Protein 4.1G Regulates Cell Adhesion, Spreading, and Migration of Mouse Embryonic Fibroblasts through the β1 Integrin Pathway*

Protein 4.1G is a membrane skeletal protein that can serve as an adapter between transmembrane proteins and the underlying membrane skeleton. The function of 4.1G remains largely unexplored. Here, using 4.1G knockout mouse embryonic fibroblasts (MEFs) as a model system, we explored the function of 4.1G in motile cells. We show that the adhesion, spreading, and migration of 4.1G−/− MEF cells are impaired significantly. We further show that, although the total cellular expression of β1 integrin is unchanged, the surface expression of β1 integrin and its active form are decreased significantly in 4.1G−/− MEF cells. Moreover, the phosphorylation of focal adhesion kinase, a downstream component of the integrin-mediated signal transduction pathway, is suppressed in 4.1G−/− MEF cells. Co-immunoprecipitation experiments and in vitro binding assays showed that 4.1G binds directly to β1 integrin via its membrane-binding domain. These findings identified a novel role of 4.1G in cell adhesion, spreading, and migration in MEF cells by modulating the surface expression of β1 integrin and subsequent downstream signal transduction.

in the activation of a number of signaling pathways (5). These pathways include protein tyrosine kinases, such as focal adhesion kinase (FAK) 3 (6), and members of the Rho family of small GTP-binding proteins, such as Cdc42, Rac1, and RhoA (7), which play important roles in regulating the organization of the cytoskeleton. Activated FAK and Rho-GTPase regulate cell adhesion, spreading, and migration (8,9).
One important feature of integrins is that they can shift between low-and high-affinity conformations for ligand binding. The shift from a low-to a high-affinity state is termed "integrin activation" (10). Because altered integrin activation is associated with many diseases, such as bleeding disorders, leukocyte adhesion deficiencies, and skin blistering, integrin activation has to be controlled stringently (11). It was originally thought that talin is the only master regulator of integrin activation (12). Later works have shown that the kindlin family of proteins is as important as talin in mediating integrin function (13,14). Both talin and kindlins belong to a family of evolutionarily conserved FERM (four-point-one, ezrin, radixin, moesin) domain-containing proteins (15). They regulate integrin function by binding directly to the cytoplasmic tail of integrin via their FERM domain, which triggers a conformational change in the extracellular ligand-binding domain, increasing its affinity for its ligand (10,16). These findings suggest that other FERM domain-containing proteins may also associate with integrin and regulate integrin function. Protein 4.1 family members (which includes 4.1R, 4.1B, 4.1G, and 4.1N) are the prototypical members of the FERM domaincontaining superfamily of proteins. We have shown recently that 4.1R binds to ␤1 integrin and modulates the surface expression of ␤1 integrin in keratinocytes (17). A study by McCarty et al. (18) has also documented the association of 4.1B with ␤8 integrin in cultured astrocytes and in the brain. In this study, we identified a novel role of 4.1G in cell adhesion, spreading, and migration of mouse embryonic fibroblasts by modulating the surface expression of ␤1 integrin through a direct association between 4.1G and ␤1 integrin. antibodies used in this study were as follows: rat 9EG7 monoclonal antibody, which preferentially recognizes the active conformation of mouse ␤1 integrins (21) (BD Biosciences); conformation-independent MB1.2 rat monoclonal antibody against mouse ␤1 integrin (22, 23) (Millipore, Billerica, MA); anti-FAK and anti-phosphotyrosine (4G10) (Millipore); anti-␣2-integrin, anti-␣5-integrin, and anti-␣6-integrin (Abcam, Cambridge, MA); and anti-␣3-integrin and ␤4-integrin (BD Biosciences). Affinity-purified rabbit polyclonal antibodies against GST and His were prepared by our laboratory. Alexa Fluor 488-conjugated and Alexa Fluor 594-conjugated secondary antibody to mouse and rabbit IgG, TO-PRO3 for nuclear staining, and Alexa Fluor 488-labeled wheat germ agglutinin for membrane staining were from Invitrogen. Goat anti-mouse HRP and goat anti-rabbit HRP were obtained from Jackson ImmunoResearch Laboratories (West Grove, PA).
