Phosphorylation of WASp is a key regulator of activity and stability in vivo

Abstract

The Wiskott-Aldrich syndrome protein (WASp) is a key cytoskeletal regulator in hematopoietic cells. Covalent modification of a conserved tyrosine by phosphorylation has emerged as an important potential determinant of activity, although the physiological significance remains uncertain. In a murine knockin model, mutation resulting in inability to phosphorylate Y293 (Y293F) mimicked many features of complete WASp-deficiency. Although a phosphomimicking mutant Y293E conferred enhanced actin-polymerization, the cellular phenotype was similar due to functional dysregulation. Furthermore, steady-state levels of Y293E-WASp were markedly reduced compared to wild-type WASp and Y293F-WASp, although partially recoverable by treatment of cells with proteasome inhibitors. Consequently, tyrosine phosphorylation plays a critical role in normal activation of WASp in vivo, and is indispensible for multiple tasks including proliferation, phagocytosis, chemotaxis, and assembly of adhesion structures. Furthermore, it may target WASp for proteasome-mediated degradation, thereby providing a default mechanism for self-limiting stimulation of the Arp2/3 complex.

Wiskott-Aldrich syndrome (WAS) is an X-linked recessive primary immunodeficiency characterized by immune dysregulation, microthrombocytopenia, and eczema (1). Mutations in the WAS gene usually lead to an absence or reduction in levels of the WAS protein (WASp) (1). A lack of WASp results in cytoskeletal defects that compromise multiple aspects of normal cellular activity including proliferation, phagocytosis, immune synapse formation, adhesion, and directed migration (2–6). Biochemical studies indicate that non-activated WASp assumes an autoinhibitory conformation in which the VCA domain interacts with the hydrophobic core of the GTPase binding domain (GBD) (7). The N-terminal WH1 domain also binds the chaperone, WASp-interacting protein (WIP) (8, 9), in the absence of which WASp is unstable (10). Activation of WASp by cooperative binding of PIP₂ and GTP-bound Cdc42 is thought to structurally disrupt the autoinhibitory confirmation allowing binding of the Arp2/3 complex to the C terminus (11–13).

Phosphorylation has also been proposed to be an important regulator of WASp activity although the physiological significance remains uncertain. A single WASp tyrosine phosphorylation site Y291 (Y293 in murine WASp) is a target for a number of non-receptor kinases such as Btk, Fyn, and Hck, and the phosphatase PTP-PEST through binding of the adapter PSTPIP1. Phosphorylation of this residue activates WASp in vitro and in vivo, possibly through disruption or destabilization of the autoinhibited conformation (14, 15). Tyrosine phosphorylation may facilitate activation by SH2 domain-binding and allow sustained activation following an initial trigger for example by GTP-bound Cdc42 (15). In some circumstances tyrosine phosphorylation may occur independently of binding to GTP-bound Cdc42 (14, 16). In model systems, phosphorylation at Y291 has been reported to be necessary for WASp effector functions downstream of the T cell receptor including efficient actin polymerization, immunological synapse (IS) formation, and T cell activation, as well as for phagocytic cup formation and generation of osteoclast sealing zones (16–18).

To explore the physiological relevance of tyrosine phosphorylation, we have generated murine knockin models that either abrogate or simulate phosphorylation at Y293, the murine equivalent. We demonstrate that phosphorylation is indispensible for multiple normal cellular responses, and that activation is associated with enhanced degradation of WASp.

Results

Generation of Knockin Mice Incorporating Y293F and Y293E.

To explore the significance in vivo of Y291 we generated murine knockin models incorporating a phospho-dead mutation (Y293F-mWASp) and a phosphomimicking mutation (Y293E-mWASp) at the equivalent conserved residue (Fig. S1). Homozygous knockin mice were viable and fertile and breeding colonies of both mutant strains were established.

Abnormalities of Cell Numbers and Proliferative Responses.

