Lymphocytes use integrin-based platforms to move and adhere firmly to the surface of other cells. The molecular mechanisms governing lymphocyte adhesion dynamics are however poorly understood. In this study, we show that in mouse B lymphocytes, the actin binding protein vinculin localizes to the ring-shaped integrin-rich domain of the immune synapse (IS); the assembly of this platform, triggered by cognate immune interactions, is needed for chemokine-mediated B cell motility arrest and leads to firm, long-lasting B cell adhesion to the APC. Vinculin is recruited early in IS formation, in parallel to a local phosphatidylinositol (4,5)-bisphosphate wave, and requires spleen tyrosine kinase activity. Lack of vinculin at the IS impairs firm adhesion, promoting, in turn, cell migration with Ag clustered at the uropod. Vinculin localization to the B cell contact area depends on actomyosin. These results identify vinculin as a major controller of integrin-mediated adhesion dynamics in B cells.
This article is featured in In This Issue, p.2023
The regulated interplay between cell adhesion and cell motility is critical for B lymphocyte function. B cells must explore entire follicles in secondary lymphoid organs, where Ags are collected and presented by various APCs (1). To do this, B cells migrate continuously by random walking in response to the chemokine CXCL13 (2–4). This chemokine is produced mainly at the network of follicular dendritic cells; they expose it on their surface in the context of integrin ligands, which might assist in B cell motility (4, 5). Specific BCR recognition of Ag above a signaling threshold leads B cells to adhere firmly to the APC; a large LFA-1 integrin cluster is assembled, and the IS is formed (6, 7). The synapse platform has an important role in several aspects of the B cell activation process (6, 8, 9). The control of integrin activation, clustering, and localization thus underlies the precise modulation of B cell behavior.
Chemokine receptor and BCR signaling activate LFA-1 by modulating integrin affinity (conformational change) and avidity (spatial distribution and clustering) for its ligand ICAM-1 (10, 11). IS formation also drives LFA-1/ICAM-1 segregation into a peripheral ring (peripheral supramolecular activation cluster [pSMAC]) that surrounds the central BCR/Ag cluster (central SMAC [cSMAC]) at the site of B cell–APC contact (6). IS assembly involves actin cytoskeleton remodeling and the formation of an F-actin—rich ring at the pSMAC, where LFA-1 anchors through the scaffold protein talin (12, 13). The strength of cognate interactions through the BCR determines synapse formation or the establishment of CXCL13-mediated LFA-1–supported migratory platforms (kinapse) that allow Ag recognition (7). The molecular mechanisms used by the CXCL13 receptor CXCR5 and the BCR to coordinate LFA-1 function, and thus, B cell adhesion and motility remain almost unexplored.
The scaffold protein vinculin is being recognized as a key regulatory element of adhesion dynamics in nonimmune cells. It controls assembly, strength, and transmission of mechanical forces at focal adhesions (FA); these specialized structures support integrin-mediated cell contact with the extracellular matrix (14–17). Vinculin links the actin cytoskeleton with integrins at the plasma membrane through its association with talin and F-actin (14). Vinculin activation and translocation from cytosol to adhesion sites require its interaction with the membrane phospholipid phosphatidylinositol (4,5)-bisphosphate (PIP2) (18). Local and temporal PIP2 production at the plasma membrane controls adhesion site assembly and actin dynamics (19). PIP2 is generated mainly through PIP (4)-phosphate phosphorylation at the 5-position by type I PIP kinases (PIPKI); of the PIPKI isoforms and splice variants, PIPKIγ targets to FA (20, 21). Vinculin is essential for embryonic development (22); cancer cells that lack vinculin are highly motile and metastatic (23–25), and vinculin overexpression reduces cell motility and enhances cell adhesion (26).
In this study, we identified a major role for vinculin in governing B cell adhesion and motility. We showed that BCR recognition of membrane-tethered Ag (tAg) led to vinculin recruitment to the immune synapse. Vinculin localized to the pSMAC to strengthen B cell adhesion; loss of vinculin impeded appropriate pSMAC assembly, and in turn, B cells moved in response to CXCL13, bearing the Ag cluster at the uropod. We demonstrated that spleen tyrosine kinase (Syk) and actomyosin control vinculin recruitment and stability at the B cell synapse.