Cell Culture-Isolation of primary mouse embryonic fibroblast (MEF) cells from 4.1G ϩ/ϩ and 4.1G Ϫ/Ϫ C57Bl/6 mice (20) was performed as described before (24). MEF cells were prepared from embryonic day 13.5 embryos. The head and internal organs were removed. The remaining embryonic tissue was minced using a pair of scissors and immersed in 0.25% trypsin overnight at 4°C. After 24 h, MEF cells were collected after centrifugation at 1500 rpm and maintained in DMEM containing 10% FBS (Gibco) and 100 g/ml penicillin/streptomycin. After two passages, the MEF cells were immortalized by retroviral transduction of the SV40 large T antigen. For serum starvation experiments, MEF cells were plated in DMEM containing 0.1% FBS and then incubated at 37°C for 18 h.
Cloning of 4.1G cDNA from MEF Cells-Total RNA was isolated from 4.1G ϩ/ϩ and 4.1G Ϫ/Ϫ MEF cells with the RNeasy mini kit (Qiagen). RNA (1 g) was reverse-transcribed into cDNA using random nonamers and M-MuLV reverse transcriptase (New England Biolabs) for 60 min at 42°C. An equivalent of 5 ng of cDNA was used for PCR. PCR was performed using Accuprime Platinum Pfx DNA polymerase (Invitrogen). The PCR primers used were as follows: forward, ATGACTAC-TGAAGTTGGCT-CTGCATCTGAA; reverse, TTATTCTT-CTC-CTTCCTCCGCCAACTCTG. Primers were designed to incorporate recognition sequences for the restriction enzymes SacII and XmaI at the 5Ј and 3Ј ends of the PCR product, respectively. N-terminal GFP fusion constructs were created by ligating SacII/XmaI-digested 4.1G cDNAs downstream of the GFP coding sequence in the pEGFP-C3 vector. The fidelity of the constructs was confirmed by sequencing.
Immunofluorescence Staining-For confocal immunofluorescence microscopy, cells were grown on MatTek glass-bottom microwell cell culture dishes (MatTek) coated with 10 g/ml fibronectin (FN), and we let the cells grow into sparse density or to ϳ90% confluence. Then the cells were fixed with 1% paraformaldehyde for 15 min and permeabilized with 0.1% Triton X-100 in 0.25% paraformaldehyde-PBS. Cells were then incubated in 10% horse serum and 0.1% Triton X-100 in PBS for 30 min to minimize nonspecific antibody binding. The cells were incubated with primary antibodies at 4°C overnight, washed three times with PBS, and incubated with the appropriate second antibody at room temperature for 30 min. The following primary antibodies were used: rabbit polyclonal antibodies to 4.1G-U3, rat monoclonal antibody against ␤1 integrin (clone 9EG7), and mouse monoclonal antibody against FAK and paxillin. Alexa Fluor-conjugated secondary antibodies were purchased from Molecular Probes and diluted 1/700. The secondary antibodies were donkey anti-rabbit, donkey anti-rat, and donkey anti-mouse IgG labeled with Alexa Fluor 488 or Alexa Fluor 594. Actin was counterstained with Rhodaminephalloidin (red). Images were collected on a Zeiss LSM510 META confocal microscope using a ϫ63 oil immersion objective.
Flow Cytometry-4.1G ϩ/ϩ and 4.1G Ϫ/Ϫ MEF cells were serum-starved for 18 h. The cells were trypsinized and washed twice with 0.5% BSA in PBS. Primary antibodies against total ␤1 integrin (catalog no. MAB1997, Millipore) and against activeform ␤1 integrin (clone 9EG7, BD Biosciences) were used to stain the cells in 0.5% BSA in PBS for 30 min on ice. The cells were washed twice and incubated with allophycocyanin-conjugated anti-rat or anti-mouse secondary antibody for an additional 30 min on ice. After further washing, flow cytometric analysis was performed on a FACSCanto flow cytometer (BD Biosciences), and flow data overlay plots were produced using CellQuest Pro software (BD Biosciences).