The size and cellularity of bone marrow, thymus and lymph nodes was equivalent between all strains (Table S1). As expected, platelet numbers were reduced in the peripheral blood of WASp-null mice compared to wild-type (wt). This reduction was also observed in Y293E and Y293F mice (Fig. S2A). Interestingly, monocyte and neutrophil cell numbers were normal for all mutant strains. This was somewhat unexpected as the Y293E mutation was predicted to produce a similar phenotype to that previously described for activating WASp mutations in human patients in whom neutropenia and monocytopenia is characteristic (Fig. S2B). There was however a reduction in the proportion of CD3 T cells in all mutants which included both CD4 and CD8 cells (Fig. S2 C and D). Total splenocyte numbers were increased in all mutant strains compared to wt (Table S1). However, there was a significant reduction in the proportion of CD3 positive cells in spleen, lymph node and thymus (Fig. 1A–F) for both WASp-null and knockin mutants. This was most pronounced for CD8-positive cells and double-positive CD4 and eight thymocytes in older animals (Fig. 1 A–F). NK cells were also present in reduced numbers in the spleen of WASp-null and both knockin mutants compared to C57BL/6 mice (Fig. 1 A and B). NKT cells in the thymus were reduced as reported recently in WASp-null animals (19), and in both mutant strains (Fig. 1 G and H). Both Y293E and Y293F mutant strains also exhibited a severe impairment of proliferation equivalent to that of WASp-null animals (Fig. 1I). Previous studies have defined defects of peripheral B cell homeostasis in WASp-null mice, with marked deficiency of marginal zone (MZ) cells in particular (6, 20, 21). No abnormalities of B cell development were detected in bone marrow of all strains. Total B220-positive cell numbers and the proportion in the spleen were significantly increased, predominantly due to increased follicular mature populations (Fig. 2A and C). However, the proportion of mature MZ/MZ precursor (MZP) and T2 cells (Fig. S3 A and B for flow cytometry gating) were decreased to similar levels observed in WASp-null animals (Fig. 2 B and D). The reduction in MZ cells was supported by quantitative analysis of cryosections stained for B220 and MOMA-1 to delineate the marginal zone (Fig. 2 E and F). In addition both Y293E and Y293F mice developed significant levels of anti dsDNA auto antibodies [previously reported for WASp-null mice (22)], albeit at a lower titer (Fig. S3C). Overall both Y293F and Y293E mutant strains share similar abnormalities to those observed in WASp-null animals indicating that phosphorylation of this amino acid is capable of producing significant functional modifications in multiple hematopoietic lineages in vivo.

Fig. 1.

Reduction in T cell numbers and proliferation in lymphoid tissue. T cell subsets in (A and B) spleen, (C and D) lymph nodes, and (E and F) thymus. NK cells in (A) spleen and (G and H) NKT cells in thymus. (I) T cell proliferation. (A–F) average ± SEM of eight mice per group. (G and H) Average ± SEM of 3–6 mice per group. (I) average ± SEM of three mice and is representative of four independent experiments. (filled circles) C57BL/6, (open circles) WAS KO, (filled triangles) Y293E, and (inverted filled triangles) Y293F. P values shown are from students t test.

Fig. 2.

B-cell subset quantification. (A–D) Total B cell and B cell subsets in spleen analyzed by flow cytometry. (E and F) MZB were visualized by confocal microscopy and the MZ quantified. Average ± SEM of eight mice per group, (filled circles) C57BL/6, (open circles) WAS KO, (filled triangles) Y293E, and (inverted filled triangles) Y293F. Average ± SEM of four mice per group. P values shown are from students t test.

Reduction in Podosome Number and Turnover.

Bone marrow dendritic cells (BMDC) were analyzed for their capacity to form podosomes. Wt BMDC (Fig. 3A) but not WASp-null BMDC made well-defined podosomes as determined by immunofluorescent staining (Fig. 3 A and B). Both Y293E and Y293F BMDC assembled podosomes (Fig. 3A) but the numbers of cells with podosomes (15% and 16%, respectively) was significantly reduced (Fig. 3B). Furthermore, in both Y293E and Y293F BMDC, there were fewer podosomes per cell (Fig. 3C). Qualitatively, podosomes formed in Y293E DC and Y293F DC were generally smaller and defective in recruitment of vinculin (Fig. 3 A, D, and E) although Y293F DC podosomes showed some variation with some cells exhibiting larger podosomes. Podosomes are dynamic adhesion structures with a short half life of about 2–12 min (3). Using time-lapse interference reflection microscopy we imaged and measured the formation and disassembly of podosomes in splenic DC (SPDC) (10, 23). Wt SPDC exhibited an increased turnover of adhesions compared to WASp-null SPDC (Fig. 3F). When Y293E and Y293F SPDC were analyzed the turnover of podosomes was reduced compared to wt, but was increased over WASp-null SPDC (turnover index normal cells = 71.50; Y293E = 37.55; Y293F = 17.94, WASP-null = 8.86) (Fig. 3F). These findings indicate that regulated phosphorylation of WASp is essential for normal podosome assembly and dynamic activity in DC.

