Research ArticleCell Biology

Triple-Color FRET Analysis Reveals Conformational Changes in the WIP-WASp Actin-Regulating Complex

See allHide authors and affiliations

Science Signaling  24 Jun 2014:
Vol. 7, Issue 331, pp. ra60
DOI: 10.1126/scisignal.2005198

Abstract

Wiskott-Aldrich syndrome protein (WASp) is a key regulator of the actin cytoskeletal machinery. Binding of WASp-interacting protein (WIP) to WASp modulates WASp activity and protects it from degradation. Formation of the WIP-WASp complex is crucial for the adaptive immune response. We found that WIP and WASp interacted in cells through two distinct molecular interfaces. One interaction occurred between the WASp-homology-1 (WH1) domain of WASp and the carboxyl-terminal domain of WIP that depended on the phosphorylation status of WIP, which is phosphorylated by protein kinase C θ (PKCθ) in response to T cell receptor activation. The other interaction occurred between the verprolin homology, central hydrophobic region, and acidic region (VCA) domain of WASp and the amino-terminal domain of WIP. This latter interaction required actin, because it was inhibited by latrunculin A, which sequesters actin monomers. With triple-color fluorescence resonance energy transfer (3FRET) technology, we demonstrated that the WASp activation mechanism involved dissociation of the first interaction, while leaving the second interaction intact. This conformation exposed the ubiquitylation site on WASp, leading to degradation of WASp. Together, these data suggest that the activation and degradation of WASp are delicately balanced and depend on the phosphorylation state of WIP. Our molecular analysis of the WIP-WASp interaction provides insight into the regulation of actin-dependent processes.

INTRODUCTION

Stimulation of the T cell receptor (TCR) by cognate antigen triggers multiple cellular processes, including the formation of signaling complexes, the initiation of transcriptional activity, and the regulation of cellular adhesion and motility (1, 2). These cellular processes are dependent on dynamic rearrangements of the actin cytoskeleton, initiated at the interface between the T cell and an antigen-presenting cell, which leads to the formation of the immunological synapse (1, 3). The hematopoietic cell–specific Wiskott-Aldrich syndrome protein (WASp) and its ubiquitously expressed homolog neural WASp (N-WASp) are actin nucleation–promoting factors that play a key regulatory role in cytoskeletal dynamics (4, 5).

WASp-interacting protein (WIP) is a ubiquitously expressed actin-binding protein that interacts with WASp and N-WASp, controlling the ability of hematopoietic and nonhematopoietic cells to undergo migration, adhesion, spreading, and invasion (69). WASp activity depends on its functional activation. In resting T cells, an intramolecular interaction between the hydrophobic patch within the guanosine triphosphatase (GTPase)–binding domain (GBD) and the verprolin homology, central hydrophobic region, and acidic region (VCA) domain at the C terminus of WASp maintains the protein in an autoinhibited conformation. After cellular activation, the Rho-family GTPase Cdc42 binds to the WASp GBD, and Tyr291 undergoes protein tyrosine kinase–mediated phosphorylation. Together, these two events induce a substantial conformational change within WASp (10), which releases the VCA domain to activate the actin nucleation complex, Arp2/3, thereby promoting actin polymerization (11, 12).

Missense mutations, molecular deletions, or frameshifts within the gene encoding WASp result in a partial or complete deficiency in WASp, which causes primary immunological disorders, such as Wiskott-Aldrich syndrome (WAS) and X-linked thrombocytopenia (XLT) (13). Most WAS- and XLT-inducing missense mutations are located in the region encoding the N-terminal WASp-homology-1 (WH1) domain (14, 15), which is recognized by WIP. Both WASp and WIP play key roles in immune cell motility, cytokine production, proliferation, and transcriptional activity (1618). Both proteins are recruited to the immunological synapse (19), and they are required for the formation of membrane structures, such as lamellipodia and podosomes (20, 21). In resting lymphocytes, more than 95% of WASp is found in complex with WIP, which protects WASp from degradation by calpain (22, 23) or as a result of ubiquitylation (24). Mutations in the region encoding the WH1 domain of WASp impair its interaction with WIP, an interaction that is crucial for WASp stability (2224). This chaperone-like function of WIP was unequivocally established by the observation that the abundance of functional WASp is reduced in WIP-deficient mice, whereas the amounts of WASp mRNA are unaffected, and that WASp activity is rescued by the addition of exogenous WIP (22, 23, 25).