Materials and Methods
Mice and cells
Primary B cells were freshly isolated from spleens of wild-type and MD4 BCR transgenic mice on the C57BL/6 genetic background by negative selection (>95% purity), as described (7). For time-lapse experiments, purified B cells were labeled before use with 0.1 μM CFSE long-term dye (Molecular Probes; 10 min, 37°C). Animal experimentation was approved by the Centro Nacional de Biotecnología-Consejo Superior de Investigaciones Científicas Bioethics Committee and conforms to institutional, national, and European Union regulations. The A20 murine B cell line was transiently transfected by electroporation with PIPKIγ-GFP [a kind gift from Rosana Lacalle (27)], the PIP2 probe PLCδ-PH-GFP [a kind gift from Isabel Mérida (28)], vinculin-GFP [a kind gift from Miguel Vicente-Manzanares (29)], and the F-actin probe Lifeact-RFP constructs [a kind gift from Mario Mellado (30)] and used 20 h later for time-lapse experiments.
We prepared artificial planar lipid bilayers containing glycosyl phosphatidylinositol (GPI)-linked mouse ICAM-1 (density 150 molecules/μm2) and when indicated, biotin-modified phospholipids at specific molecular densities (7). Membranes were assembled on FCS2 closed chambers (Bioptechs) and blocked with PBS/2% FCS (1 h, room temperature [RT]). Ag (density 20 molecules/μm2) was tethered to the membranes by incubating with Alexa Fluor 647 or Alexa Fluor 555 streptavidin (Molecular Probes), followed by monobiotinylated hen egg lysozyme (Sigma-Aldrich) for MD4 B cells or monobiotinylated anti-κ L-chain mAb (BD Biosciences) for wild-type B cells and the A20 B cell line. Before imaging, membranes were coated with 100 nM recombinant murine CXCL13 (PeproTech). We estimated the density of chemokine deposited on the lipid bilayer by an established immunofluorometric assay (6) using anti-mouse CXCL13 Ab (R&D Systems) and, for standard values, microbeads with calibrated IgG-binding capacities (Quantum Simply Cellular Kit, Bangs Laboratories); the value obtained was in the range of 30–40 molecules/μm2. Unlabeled or CFSE-labeled primary B cells (2 × 106) and transfected A20 B cells (1 × 106) were injected into the warmed chamber (37°C), and imaging was started. Confocal fluorescence, differential interference contrast (DIC), and interference reflection microscopy (IRM) images were acquired every 30 s for 20 min; when indicated, consecutive videos were acquired. Assays were performed in PBS/0.5% FCS/0.5 g/l d-glucose/2 mM MgCl2/0.5 mM CaCl2. For soluble Ag (sAg) stimulation, F(ab′)2 anti-IgM Ab (Jackson ImmunoResearch Laboratories) was added at 1 μg/ml final concentration to the B cell suspension immediately before injection into the FCS2 chamber. When indicated, primary B cells were treated with specified doses of the chemical inhibitor BAY 61-3606 (BAY; Calbiochem; 20 min, 30°C) before injection. B cells were treated with blebbistatin (50 μM; Calbiochem) in situ by injection into the FCS2 chamber, incubated 5–10 min, and imaged. Images were acquired on an Axiovert LSM 510-META inverted microscope with a 40× oil immersion objective (Zeiss) and analyzed with Imaris 7.0 software (Bitplane).
Primary B cells were in contact with planar lipid bilayers containing GPI-linked ICAM-1 and CXCL13 coating, alone or with tAg or with sAg (30 min), fixed with 4% paraformaldehyde (10 min, 37°C), permeabilized with PBS/0.1% Triton X-100 (5 min, RT), blocked with PBS/2% FCS/2% BSA (overnight, 4°C), and stained with Alexa Fluor 647 phalloidin (Molecular Probes), rabbit anti–phospho-Syk (Tyr352) (Cell Signaling Technology) plus Alexa Fluor 488 goat anti-rabbit IgG (Southern Biotechnology Associates), and mouse anti-talin or -vinculin (clones 8d4 and hVIN-1, respectively; Sigma-Aldrich) plus FITC goat-anti-mouse IgG1 (BD Biosciences) (30 min, RT). FCS2 chambers were imaged by confocal fluorescent microscopy on a Zeiss Axiovert inverted microscope (Zeiss) as above.