Fig. 3.

Decreased podosome number, appearance and turnover. (A) Podosomes were visualized by confocal microscopy in BMDC, (B) percentage of cells with podosomes, and (C) number of podosomes per cell determined. (D and E) The size of podosomes was analyzed from reconstructed 3D images. (F) Turnover of adhesion structures was analyzed by IRM. (A) Cells were stained with rhodamine phalloidin (red), vinculin (green), and DAPI (blue), insets show detailed podosome assembly (Scale bars, 25 μm.) (B) are averages ± SEM of 4–5 mice with 100–200 cells counted per mouse, in (C) of 30–70 cells containing podosomes and in (D) of 250–500 individual podosomes. (E) Representative images show how the podosome volume (green) was visualized. (F) Average ± SEM of three experiments, (filled circles) C57BL/6, (open circles) WAS KO, (filled triangles) Y293E, and (inverted filled triangles) Y293F. P values shown are from students t test.

Functional Defects of Cell Migration and Phagocytosis.

Chemotaxis abnormalities were determined using Dunn chambers, which allowed the analysis of individual DC. There was a general reduction in the velocity in response to CCL3 compared to wt BMDC (Fig. 4A). Y293F DC velocity was similar to WASp-null cells, whereas that of Y293E DC was intermediate. Individual cell trajectory, analyzed in circular histogram plots, showed that directional chemotaxis of WASp-null, Y293E, and Y293F DC was reduced compared to wt DC (Fig. 4B). Chemotaxis of WASp-null, Y293E and Y293F DC was also impaired in a transwell system compared to wt cells (Fig. 4C). WASp-null mtk;4acrophages and DC have been reported to display defective phagocytosis of particulate antigens (24, 25). There was a significant increase in phagocytosis of latex beads at both 15 min and 60 min compared to the negative control in wt BMDC that was not observed in WASp-null, Y293E, or Y293F BMDC (Fig. 4D).

Fig. 4.

Impaired chemotaxis, migration and phagocytosis in vitro. (A and B) BMDC chemotaxis toward CCL3 was analyzed in a Dunn chamber and (C) in transwells. (D) Phagocytosis of fluorescent latex beads was determined by flow cytometry. (A) shows average velocity ± SEM of five independent experiments. (B) Circular histograms of 100–150 cells with C57BL/6 showing significant chemotaxis (P < 0.001, Rayleigh V Test). The chemokine gradient is zero degrees. Each point in (C) represents one animal performed in triplicate and (D) average ± SEM of 0 (60 min on ice), 15, and 60 min incubation at 37 °C presented as an index compared to phagocytosis observed after 60 min on ice of 6–8 animals per group. P values shown are from students t test.

Abnormal Actin Polymerizing Activity In Vitro and In Vivo.

Y291E-hWASp has previous been shown to exhibit enhanced actin polymerization activity in vitro (14). Expression of this mutant in U937 cells resulted in increased levels of dysregulated F-actin, and higher levels of apoptosis and multinucleation compared to wtWASp as has been described previously for other activating mutants (26) (Fig. S4 A–C). Both Y293F and Y293E were tested in an in vitro actin polymerization assay. Formation of F-actin was determined by SDS/PAGE or visualized on latex beads stained with phalloidin. Human mutant Y291E exhibited an enhanced ability to stimulate actin polymerization compared to wtWASp, although not as potently as the constitutively active mutant S272P-hWASp or a mutant containing a deletion of the WASp sequence essential for VCA autoinhibitory binding (Fig. 5A). A similar increase in actin polymerization compared to wtWASp was observed for Y293E-mWASp (Fig. 5B and Fig. S5A) although not as potent as another constitutively active mutant, L272P-mWASp (murine equivalent of L270P-hWASp). Y293F-mWASp was able to mediate actin polymerization in this in vitro system at levels equivalent to wtWASp (Fig. 5B and Fig. S5A) (14). To assess actin polymerization in vivo BMDC were stimulated with CCL3 or LPS and cellular F-actin levels quantified by flow cytometry. The increase in actin polymerization in response to CCL3 for wt cells was significantly increased over WASp-null cells (Fig. 5C). When the more potent stimulus LPS was added, there was a further increase in actin polymerization in wt DC (Fig. 5D). By comparison, actin polymerization in Y293E mutant cells was similar to that observed in wt LPS stimulated cells regardless of the stimulation (Fig. 5 C and D) and Y293E mutant baseline F-actin level increased compared to wt cells (Fig. S5B). Y293F mutant mice cells responded in a similar manner to WASp-null mice showing little or no increase in actin polymerization with either stimulus in vivo (Fig. 5 C and D). The increase in actin polymerizing activity conferred by Y293E-mWASp was also evident in BMDCs (Fig. 5E). Therefore, Y291E-hWASp and Y293E-mWASp both support enhanced actin polymerization.

