The activation of the BCR, which initiates B cell activation, is triggered by Ag-induced self-aggregation and clustering of receptors at the cell surface. Although Ag-induced actin reorganization is known to be involved in BCR clustering in response to membrane-associated Ag, the underlying mechanism that links actin reorganization to BCR activation remains unknown. In this study, we show that both the stimulatory Bruton’s tyrosine kinase (Btk) and the inhibitory SHIP-1 are required for efficient BCR self-aggregation. In Btk-deficient B cells, the magnitude of BCR aggregation into clusters and B cell spreading in response to an Ag-tethered lipid bilayer is drastically reduced, compared with BCR aggregation observed in wild-type B cells. In SHIP-1−/− B cells, although surface BCRs aggregate into microclusters, the centripetal movement and growth of BCR clusters are inhibited, and B cell spreading is increased. The persistent BCR microclusters in SHIP-1−/− B cells exhibit higher levels of signaling than merged BCR clusters. In contrast to the inhibition of actin remodeling in Btk-deficient B cells, actin polymerization, F-actin accumulation, and Wiskott–Aldrich symptom protein phosphorylation are enhanced in SHIP-1−/− B cells in a Btk-dependent manner. Thus, a balance between positive and negative signaling regulates the spatiotemporal organization of the BCR at the cell surface by controlling actin remodeling, which potentially regulates the signal transduction of the BCR. This study suggests a novel feedback loop between BCR signaling and the actin cytoskeleton.
The BCR induces signaling cascades and Ag processing and presentation in response to Ag binding. These BCR-induced cellular activities combine with signals from the microenvironment to determine the fate of B cells. Biochemical and genetic studies in the past two decades (1–3) have shown that upon cross-linking by Ag, surface BCRs aggregate and associate with lipid rafts (4), where they are phosphorylated by Src kinases, such as Lyn. The binding of tyrosine kinase Syk to phosphorylated ITAMs in the cytoplasmic tails of the BCR activates Syk, which in turn activates downstream signaling components including phospholipase Cγ2 (PLCγ2), Ras, PI3Ks, and Bruton’s tyrosine kinase (Btk). Ag binding to the BCR also activates negative signaling components, in particular, SHIP-1 (5–7). SHIP-1 converts phosphatidylinositol-3,4,5-triphosphate into phosphatidylinositol-3,4-biphosphate, eliminating the docking sites of PLCγ2, Btk, and Akt at the plasma membrane and turning down BCR signaling (7, 8).
Recent studies using advanced cell imaging technologies have begun to reveal the molecular details of the initiation events in BCR activation (9–11). Ag binding induces conformational changes of the BCR, which potentially expose the Cμ4 domain of membrane IgM for BCR self-aggregation (12) and ITAMs for signaling molecules to bind (13). Self-aggregation reduces the lateral mobility of the BCR and induces the formation of BCR microclusters (12). Newly formed BCR microclusters reside in lipid rafts (14) and recruit signaling molecules, including Lyn, Syk (13), PLCγ2, Vav (15), and the costimulatory receptor CD19 (16). BCR microclusters grow in size by trapping more BCRs and merging into each other. This leads to the formation of a polarized central cluster, similar to the immunological synapse formed between T cells and APCs (17). Therefore, the control of BCR mobility and self-aggregation is essential for signal initiation and transduction.
The surface mobility and aggregation of the BCR has been shown to require Ag-induced actin reorganization. The actin cytoskeleton is known to control cell morphology (18, 19) and lateral diffusion of transmembrane proteins (19). Recent studies have shown that membrane-associated Ags induce B cell spreading, which is followed by cell contraction. These morphological changes of B cells enhance the formation of BCR clusters. Disrupting the actin cytoskeleton inhibits this enhanced BCR cluster formation (20). However, in the absence of Ag, actin disruption increases the lateral diffusion rate of surface BCRs and induces spontaneous signaling in B cells (21). These findings suggest that Ag-induced actin remodeling can regulate BCR self-aggregation by controlling B cell morphology and BCR lateral mobility at the cell surface.
Ag-induced actin reorganization, BCR microcluster formation and B cell spreading all are signaling-dependent processes. Multiple BCR signaling molecules, including CD19, PLCγ2, Vav, and Rac2, promote BCR cluster formation and B cell spreading (15, 16, 22). In contrast, coengagement of the BCR and FcγRIIB, which activates SHIP-1, inhibits the formation of BCR clusters and BCR signaling (23, 24). We have previously shown that Btk can relay BCR signaling to the actin cytoskeleton by activating an actin nucleation promoting factor, Wiskott–Aldrich symptom protein (WASP), via Vav and phosphatidylinositol (25). These findings point to an intimate cooperativity between BCR signaling and the actin cytoskeleton. However, the underlying mechanism for the signaling–actin interplay during BCR activation remains unknown.