We used Imaris 7.0 software for qualitative and quantitative analysis of fluorescence signals at distinct cell planes and in the whole cell volume, as well as IRM area measurements. Ratios were obtained dividing the total fluorescence of the indicated protein at the synapse/contact plane between the total fluorescence at the midplane. Total fluorescence at the entire cell volume was obtained from z-stack images (optical slice thickness 1 μm) using Imaris 7.0 software.
Infection using lentiviral vectors
Recombinant lentiviral particle stocks were obtained from HEK 293T cells by cotransfecting the short hairpin RNA (shRNA)-coding vector (pLKO.1, pGIPZ), the pMD.2G envelope vector, and the pCMVR8.91 packaging vector (31). We used two types of shRNA coding vectors: mouse GIPZ lentiviral shRNAmir vectors (clones V2LMM_45006, V2LMM_56452, and V3LMM_437636, coding for mouse vinculin-specific shRNA, and a nonsilencing GIPZ lentiviral shRNAmir control; Thermo Scientific) and mouse vinculin shRNA in pLKO.1 vector backbones (clones NM_009502.3-3466s1c1, NM_009502.3-1331s1c1, and NM_009502.3-3154s1c1; Mission shRNA; Sigma-Aldrich). Briefly, 2 × 106 cells were plated on p150 dishes (48 h) and then transfected with 2 μg envelope vector, 5 μg packaging vector, and 7 μg shRNA-coding vector by precomplexing with JetPEI (0.1 mM final concentration; Polyplus Transfection; 30 min, RT) in OptiMEM (Life Technologies). After 4 h at 37°C, we replaced medium with fresh DMEM/2% FCS; virus particles were harvested 48 or 72 h posttransfection. The virus suspension was filtered (0.45 μM pore size) and concentrated by ultracentrifugation (23,000 rpm, 2 h, 4°C). The pellet was resuspended in RPMI 1640 and stored at −80°C. Freshly isolated primary B cells (2 × 106) were infected with lentiviral particles (multiplicity of infection 1–10) in 500 μl RPMI 1640/10% FCS, alone or with recombinant murine IL-4 (50 ng/ml; PeproTech), CpG (1 μg/ml; Invivogen), or LPS (2.5 μg/ml; Sigma-Aldrich) for 6 h at 37°C. Medium was replaced with fresh RPMI 1640/10% FCS, alone or with the specified stimuli, and infected B cells cultured (48 h) to allow shRNA and gene reporter expression. Vinculin protein levels and GFP reporter expression were analyzed in Western blot.
Western blot analysis
Freshly isolated primary B cells (5 × 106) were cultured in depletion medium (RPMI 1640/0.5% FCS; 1 h, 37°C) and then stimulated with F(ab′)2 anti-IgM Ab (1 μg/ml; with shaking) or with ICAM-1/CXCL13 membranes in absence or presence of tAg for 30 min at 37°C. Ice-cold PBS was added and B cells centrifuged (2000 rpm, 5 min, 4°C) and lysed in RIPA lysis buffer with protease and phosphatase inhibitors (Roche; 30 min, 4°C). Lysates were centrifuged (14,000 rpm, 30 min, 4°C) and supernatants collected and stored at −80°C. Lentiviral particle–infected B cell lysates were obtained similarly. Total protein was quantified with the Micro BCA Protein assay kit (Thermo Scientific), separated by SDS-PAGE, and transferred to polyvinylidene difluoride membranes (Bio-Rad). Blots were blocked with 2% BSA in TBST (10 mM Tris-HCL [pH 8], 150 mM NaCl, and 0.1% Tween-20) (1 h, RT) and incubated with rabbit anti-Syk, rabbit anti–phospho-Syk (Tyr352), rabbit anti–phospho-ERK1/2 (Cell Signaling Technology), mouse anti-vinculin (clone hVIN-1; Sigma-Aldrich), mouse anti–β-actin (Sigma-Aldrich), or mouse anti–α-tubulin (clone DM1A; Sigma-Aldrich (overnight, 4°C), followed by HRP-conjugated secondary Abs (1 h, RT); the signal was detected with the ECL detection system (GE Healthcare). Signal intensity values in arbitrary units for each protein (p-Erk, vinculin) were quantified using ImageJ software (National Institutes of Health), normalized to tubulin or β-actin signal, relative to individual controls.