Fig. 5.

Increased actin polymerisation of phospho-mimetic mutant in vitro and in vivo. Actin polymerization of (A) human and (B) murine WASp constructs was analyzed by Western blot. Actin polymerization of BMDC of knockin mice in response to (C) CCL3 and (D) LPS was determined by staining with rhodamine phalloidin and flow cytometry. (E) Confocal analysis of phalloidin stained BMDC in pseudocolour (Scale bar, 25 μm.) (A and B) immunoblots with anti-GST or anti-actin antibodies of at least four independent experiments. Z scores show changes in actin polymerizing ability where values outside of shaded area are statistically significant from WT WASp as determined by ANCOVA, ∗, P < 0.05, ∗∗, P < 0.01, Stan = standard and CyD = cytochalasin D. (C and D) relative to polymerization observed in unstimulated DC.

Analysis of Protein Levels and Stability.

Despite enhanced actin polymerization in vitro, Y293E mutant mice exhibit a cellular phenotype that closely resembles WASp-null mice and the phospho-dead Y293F mutant. Immunoblotting of spleen and BMDC lysates for WASp revealed a specific band as predicted of approximately 60 kDa in wt mice which was absent in WASp-null mice. In comparison to wtWASp, Y293F-mWASp was expressed at normal levels, whereas the levels of Y293E-mWASp were markedly reduced (Fig. 6A). Real-time quantitative PCR revealed a reduction in the mRNA in WASp-null mice as been previously reported (27). However, mRNA levels were equivalent to normal for both Y293E and Y293F mutants (Fig. 6B) indicating a posttranslational mechanism for reduction in protein. Immunoprecipitation experiments to determine WASp/WIP interaction revealed that WIP bound normally to both the Y291E-hWASp mutant and a S272P-hWASp constitutively active human mutant, but not to an EVH1 domain deletion mutant (Fig. 6 C and D). Previous studies have suggested WASp is susceptible to degradation through calpain and ubiquitin-proteasome-mediated pathways (particularly in the absence of WIP stabilization). When cells from mutant mice were treated with the proteasome inhibitors bortezomib or MG-132 there was a variable, but generally reproducible significant increase in the amount of full-length Y293E-mWASp that was not seen with the calpain inhibitor ALLN (Fig. 6 E and F). In contrast, there was little increase in the levels of Y293F-mWASp with either treatment. In addition there was an increase in the turnover of Y293E compared to both wt and Y293F in transfected 293T cells (Fig. S6). These findings suggest that ubiquitin-proteasome-mediated degradation is a dominant mechanism for the Y293E-mWASp mutant, and by implication for tyrosine phosphorylated wtWASp.

Fig. 6.

Mutation of Y293 dramatically alters WASp protein levels. (A) WASp levels were determined by Western blot of spleen and BMDC lysate and (B) mRNA by RT-QPCR. (C) WIP binding was determined by WIP immunoblot after GST-WASp immunoprecipitation and (D) z score of WASp-WIP affinity determined, outside the shaded area are significant determined by ANCOVA, ∗, P < 0.05, ∗∗, P < 0.01. (E) WASp degradation was analyzed by immunoblot after incubation with proteasome and calpain inhibitors. (F) the relative increase of full length WASp when treated with (light gray squares) bortezomib or MG-132 and (dark gray squares) ALLN compared to untreated samples. (A) Representative of at least 10 experiments. (E) Representative of at least 10 experiments. P values shown are from students t test.

Discussion

Recent data indicates that WASp (and N-WASp) is at least in part regulated by release of an autoinhibited conformation upon binding of Cdc42 and PIP₂ with contribution from other factors such as SH3 domain containing adaptor molecules such as Nck and WISH (11–13). Phosphorylation of a conserved tyrosine residue within the GBD of WASp has emerged as an additional potential regulator of activity as the introduction of a negative charge by phosphorylation would destabilize the autoinhibited conformation and facilitate binding of the Arp2/3 complex. In support of this, a synthetic phosphomimicking mutant (Y291E) has been shown to enhance actin polymerization both in vitro and in vivo (14). It has been suggested that efficient phosphorylation and dephosphorylation at Y291 are contingent on binding of GTP-loaded Cdc42, as the structure of the autoinhibited conformation would protect this residue from modifying enzymes (15). There is also some evidence for separation of these events (14, 28). It has been reported that tyrosine phosphorylation is absolutely required for normal T cell activation occurring independently of Cdc42 binding (29). However, the GBD mutants used for these experiments could have resulted in a disturbance of the autoinhibited confirmation. Phosphorylation of Y291 has also been implicated in formation of filopodia and phagocytic cups in macrophages, and osteoclast sealing zones (17, 18). However, the physiological significance and importance for normal activation in vivo is uncertain.

