Tetraspanin CD151 regulates α6β1 integrin adhesion strengthening

Abstract

The tetraspanin CD151 molecule associates specifically with laminin-binding integrins, including α6β1. To probe strength of α6β1-dependent adhesion to laminin-1, defined forces (0–1.5 nN) were applied to magnetic laminin-coated microbeads bound to NIH 3T3 cells. For NIH 3T3 cells bearing wild-type CD151, adhesion strengthening was observed, as bead detachment became more difficult over time. In contrast, mutant CD151 (with the C-terminal region replaced) showed impaired adhesion strengthening. Static cell adhesion to laminin-1, and detachment of beads coated with fibronectin or anti-α6 antibody were all unaffected by CD151 mutation. Hence, CD151 plays a key role in selectively strengthening α6β1 integrin-mediated adhesion to laminin-1.

Adhesion receptors in the integrin family bind simultaneously to extracellular matrix ligands and to cytoskeletal proteins, thereby transducing external biomechanical stimuli into internal biochemical responses. Biomechanical forces mediated through integrins regulate cell migration, extracellular matrix assembly and remodeling, wound healing, and tissue morphogenesis (1–5). The application of defined forces on direct engagement of specific integrins with fibronectin (6), laminin (7), or antibody to the integrin β1 subunit (8) results in strengthening of local cytoskeletal linkages. Also, agents such as thrombin may indirectly induce stimulation of integrin-cytoskeletal stiffness, as measured by using fibronectin-coated magnetic beads (9). Because different extracellular matrix protein ligands trigger distinct integrins, coupled to distinct signaling pathways (10–12), mechanisms for regulating cell adhesion-related events could vary considerably. For example, adenocarcinoma cells adhering to fibronectin preferentially develop stress fibers and focal contacts, whereas the same cells adhering to laminin form lamellipodia (10–12). Such results suggest that laminin-binding and fibronectin-binding integrins could have fundamental mechanistic differences. In this regard, only the laminin-binding integrins (α3β1, α6β1, α6β4, and α7β1) show strong lateral association with CD151, a transmembrane protein in the tetraspanin family (13–16). mAb perturbation studies indicate that CD151 modulates integrin-dependent migration, neurite outgrowth, and cell morphology on Matrigel (17–20). The short C-terminal cytoplasmic domain of CD151 was particularly important for α6β1 integrin-mediated spreading, migration, and cellular cable formation on Matrigel (20). Besides CD151, several other members of the tetraspanin protein family (such as CD9, CD81, and CD63) also regulate integrin-dependent cell migration. Although tetraspanin proteins may associate with signaling enzymes and regulate signaling pathways (14, 21–23), the mechanisms whereby they affect cell migration and spreading have not been established.

The preponderance of evidence suggests that CD151 and other tetraspanins do not modulate integrin-dependent static cell adhesion (22). Because CD151 strongly influences α6β1 integrin-dependent cellular cable formation on Matrigel (20), we hypothesized that CD151 could be regulating strength of adhesion mediated through the integrin. To test this hypothesis directly, we engaged α6β1 integrin with laminin-coated beads, exerted a defined mechanical force on the beads, and analyzed bead responses in terms of bead detachment. Magnetic traps have been used previously to explore mechanical properties of cells and to study mechanotransduction events (8, 9, 24–30). Here, we extend the range of magnetic trap application to include probing the adhesion strength of beads coated with either laminin or fibronectin to integrin receptors on the cell surface. Results from this application demonstrate clearly that CD151 plays a key role in regulating the time-dependent gain of adhesion strength for the α6β1 integrin, with the C-terminal region of CD151 being particularly important.

Materials and Methods

Cells and Antibodies. The CD151 C (217)-N mutant was generated (31) and later renamed as CD151-c2-NAG2 (20). For this mutant, the C-terminal HLRVIGAVGIGIACVQVFGMIFTCCLYRSLKLEHY of CD151 was replaced by the corresponding NLLAVGIFGLCTA LVQILGLTFA MTMYCQVVKADTYCA from NAG2 tetraspanin protein. For stable expression of wild-type CD151(CD151-WT) and mutant CD151, plasmid DNAs were transfected into NIH 3T3 cells by using Lipofectamine (Life Technologies, Grand Island, NY). After 48 h, cells were cultured in media containing 200 μg/ml Zeocin (Invitrogen). After 2 weeks of selection, colonies were pooled and CD151-positive cells were sorted by flow cytometry. Surface expression of CD151 and α6 integrin on NIH 3T3 transfectants was assessed by flow cytometry as described (32). Cells were maintained in DMEM (GIBCO) with 10% FCS and antibiotics.