Graphs and statistical analysis were done using Prism 4.0 software (GraphPad). Two-tailed unpaired Student t test was applied: *p < 0.05, **p < 0.001, ***p < 0.0001.
B cell motility arrest requires IS assembly triggered by membrane-tAg recognition
We tested whether, in the absence of IS establishment, BCR-induced LFA-1 activation was sufficient to halt CXCL13-mediated B cell migration. BCR recognition of membrane tAg leads to B cell synapse formation; to trigger LFA-1 activation in the absence of IS assembly, we used Ag in soluble form (sAg). To study B cell dynamics in response to CXCL13 and Ag, we used a two-dimensional model combined with time-lapse confocal microscopy; this experimental system is based on the assembly of planar lipid bilayers containing GPI-linked ICAM-1, a chemokine coating, and membrane-tAg. The model mimics the surface of an APC and reproduces B cell dynamics to those observed in vivo by multiphoton microscopy (7).
Freshly isolated naive B cells from wild-type mouse spleen were labeled with the fluorescent probe CFSE and allowed to settle on artificial planar lipid bilayers containing GPI-linked ICAM-1 and a CXCL13 coating (ICAM-1/CXCL13 membranes), alone or with sAg [1 μg/ml F(ab′)2 anti-IgM Ab] or tAg (anti-κ L-chain at 20 molecules/μm2). We monitored B cell dynamics by time-lapse microscopy. B cells migrated by random walking across CXCL13 coated ICAM-1–containing artificial membranes (Supplemental Video 1). CXCL13 alone promoted LFA-1/ICAM-1 interactions (65%; detected by IRM; Fig. 1A) and a high frequency of cell polarization (70%), estimated by DIC as the fraction of cells with membrane protrusion activity (membrane ruffles) (Fig. 1B), as reported (7). Half of the polarized cells migrated across the substrate at a mean velocity of 4 μm/min (Fig. 1B, 1C, Supplemental Video 2). Although the presence of sAg increased the fraction of B cells with active LFA-1 (85% IRM+), it did not alter CXCL13-mediated cell polarization and migration. Motile B cells nonetheless showed a significant reduction in mean velocity, suggesting that sAg/BCR signaling impairs chemokine-mediated B cell motility (Fig. 1A, 1C). Recognition of tAg induced the highest frequencies of LFA-1–active B cells (95% IRM+); however, it completely abolished CXCL13-mediated B cell migration (Fig. 1A, 1C, Supplemental Video 3), as expected. We obtained similar results using MD4 BCR-transgenic B cells, their specific Ag hen egg lysozyme in membrane-tethered form and F(ab′)2 anti-IgM Ab as sAg (Supplemental Fig. 1A, 1B). These data thus indicated that tAg/BCR-mediated IS assembly is needed for B cell arrest, as LFA-1 activation triggered by sAg allowed motility.
Syk is important for integrin activation by the BCR (32) and by chemokine receptor stimulation (33). We compared the intensity of signals transmitted through the BCR after tAg and sAg stimulation by measuring p-Syk levels in B cells. Each stimulation condition led to distinct p-Syk patterns at the B cell–target membrane contact plane; with tAg, p-Syk concentrated at the IS cSMAC, whereas distribution was homogeneous in the case of sAg (Fig. 1D). Comparison of p-Syk fluorescence intensity at the B cell contact plane and the midplane showed polarization at the contact plane only in tAg conditions (Fig. 1D). p-Syk values at the B cell contact plane were significantly higher in presence of tAg than with sAg; results were similar for p-Syk quantified in the entire B cell volume (Fig. 1E, 1F). Syk protein levels were comparable in all conditions analyzed (Supplemental Fig. 1C). Greater Syk activation and localization at the B cell contact plane promoted by tAg/BCR signaling might be important for LFA-1 activation, synapse assembly, and B cell arrest.