In this study, we used live cell imaging, total internal reflection fluorescence microscopy (TIRFm), interference reflection microscopy (IRM), and genetically altered mice to examine the molecular mechanism by which the actin cytoskeleton cooperates with early BCR signaling at the cell surface during BCR activation. Our results show that the positive and negative downstream signaling molecules of the BCR, Btk and SHIP-1, have distinct roles in self-aggregation and cluster formation of surface BCRs, B cell morphology, and actin reorganization. The activation of Btk promotes B cell spreading, BCR microcluster formation, and actin polymerization and accumulation. In contrast, SHIP-1 promotes the merger of BCR microclusters and B cell contraction but inhibits actin polymerization and accumulation. Our results suggest a balance of Btk and SHIP-1 activation controls the nature of actin reorganization, which in turn regulates the spatiotemporal organization of surface BCRs.
Materials and Methods
Mice and cells
Wild-type (wt) (CBA/CaJ), xid (CBA/CaHNBtkxid/J), and CD19Cre/+ (B6.129P2(C)-Cd19tm1(cre)Cgn/J) mice were purchased from The Jackson Laboratory (Bar Harbor, ME). B cell-specific SHIP-1–knockout mice CD19Cre/+SHIP-1Flox/Flox were generated by crossing CD19Cre/+ with SHIP-1Flox/Flox mice (26). WASP−/− mice on a 129 SvEv background were provided by Dr. S. Snapper (Harvard Medical School, Boston, MA) (27), and 129 SvEv wt mice were from The Jackson Laboratory. Splenic B cells were isolated as described previously (25). All animal work was reviewed and proved by the Institutional Animal Care and Usage Committee of University of Maryland.
Preparation of monobiotinylated Fab′ Ab
Monobiotinylated Fab′ fragment of anti-mouse IgM+G Ab (mB-Fab′–anti-Ig) was generated from the F(ab′)2 fragment (Jackson ImmunoResearch Laboratories, West Grove, PA) using a published protocol (28). The disulfide bond that links the two Fab′ was reduced using 20 mM 2-mercaptoethylamine, and the reduced cysteine was biotinylated by maleimide-activated biotin (Thermo Scientific, Odessa, TX). Fab′ was further purified using Amicon Ultracentrifugal filters (Millipore, Temecula, CA). One biotin per Fab′ was confirmed by a biotin quantification kit (Thermo Scientific). Fab′ was labeled with Alexa Fluor (AF) 546 (Invitrogen, Carlsbad, CA).
Preparation of Ag-tethered planar lipid bilayers
The planar lipid bilayer was prepared as described previously (14, 29). Liposomes were made by sonicating 1,2-dioleoyl-sn-glycero-3-phosphocholine and 1,2-dioleoyl-sn-glycero-3-phosphoethanolamine-cap-biotin (Avanti Polar Lipids, Alabaster, AL) in a 100:1 molar ratio in PBS at a lipid concentration of 5 mM. Aggregated liposomes were removed by ultracentrifugation and filtration. Coverslip chambers (Nalge Nunc International, Rochester, NY) were coated with the planar lipid bilayer by incubating with the liposomes (0.05 mM) for 10 min. After extensive washes, the coated coverslip chamber was incubated with 1 μg/ml streptavidin (Jackson ImmunoResearch Laboratories), followed by 2 μg/ml AF546-mB-Fab′–anti-Ig mixed with 8 μg/ml mB-Fab′–anti-Ig Ab. As a nonspecific Ag (NS-Ag), the coated coverslip was incubated with 1 μg/ml streptavidin (Invitrogen), followed by 10 μg/ml AF546-labeled biotinylated Fab-anti–rabbit IgG Ab. For a nonantigenic control, surface BCRs were labeled by incubating with AF546-Fab–anti-Ig (2 μg/ml) on ice for 30 min. The labeled B cells were then incubated with biotinylated holo-transferrin (Tf; 16 μg/ml, which gave an equal molar concentration of 10 μg/ml mB-Fab′–anti-Ig; Sigma-Aldrich, St. Louis, MO) tethered to lipid bilayers by streptavidin.
