin interaction during cell spreading and adhesion
It was hypothesized that the Lu/BCAM–spectrin interaction might be critical for signalling and laminin 511/521 receptor function rather than in the maintenance of shape of the RBC. Lu/BCAM is involved in the abnormal adhesion of RBCs to laminin 511/521 and endothelium in pathological situations [2,5,16,17]. We and others [15,18] have demonstrated that the interaction of Lu/BCAM with erythroid spectrin negatively regulated its adhesive receptor function in normal and spectrin-deficient HS RBCs. In non-erythroid adherent cells, partial αII-spectrin depletion was associated with loss of cell spreading, defective adhesion, and a decrease and irregularity of focal adhesion points .
The results of the present study have shown that Lu/BCAM plays a critical role during epithelial cell adhesion to laminin 511/521. Cell adhesion and spreading to laminin 511/521 were significantly reinforced in cells expressing the Lu/BCAM RK573-574AA mutant, which could be a consequence of its extended half-life at the cell surface. Other proteins of the spectrin-based cytoskeleton, such as the adapter protein ankyrin G, can also modulate the function of adhesion molecules, including cell adhesion, migration and membrane distribution. The membrane expression of L1CAM (L1 cell-adhesion molecule), involved in the migration of neuronal growth cones and the static adhesion between adjacent axons, is regulated by its interaction with ankyrin G. Inhibition of this interaction led to retrograde movements of L1CAM in the cell membrane and stimulated L1CAM-mediated neurite outgrowth .
To date, the function of the Lu/BCAM laminin 511/521 receptor was known to be activated by two pathways in pathological RBCs: one involving the phosphorylation of its cytoplasmic domain by PKA , and the other mediated by a Lu/BCAM cytoplasmic domain interaction with the spectrin-based cytoskeleton [15,18]. The present study indicates that interaction between Lu/BCAM and spectrin regulates epithelial cell adhesion and spreading to laminin 511/521. This regulation is not related to an increased Lu/BCAM phosphorylation of the Lu RK573-574AA mutant. Conversely, laminin 511/521 binding could induce Lu/BCAM cytoplasmic tail phosphorylation, thus modulating its interaction with spectrin. Further experiments are needed to investigate this question.
Consequences on actin dynamics
Cell adhesion to extracellular matrix triggers outside–in signalling pathways leading to actin skeleton reorganization with extension of lamellipodia and filopodia. In the present study, we have shown that Lu/BCAM induced stress fibre formation during the early steps of cell adhesion and spreading to laminin 511/521. These changes were due to the Lu/BCAM–laminin 511/521 interaction, as cells expressing the D343R mutant, which is defective for laminin 511/521 binding, were unable to induce stress fibre formation. This rearrangement of the actin skeleton was associated with an activation of the small GTP-binding protein RhoA, which is known to regulate actin polymerization and stress fibre formation . Similarly, RhoA activation was only observed in cells expressing wt Lu/BCAM. We can conclude that Lu/BCAM binding to laminin 511/521 drives actin filament formation.
The involvement of laminin α5 in cell spreading and in polymerization of parallel actin filaments, named filipodia-like microspikes, has also been demonstrated in primary dental epithelial cells in which laminin α5 interacts with α6β4 integrins. This interaction participates in the activation of PI3K (phosphoinositide 3-kinase)-Cdc42/Rac pathways . We have shown that Cdc42/Rac pathways are not involved in the Lu/BCAM-induced actin reorganization.
The signalling pathway induced by Lu/BCAM binding to laminin 511/521 and leading to actin reorganization requires the Lu/BCAM–αII-spectrin interaction. Disruption of this interaction reduced both stress fibre formation and RhoA activation. We have shown that both Lu/BCAM isoforms acted similarly with regard to actin reorganization. The ability of the short isoform Lu(v13) to induce stress fibre formation indicated that this was independent of the proline-rich SH3-binding domain and the phosphorylation sites in the 40 C-terminal amino acids present in the long isoform. The results of the present study strongly suggest that the signalling cascade leading to actin reorganization upon Lu/BCAM binding to laminin 511/521 is vehicled by spectrin. All of these results provide new insights into a novel signalling pathway regulating the actin cytoskeleton via the Lu/BCAM–spectrin interaction which links the upstream laminin 511/521-binding signal to downstream RhoA activation.
The role of αII-spectrin in actin organization has been recently demonstrated, as αII-spectrin-knocked-down melanoma WM-266 cells exhibited modifications of the actin cytoskeleton, such as a loss of stress fibres . αII-Spectrin, via its SH3 domain, has also been implicated in initiating Rac activation in the specialized β3-integrin clusters that initiate cell adhesion and spreading . Moreover, the SH3 domain of αII-spectrin was demonstrated to bind proteins involved in the regulation of actin cytoskeleton dynamics, such as two members of the Mena-VASP (vasodilator-stimulated phosphoprotein) family [EVL (Enabled/VASP-like, VASP)] and Tes [36–39]. VASP participates in actin fibre formation, and the αII-spectrin–VASP complexes regulate cortical actin cytoskeleton assembly with implications in cell–cell contact formation. Overexpression of Tes resulted in increased cell spreading and decreased cell motility, whereas knockdown of Tes in HeLa cells resulted in loss of actin stress fibres and reduced RhoA activity . All of these results reinforce the idea of a pivotal role of spectrin in actin-dependent processes.
In conclusion, spectrin is involved in regulating Lu/BCAM function and expression in non-erythroid cells. The results of the present study provide new evidence that spectrin plays a novel role as a linker between extracellular signals, triggered by laminin 511/521, and intracellular events modulating actin dynamics.
Emmanuel Collec designed and performed research, analysed data and wrote the paper. Marie-Christine Lecomte, Wassim El Nemer and Yves Colin discussed the results, gave advice and commented on the manuscript at all stages. Caroline Le Van Kim supervised the project, designed and performed research, and wrote the paper.
This work was supported, in part, by the Institut National de la Transfusion Sanguine (INTS), the Institut National de la Santé et de la Recherche Médicale (INSERM) and Université Paris Diderot -Paris7.
We thank Julien Picot for helpful advice on flow cytometry experiments.
Abbreviations: BCAM, basal cell-adhesion molecule; calcein/AM, calcein acetoxymethyl ester; Cdc42, cell division cycle 42; DMEM, Dulbecco's modified Eagle's medium; F-actin, filamentous actin; FBS, fetal bovine serum; GST, glutathione transferase; HS, hereditary spherocytosis; L1CAM, L1 cell-adhesion molecule; Lu, Lutheran; mAb, monoclonal antibody; MDCK, Madin–Darby canine kidney; pAb, polyclonal antibody; PIC, phosphatase inhibitor cocktail; PKA, protein kinase A; RBC, red blood cell; SABC, specific antibody-binding capacity; SH3, Src homology 3; sulfo-NHS-LC-biotin, sulfo-succinimidyl-6-(biotinamido)hexanoate; Ubc9, ubiquitin-conjugating enzyme 9; VASP, vasodilator-stimulated phosphoprotein; wt, wild-type
- © The Authors Journal compilation © 2011 Biochemical Society