Cell Cable Formation and Adhesion Assays. To observe cellular cable formation, cells were plated on a thick layer of Matrigel in 5% FBS/DMEM at 5 × 10⁴ cells per well in a 24-well plate, analyzed by using a Zeiss Axiovert 135 microscope, and photographed after 18 h as described (20). To measure static cell adhesion, cells were incubated on laminin-1 (coated at 20 μg/ml) for varying times, washed, and attached cells were quantitated by using the Cytofluor 2300 measurement system (Millipore) as described (33).

Particle Tracking and Magnetic Trap Calibration. Cells and magnetic beads were imaged through an inverted light microscope (Olympus, IX-70) at ×20 and ×30 magnification. Images were recorded with a Megaplus ES310/T digital camera (Roper Scientific MASD, San Diego) at 25 frames per sec and stored on a computer. Custom-written MATLAB software was used to track bead position with a spatial resolution of ≈30 nm at ×30 magnification. The magnetic trap was calibrated by suspending 4.5-μm diameter magnetic beads (Dynabeads M-450, Dynal, Great Neck, NY) in dimethylpolysiloxane (DMPS-12 M; Sigma) and tracking bead positions as they were attracted to the magnetic trap operated at a range of electrical currents (0.3–1.5 amps). The force for a given current was computed based on Stokes Law, F = 3πνDu, where ν denotes the viscosity of the fluid, D the bead diameter, and u the relative velocity between fluid and bead, determined from numerically differentiating bead position with respect to time. Measurements were repeated for at least three beads at each setting and an exponential regression of type F = a (x + b)^c was computed as the least-square fit, where F denotes force, x denotes the distance from the magnetic trap and a, b, and c the coefficients to be fitted.

Bead Coating and Measurements of Detachment and Displacement. Magnetic beads (M-450, Dynal) were coated for 18 h at 37°C with laminin (L-2020, Signma) in acetate buffer (pH 4.0) or fibronectin (33016-023, GIBCO/BRL) in borate buffer (pH 8.5), with each protein at 500 μg/ml. Beads were also coated with anti-α6 mAb (555734, PharMingen). Cells were plated in DMEM with 10% FCS, penicillin/streptomycin, and Zeocin (200 μg/ml) at 3 ml per dish on gelatin-coated (0.1% gelatin in PBS, overnight at 4°C) 35-mm polystyrene cell culture dishes (430588, Corning) at a density of 150,000 cells per dish and incubated at 37°C overnight. The medium was replaced the next day with DMEM containing 5% FCS and 3–5 μl of laminin-coated, fibronectin-coated, or anti-α6 mAb-coated magnetic bead suspension (5 × 10⁵ beads per dish) and incubated at 37°C for 120 min to guarantee sufficient bead attachment. On a temperature-controlled stage, cells were imaged at ×30 magnification by using an inverted light microscope (IX-70, Olympus). Nonconfluent cells with a single bead firmly attached (confirmed by a lack of bead motion) were selected for magnetic trap experiments. The magnetic trap was brought into a parfocal position 115 μm from the bead. Digital video acquisition was started while electric current powering the magnetic trap was increased from 0 to 1.5 amps in steps of 0.3 amps every 2 sec (95% of detachment events occur within 1.8 sec). Video acquisition was continued for 2 sec after termination of force application (see Fig. 3D) to monitor bead relaxation. After force application, a new cell was selected at least 5 mm away from any previous force application sites to avoid studying preconditioned cells. Fifteen cells were selected in each dish and experiments were concluded within 30 min per dish. Bead displacement and detachment were evaluated offline using digitally recorded videos. Maximal displacement was defined as the difference in mean bead position between the last 10 frames (= 0.4 sec) of force application at each force level and the initial position, estimated as the mean bead position during the last 25 frames (= 1 sec) before the start of force application.