Total internal reflection fluorescence microscopy
Images were acquired using a Nikon laser TIRF system on an inverted microscope (Nikon TE2000-PFS), equipped with a 60×, NA 1.49 Apochromat TIRF objective (Nikon Instruments, Melville, NY), a Coolsnap HQ2 charge-coupled device camera (Roper Scientific, Sarasota, FL), and two solid-state lasers of wavelengths 491 and 561 nm. For live cell imaging, time lapse images were acquired at the rate of one frame every 3 s. Image acquisition started upon the addition of B cells onto an Ag-tethered lipid bilayer and continued for 5–10 min at 37°C. Interference refection images (IRM) and AF488 and AF546 images were acquired sequentially.
To image intracellular molecules, B cells were incubated with an Ag-tethered lipid bilayer at 37°C for varying lengths of time. Cells were then fixed with 4% paraformaldehyde, permeabilized with 0.05% saponin, and stained for phosphotyrosine (pY) (Millipore), phosphorylated Btk (pBtk, Y551; BD Biosciences, San Jose, CA), Akt (S473; Cell Signaling Technology, Danvers, MA), and WASP (S483/S484; Bethyl Laboratory, Montgomery, TX). F-actin was stained using AF488-phalloidin. In the case of Btk inhibition, splenic B cells were pretreated with LFM A-13 [2-cyano-N-(2,5-dibromophenyl)-3-hydroxy-2-butenamide, 50 or 400 μg/ml; EMD Bioscience, Gibbstown, NJ] for 1 h at 37°C before incubation with lipid bilayer-tethered Ag. The inhibitor was also included in the incubation media. The B cell contact area was determined using IRM images and MATLAB software (MathWorks, Natick, MA). The total fluorescence intensity (TFI) and mean fluorescence intensity (MFI) of each staining in the B cell contact zone and relative fluorescence and IRM intensity along a line across cells were determined using Andor iQ software (Andor Technology, Belfast, U.K.). Background fluorescence generated by Ag tethered to lipid bilayers in the absence of B cells or secondary Ab controls was subtracted. For each set of data, >20 individual cells from two or three independent experiments were analyzed.
Analysis of the mobility and fluorescence intensity of BCR clusters
To analyze the mobility of BCR microclusters, kymographs of time lapse images by TIRFm were generated using Andor iQ software (Andor Technology). The moving velocity of BCR microclusters were calculated using the slope of moving streaks of individual clusters in kymographs. The length of time that each emerging microcluster required to merge with the central cluster was calculated as the life span.
The TFI of BCR and pY staining in individual BCR clusters was determined using TIRFm images of B cells that were incubated with an Ag-tethered lipid bilayer for 3 and 7 min, using Andor iQ software. The data were plotted as the TFI ratio of pY to the BCR in individual BCR clusters versus the TFI of the BCR in individual clusters. Smooth curves of the plot were generated using a nonparametric regression method, locally weighted scatterplot smoothing (LOWESS) (30, 31), by Stata software (StataCorp, College Station, TX). The bwidth or the smoothing factor used for the LOWESS analysis is 0.8.
Analysis of actin nucleation sites
Actin nucleation sites were detected as described previously (32). B cells were incubated with AF546-mB-Fab′–anti-Ig tethered to lipid bilayers in the presence of AF488-G-actin (Invitrogen) and 0.025% saponin at 37°C. Time lapse images were acquired for 5 min using TIRFm. The TFI and MFI of incorporated AF488-G-actin in the B cell contact zone and relative fluorescence and IRM intensity along a line across cells were determined using Andor iQ software (Andor Technology).
Statistical significance was assessed using the Mann–Whitney U test by Prism software (GraphPad Software, San Diego, CA). The p values were determined in comparison with wt or control B cells.
Btk and SHIP-1 regulate B cell spreading and BCR aggregation
Ag engagement of the BCR induces the activation of both Btk (2, 33) and SHIP-1 (5–7), the key positive and negative signaling molecules downstream of the BCR. To determine how BCR signaling regulates the organization of surface BCRs, we used xid mice, which express inactive Btk that has a point mutation in its pleckstrin homology domain (34), and B cell-specific SHIP-1–knockout mice (26). We examined the effects of Btk or SHIP-1 deficiency on BCR aggregation at the cell surface and B cell spreading in response to membrane-associated Ag in the absence of adhesion molecules. The model Ag consisted of a fluorescently labeled, monobiotinylated Fab′ fragment of anti-mouse IgG+M Ab (mB-Fab′–anti-Ig) tethered to planar lipid bilayers by streptavidin. TIRFm was used to evaluate the aggregation of surface BCRs and IRM to identify and determine the areas of B cells contacting an Ag-tethered lipid bilayer (B cell contact zone).