Immunoblot Analysis-Cells were trypsinized, washed with PBS, and lysed with ice-cold lysis buffer (50 mM HEPES (pH 8.3), 420 mM KCl, 0.1% Nonidet P-40, and 1 mM EDTA) for 30 min on ice in the presence of proteinase inhibitor mixture (Sigma) and phosphatase inhibitor (Roche). After centrifugation at 16,000 ϫ g at 4°C for 10 min, the supernatant was collected. Protein concentration was measured by the Bradford method using BSA as standard. 30 g of protein was separated by 10% SDS-PAGE and transferred to a nitrocellulose membrane. After blocking for 1 h in blocking buffer (10 mM Tris-HCl (pH 7.4), 150 mM NaCl, 0.5% Tween 20, and 5% nonfat dried milk powder), the blot was probed for 1 h with the desired primary antibodies. After several washes, the blot was incubated with anti-rabbit or anti-mouse IgG coupled to HRP and developed with the SuperSignal West Pico chemiluminescence detection kit (Molecular Probes). All steps were performed at room temperature.
Co-immunoprecipitation-MEF cells were lysed with icecold lysis buffer (50 mM HEPES (pH 8.3), 420 mM KCl, 0.1% Nonidet P-40, and 1 mM EDTA) for 30 min on ice. The supernatant was collected after centrifugation at 16,000 ϫ g at 4°C for 10 min, and the concentration of protein in the supernatant was determined by the Bradford method using BSA as standard (Bio-Rad). 500 g of extract was incubated with either 5 g of anti-4.1G-HP or anti-␤1 integrin antibody or preimmune IgG in 500 l of co-immunoprecipitation buffer (Active Motif) at 4°C overnight with rotation. The immunoprecipitates were isolated on protein G beads and separated by 10% SDS-PAGE, followed by transfer to a nitrocellulose membrane. The membrane was probed with antibodies against ␤1 integrin, ␣6 integrin, or 4.1G-HP.
Wound Healing Assay-MEF cells were grown at equivalent confluence for 18 h on a MatTek glass-bottomed chamber precoated with 10 g/ml FN. Confluent 4.1G ϩ/ϩ and 4.1G Ϫ/Ϫ MEF cells were arrested mitotically by incubation with 8 g/ml mitomycin C (Roche) in DMEM for 2 h under normal culture conditions. Mitomycin C was removed by three washes in PBS. A pseudowound was introduced in an equivalent confluent monolayer of cells by lightly scratching with a 10-l pipette tip across the cell layer. Cell debris was removed by two washes with culture medium. A minimum of six "wounded areas" was filmed for each sample by obtaining images every 15 min for 16 h. Images were collected on a Zeiss LSM510 META confocal microscope using a ϫ25 phase-contrast objective. The wounded area was measured using LSM510 software for each representative time point.
Cell Spreading Assay-MEF cells were trypsinized and replated on coverslips precoated with 10 g/ml FN and allowed to spread for 1 or 3 h at 37°C in the presence of complete media. Cells were fixed and labeled with Alexa Fluor 488-conjugated wheat germ agglutinin (Invitrogen) for 30 min to better visualize the cell outlines. Cell surface boundaries were outlined for 35 individual cells chosen randomly, and LSM 510 software was used to calculate the mean surface area and standard deviation of each population. One-tailed Student's t tests were applied to test the statistical significance of the data. We would like to note that, in our preliminary experiments, we checked the spreading area of MEF cells at 1, 3, 6, and 12 h. Although the spreading areas increased from 1 to 3 h, there was no significant difference between 3, 6, and 12 h, demonstrating that MEF cells are fully spread after 3 h. Therefore, in this study, we chose the 1-and 3-h time points to perform the cell spreading assay.
Transwell Migration Assay-For migration assays, 8-m-diameter pore transwell cell culture inserts (BD Biosciences) were placed in 6-well plates. The underside of the insert and the bottom of the well were coated with 10 g/ml of FN at 37°C for 1 h. Cells suspended in serum-free media were seeded into the upper chamber of the insert (4 ϫ 10 5 /well), and complete medium was added to the lower chamber. Cells were then incubated for another 6 h, during which cells migrated through the pores in the insert to the lower side of the membrane insert. At the end of cell migration, we cleansed the upper side of the chamber with a cotton swab, stained the filter for 1 h with crystal violet (Sigma) in 2% ethanol, and then rinsed it in water. The filters were then imaged with a Leica inverted microscope. Five representative images (ϫ10 magnification) were captured randomly for each insert and used to manually count the number of cells present. The results were presented as mean number of cells per field Ϯ S.D.