Fig. 3.

Magnetic trap characterization. (A) Magnetic trap with temperature controlled stage. (B) Shown is force as a function of electric current at a distance of 115 μm from the magnetic trap tip. Line shows linear regression forced through origin after the equation force = 0.965 nN/amp (R² = 0.9948). (C) Bright-field image of cell with bead (black arrow) and magnetic trap tip (white arrow) at ×30 magnification. (D) Force-displacement profile for two single beads, one detaching at 1.2 nN.

Antibody Measurements. CD151-WT cells were plated as described in the protocol for detachment studies. The next day, media was replaced with DMEM containing 5% FCS and 8 μl of laminin-coated magnetic bead suspension (final concentration of 8 × 10⁵ beads per dish) in the absence or presence of purified anti-human integrin α6 antibody (PharMingen, 555734, mAb GoH3, final concentration of 5–10 μg/ml) and incubated at 37°C for ≈120 min. Detachment measurements were carried out as described above.

Attachment Measurements. Cells were plated according to the protocol given for detachment experiments, but at a higher density (≈300,000 cells per dish) to achieve a confluent monolayer of cells. On the next day, media was replaced with DMEM containing 5% FCS and magnetic beads at a final concentration of 5 × 10⁵ beads per dish. The cells were then incubated at 37°C for 120 min. Bead attachment was measured by subjecting a large section of cells to low forces (<0.6 nN) for 4 sec and counting the fraction of beads that remained firmly attached during that time. In contrast to the detachment study, we did not select for firmly attached beads but included all beads located on the cell surface. For each dish, 10–12 sections were evaluated, with a distance of at least 5 mm between sections. Experiments were repeated on at least two dishes per cell type for a total number of ≈100–200 cells.

Statistics. Statistical analysis was performed by using the INSTAT software (GraphPad, San Diego). Differences in adhesion and detachment events were evaluated by using a 2-by-2 contingency table and applying Fisher's exact test. For the displacement measurements, the nonparametric Mann–Whitney test was used to detect differences in median bead displacement, while an unpaired t test (allowing different SD) was used for the log-transformed data. Results are expressed as mean ± SE. A two-tailed P value of <0.05 was considered significant.

Results

Characterization of CD151 Mutant. Previously we demonstrated that deletion or exchange of the CD151 C-terminal cytoplasmic tail region did not alter cell surface expression, CD151-integrin association, or integrin-dependent static cell adhesion to laminin-1 (20). However, such mutations did markedly impair α6β1 integrin-dependent cell spreading and formation of cellular cables (20). Confirming and extending those results, we show here that mutant CD151-c2-NAG2 (human CD151 with C-terminal region replaced by corresponding sequence from NAG2 tetraspanin) failed to support cellular cable formation by NIH 3T3 cells when plated on Matrigel (Fig. 1 Bottom). In contrast, NIH 3T3 cells that were mock-transfected (Fig. 1 Top), or expressing CD151-WT (Middle), showed abundant cable formation. Nearly complete inhibition by anti-α6 mAb GoH3 (Fig. 1 Left) and strong inhibition by anti-human CD151 mAb (Fig. 1 Center) confirm that cellular cable formation depends on the α6β1 integrin, and on human CD151, when it is present. In the absence of human CD151, NIH 3T3 cells were not affected by anti-CD151 mAb 5C11 (Top and Middle). As seen (20), the CD151-c2-NAG2 mutant (present at 2- to 3-fold above endogenous CD151) is likely exerting a dominant negative effect on the function of endogenous murine CD151.

Fig. 1.

CD151–α6β1 integrin complex contributes to NIH 3T3 cell assembly into cables. As described (20), mock-transfected or CD151-transfected NIH 3T3 cells were plated for 18 h on the surface of a thick layer of Matrigel in 5% FBS/DMEM at 5 × 10⁴ cells per well in a 24-well plate, analyzed using a Zeiss Axiovert 135 microscope, and photographed. In some cases, anti-CD151 mAb 5C11 or anti-α6 mAb GoH3 were added (7.5 μg/ml) at the beginning of the experiment.