Upon incubation with an Ag-tethered lipid bilayer, the contact zone of wt and CD19Cre/+SHIP-1+/+ control B cells rapidly expanded, peaked at ∼6 and ∼3 min, respectively, and then slightly decreased (Fig. 1A–D, Supplemental Videos 1, 2). Concurrently, surface BCRs formed microclusters, appearing as puncta, in the first few minutes, which then merged into each other, forming a central cluster in the B cell contact zone (Fig. 1A, 1B, Supplemental Videos 1, 2). The TFI of Ag in the contact zone of wt and control B cells increased over time and reached a plateau at ∼7 or 5 min, respectively (Fig. 1A, 1B, 1E, 1F, Supplemental Videos 1, 2). Fab-anti–rabbit Ab tethered to lipid bilayers was used as an NS-Ag control, and Tf tethered to lipid bilayers, which binds to Tf receptor on the B cell surface, was used as a nonantigenic control. For the NS-Ag control, wt B cells established limited contact with lipid bilayers, but neither spread further nor recruited the Fab to the B cell contact zone (Fig. 1A, 1C, 1E). For the nonantigenic control, the B cell contact area was larger than that observed with nonspecific Fab, but smaller than that observed with specific Ag (Fig. 1A–D). Similarly, BCR staining in the contact zone on a Tf-tethered lipid bilayer was much lower than that on an Ag-tethered lipid bilayer (Fig. 1A, 1B, 1E, 1F). Because only BCR-specific Ag induces cluster formation, the Ag accumulation shown in this study reflects the aggregation of surface BCRs. Therefore, BCR cluster formation and B cell spreading are events induced by specific Ag.
When incubated with an Ag-tethered lipid bilayer, the xid splenic B cells only established loose attachment to the lipid bilayer and failed to spread (Fig. 1A, 1C, Supplemental Video 1). Compared with that of wt B cells, the amount of the BCR accumulated in the contact zone of xid B cells was reduced drastically (Fig. 1A, 1E). The behavior of xid B cells was similar to that of wt when interacting with a Tf-tethered lipid bilayer (Fig. 1A, 1C, 1E). This indicates that Btk deficiency inhibits Ag-induced BCR cluster formation and B cell spreading but not Tf-induced B cell spreading. In contrast to xid B cells, SHIP-1−/− B cells spread more extensively than CD19Cre/+SHIP-1+/+ control B cells (Fig. 1B, 1D, Supplemental Video 2). Despite the increase in the contact area, the amount of the BCR in the contact zone of SHIP-1−/− B cells was substantially reduced (Fig. 1B, 1F, Supplemental Video 2), showing that BCR aggregation efficiency is disassociated from the B cell spreading process. Moreover, the BCR at the contact zone of SHIP-1−/− B cells appeared as punctate clusters and failed to form a central cluster as in control B cells (Fig. 1B, Supplemental Video 2).
To determine whether B cell developmental defects caused by Btk mutation affect BCR microcluster formation and B cell spreading, we sorted mature and immature B cells from the spleens of wt and xid mice based on the CD93 expression level. When incubated with an Ag-tethered lipid bilayer, mature (CD93−) and immature (CD93+) B cells spread and accumulated BCRs in the contact zone in a similar kinetics and to a similar magnitude. Furthermore, Btk deficiency inhibited BCR microcluster formation and B cell spreading in both mature and immature B cells (Supplemental Fig. 1). B cell-specific SHIP-1 knockout did not cause any major alterations in immature transitional and mature follicular B cell subsets in the spleen (W.-H. Leung, T. Tarasenko, Z. Biesova, H. Kole, and S. Bolland, manuscript in preparation). This indicates that the effect of Btk and SHIP-1 deficiency on BCR cluster formation and B cell spreading are not due to alterations of B cell subsets in the spleens of xid and SHIP-1−/− mice.
Taken together, these results indicate that both positive and negative signaling mediated by Btk and SHIP-1 are involved in regulating Ag-induced B cell spreading and BCR aggregation and accumulation in the contact zone. Btk induces B cell spreading and BCR microcluster formation, and SHIP-1 is involved in inhibiting B cell spreading and promoting the formation of the central cluster.
The centripetal movement of BCR microclusters is inhibited in SHIP-1−/− B cells
To understand why SHIP-1−/− B cells fail to form BCR central clusters, we analyzed the lateral movement of BCR microclusters using kymographs. Kymographs generated from time lapse images by TIRFm (Supplemental Videos 1, 2) provide graphical representations of spatial positions of BCR microclusters in the B cell contact zone. Using kymographs, we determined the mobility of individual BCR microclusters and the time span that an emerging BCR microcluster takes to merge with the central cluster (life span). The results show that in CD19Cre/+SHIP-1+/+ control B cells, BCR microclusters moved at an average mobility of ∼10 nm/s or ∼0.6 μm/min toward the center of the B cell contact zone (Fig. 2A, 2B). It took ∼1 min for emerging BCR microclusters to merge with central clusters (Fig. 2C). In SHIP-1−/− B cells, the mobility of BCR microclusters was reduced to 1.2 nm/s or 0.072 μm/min, and therefore, they took six times longer to merge into the central cluster (Fig. 2). These data demonstrate a critical role for SHIP-1 in the centripetal movement of BCR microclusters and the merger of BCR microclusters into the central cluster.