Recently, several human WASp mutations have been identified that result in constitutive activation through disruption of the autoinhibitory conformation, resulting in a phenotype of myeloid cytopenia (30–32). In all cases, the levels of WASp within peripheral blood cells have been reported to be normal, but they have displayed cytoskeletal abnormalities associated with enhanced and delocalized actin polymerization (26, 31). In some ways, these mutations (particularly I294T-hWASp) would be predicted to mimic effects of constitutive WASp phosphorylation.

To define the role of WASp phosphorylation in vivo, we have generated mice with mutations that either prevent (Y293F) or mimic phosphorylation (Y293E) at least in terms of charge alteration without disrupting the binding sites of other factors such as Cdc42. Our findings underline the absolute requirement for phosphorylation during multiple cellular tasks, including migration, phagocytosis, and proliferation. In many ways, the Y293F mutant resembles a WASp-null mutant, with no obvious redundancy. What is more surprising is that the Y293E mutant has a similar cellular phenotype. Biochemically, Y293E behaves as expected in that actin polymerization is enhanced in vitro, and in cells derived from mice. However, the in vitro activity is not as pronounced as that of an alternative activating mutant, suggesting that other physiological ligands are required for full activity. Interestingly, the most active mutant in vitro contains a wider deletion of the VCA-binding region of WASp, yet retains the Cdc42-binding site. It is therefore reasonable to suggest that the autoinhibited confirmation can be disrupted to different degrees depending on the type and combination of stimuli and that phosphorylation may act to lower the threshold for full activation.

Surprisingly, the levels of Y293E-mWASp were severely diminished as a result of post-translational degradation, whereas Y293F-mWASp was expressed at normal levels. It is therefore difficult to determine whether the cellular dysfunction in this mutant strain is due to deficiency of WASp or abnormal subcellular localization of actin polymerization as a result of constitutive activity (Fig. S4D). Although there is an increase in total actin polymerization within the cell as shown in DC, there may be relative deficiency at localized sites within the cell where it is needed for normal biological function. This is supported by the observation that cells from patients with activating mutations often show a similar end-phenotype to that of classical WAS patients, such as compromised CD3-mediated lymphocyte proliferation, and abnormal cell motility and phagocytosis (31). As patients with constitutively active mutations appear to express relatively normal levels of WASp, phosphorylation and degradation may be intimately linked. The levels of WASp within a cell have been shown to be regulated by calpain and the ubiquitin-proteasome (8, 10). In normal cells, the majority of WASp is complexed with its chaperone WIP. Mutations in the WH1 domain abrogate WIP binding, and levels of WASp are markedly diminished as a result of degradation. Similarly, WIP-/- mice have very low levels of WASp (8, 33). Treatment of such cells with calpain inhibitors and/or proteasome inhibitors results in partial restoration of protein levels, suggesting that calpain and the ubiquitin-proteasome play an important role in turnover. Both Y293 mutants tested in this study appear to bind WIP normally, but it is possible that the degradative mechanisms are separate from those implicated in situations where WIP does not bind, or is released from binding. In fact, it remains unclear whether dissociation of WIP from WASp is a requirement for activation of the Arp2/3 complex, and therefore whether degradation in the absence of WIP binding is at all physiologically relevant during normal cell function (34, 35).

The ubiquitin-proteolysis system has been implicated in control of the abundance of many regulatory proteins following polyubiquitination and in control of immune signaling pathways. Phosphorylation of Y256 in N-WASp in epidermal growth factor-treated PC12 cells initiates degradation by the proteasome via a ubiquitin mediated pathway, although the persistence of phosphorylated N-WASp was modulated by treatment with nerve growth factor to stimulate neurite outgrowth (36). N-WASp proteasome mediated degradation has also been shown to be prevented by binding of the adaptor molecule HSP90 (37). It is likely that phosphorylation of hWASp at Y291 has a similar functional activity, and is consistent with findings in this study.