CD151-WT and mutant CD151 supported similar levels of static cell adhesion to laminin-1 (at 15 or 40 min; Fig. 2A). This adhesion was almost entirely dependent on the α6β1 integrin, as evidenced by the strong inhibitory effects of mAb GoH3 on all NIH 3T3 cells, regardless of presence of CD151-WT or mutant CD151 (Fig. 2 A). As measured by flow cytometry, surface levels of mutant CD151 and CD151-WT were comparable (Fig. 2B Right) as were levels of endogenous α6 in the same cells (Fig. 2B Left). As seen previously, the CD151-c2-NAG2 mutant retained full association with laminin-binding integrins α6β1 (in NIH 3T3 cells) and α3β1 (in other cells) (20, 34).

Fig. 2.

Mutant CD151 and CD151-WT effects on static cell adhesion and cell surface expression. (A) Transfected NIH 3T3 cells were labeled with BCECF-AM, plated for 15 or 40 min on a plastic surface coated with laminin-1, and after washing, adhesion was quantitated as described (20). In some experiments, anti-α6 mAb GoH3 (7.5 μg/ml) was added. (B) NIH 3T3 cells transfected with CD151-WT or CD151-c2-NAG2 were analyzed by flow cytometry for α6 integrin (mAb GoH3) or CD151 (mAb 5C11) as described (20).

Magnetic Trap Calibration. To investigate possible influence of CD151 on strength of integrin adhesion, we used a magnetic trap (Fig. 3A). The magnetic force applied to a single paramagnetic bead is proportional to the product of the magnetic field and its gradient. Therefore, the force depends on the electric current powering the magnetic trap and on the distance of the bead from the pole tip, as the magnetic field decays exponentially with the distance from the tip (see Fig. 7A, which is published as supporting information on the PNAS web site, www.pnas.org). At a given distance from the tip of our magnetic trap (Fig. 3A), the force varies almost linearly with the applied current (Figs. 3B and 7B). For example, when the magnetic trap was positioned 115 μm away from the bead and operated at electric currents of 0.3, 0.6, 0.9, 1.2, and 1.5 amps, this resulted in precisely controlled linear forces of 0.24 ± 0.002, 0.57 ± 0.03, 0.85 ± 0.003, 1.23 ± 0.04, and 1.42 ± 0.04 nN, respectively, on a single 4.5-μm paramagnetic bead. Nonconfluent cells with single-laminin-coated or fibronectin-coated magnetic beads attached to the cell surface were selected for magnetic trap experiments and firm attachment of bead to cell was confirmed (Fig. 3C). The bead was then subjected to a magnetic force that increased from 0 to nearly 1.5 nN in steps of 0.3 nN, using a step duration of 2 sec. Bead position and attachment were continuously monitored during the force application (Fig. 3D). To avoid possible preconditioning effects of bead-cell attachment, subsequent beads were selected at least 5 mm away from previous force application sites. These results establish that bead detachment and bead displacement can be measured in response to carefully defined forces.

CD151 Mutation Effects on Bead Detachment. To assess CD151 C-terminal mutation effects on α6β1 integrin function, we added beads coated with laminin to CD151–2c-NAG2-transfected NIH 3T3 cells, waited 2 h, and then measured bead detachment. In an initial experiment (at force = 1 nN), the majority of laminin beads (24 of 37) detached from CD151–2c-NAG2 transfectants, whereas only a few (5 of 43) detached from mock-transfected NIH 3T3 cells (Fig. 4A). In a control experiment, only a few beads coated with fibronectin (which engages the α5β1 integrin) detached from either CD151–2c-NAG2 (5 of 28) or mock (2 of 21) transfectants (Fig. 4B), indicating that the CD151–2c-NAG2 mutation was selectively affecting laminin bead detachment. A more comprehensive experiment was then carried out (Fig. 4C) over a range of forces (0–1.5 nN). Again the CD151–2c-NAG2 mutant showed significantly easier detachment of laminin-coated beads. This was especially evident for forces of 1.2 nN and higher (P < 0.001, Fig. 4C). Mock-transfected and CD151-WT transfected NIH 3T3 cells showed comparable bead detachment across the whole range of forces. Among the three transfectants, no significant differences in bead detachment were observed for fibronectin-coated beads at any force level (Fig. 4D).

Fig. 4.