The growth of BCR microclusters is inhibited, but the signaling capability of BCR microclusters is enhanced in SHIP-1−/− B cells
BCR self-aggregation is essential for receptor activation. Although the contribution of SHIP-1 to BCR signaling via its phosphatase activity is well-known, the effects of SHIP-1 deficiency on surface BCR aggregation suggest a role for SHIP-1 beyond its enzymatic activity. To investigate this hypothesis, we examined the levels and distribution of pY, pBtk, and pAkt in relation to BCR microclusters in the B cell contact zone. In CD19+/+SHIP-1Flox/Flox control B cells, the level of pY staining in the contact zone increased over time and peaked at ∼3 min (Fig. 3A). As the pY level rose, the pY staining largely colocalized with BCR microclusters (Fig. 3D, 3F). After 3 min, the pY level in the B cell contact zone decreased (Fig. 3A). Concurrent with this decrease, BCR microclusters merged into a central cluster, and the pY staining redistributed away from BCR clusters to the outer edge of the B cell contact zone (Fig. 3D, 3F). Similarly, the levels of pBtk and pAkt in the contact zone of control B cells peaked ∼3 min and decreased afterward (Fig. 3B, 3C). This suggests a negative correlation between signaling activity and the merger of BCR microclusters at the cell surface. In SHIP-1−/− B cells, the pY level in the contact zone increased at a rate similar to that of control B cells, but the peak level was sustained ∼2 min longer than that in the control B cells (Fig. 3A). The higher level of pY in the contact zone was not simply due to an increase in cell spreading, because the MFI of pY staining in the contact zone of SHIP-1−/− B cells was also higher than that in control B cells (data not shown). In addition, the levels of pBtk and pAkt in the contact zone of SHIP-1−/− B cells were markedly higher than those in control B cells (Fig. 3B, 3C). Moreover, the pBtk level in the contact zone of SHIP-1−/− B cells continuously increased until ∼5 min (Fig. 3B), and the high level of pAkt persisted rather than returning to the basal level like in the control B cells (Fig. 3C). The surface distribution of pY was also altered in SHIP-1−/− B cells, appearing as puncta and colocalizing with BCR clusters at all the time points tested (Fig. 3E, 3G).
To further investigate the relationship between the growth of BCR clusters and their signaling activities, we determined the relative size of BCR clusters based on the TFI of BCR labeling in individual clusters and the relative signaling levels of BCR clusters based on the fluorescence intensity ratio of the pY to the BCR in individual clusters. We used a nonparametric regression method, LOWESS, to analyze the trend of the data. In CD19+/+SHIP-1Flox/Flox control B cells, we found a two-phase correlation between the sizes of BCR clusters and their tyrosine phosphorylation activity. First, when the sizes of BCR clusters were relatively small, the pY-to-BCR ratio increased as emerging microclusters grew (Fig. 3H). However, after the sizes of the BCR clusters reached a certain level, the pY-to-BCR ratio decreased as BCR clusters further expanded, likely via the merger of BCR microclusters (Fig. 3H). In contrast, BCR clusters formed in the contact zone of SHIP-1−/− B cells were limited to smaller sizes in comparison with those in control B cells (Fig. 3I). The fluorescence intensity ratio of pY to BCR in individual microclusters of SHIP-1−/− B cells was much higher than that of control B cells. These results suggest that SHIP-1 promotes the growth of BCR clusters, which contributes to signaling downregulation.