Preparation of Recombinant Proteins-The plasmid DNA encoding various recombinant proteins was transformed into Escherichia coli BL21(DE3) for protein expression. The recombinant proteins were expressed at 16°C in the presence of 0.1 mM isopropyl-␤-D-thiogalactopyranoside. GST-tagged 4.1G domains were purified by a glutathione-Sepharose 4B affinity column, and the maltose binding protein (MBP)-tagged cytoplasmic domain of ␤1 integrin was purified by amylose resin.
GST Pulldown Assay-For the pulldown assay, various domains of 4.1G were cloned into pGEX 4T-2, and the cytoplasmic domain of ␤1 integrin was cloned into the pMal-c2X vector. GST-tagged proteins were coupled to glutathione-Sepharose 4B beads, and the MBP-tagged cytoplasmic domain of ␤1 integrin was coupled to amylose resin at room temperature for 1 h. The beads were pelleted and washed. GST-tagged 4.1G domains or the MBP-tagged cytoplasmic domain of ␤1 integrin were added to the coupled beads in a final volume of 100 l. The final concentration of the coupled protein was 2 M. The mixture was incubated for 1 h at room temperature, pelleted, washed, and eluted with 10% SDS. The pellet was analyzed by SDS-PAGE. The binding of GST-tagged 4.1G domains to the MBP-tagged cytoplasmic domain of the ␤1 integrin fragment was detected by Western blot using anti-GST or anti-MBP antibody.
FAK Phosphorylation Assay-MEF cells were serum-starved by culturing in DMEM without FBS for 24 h. Following trypsinization and replating on coverslips precoated with 10 g/ml FN, cells were allowed to recover for various time periods. The cells were then collected and lysed with ice-cold lysis buffer (50 mM HEPES (pH 8.3), 420 mM KCl, 0.1% Nonidet P-40, and 1 mM EDTA) for 30 min on ice. The supernatant was collected, and protein concentration was determined by the Bradford method using BSA as standard (Bio-Rad). 500 g of extract was incubated with 5 g of anti-FAK in 500 l of coimmunoprecipitation buffer (Active Motif) at 4°C overnight with rotation. The immunoprecipitated proteins were isolated on protein G beads and separated by 10% SDS-PAGE, followed by transfer to a nitrocellulose membrane. The membrane was probed with anti-FAK antibody or 4G10 antibody, which specifically recognizes phosphorylated tyrosine.

Results
Expression of 4.1G in MEF Cells-It has been well established that all four genes encoding the family of 4.1 proteins undergo extensive alternative splicing that leads to the generation of multiple isoforms. As the first step toward investigating the function of 4.1G in MEF cells, we examined the expression of 4.1G by RT-PCR and Western blotting. RT-PCR using a 4.1Gspecific primer set amplified a band around 3000 bp from 4.1G ϩ/ϩ MEF cells but not 4.1G Ϫ/Ϫ MEF cells (Fig. 1A). Sequencing of the PCR product of six distinct clones revealed that it contains 2967 bp. The sequence was identical to the coding sequence of RefSeq entry NM_001199265.1. The corresponding protein database entry is NP_001186194.1. The exon composition of this isoform is shown in Fig. 1B. The 4.1G transcript contains exons encoding the head piece (U1, exon 2), the FERM domain (exons 4 -12), the U2 region (exon 13), spectrinactin binding domain (exons 16,17), the U3 region (exon 17D), and the C-terminal domain (exons 18 -21).
The expression of 4.1G protein was examined by Western blot analysis using three 4.1G-specific antibodies. Fig. 1C shows that all antibodies detected one band with a molecular mass of ϳ160 KD. The specificity of the band was validated by the finding that it is not detected in 4.1G Ϫ/Ϫ MEF cells. We also examined the expression of the other three protein 4.1 family members by Western blot analysis. Localization of 4.1G in MEF Cells-We then examined the localization of endogenous 4.1G by immunofluorescence staining using anti-4.1G U3 antibody. Fig. 1E, a, shows both membrane and cytoplasmic localization of 4.1G. Furthermore, as shown in Fig. 1E, b, 4.1G appears to localize at the leading edge of a motile cell. No staining is seen in 4.1G Ϫ/Ϫ MEF cells. The localization profile of 4.1G strongly suggests the potential involvement of 4.1G in cell spreading and migration.