Detachment of laminin-coated and fibronectin-coated beads from NIH 3T3 transfectants. (A) Detachment of laminin-1-coated beads at 1 nN (*, P < 0.0001). (B) Detachment of fibronectin-coated beads at 1 nN. (C) Bead detachment fraction at increasing force levels for laminin-coated beads (*, P < 0.01 vs. CD151-WT cells; **, P < 0.001). (D) Shown is the bead detachment fraction at increasing levels for fibronectin-coated beads. At lower fibronectin-coating levels, we could decrease bead attachment, but those beads that attached firmly withstood even the highest forces (>3 nN) without detaching.

To ascertain the extent to which resistance to laminin-bead detachment is α6β1 integrin-dependent, CD151-WT transfectants were incubated with laminin beads in the presence of purified anti-murine α6 antibody (5–10 μg/ml) and then bead detachment was measured. As indicated (Fig. 5A), bead detachment in the presence of anti-α6 antibody was significantly enhanced over the entire 0–1.5 nN force range. The difference between the two curves defines the magnitude of bead detachment that depends on the α6β1 integrin. Remarkably, the magnitude of this antibody effect (Fig. 5A) is very similar to the magnitude of the CD151 tail-mutation effect (Fig. 4C).

Fig. 5.

Bead detachment effects of anti-α6 antibody and bead attachment by laminin and fibronectin beads. (A) NIH 3T3-CD151-WT cells were incubated in the presence (▪, n = 89) or absence (•, n = 105) of mAb GoH3 (5–10 μg/ml) at the time of bead attachment (*, P < 0.05). The nature of the detachment assay requires selection of beads that have already attached and hence are more difficult to displace due to antibody inhibition. In contrast, during the static cell adhesion assay, integrin contacts with laminin-1 are not allowed to develop when the inhibitory antibody is present. Hence, GoH3 shows only a partial effect in A, compared with a more complete inhibition in Fig. 2 A. (B) For laminin-1-coated beads, attachment fractions = 0.911 ± 0.021, 0.912 ± 0.024, and 0.973 ± 0.013 for CD151-c2-NAG2, CD151-WT, and pZeo, respectively (*, P < 0.05 vs. CD151-c2-NAG2 and CD151 WT). For fibronectin-coated beads, attachment fractions = 0.951 ± 0.019, 0.973 ± 0.015, and 0.946 ± 0.023 for CD151-c2-NAG2, CD151-WT, and pZeo, respectively.

To confirm that the difference in bead detachment is caused by a lack of adhesion strengthening and not a deficiency in initial bead attachment, we measured the adhesion of laminin-coated and fibronectin-coated beads to the transfected cells. Bead adhesion was defined as the fraction of beads located on the cell surface that withstood a weak force application (0.3–0.6 nN) for 4 sec. This force level was sufficient to wash off undetached beads, but was insufficient to detach beads that had firmly attached to the cell surface (compare with results of Fig. 4 C and D). The C-terminal mutation did not affect adhesion of laminin-coated beads as attachment fractions were comparable for CD151-c2-NAG2 and CD151-WT with the mock-transfectants exhibiting a slightly higher attachment fraction (P < 0.05) compared with both c2-NAG2 and CD151-WT (Fig. 5B). Attachment of fibronectin-coated beads was also uniformly high with no significant differences among CD151 transfectants (Fig. 5B). Beads coated with BSA showed no attachment to any of the cells (data not shown).

Whereas bead detachment was significantly different for CD151-WT and CD151 mutant cells after 2 h (Fig. 4C), at an earlier time point (beads attached for 30 min instead of 2 h), CD151-WT and CD151 mutant cells showed a similar ease of detachment of laminin-coated beads (Fig. 6A, with key curves from Fig. 4C superimposed). In contrast to the laminin-coated beads, anti-α6 antibody-coated beads did not detach very readily at 2 h (Fig. 6B) or 30 min (data not shown), and this low level of detachment was not affected by mutant CD151 (Fig. 6B).

Fig. 6.

CD151 mutant effects are time-dependent and ligand-dependent. (A) Shown is the bead detachment fraction at increasing force levels for laminin-coated beads after only 30 min of attachment (squares). For comparison, detachment results after2hof bead attachment are also shown (circles, as in Fig. 4C). (B) Shown is the bead detachment fraction at increasing force levels for anti-α6 antibody-coated beads that had been allowed to attach for 2 h.