Btk and SHIP-1 play opposing roles in actin reorganization
Both BCR aggregation and B cell spreading are shown to depend on the actin cytoskeleton. The effects of Btk and SHIP-1 deficiency on BCR cluster formation and B cell spreading imply a role for these signaling molecules in regulating actin dynamics. We have previously shown that Btk deficiency inhibits actin polymerization and WASP activation in response to soluble Ag (25). In this study, we examined the effect of SHIP-1 gene knockout on the levels and distribution of F-actin and actin polymerization in the contact zone. Although the level of F-actin in the contact zone of both CD19+/+SHIP-1Flox/Flox control and SHIP-1−/− B cells increased in response to an Ag-tethered lipid bilayer, the increase was significantly greater in SHIP-1−/− B cells (Fig. 4A–C). In the contact zone of control B cells, F-actin was largely colocalized with BCRs upon B cell interaction with an Ag-tethered lipid bilayer (Fig. 4A). When BCR microclusters merged into each other, F-actin moved away from BCR clusters to BCR poor regions and the outer edge of the contact zone (Fig. 4A, 4D). In SHIP-1−/− B cells, F-actin did redistribute away from BCR clusters, but it formed a wide ring in the middle of the contact zone and did not accumulate at the outer edge of the contact zone (Fig. 4B, 4E).
F-actin accumulation suggests an increase in actin polymerization. To compare actin polymerization activity of Btk and SHIP-1–deficient B cells with wt and CD19+/+SHIP1Flox/Flox control B cells, we used G-actin incorporation assay. In this assay, the incorporation of fluorescently labeled G-actin to the polymerizing end of F-actin indicates the location and level of actin polymerization (32). We found that in xid B cells, the MFI of incorporated AF488-G-actin in the contact zone was significantly decreased, compared with that in wt B cells. In contrast, G-actin incorporation was significantly increased in the contact zone of SHIP-1−/− B cells, compared with that in control B cells (Fig. 4F). Despite the enhanced actin polymerization in SHIP-1−/− B cells, actin polymerization sites were distributed in a pattern similar to what was seen in control B cells (Fig. 4G, 4H).
Actin polymerization is activated by actin nucleation promoting factors. Our previous study shows that Btk promotes actin polymerization by activating a hematopoietic-specific actin nucleation promoting factor, WASP (25). In this study, we determined the effect of SHIP-1 deficiency on WASP activation using an Ab specific for the phosphorylated WASP (pWASP, S483/S484). Similar to F-actin, the level of pWASP in the contact zone of CD19+/+SHIP-1Flox/Flox control B cells increased over time in response to an Ag-tethered lipid bilayer and reached a peak at ∼5 min (Fig. 5A, 5F). In SHIP-1−/− B cells, the pWASP level in the contact zone was significantly increased, especially at 1–3 min (Fig. 5B, top panels, and 5F). The pWASP staining in the contact zone of both control and SHIP-1−/− B cells appeared as puncta and colocalized with BCR microclusters early during the stimulation (Fig. 5A, 5C; data not shown). As BCR microclusters merged into each other, pWASP in control B cells redistributed away from BCR clusters to the outer edge of the contact zone and to BCR poor regions (Fig. 5A, 5D), but pWASP in SHIP-1−/− B cells failed to do so (Fig. 5B, top panels, 5E).
These results collectively indicate that Btk and SHIP-1 positively and negatively regulate actin reorganization, respectively. Btk induces whereas SHIP-1 inhibits WASP activation, actin polymerization, and F-actin accumulation in the B cell contact zone.
SHIP-1 regulates WASP activation, B cell spreading, and BCR cluster formation in a Btk-dependent manner
Btk is one of the downstream targets of SHIP-1. SHIP-1 dephosphorylates phosphatidylinositol-3,4,5-triphosphate at its five position, removing the docking site of Btk at the plasma membrane. Therefore, SHIP-1 deficiency increases Btk activation (Fig. 3B) (8). To examine the functional interrelationship between SHIP-1 and Btk in initiation steps of BCR activation, we reduced Btk activity in SHIP-1−/− B cells using a Btk-specific inhibitor, LFM A-13. Our previous study shows that this inhibitor causes reductions in BCR and WASP activation similar to Btk deficiency (25). An intermediate and a high concentration of LFM A-13 were used to manipulate the level of the kinase activity of Btk. At an intermediate concentration, the Btk inhibitor restored the magnitudes of B cell spreading (Fig. 5G), BCR accumulation (Fig. 5H), and WASP phosphorylation (Fig. 5B, 5F) in the contact zone of SHIP-1−/− B cells to levels similar to those in CD19+/+SHIP-1Flox/Flox control B cells. At a high concentration, the Btk inhibitor reduced B cell spreading (Fig. 5G), BCR accumulation (Fig. 5H), and WASP phosphorylation in the contact zone (Fig. 5B, 5F) to levels close to those in Btk-deficient B cells. These data indicate that SHIP-1 can regulate actin reorganization, B cell spreading, and BCR aggregation and accumulation via inhibiting Btk.