Targeted knockin models have the advantage of allowing the study of molecules in an appropriate cellular and regulatory environment. Our findings suggest that phosphorylation is a key event in the activation of WASp during many cellular activities. Furthermore, phosphorylation may target WASp for proteasomal degradation, thereby providing a default pathway for switching off Arp2/3 mediated actin polymerization in a timely fashion.

Materials and Methods

In Vitro Actin Polymerization and Immunoblots.

COS-7 and U937 cells were transfected with GST-hWASp-WT, GST-mWASp-WT, GST-hWASp-Y291E, GST-hWASp-S272P, GST-mWASp-Y293F, GST-mWASp-Y293E, GST-mWASp-L272P, or GST vector. WASp-coated beads were incubated in U937 lysate as previously described (14). Cytochalasin D (2 μM; Sigma) was used as negative control. Beads were resolved by SDS/PAGE and stained for GST (B-14; Santa Cruz) or actin (AC-15; Sigma). Similarly BMDC generated as previously described (6, 10) and 1 × 10⁷ splenocytes treated as described in Results were lysed in buffer containing sodiumorthovanadate and protease inhibitors resolved by SDS/PAGE then probed for WASp (B-9, Santa Cruz), GAPDH (6C5, Santa Cruz) or β actin (AC-74, Sigma).

Mice.

C57BL/6 wild-type (Charles River), WAS KO mice (kindly supplied by T. Strom, Memphis, TN) and knockin mice (see SI Text for generation methods) were housed in SPF conditions and used at 6–18 weeks of age. Experiments were performed according to Home Office animal welfare legislation.

Immunofluorescence Staining and Imaging.

Splenocytes, lymph node cells, leukocytes, and splenic cryosections were incubated with antibodies (see SIText for details) and cells analyzed using a CyAn flow cytometer (Dako) and Summit (Dako) software. BMDC were stained with rhodamine phalloidin (Invitrogen) and mouse anti-vinculin (hVIN-1, Sigma). Images were captured on a Leica (TCS-SP2) confocal microscope and processed in Adobe Photoshop. Marginal zone thickness was determined as described previously (38). Volume of podosomes was quantified by automated analysis using Volocity (Improvision). Volume and surface area of podosomes were recorded from xyz (3D) images.

Analysis of Adhesion Turnover.

Interference reflection microscopy (IRM) was used to visualize the adhesion-substratum interface of cells as described previously (39, 40). SPDCs were plated on coverslips and mounted onto viewing chambers. Interference reflection micrographs were collected and analyzed as previously described (41).

In Vitro Migration.

Fibronectin coated (10 μg/mL; Sigma) transwells (5-μm pores; Costar, Corning) had BMDC (1.5 × 10⁵ cells) added to the upper and CCL3 (100 ng/mL; Peprotech) or culture medium to the lower compartment. After a 90-min incubation at 37 °C, cells from the lower compartment were counted by flow cytometry. 25 × 10³ BMDC, on fibronectin-coated coverslips were mounted on Dunn chambers (Hawksley) and migration toward CCL3 recorded at 37 °C by microscopy (Zeiss Axiovert 135; Zeiss). Images were acquired every 10 min over 5 h and migration analyzed with Volocity .

T-Cell Proliferation.

Thymocytes were stimulated in triplicate for 72 h in RPMI with 10% FCS with anti CD3e antibody (BD Biosciences) coated onto 96-well plates. For the last 16 h of culture, 3H-thymidine (0.5 μCi [0.0185 MBq]; GE Healthcare) was added, plates harvested and proliferation measured as 3H-thymidine incorporation.

In Vivo Actin Polymerization and Phagocytosis.

BMDC were stimulated for 90 sec at 37 °C with LPS (100 ng/mL; Sigma) or CCL3 (100 ng/mL), fixed, permeabilized (BD Cytofix/Cytoperm), stained with rhodamine-phalloidin and analyzed by flow cytometry. For phagocytosis, BMDC were incubated with FITC-labeled beads (Sigma) for 15 and 60 min at 37 °C or 60 min on ice, washed in PBS and fixed. Cells were analyzed by flow cytometry and phagocytosis expressed as an index relative to phagocytosis after a 60-min incubation on ice.

Real-Time Quantitative PCR.

Total RNA was reverse transcribed to form a DNA/RNA hybrid. Expression of WAS was compared to pgk-1 as internal control (Applied Biosystems, Taqman Gene Expression Assays). Relative expression is represented as 2^−ΔΔCT between samples compared to a normal control.