Impaired Adhesion and Spreading of 4.1G Ϫ/Ϫ MEF Cells-Having characterized the expression and localization of 4.1G in MEF cells, we further explored the function of 4.1G in these cells. First we compared the adhesion of 4.1G ϩ/ϩ and 4.1G Ϫ/Ϫ MEF cells to the FN-coated surface. The adhesion was examined 0.5, 1, and 3 h after plating. Fig. 2A shows that, at all time points, the adhesion of 4.1G Ϫ/Ϫ cells to the FN-coated surface was less than that of 4.1G ϩ/ϩ MEF cells (ϳ20% less after 0.5 h and ϳ35% less after 1 and 3 h). We also examined the extent of cell spreading on an FN-coated surface. Fig. 2B shows that, after 1 or 3 h of incubation, the extent of spreading of 4.1G Ϫ/Ϫ cells was much less than that of 4.1G ϩ/ϩ cells. Quantitative analysis revealed an ϳ50% reduction in the spreading area of 4.1G Ϫ/Ϫ cells compared with that of 4.1G ϩ/ϩ cells at both time points (Fig. 2C).
To confirm a direct role of 4.1G in MEF cell spreading, we transfected 4.1G Ϫ/Ϫ MEF cells with a plasmid encoding GFP-4.1G or GFP only. Western blot analysis showed that both GFP and GFP-4.1G proteins are expressed in the transfected cells (Fig. 2D). Fig. 2E shows that the spreading area of 4.1G Ϫ/Ϫ MEF cells transfected with a GFP-4.1G plasmid is increased significantly compared with the neighboring untransfected 4.1G Ϫ/Ϫ MEF cells, whereas the spreading area of cells transfected with GFP showed no significant change. Quantitative analysis  JANUARY 29, 2016 • VOLUME 291 • NUMBER 5 revealed an ϳ220% increase in the spreading area of 4.1G Ϫ/Ϫ cells transfected with GFP-4.1G compared with that of nontransfected cells or cells transfected with GFP (Fig. 2F).

4.1G and ␤1 Integrin in Mouse Embryonic Fibroblasts
Western blot analysis was also performed to compare the expression of GFP-4.1G in 4.1G Ϫ/Ϫ cells with endogenous 4.1G in wild-type cells and to examine whether the levels of 4.1N decrease upon 4.1G rescue. These results are shown in Fig. 2G, a. In the situation that the transfection efficiency of GFP-4.1G Ϫ/Ϫ is about 20 -30%, the expression level of GRP-4.1G is about 40% of that of the 4.1F ϩ/ϩ cells. We estimated that GFP-4.1G, when the transfection efficiency is 100%, is about 1.5 to 2 times higher than that of endogenous 4.1G (Fig. 2G, b).  Fig. 3C, and the rate of wound closure is shown in Fig. 3D. Although the closure of the wound area of 4.1G ϩ/ϩ cells was nearly complete 8 h after wounding, more than 30% of the wounded area of 4.1G Ϫ/Ϫ cells was still not closed. We also examined the effect of 4.1G on random motility of MEF cells using live-cell video microscopy. Track plots of the randomly migrating MEF cells revealed no detectable difference in either distance moved or velocity of migration between 4.1G ϩ/ϩ and 4.1G Ϫ/Ϫ MEF cells (Fig. 3, E  and F). Therefore, 4.1G plays a role in directional migration but not in random motility of MEF cells.