Vinculin is recruited to the integrin-rich domain of the B cell IS
To determine the role of vinculin in stabilizing the B cell IS and arresting B cell motility, we analyzed vinculin at the B cell synapse. B cells in contact with ICAM-1/CXCL13 membranes with tAg (20 min) were fixed and stained for talin, vinculin, and F-actin. We detected vinculin at the B cell IS; it accumulated markedly in the ring-shaped pSMAC structure that matched the F-actin–rich domain (Fig. 2A). Its binding partner talin was also found at the IS, which colocalized mainly with the F-actin–rich pSMAC, but also in other parts of the contact area such as membrane ruffles (Fig. 2A). Comparison of fluorescent signals at the B cell contact plane with those at the midplane indicated that both vinculin and talin adaptor proteins were recruited to the IS. As we found that sAg/BCR signaling did not halt B cell motility, we tested the implication of vinculin in this observation by analyzing its localization to the contact site in sAg stimulation conditions. Vinculin did not localize nor was it distributed in a specific pattern at the plane of B cell contact with the target membrane; it accumulated mainly near the F-actin–rich cell edges (Fig. 2B). Quantification of vinculin at the contact area showed significantly lower values after BCR stimulation with sAg than with tAg; results were comparable when we analyzed total F-actin at the contact area (Fig. 2C).
Thus, membrane-tAg/BCR stimulation leads to vinculin recruitment to the synapse. Vinculin localization at the pSMAC suggested a role for this scaffold protein in adhesion strength and stability of the B cell IS.
Vinculin recruitment coincides with a PIPKIγ-produced local PIP2 wave at the synapse
To study the molecular dynamics of vinculin localization and F-actin polymerization at the IS, we performed time-lapse confocal microscopy experiments with A20 B cells expressing a vinculin-GFP construct and the F-actin probe Lifeact-RFP. Vinculin recruitment followed the formation of nascent Ag clusters and LFA-1/ICAM-1 interactions (Supplemental Fig. 2, Supplemental Video 4). Vinculin accumulated gradually and segregated to the pSMAC in the first 5 min of synapse formation, after which its levels remained almost constant over time (1 h). The Lifeact-RFP profile showed an acute F-actin polymerization phase at the IS in the first 2.5 min, followed by formation of the ring structure in which it merged with vinculin (Supplemental Fig. 2, Supplemental Video 4).
Vinculin activation and translocation from cytosol to adhesion sites require its interaction with PIP2 produced by PIPKI in other nonimmune cells. We studied PIPKIγ and PIP2 dynamics at the B cell–target membrane contact plane by time-lapse confocal microscopy. We used a PIPKIγ-GFP construct and the PIP2 probe PLCδ-PH–GFP construct to transfect A20 B cells; transfectants were tracked for synapse formation in contact with ICAM-1/CXCL13 membranes and tAg (anti-κ; 20 molecules/μm2). We discarded those transfected cells showing high GFP levels for analysis. We detected short-lived PIPKIγ recruitment at early stages of synapse formation, coinciding with nascent LFA-1/ICAM-1 interactions (detected by IRM) and Ag clusters; with time, PIPKIγ persisted at the contact plane at lower levels, comparable to those in the rest of the B cell (Fig. 3A, 3B, Supplemental Video 5). PIP2 production followed a similar pattern, with a maximum immediately after the PIPKIγ peak early in synapse formation (Fig. 3C, 3D, Supplemental Video 6). Both PIPKIγ and PIP2 dynamics were triggered by tAg recognition; neither PIPKIγ recruitment nor marked changes in PIP2 levels were detected at the B cell–target membrane contact area in the presence of the CXCL13 coating only (Fig. 3B, 3D). We also observed that PIPKIγ and PIP2 localized mainly at the pSMAC of the mature B cell synapse, where vinculin accumulated (Fig. 3E, 3F).
Then, after PIPKIγ produced the local PIP2 increase, vinculin localized gradually to the synapse; the local PIP2 wave might be needed for vinculin recruitment to the site of B cell–target membrane contact.