B cell spreading, BCR cluster formation, and tyrosine phosphorylation are reduced in WASP−/− B cells
Our finding that WASP activation is regulated by both Btk and SHIP-1 suggests that WASP contributes to BCR activation. To determine the role of WASP in the early event of BCR activation, we compared B cell spreading, BCR cluster formation, and tyrosine phosphorylation in WASP−/− B cells with those in wt B cells. In WASP−/− B cells, the magnitude of B cell spreading, the extent of BCR clustering, and the level of tyrosine phosphorylation in the B cell contact zone were significantly reduced, compared with wt B cells (Fig. 6). However, the inhibitory effects of WASP deficiency were less remarkable than those of Btk deficiency and Btk inhibition (Figs. 1, 5). These results suggest that WASP is important, but not essential, for B cell spreading and BCR aggregation in response to membrane-associated Ag.
In this study, we have examined the underlying mechanism for the functional interplay between signaling and the actin cytoskeleton during BCR activation. Our results show that the stimulatory kinase Btk promotes B cell spreading, BCR microcluster formation, and actin polymerization and recruitment in response to membrane-associated Ag. In contrast, the inhibitory phosphatase SHIP-1 inhibits B cell spreading and actin polymerization and recruitment but promotes the centripetal movement and merger of BCR microclusters. These results suggest that a balance between positive and negative signaling regulates the dynamics and magnitude of BCR self-aggregation at the B cell surface via controlling actin reorganization.
The unique finding of this study is that positive and negative signals have distinct roles in regulating BCR aggregation. Previous studies show that the activation of proximal signaling molecules, including PLCγ2, Vav, Rac2, and costimulatory molecule CD19 (15, 16, 22), can have significant impact on BCR aggregation on the cell surface. Consistent with these findings, we show in this study that Btk, a downstream molecule of CD19 and an upstream molecule of Vav and Rac, is required to induce the formation of BCR microclusters in response to membrane-associated Ag. However, the inhibitory phosphatase SHIP-1 is not essential for the formation of BCR microclusters, but it is involved in the centripetal movement and merger of BCR microclusters after their formation. Our results suggest that the relative activation levels and timing of stimulatory kinases, such as Btk, and inhibitory phosphatases, such as SHIP-1, can control the formation and growth of BCR microclusters. Therefore, signaling triggered by Ag-induced aggregation of BCRs in turn regulates the kinetics and magnitude of BCR self-aggregation, providing feedbacks to signal initiation.
BCR activation is triggered by self-aggregation of surface receptors into microclusters. The formation and growth of BCR microclusters has been positively linked to the signaling capability of the receptor (35). Contrary to this positive link, proximal signaling molecules have been shown to be primarily recruited to TCR microclusters in the periphery but not the center of the immunological synapse (36). In line with the previous study, we detected tyrosine phosphorylation primarily at BCR microclusters and the outer edge of the BCR central cluster. This supports the notion that signal initiation occurs at receptor microclusters rather than the central cluster. By analyzing the ratio between the amount of BCRs and the level of tyrosine phosphorylation in individual clusters, our results suggest a two-phase relationship between the size and signaling capability of BCR clusters. In the first phase, the tyrosine phosphorylation level increases as BCR microclusters emerge and undergo initial growth, which can be considered as the signal activation phase. In the second phase, the tyrosine phosphorylation level decreases as the size of the BCR clusters further expands, which can be considered as the signal transitional or downregulation phase. We found that SHIP-1 deficiency inhibited the growth but not the formation of BCR microclusters, which seems to limit the size of BCR microclusters to the signal activation phase and to prevent BCR microclusters from moving to the signal downregulation phase. Therefore, Btk plays a dominant role in the signal activation phase and SHIP-1 in the signal downregulation phase of BCR clusters. This suggests that in addition to their enzymatic activity, Btk and SHIP-1 could upregulate and downregulate BCR activation by promoting the formation of BCR microclusters or the growth and merger of BCR microclusters into the central cluster. How the merger of BCR microclusters contributes to signaling downregulation remains to be elucidated.