Decreased Surface Expression of ␤1 Integrin in 4.1G Ϫ/Ϫ MEF Cells-Integrins play important roles in cell adhesion, spreading, and migration. Members of FERM domain-containing proteins such as talin (25), kindlin (13), 4.1B (18), and 4.1R (17) have been shown to interact with integrins and modulate their functions. These findings promoted us to examine how deletion of 4.1G in MEF cells affects integrins. We first examined the surface expression of ␤1 integrin by flow cytometry, using antibodies specific for the total or activated form of ␤1 integrin. Representative flow cytometric profiles and the quantitative data from three independent experiments are shown in Fig. 4, A and B. These results reveal that the surface expression of total ␤1 integrin was decreased by ϳ40% and that the active form of ␤1 integrin was decreased by ϳ50% in 4.1G Ϫ/Ϫ MEF cells. Interestingly, Western blot analyses showed no significant change in the expression levels of ␤1 integrin in 4.1G ϩ/ϩ and 4.1G Ϫ/Ϫ MEF cells. These findings suggest that 4.1G, although not affecting the expression levels of ␤1 integrin, plays an important role in the surface expression ␤1 integrin. Because ␤ integrin pairs with ␣ integrin to form ␣/␤ heterodimers, we also examined the expression of several ␣ integrins (potential ␤1 partners) by Western blot analysis. Although no significant difference in the expression levels of ␣2 and ␣6 integrin, but a significant difference in the expression level of ␣3 integrin, was observed between 4.1G ϩ/ϩ and 4.1G Ϫ/Ϫ MEF cells, surprisingly, a more than 3-fold increase in ␣5 integrin expression was observed in 4.1G Ϫ/Ϫ cells (Fig. 4, C and D).
Association of Protein 4.1G with ␤1 Integrin-Members of the protein 4.1 superfamily have been shown to bind to integrins and to engage in regulation of the integrin-mediated signaling pathway (13, 17, 18, 25). We searched for a similar activity on the part of 4.1G. Double staining of 4.1G and ␤1 integrin was performed to examine whether the two molecules co-localize. As shown in Fig. 5A, under confluent conditions, 4.1G and ␤1 integrin co-localized in the cytoplasm and on the plasma membrane. Co-immunoprecipitation and GST pulldown assays were performed to test for direct binding of 4.1G to ␤1 integrin. As shown in Fig. 5B, protein 4.1G could be co-immunoprecipitated with ␤1 integrin by anti-␤1 integrin antibody. Conversely, ␤1 integrin was pulled down with protein 4.1G by anti-4.1G HP antibody. In a negative control, ␣6 integrin did not co-immunoprecipitate with 4.1G (Fig. 5C). GST pulldown assays were performed with a set of distinct 4.1G  (Fig. 5E). These results demonstrate that protein 4.1G interacts directly with the cytoplasmic domain of ␤1 integrin in MEF cells through its membrane binding domain.
Diminished Phosphorylation of FAK in 4.1G Ϫ/Ϫ MEF Cells-FAK is known to be a crucial component in the transduction of signaling pathways initiated by integrin ligation (26). This, in turn, regulates cell spreading and migration (27). The impaired cell spreading and motility of cells lacking 4.1G raises the question of whether the protein plays a role in the phosphorylation of FAK. We therefore compared the phosphorylation levels of FAK in 4.1G ϩ/ϩ and 4.1G Ϫ/Ϫ MEF cells at different times after FN stimulation. Fig. 6 shows that, in 4.1G ϩ/ϩ cells, phosphor- ylation of FAK is increased in response to FN stimulation. In 4.1G Ϫ/Ϫ cells, by contrast, the phosphorylation level upon FN stimulation is significantly less than that in 4.1G ϩ/ϩ cells (Fig. 6,  A and B). Therefore, protein 4.1G is required for the phosphorylation of FAK. In support of a role of 4.1G in FAK phosphorylation, Fig. 6C shows that, similar to ␤1 integrin, 4.1G co-localizes at the focal adhesion site.

Discussion
4.1G is a member of the protein 4.1 family, which includes 4.1R, 4.1G, 4.1B, and 4.1N (28,29). In contrast to an extensive understanding of the function of the prototypical member 4.1R, the knowledge regarding 4.1G is very limited. Earlier in vitro biochemical studies have shown the association of 4.1G with several transmembrane receptors (30 -32). Using 4.1G Ϫ/Ϫ mice, we and others have recently documented the role of 4.1G in male fertility (20) and in the organization of the internodes in peripheral myelinated nerves (33). In this study, using MEF cells derived from 4.1G Ϫ/Ϫ mice, we identified a previously unrecognized role for 4.1G in the motile behavior of MEF cells. We further documented that 4.1G affects cell adhe-sion, spreading, and migration through the ␤1 integrin pathway.