Impaired vinculin recruitment to the synapse allows B cell motility
To determine the relevance of vinculin in arresting motile B cells after tAg encounter, we attempted to knock down its expression in primary B cells. Lentiviral vectors coding for distinct mouse vinculin-specific shRNA were used to generate lentiviral particles and to infect B cells in several conditions (no stimulus, IL-4, CpG, or LPS; see 2Materials and Methods). At 48 h postinfection, we used Western blot to analyze vinculin levels in infected B cells. GFP reporter expression confirmed primary B cell infection by the lentiviral particles, with distinct efficiency depending on the stimulation (Supplemental Fig. 3A). We nonetheless found no clear reduction in vinculin levels with any of the shRNA used (Supplemental Fig. 3A, 3B). We did not analyze longer time points because infected primary B cells died or differentiated into plasma cells.
As vinculin recruitment was associated with strong local Syk activation through tAg/BCR signaling (Fig. 1D, 2B), we tested another approach to impair vinculin function. We used the specific chemical inhibitor BAY to interfere with Syk activity. B cells were treated with several BAY doses (1 to 0.1 μM); after BCR stimulation, we evaluated Syk activity by detection of p-ERK, a downstream effector. We also assessed the ability of BAY-treated B cells to migrate on ICAM-1/CXCL13 membranes to determine Syk inhibition downstream of CXCR5. High BAY doses (1 and 0.6 μM) abolished BCR-mediated Syk activation and reduced chemokine-triggered B cell motility (Supplemental Fig. 4A, 4B). Treatment with 0.3 μM BAY impaired BCR-triggered Syk activity, but allowed a higher frequency of migrating B cells, with no significant alteration in mean velocity compared with controls (Supplemental Fig. 4A, 4B). We evaluated the effect of this BAY dose on vinculin recruitment to the IS. BAY-treated B cells showed lower vinculin levels at the contact plane and a distribution pattern distinct from that of untreated B cells (Fig. 4A, Supplemental Fig. 4C). Absence of the F-actin–rich ring indicated profound alterations in pSMAC assembly in BAY-treated B cells (Fig. 4A). The ratio of vinculin at the contact plane with those at the midplane indicated almost no vinculin recruitment to the IS in BAY-treated B cells; F-actin polymerization was also impaired (Fig. 4B, Supplemental Fig. 4C). BAY treatment did not alter cSMAC formation or talin polarization to the contact plane (Fig. 4C). B cells treated with a lower BAY dose (0.1 μM) did not show any change in vinculin recruitment and localization at the IS (Supplemental Fig. 4C, 4D).
We used time-lapse microscopy to monitor the behavior of CFSE-labeled B cells, untreated or treated with 0.3 μM BAY, in contact with ICAM-1/CXCL13 membranes with tAg. Untreated B cells showed the predicted IS establishment and CXCL13-driven membrane ruffles; cells hardly moved from their position (Fig. 5A, Supplemental Video 7). A large percentage of BAY-treated B cells (40%) assembled an Ag cluster and migrated across the membrane (Fig. 5A, 5B, Supplemental Video 8). Motile BAY-treated B cells extended a clear lamellipodium at the cell front and carried the Ag cluster at the back uropod (Fig. 5A, Supplemental Video 8); they reached mean velocity values at ∼3 μm/min (Fig. 5C). These data indicated that vinculin recruitment is important for adhesion strength at the IS and to arrest chemokine-mediated B cell motility.
Non–muscle myosin-II activity is necessary for vinculin function at the B cell synapse
In nonimmune cells, vinculin recruitment to and function at focal adhesions requires non–muscle motor protein myosin-II (NM-II) activity (34). Active NM-II is present at the B cell synapse (7). We studied the role of NM-II in vinculin function at the B cell IS using the specific chemical inhibitor blebbistatin to interfere with NM-II activity. B cells were allowed to settle and establish an IS in contact with ICAM-1/CXCL13 membranes and tAg. After blebbistatin treatment (20 min), we fixed cells and stained for vinculin and F-actin. NM-II inhibition resulted in vinculin ring disorganization and loss of vinculin localization to the synapse (Fig. 6A, 6B). At the time analyzed, F-actin distribution was maintained surrounding the cSMAC, and its levels at the IS contact plane were lost (Fig. 6A, 6B). Before fixation, blebbistatin-treated B cells remained adhered to the membrane (detected by IRM), although the contact area was significantly reduced (Fig. 6C), suggesting internal disorganization of the pSMAC structure. Blebbistatin-treated B cells showed no sign of motility on the membranes, as full NM-II activity is needed for CXCL13-mediated B cell migration.