The effects of Btk and SHIP-1 deficiency on the organization of the actin cytoskeleton suggest that these signaling molecules can regulate BCR aggregation by controlling actin dynamics. In addition to determine cell morphology, the actin cytoskeleton also creates a barrier for the lateral movement of transmembrane proteins whose cytoplasmic tails are extended into the cortical actin network (19). Perturbing this barrier has been shown to increase the lateral mobility of surface BCRs and to induce signaling without Ag (21). We found in this study that Btk and SHIP-1 had opposite roles in actin reorganization. Btk and SHIP-1 deficiency decrease and increase actin polymerization activity and F-actin accumulation in the B cell contact zone, respectively. The decreased actin polymerization in Btk-deficient B cells is associated with a marked reduction in B cell spreading and surface BCR cluster formation. This suggests that Btk-induced actin polymerization is required for BCR aggregation and B cell spreading. The inhibition of BCR aggregation in Btk-deficient B cells is likely caused by the reduction in B cell spreading, because B cell spreading has been shown to enhance the formation of BCR microclusters by increasing the number of surface BCRs engaging Ag (20). In SHIP-1−/− B cells, enhanced actin polymerization and F-actin accumulation are concurrent with increased B cell spreading and reduced centripetal movement and merger of BCR microclusters. This suggests that in response to antigenic stimulation, SHIP-1–suppressed actin polymerization is important for preventing B cells from further spreading and for driving the centripetal movement of BCR microclusters. The timing and level of SHIP-1 activation may negatively control the magnitude of B cell spreading and the mobility of BCR aggregates by altering the dynamics and organization of the actin cytoskeleton. Therefore, a balance of positive and negative signals can regulate the nature of actin reorganization, which in turn controls the morphology of B cells and the mobility and growth of BCR aggregates.
How the proximal signaling molecules regulate actin dynamics is not fully understood. In this study, we show that WASP, an actin nucleation promoting factor, is a common downstream target of Btk and SHIP-1 during antigenic activation of the BCR. WASP has been shown to be involved in the formation of immunological synapses in both B and T cells (37–39). We have previously shown that Btk can induce the activation of WASP via increasing the phosphorylation of Vav and the level of phosphatidylinositol-4,5-biphosphate (25). This study shows that SHIP-1 inhibits WASP phosphorylation, counteracting against Btk. The opposing effect of Btk and SHIP-1 on WASP activation is consistent with their opposite impact on actin polymerization. Moreover, phosphorylated WASP and F-actin were found to exhibit similar distribution patterns in the B cell contact zone. These results suggest that WASP controls de novo actin polymerization induced by BCR activation. Our finding that the SHIP-1–mediated inhibition of WASP depends on its ability to inhibit Btk indicates that SHIP-1 prevents WASP from activating by turning off the kinase that induces WASP activation. Therefore, BCR signaling can regulate actin reorganization by controlling the activity of actin regulators like WASP.
In support of a role for WASP in BCR activation, this study found that B cell spreading, BCR cluster formation, and tyrosine phosphorylation in the contact zone induced by membrane-associated Ag were inhibited by WASP deficiency. However, the inhibition was only partial, suggesting that there are additional actin regulators that have overlapping functions with WASP in the process of signal initiation. WASP belongs to a family of proteins (40). The other members of WASP family proteins, such as N-WASP and WAVE, may compensate for the function of WASP in WASP−/− B cells. Such compensational roles could explain the partial inhibition observed in this study as well as the mild defects in B cell activation and B cell responses in WASP−/− mice reported previously (27, 37). It is also noted that the inhibitory effects of WASP deficiency on B cell spreading and BCR clustering are less dramatic than those of Btk deficiency. This suggests that Btk can regulate these processes through multiple downstream molecules in addition to WASP. Identification of additional actin regulators downstream of Btk and SHIP-1 that are involved in actin remodeling during BCR activation is a subject of our future investigation.
In summary, this study demonstrates a close cooperation between signaling and the actin cytoskeleton during BCR activation. Positive and negative signals mediated through Btk and SHIP-1 regulate B cell membrane dynamics and spatiotemporal organization of surface BCRs via controlling actin reorganization. The magnitude of BCR aggregation and the mobility of BCR aggregates regulate the signaling capability of the receptor. This interplay between actin reorganization with signaling forms a mechanistic basis for feedback regulation of BCR signaling. Such a feedback is potentially a general regulatory mechanism of receptor signaling. Further studies are required to identify additional molecular linkers and to define the molecular nature of the interaction between the actin cytoskeleton and BCR signaling pathway.
We thank Drs. Hae Won Sohn and Wanli Liu at the National Institute of Allergy and Infectious Diseases, National Institutes of Health, for technical assistance on generating lipid bilayers. We also thank Kenneth Class and the Maryland Pathogen Research Institute Flow Cytometry Facility for technical assistance on cell sorting.
This work was supported by Grant AI059617 (to W.S.) and Intramural Research Program (to S.B.) of the National Institute of Allergy and Infectious Diseases, National Institutes of Health.
The online version of this article contains supplemental material.
Abbreviations used in this article:
Bruton’s tyrosine kinase
interference reflection microscopy
locally weighted scatterplot smoothing
mean fluorescence intensity
phosphorylated Wiskott–Aldrich symptom protein
total fluorescence intensity
total internal reflection microscopy
The authors have no financial conflicts of interest.