Formation of the WIP-WASp complex is crucial for the stability and function of WASp and for the overall immune response (2, 4). WIP is essential for the recruitment of N-WASp to the adapter protein Nck at sites of actin polymerization (26, 27). However, the molecular nature of the WIP-WASp interaction and the manner by which WIP shields WASp from degradation are the subject of intense debate (24, 2831), with several core issues remaining unresolved. One of the most intriguing questions is how WIP protects WASp from degradation while enabling its release from autoinhibition. The domains and structural elements within WASp and WIP that are involved in their physical interaction are not fully characterized (6, 7, 27, 28, 32, 33). In contrast to the well-accepted interaction between the WH1 domain of WASp and the C terminus of WIP (6, 32, 33), there are conflicting reports regarding the inhibitory role suggested for the N terminus of WIP as a possible additional domain affecting the transcriptional activity of WASp (28). Another insufficiently understood event is the phosphorylation of WIP at its consensus protein kinase C θ (PKCθ) phosphorylation motif (RxxSxR, residues 485 to 490) after T cell activation. In addition, the role of PKCθ in the formation of the WIP-WASp complex is unclear. Finally, little if any information is available on WIP-WASp binding at the cellular spatiotemporal level, which is difficult to address with standard biochemical methods. Thus, questions such as the timing during TCR signaling at which these two proteins become associated or are released, and the site within the activated cell at which this interaction occurs, remain unanswered.

We investigated the role of WIP phosphorylation in the dynamics of WIP-WASp complex formation and demonstrated an interaction between WIP and PKCθ at the early stages of T cell activation. Furthermore, with triple-color fluorescence resonance energy transfer (3FRET) analysis (34), we demonstrated that WIP was present in an autoinhibited conformation in resting T cells in which it tightly bound to WASp and prevented its degradation. This autoinhibitory conformation affected the movement of WIP and its function in actin rearrangement. After TCR activation and the phosphorylation of WIP on Ser488 by PKCθ, a conformational change released WIP from its autoinhibited state and altered, but did not abolish, the WIP-WASp interaction. This conformational change exposed the WH1 domain of WASp to ubiquitylation and promoted its degradation. We defined the nature of the interaction and the molecular determinants of complex formation, and identified the domains of WIP and WASp that were essential for a functional interaction between both proteins. We demonstrated the involvement of actin, whose nucleation and polymerization are mediated by the WIP-WASp complex, in the formation of this complex. Our findings provide a possible explanation for the susceptibility of WASp to degradation in hematopoietic cells of WAS patients, and provide evidence of the regulated nature of the WIP-WASp interaction, which controls actin polymerization.

RESULTS

WIP and WASp constitutively associate in T cells

The effect of T cell activation on the molecular nature of the WIP-WASp interaction has been a source of controversy (28, 31). One proposed mechanism suggests that after TCR engagement, WASp is recruited to the immunological synapse and is released from its inhibitory interaction with WIP (31), which results in WASp-mediated activation of Arp2/3 and local actin polymerization. However, subsequent studies contradicted this paradigm and demonstrated that dissociation of the WIP-WASp complex is not required for WASp function, including its role in transcriptional activity (28). We used immunoprecipitation to determine how the interaction of WIP and WASp was affected by TCR activation and the phosphorylation of WIP. We generated two WIP mutants: the R485K, R490K double mutant (WIP R485,490K) and the S488D single mutant (WIP S488D). The former is inaccessible to phosphorylation because it is not recognized by PKCθ, and the latter mimics constitutively phosphorylated WIP (31). We immunoprecipitated wild-type WIP and both mutant WIP proteins from activated and nonactivated Jurkat E6.1 cells (a human CD4+ T cell leukemia cell line) and analyzed the samples by Western blotting with an anti-WASp antibody (fig. S1A). We observed no change in the WIP-WASp interaction as a consequence of either TCR activation or WIP phosphorylation state, suggesting that WASp constitutively associates with WIP (fig. S1A).