Members of the FERM protein superfamily have been reported to participate in integrin-linked downstream functions. For example, by binding to the cytoplasmic tails of ␤1 or ␤3 integrin, talin induces conformational changes in the extracellular domain of integrins, increasing their affinity for ligands (10). Members of another group of FERM proteins, the kindlin family, are also known to be involved in integrin transmembrane signaling, and expression of kindlin is necessary for integrin activation (13,34). Recent studies have revealed that members of the protein 4.1 family also play important modulatory roles in integrin-related processes. Protein 4.1B has been found to interact selectively with ␣v␤8 integrin and plays an important functional role in the development and maintenance of the CNS (18). We recently documented that 4.1R plays an important role in cell adhesion, spreading, and migration of keratinocytes by modulating the surface expression of ␤1 integrin (17).
Although several members of the protein 4.1 superfamily have been implicated in mediating integrin functions, it is interesting to note that they probably do so through different mech- anisms. Although the binding of talin or kindlin to cytoplasmic tails of integrin induces conformational changes in the extracellular domain of integrin, leading to integrin activation, the association of integrin to members of protein 4.1 is required for the surface expression of integrin.
One consensus role of the protein 4.1 family members is their ability to associate with a variety of transmembrane proteins and regulate the expression of these proteins. However, it should be noted that there are some differences. In red cells, deficiency of 4.1R leads to decreased expression levels of its binding partners glycophorins C, band 3, XK, and Duffy (35). In stomach epithelial cells, lack of 4.1R resulted in decreased expression of the binding partner ␤-catenin (36). Similarly, the expression of nectin-like 4 is reduced in 4.1G-deficient testis Sertoli cells (20). On the other hand, lack of 4.1R in keratinocytes only impaired the surface expression of ␤1 integrin, which was accompanied by an increased in overall ␤1 integrin expression (17). Here we show that, in MEF cells, lack of 4.1G leads to decreased surface expression of ␤1 integrin without affecting overall ␤1 integrin expression. Given the findings that some other protein 4.1 members, such as 4.1N and 4.1R, play an important role in intracellular protein traffic (37,38) and that integrins undergo extensive intracellular trafficking (39,40), it is reasonable to speculate that protein 4.1 family members affect the expression of their binding partners through different mechanisms.
Another emerging role of protein 4.1 family members is in mediating intracellular signal transduction. For example, we have documented that, in CD4 ϩ T cells, lack of 4.1R results in hyper-phosphorylation of the adapter protein linker of activation of T cells (LAT) and enhanced downstream signal transduction, implying that 4.1R negatively regulates signal transduction in CD4 ϩ cells (38). In contrast, here we show that lack of 4.1G in MEF cells led to impaired phosphorylation of FAK.
Although in vitro studies have suggested that protein 4.1 family members may play many different functional roles (41,42), the phenotypic changes in various individual 4.1 knockout mouse models are not as severe as would be expected from in vitro studies. It is possible that members of the protein 4.1 family may compensate function for each other. Indeed, we have shown previously that all 4.1 family members are expressed in CD4 ϩ cells and that the expression of 4.1N is significantly upregulated in 4.1R-deficient CD4 ϩ T cells (41). We have also shown that all 4.1 family members are expressed in keratinocytes and that, in 4.1R-deficient keratinocytes, both 4.1N and 4.1G are up-regulated (17). In addition, all protein 4.1 members are expressed in the adrenal gland (43). We now show that all members of the protein 4.1 family are also expressed in MEF cells and that 4.1N is up-regulated significantly in 4.1G Ϫ/Ϫ MEF cells, suggesting that 4.1N may partially compensate for the loss of function of 4.1G in vivo. The compensatory effect of 4.1N was also supported by our findings showing that the upregulation of 4.1N in 4.1G Ϫ/Ϫ MEF cells was decreased upon 4.1G rescue (Fig. 2G) and that the attempted knockdown of 4.1N in 4.1G-deficient MEF cells caused severe cell death (data not shown). To further clarify whether protein 4.1 family members have redundant or distinct function, studies on double or triple knockout mice or cells derived from these mice should help to address these issues.