Increased Bead Displacement for CD151 Mutant. Beads remaining attached could nonetheless show variable bead displacement. To assess whether connection between the extracellular matrix and the cytoskeleton is impaired in CD151 C-terminal mutants, we measured the maximal displacement of magnetic beads during force application. To avoid artifacts such as bead tethering or partial detachment associated with high force levels, we limited our evaluation to bead displacements at the low force level of 0.3 nN. The C-tail exchange mutation (CD151-c2-NAG2) resulted in a significantly increased median bead displacement for laminin-coated beads (0.32 μm vs. 0.20 μm, P = 0.0022; see Fig. 8 A and C, which is published as supporting information on the PNAS web site). No significant difference in bead displacement was observed for fibronectin-coated beads (0.15 vs. 0.10 μm, P = 0.131; see Fig. 8 B and D).

Discussion

Our results establish that CD151 plays a key role as a regulator of α6β1 integrin adhesion strengthening. Mutation of the CD151 C-terminal region had little or no effect on static cell adhesion to laminin-1, bead attachment, or laminin-coated bead detachment after bead attachment for only 30 min. However, after beads had attached for 2 h, the CD151 mutant did not show nearly as much adhesion strengthening as CD151-WT. This finding was manifested as increased detachment differential between mutant CD151 and CD151-WT (especially at 0.9–1.5 nN), with mutant beads being more easily detached. Bead displacement was also significantly increased for mutant CD151, consistent with decreased cellular stiffness. Because the cytoskeleton is the main contributor to cellular stiffness, we conclude that our mutation of CD151 has disrupted dynamic cytoskeleton-dependent processes, critical for α6β1 adhesion strengthening. Previous results (17) showed that CD151 associated strongly with α6β1 (a laminin receptor), but not α5β1 (a fibronectin receptor). Those results are entirely consistent with our CD151 mutation affecting detachment and displacement of laminin-1 beads, but not fibronectin beads. We predict that CD151 should also regulate adhesion strengthening through the α3β1, α6β4, and α7β1 integrins, because those integrins also associate strongly with CD151 (16). It remains unclear why laminin-binding integrins, but not fibronectin-binding integrins or other integrins would need a tetraspanin protein to augment their adhesion strengthening capabilities.

Elsewhere we observed CD151-dependent cell spreading and cellular cable formation on Matrigel, but the role of CD151 was not clear. Our data now support a ``transmembrane linker'' model in which (i) the integrin contacts laminin, (ii) CD151 forms an extracellular contact with the α subunit of laminin-binding integrins (31, 35), and (iii) CD151 uses its short cytoplasmic tail to engage as yet unidentified membrane-proximal elements. These unknown elements are critical for strengthening integrin-mediated adhesion, such that it withstands larger mechanical forces. By mutating the CD151 C-terminal region, adhesion strengthening is impaired (this article), and cells cannot spread or form cables (20), likely because adhesion strengthening is needed to transmit the mechanical forces that play an essential role in the process of integrin-dependent cellular cable formation (36, 37) and cell spreading. Consistent with our transmembrane-linker model, mutation of either the CD151 C-terminal region, or an extracellular ``QRD'' site needed for integrin association, resulted in severely impaired cellular cable formation (20, 34). Thus the extracellular CD151 integrin-association site may be just as important as the intracellular C-terminal site with respect to adhesion strengthening, although this remains to be tested in the magnetic bead assay. Providing further support for our transmembrane-linker model, inhibition of extracellular integrin-ligand binding by an anti-α6 mAb, and mutation of the intracellular CD151 C-terminal region yielded remarkably similar bead detachment profiles over a range of increasing forces. If we accept that the antibody inhibition experiment (Fig. 5A) roughly defines the window of bead detachment that is integrin-dependent, then by comparison (Fig. 4A), the C-terminal region of CD151 is at least as critical as, and perhaps even more critical than, antibody-sensitive extracellular contacts. Stated another way, the entire window of bead detachment that is α6 integrin-dependent appears to be eliminated on C-terminal tail deletion of CD151.