The WIP-WASp interaction is modulated by TCR activation and phosphorylation

Whereas standard biochemical techniques offer reliable, but limited, under?????standing (in terms of sensitivity and time resolution) of signaling complex formation and regulatory mechanisms, imaging experiments with multiple tagging of proteins and FRET measurements provide a technology capable of overcoming the fundamental limitations of biochemical approaches. Thus, to understand their interaction within a cellular context, we fluorescently tagged wild-type WASp and WIP with cyan fluorescent protein (CFP) and yellow fluorescent protein (YFP), respectively. Cells stably expressing CFP-WASp and YFP-WIP were plated on either nonstimulatory coverslips (coated with an anti-CD43 antibody) or stimulatory coverslips (coated with an anti-CD3 antibody) and then were fixed after 2 or 5 min (Fig. 1A). Under nonstimulatory conditions, no clusters of WIP or WASp were observed in the cells, and a dispersed pattern of high FRET efficiency was detected throughout T cell spreading (after 2 and 5 min, we observed 27 ± 3.4% and 19.8 ± 1.4% FRET efficiency, respectively; P ≥ 0.08), indicating a close and stable interaction between WIP and WASp. In contrast, under stimulatory conditions, clustering of WIP and WASp occurred, and the high average FRET efficiency observed after 2 min (33.6 ± 6.5%) was statistically significantly reduced after 5 min of stimulation (7.5 ± 3.2%, P ≤ 0.001) (Fig. 1, A and B). During this time, the clusters translocated to the cell perimeter and rapidly dissipated (movie S1).

Fig. 1 The WIP-WASp interaction is modified after TCR engagement and WIP phosphorylation.

(A) Jurkat E6.1 cells expressing CFP-WASp and YFP–WIP wild type (WT) were plated over a nonstimulatory coverslip coated with anti-CD43 antibody (left) and were fixed after 2 min (n = 43 cells) or 5 min (n = 46 cells), or were plated over a stimulatory coverslip coated with anti-CD3 antibody (right) and were fixed after 2 min (n = 45 cells) or 5 min (n = 41 cells) of activation. Cells were imaged by confocal microscopy, and FRET efficiency was measured by the donor-sensitized acceptor emission technology (see Materials and Methods for details). (B) Graph summarizing the percentage FRET efficiency in cells plated over the indicated coverslips for the indicated times. Data are means ± SEM from at least four independent experiments. (C) Jurkat E6.1 cells transfected with plasmids encoding CFP-tagged WASp and YFP-tagged WT WIP or WIP phosphorylation site mutants (R485,490K or S488D) were plated over a stimulatory coverslip and fixed after 2 or 5 min of activation. The numbers of cells analyzed are as follows. WASp and WT WIP: n = 45 and 41 cells for 2 and 5 min, respectively; for WASp and WIP R485,490K: n = 37 and 42 cells for 2 and 5 min, respectively; and for WASp and WIP S488D: n = 49 and 50 cells for 2 and 5 min, respectively. (D) Mean FRET efficiency in cells expressing WASp and the indicated form of WIP. Data are means ± SEM from three independent experiments.

To investigate whether phosphorylation induced a structural change in the WIP-WASp molecular complex, we repeated this experiment with cells stably expressing either of the two mutant forms of YFP-WIP. We performed flow cytometry and Western blotting analyses to show that cells contained similar amounts of the tagged wild-type and mutant WIP proteins (fig. S1, B and C). Cells expressing WIP mutant protein were stably transfected with plasmid encoding CFP-WASp, and both cluster distribution and FRET efficiency in these cells were compared to those observed in cells expressing YFP-tagged WIP wild type (Fig. 1C). During the first 2 min of activation, similar distributions of signaling clusters were recruited to the stimulatory plane, the site of contact between the TCR-CD3 complex and the anti-CD3 antibody, for all three forms of WIP (Fig. 1C). However, despite the previously demonstrated constitutive association of WIP and WASp, FRET efficiencies between WASp and either of the two mutant WIP proteins indicated marked differences in the molecular nature of their interaction (Fig. 1D). The FRET efficiency for WIP S488D was negligible at 2 and 5 min (1.3 ± 0.9% and 0%, respectively), whereas that of the WIP R485,490K mutant was high and persistent at 2 and 5 min after activation (27 ± 7.8% and 26.2 ± 6%, respectively). Together with the immunoprecipitation results, these findings suggest that whereas WIP and WASp interacted constitutively, both T cell activation and phosphorylation modulated this association, which caused a molecular change in the interaction between the two proteins, as observed by the change in the FRET efficiency. These results also suggest that WIP phosphorylation at Ser488 by PKCθ decreases the WIP-WASp interaction during TCR activation.