We used A20 B cells expressing a vinculin-GFP construct and Lifeact-RFP to monitor the effect of NM-II inhibition in time-lapse experiments. A20 B cells formed a mature IS in contact with ICAM-1/CXCL13 membranes and tAg; vinculin and F-actin distributed in a ring surrounding the cSMAC (Fig. 6D). After adding blebbistatin, we tracked the molecular dynamics of vinculin and F-actin at the IS by confocal microcopy. By 20 min posttreatment, the vinculin pattern was completely disorganized and its fluorescent signal declined; F-actin polymerization was reduced, and no longer confined to the vicinity of the cSMAC (Fig. 6D, Supplemental Video 9).
The stability of the vinculin-rich domain at the B cell IS thus depends on appropriate NM-II activity. The data also indicated that loss of vinculin is accompanied by diminished F-actin polymerization and F-actin ring disassembly at the contact plane.
Our study showed that vinculin is recruited to the B cell IS and distributes in the LFA-1–rich pSMAC domain together with F-actin, talin, PIPKIγ, and the lipid PIP2. tAg/BCR-triggered Syk activity is needed for vinculin localization to the B cell–APC contact site; absence of vinculin recruitment allows B cells to continue moving in response to CXCL13 while assembling and carrying the synapse-characteristic Ag cluster at the uropod. Loss of vinculin also reduced F-actin polymerization, but not talin recruitment to the synapse. The motor protein NM-II is implicated in maintaining vinculin at the IS. These data identify vinculin as a key regulatory element of the assembly and stability of the LFA-1–mediated platforms that support B cell dynamics (i.e., synapse and kinapse).
Vinculin is found in the synapse of Jurkat T cells, in a complex containing WAVE-2, Arp2/3 and talin; vinculin is necessary for talin recruitment, but not for F-actin polymerization or integrin accumulation (35). Talin-deficient T cells do not adhere firmly to APC or arrest migration; they assemble some LFA-1 clustering, but did not recruit vinculin or F-actin to the short-lived synapse; the authors highlighted the importance of talin for T cell IS stability (36). In NK cells, LFA-1/ICAM-1 interaction leads to vinculin accumulation at the synapse; talin is needed for vinculin and F-actin localization to the contact site (37). In this study, we report that vinculin is recruited to the B cell synapse, specifically to the LFA-1–rich pSMAC domain, where it colocalizes with talin and F-actin. Talin might be also important for vinculin recruitment to the B cell synapse, a subject for further study. We nonetheless found that, without affecting talin localization, vinculin at the B cell synapse was essential for F-actin accumulation, pSMAC assembly and synapse stability. Although chemical inhibitors might have undesired effects, they were important tools to assess vinculin function in this study, due to the limitations we found to diminish vinculin expression in B cells by shRNA techniques and the nonviability of vinculin-deficient mouse models (22).
PIPKI localization and activity regulates the targeted, limited production of PIP2, which in turn governs the temporal and spatial requirements of adhesion site dynamics. In neutrophils, correct distribution of distinct PIPKI regulates cell polarity and migration (27, 38). PIPKIγ deficiency at FA impairs talin and vinculin recruitment to nascent adhesion sites, which reduces integrin-mediated cell adhesion and force coupling (39). Another report does not implicate PIP2 in vinculin recruitment to adhesion sites, but rather in vinculin release and FA disassembly (40), which is reinforced by the finding that PIPKIγ-deficient T cells show increased integrin-mediated adhesion (41). Our findings coincide with the first model; tAg/BCR stimulation promoted PIPKIγ localization, and thus, local PIP2 production at the nascent synapse that could support vinculin recruitment. They remain detectable at the pSMAC of the mature synapse, possibly assisting the active vinculin conformation. PIPKIγ is also found at the T cell synapse; the spatiotemporal regulation of PIP2 synthesis appears to control T cell rigidity and signaling organization (42).