Historically, studies of integrin transmembrane-linker functions have focused on cytoplasmic tails of integrins themselves. Our results now expand the focus to include the C-terminal tail of CD151. Elsewhere, three different types of CD151 C-terminal mutations were shown to eliminate cellular cable formation (20). As confirmed here, exchange of the C-terminal region with a corresponding region from tetraspanin NAG2 eliminated cable formation and adhesion strengthening. A shorter exchange with the C-tail of tetraspanin A15/TM4SF6, or a deletion of the CD151 tail also abolished cable formation (20). Here we have carried out magnetic bead studies for only one CD151 mutant, but we expect that other CD151 tail mutations that disrupt cellular cable formation and cell spreading should also affect detachment and displacement of laminin-coated beads. Previous and current results focus attention on the C-terminal 7 aa of CD151 (SLKLEHY) as being critical for adhesion strengthening. It remains to be seen what other proteins might be specifically engaged by this region of CD151. The CD151 molecule could potentially modulate strength of adhesion by affecting integrin clustering. Because CD151 mutations have not altered detachment of anti-α6 antibody-coated beads, static cell adhesion, or early (30 min) bead detachment, we do not suspect alterations in constitutive integrin clustering. However, CD151 tail mutation could affect adhesion strengthening by altering the dynamic integrin clustering that occurs after contact with laminin-coated beads. In this regard, it is ligand occupancy plus clustering rather than clustering alone that promotes recruitment of cytoskeletal proteins to the α5β1 integrin (38).

Results were obtained by using an assay that measures magnetic bead detachment and displacement. Magnetic bead experiments have been previously used to measure the mechanical properties of the cytoskeleton and the cell surface (8, 9, 24–26, 28, 30) and to apply mechanical stimulation to study mechanotransduction events and signaling (27, 29). Others observed (26) that vinculin-deficient cells showed increased bead displacements compared with wild-type cells, demonstrating the capabilities of the magnetic trap system to measure the changes in stiffness of the transmembrane integrin linkages to the cytoskeleton associated with specific proteins. Here, we use high-resolution (<10 nm) single-particle tracking at very low force levels to detect small differences in the mechanical coupling between extracellular matrix proteins and the cytoskeleton caused by a mutation in the CD151 tetraspanin. In addition, we further extend the range of magnetic trap applications. We measure the adhesion strength and cellular mechanics of the cell/extracellular matrix interaction by applying a wide range of forces (0–1.5 nN) and measuring bead detachment or displacement as a function of applied force. In the study of cell adhesion, this approach differs from other quantitative cell attachment methods that are designed to integrate adhesion results from many cells at once (39, 40).

Techniques such as atomic force microscopy, micropipette aspiration, or optical traps have been traditionally used for the study of single integrin-ligand molecular adhesion forces in the range of 30–150 pN (41–45). By contrast, our method can apply much higher force levels (0–1.5 nN), well suited to study the concerted action of many adhesion receptors simultaneously. At present, there is no evidence that tetraspanins can modulate adhesion forces for individual integrin heterodimers. Rather, the current results are consistent with CD151 affecting adhesion strengthening that is mediated through many integrins (e.g., through effects on cytoskeletal linkages and/or integrin clustering). Notably, the magnitude of cellular force regulated by CD151 (1–1.5 nN) is consistent with the ≈1–5 nN forces exerted by keratocytes during cell migration (46) and fibroblasts during collagen gel contraction (47). Hence, we can now better understand why CD151 mutation caused alterations in cell migration, spreading, and cellular cable formation.

In conclusion, we have demonstrated that mutation of the C-terminal region of tetraspanin CD151 markedly reduces integrin α6β1 adhesion strengthening. The concept of tetraspanins as regulators of adhesion strengthening emerges as a useful paradigm to help explain many previous examples in which tetraspanins were shown to regulate cell migration, spreading, morphology, and cellular cable formation on Matrigel basement membrane. Our results focus attention on the CD151 cytoplasmic tail as being important for modulation of cytoskeletal engagements. Notably, CD151 complexes with laminin-binding integrins are present in a variety of tissue locations, including smooth muscle costameres (16) and endothelial cell–cell junctions (18, 48). Thus, CD151 modulation of adhesion strengthening could play a key role in force-dependent processes such as muscle contraction and angiogenesis.