Phosphorylation of WIP at Ser488 is required for its movement and for actin rearrangement

To more precisely measure the effects of WIP phosphorylation on cluster stability and dynamics during T cell activation, we incubated cells expressing YFP-tagged wild-type and mutant WIP proteins on stimulatory plates and continuously visualized them by confocal microscopy. The resulting images were rendered into kymographs, and the movement of individual YFP-WIP clusters in multiple live cells was traced from the time they formed until they dissipated. Such cluster analysis enhances the temporal resolution and gives valuable information on cluster lifetime and kinetics.

Upon T cell activation, wild-type WIP was recruited to the contact site and then translocated to actin-rich sites at the perimeter of the cell (Fig. 2A, top panel), as was previously described for WASp (35). Whereas the WIP S488D mutant, which only loosely associates with WASp, formed clusters similar to those formed by wild-type WIP, WIP R485,490K formed smaller clusters that had slower movement than did those containing wild-type WIP (Fig. 2A, middle and bottom panels). We quantified the differences in clustering by following individual cluster movement rates and lifetimes (Fig. 2B), and confirmed that the behavior of WIP-containing clusters was affected by the phosphorylation state of WIP (Fig. 2, C and D). Clusters containing the constitutively unphosphorylated WIP R485,490K mutant exhibited statistically significantly longer lifetimes (with an average persistence time of 182 s) and slower movement (average speed of 0.23 μm/s) compared to those containing wild-type WIP, which had an average persistence time of 126.42 s (P ≤ 0.006) and an average speed of 2.64 μm/s (P ≤ 0.0008). Clusters containing the constitutively phosphorylated WIP S488D mutant had an opposing phenotype, exhibiting shorter lifetimes (average persistence of 65.10 s) and faster movement (4.46 μm/s) toward the center of the cells (Fig. 2, B to D, and movies S2 to S4). These findings suggest that WIP phosphorylation status affects not only the nature of the WIP-WASp interaction but also the stability and motility of WIP.

Fig. 2 Phosphorylation of WIP reduces the stability and enhances the mobility of WIP and WASp clusters.

(A) Live Jurkat E6.1 cells transfected with plasmids encoding YFP–WIP WT (top), YFP–WIP R485,490K (middle), or YFP–WIP S488D (bottom) were plated onto stimulatory coverslips and were imaged at the indicated times. White arrows indicate WIP clusters. (B) Average traces demonstrating the movement of the indicated YFP-WIP clusters. The y axis shows how far the clusters moved (μm), whereas the x axis shows how long the clusters remained visible (s). Ten individual clusters in each of 20 cells from each group were analyzed to give a total of at least 200 clusters from each group. (C and D) The data sets presented in (A) and (B) were used to determine (C) the mean lifetimes (s) of the indicated WIP clusters and (D) the average speed (μm/s) at which WIP clusters moved. (E) Jurkat E6.1 cells were transfected with plasmid encoding CFP-WASp together with plasmid encoding WT WIP, WIP R485,490K, or WIP S488D. Clusters containing CFP-WASp were imaged and analyzed as described earlier to determine the lifetime(s) of visible clusters. Ten individual clusters from each of 10 cells from each group were analyzed to give a total of at least 100 clusters from each group. Data are representative of four independent experiments.

With a similar approach, we examined the dynamics of CFP-WASp in cells that also expressed YFP–WIP S488D (Fig. 2E). Whereas WASp-containing clusters in cells expressing the nonphosphorylated WIP R485,490K mutant persisted for longer than did those formed in cells coexpressing wild-type WIP (average persistence of 180.19 s, compared to 130.2 s for those in cells coexpressing wild-type WIP; P ≤ 0.00001), WASp clusters were less persistent in cells coexpressing the constitutively phosphorylated WIP S488D mutant (average persistence time of 73.9 s, P ≤ 0.0033). This rapid dissipation of WASp-containing clusters suggests that WIP phosphorylation affects both the dynamics and stability of WASp. WIP also interacts with other proteins to modulate WASp-independent functions (18, 36). Thus, short-lived clusters containing WIP S488D may represent WIP molecules that form transient clusters with other proteins or, alternatively, WIP molecules that are loosely associated with WASp.