FA are mechanosensitive structures that transmit cell forces to the extracellular matrix. Cell forces are generated as a consequence of NM-II action on the actin cytoskeleton (43). NM-II–mediated contractility controls the localization to FA of vinculin and other adaptor proteins (34). The ability of vinculin to bear force determines the assembly or disassembly of adhesion sites under tension (16, 17). In the podosome, another type of actomyosin-based integrin-rich platform, NM-II activity is not necessary for adaptor protein composition, which is controlled by the actin network (44). NM-II participates in synapse and migratory junctions in lymphocytes (7, 45–47). Our data show that NM-II activity is important for vinculin localization at the B cell synapse, as described for FA. Lack of vinculin due to NM-II inhibition led to reduced, mislocalized F-actin polymerization and pSMAC disassembly. We propose that vinculin is the mechanical sensor also at the lymphocyte synapse; it regulates assembly and disassembly of the adhesion structure based on cell force input. Two recent reports in T cells also highlight the relevance of NM-II–generated mechanical forces on modulating synapse assembly and cell activation (48, 49); both studies involved CasL, a member of the mechanosensing Cas protein family and predominantly expressed in T cells, in the mechanical sensing.
We found accumulation of active Syk (p-Syk) at the cSMAC of the mature B cell synapse. It has been previously showed some colocalization of p-Syk with the BCR-Ag central cluster at the mature synapse of naive B cells and localization of GFP-Syk at the cSMAC in A20 B cells by TIRFM (50). Although some reports in T cells showed that Ag receptor proximal signaling does not occur in the cSMAC (51, 52), it has been proposed that Ag quality (i.e., Ag affinity and abundance) determines the balance between signaling and receptor degradation at the cSMAC (53). The p-Syk enrichment at the cSMAC of the B cell IS detected in our study might be related to the latest model; more studies combining Ag titration and p-Syk measurements need to be done. In addition, Syk has an important role in regulating trafficking and processing of Ag internalized through the BCR (54). Active Syk might thus be required at the cSMAC to perform B cell–specific functions, not shared with its homolog in T cells, Zap70.
Syk kinase has a role in integrin activation downstream of the BCR and of chemokine receptors (32, 33). Our data indicate that the Syk activity level and localization determine vinculin recruitment at the B cell synapse. Syk promotes Bruton's tyrosine kinase (Btk) recruitment to the plasma membrane (55). Btk associates with and transports PIPKI to the cell membrane to produce PIP2, the substrate of the Btk upstream activator PI3K. In this study, we show that vinculin recruitment parallels the PIP2 wave generated by PIPKIγ early in synapse formation. Btk-mediated PIP2 production might thus support PI3K activity, but also assists vinculin localization to the synapse. In our model, above a certain threshold of tAg/BCR-promoted Syk activity, there is a local increase in Btk and PIPKIγ-dependent PIP2 at the B cell–APC contact site; the PIP2 wave leads to vinculin recruitment and thus to pSMAC assembly and adhesion strength. This model explains our previous observations that below a tAg/BCR signaling threshold, B cells remain motile in response to chemokines and integrate BCR signals through the LFA-1–mediated migratory junction, the kinapse (7).
We thank I. Antón and M. Vicente-Manzanares for critical reading of the manuscript and C. Mark for editorial assistance.
This work was supported by grants from the European Union (Framework Programme 7–integrated project Masterswitch 223404 FP7) and the Spanish Ministry of Economy and Competitiveness (BFU2011-30097 to Y.R.C.). J.S.d.G. is supported by a contract from the Comunidad Autónoma de Madrid. L.B. is supported by a contract associated with Project Grant BFU2008-01194 from the Spanish Ministry of Economy and Competitiveness.
The online version of this article contains supplemental material.
Abbreviations used in this article:
Bruton's tyrosine kinase
central supramolecular activation cluster
differential interference contrast
interference reflection microscopy
non–muscle motor protein myosin-II
type I phosphatidylinositol (4,5)-phosphate kinase
peripheral supramolecular activation cluster
short hairpin RNA
spleen tyrosine kinase
The authors have no financial conflicts of interest.