Next, we investigated the influence of WIP phosphorylation on its function. WIP is involved in the regulation of the actin cytoskeleton by directly binding to actin or actin-associated proteins, such as cortactin, profilin, and Nck, which mediate the polymerization or rearrangement, or both, of actin (3638). Immunoprecipitation analysis did not show substantial differences in the ability of the different WIP Ser488 mutant forms to bind to actin (fig. S1A). However, analysis of FRET between CFP-tagged actin and the YFP-tagged WIP mutants showed that the binding of WIP S488D to actin was similar to that of wild-type WIP (FRET efficiencies of 40.1 ± 0.9% and 35.9 ± 3.1%, respectively; P ≥ 0.37, fig. S1D), whereas WIP R485,490K exhibited statistically significantly reduced actin binding (FRET efficiency of 16.3 ± 2.4%, P ≤ 0.0001, fig. S1D and movie S5). Furthermore, WIP S488D induced normal rearrangement of actin, as indicated by the actin shape index (fig. S1E) (39), whereas cells expressing WIP R485,490K displayed aberrant cell spreading behavior (fig. S1E, P ≤ 0.0001). These findings suggest that the phosphorylation status of WIP affects normal cytoskeletal behavior.

PKCθ is required for the conformational change in WIP-WASp

PKCθ, a central regulator of immune cell activation and proliferation, phosphorylates WIP at its consensus phosphorylation motif (R485xxSxR490) (31). To show that this phosphorylation event is the molecular switch that alters the WIP-WASp interaction in activated T cells, we transfected T cells stably expressing YFP-WIP and CFP-WASp with a small interfering RNA (siRNA) pool specific for PKCθ and then analyzed the cells for FRET efficiency between WIP and WASp (Fig. 3, A to C). The same cells transfected with a nonspecific scrambled siRNA pool served as a negative control. Transfection with the PKCθ-specific siRNA led to a marked decrease in the abundance of PKCθ (Fig. 3C). In addition, we found that the typical reduction in the FRET efficiency between WASp and WIP after 5 min of cell stimulation was no longer detected (Fig. 3, A and B; P ≥ 0.82). This result suggests that PKCθ is required for the molecular change in WIP that results in distancing it from WASp after TCR stimulation. Because our immunoprecipitation analysis failed to detect this change (fig. S2A), we hypothesized that this distancing might indicate a partial dissociation, occurring at the C-terminal region of WIP that contains the PKCθ phosphorylation site, but that both proteins might remain bound together in the same signaling complex. Knockdown of PKCθ did not affect the interaction of WIP with actin (fig. S2A), which binds to the N terminus of WIP, and was unaffected by WIP phosphorylation (fig. S1A).

Fig. 3 PKCθ reduces the association between WIP and WASp.

(A) Jurkat cells transfected with plasmids encoding YFP-WIP and CFP-WASp were treated with PKCθ-specific siRNA. Twenty-four hours later, cells were plated over a stimulatory coverslip and were fixed after 2 min (n = 44 cells) or 5 min (n = 39 cells) of activation. Analysis was performed to determine the FRET efficiency in the indicated cells. Cells treated with nonspecific siRNA served as a negative control for the 2-min (n = 45 cells) and 5-min (n = 41 cells) time points. (B) Graph summarizing the percentage FRET efficiency in cells plated over activating coverslips for the indicated times. Data are means ± SEM from at least three independent experiments. (C) Jurkat cells were transfected with PKCθ-specific siRNA or with nonspecific (NS) siRNA and then were analyzed by Western blotting (IB) with the indicated antibodies. Relative PKCθ abundance is presented in the bottom panel (P ≤ 0.0001). Data are representative of three independent experiments. (D) Peripheral blood lymphocytes (PBLs) treated with PKCθ-specific siRNA or nonspecific siRNA (negative control) were subjected to immunoprecipitation (IP) with anti-WASp antibody and then were analyzed by Western blotting with anti-ubiquitin (ub) and anti-WASp antibodies. (E) Jurkat cells transfected with plasmids encoding the indicated YFP-tagged WIP variants were subjected to immunoprecipitation with anti-WASp antibody and were analyzed by Western blotting with anti-ubiquitin and anti-WASp antibodies. Ubiquitylated WASp band appears at the molecular mass of 81 kD. All Western blots are representative of three independent experiments.

Partial dissociation of WIP from WASp after TCR activation results in the ubiquitylation of WASp

We previously showed that WASp is ubiquitylated on lysine residues 76 and 81 after TCR activation. Furthermore, we demonstrated that the ubiquitylation of WASp leads to its degradation (24). Thus, our next step was to examine the effect of WIP phosphorylation by PKCθ on the functional effects of the WIP-WASp molecular complex and, specifically, on WASp degradation. We transfected primary T lymphocytes with PKCθ-specific siRNA or a nonspecific control siRNA, and examined WASp ubiquitylation in TCR-stimulated cells by immunoprecipitation analysis (Fig. 3D). First, we verified that there was a >90% decrease in PKCθ abundance in the cells transfected with PKCθ-specific siRNA compared to that in cells transfected with nonspecific siRNA (fig. S2B). The immunoprecipitation analysis revealed differences in the extent of WASp ubiquitylation between cells transfected with PKCθ-specific siRNA and those transfected with control siRNA, as detected by a smear of bands recognized by an anti-ubiquitin antibody, which represented ubiquitylated WASp (Fig. 3D). Knockdown of PKCθ led to decreased amounts of ubiquitylated WASp, in comparison to those in control cells, after TCR activation (Fig. 3D). Accordingly, the amount of total WASp protein was greater in cells transfected with PKCθ-specific siRNA than in control cells, as determined by Western blotting analysis (fig. S2A).

To more specifically explore the role of WIP phosphorylation status on the stability of WASp, we performed a similar immunoprecipitation analysis with cells expressing wild-type WIP or mutant forms of WIP (Fig. 3E). Wild-type WASp was ubiquitylated after TCR activation, as expected, and WASp was ubiquitylated to a greater extent in cells expressing WIP S488D (Fig. 3E). The amount of ubiquitylated WASp observed in cells expressing WIP R485,490K was reduced and was similar to that in unstimulated cells. We obtained similar results from experiments with primary T lymphocytes expressing wild-type or mutant WIP proteins (fig. S2C). These data suggest that WASp degradation depends on the phosphorylation status of WIP and its molecular interaction with WASp.

3FRET analysis of live activated T cells monitors the three-way interactions between WIP, WASp, and PKCθ

The kinetics and dynamics governing the interaction between PKCθ and WIP and the effects of WIP phosphorylation on the stability and function of the WIP-WASp complex are unclear. We applied FRET-based imaging methods to investigate the interplay between PKCθ, WIP, and WASp throughout T cell activation. In cells stably expressing YFP-WIP and PKCθ-CFP (fig. S2D), both WIP and PKCθ were colocalized after 1 min of activation, and FRET efficiency between them was high (20.2 ± 5.3%) (Fig. 4, A and B). Soon thereafter, we observed a marked reduction in FRET efficiency, which corresponded to a reduced association between WIP and PKCθ (2.7 ± 1.4% and 0% FRET efficiency after 3 and 5 min, respectively), although both proteins were colocalized at the cell perimeter (Fig. 4, A and B).

Fig. 4 Simultaneous tracking of the dynamic interactions between WIP, WASp, and PKCθ by 3FRET analysis.

(A) Cells expressing YFP-WIP were transfected with plasmid encoding PKCθ-CFP. Forty-eight hours later, cells were plated over a stimulatory coverslip and fixed after 1 min (n = 40 cells), 3 min (n = 42 cells), or 5 min (n = 36 cells) of activation. FRET efficiency between WIP and PKCθ was measured as described earlier. (B) Graph summarizing the percentage FRET efficiency in cells plated over activating coverslips for the indicated times. Data are means ± SEM from at least three independent experiments. (C) Schematic illustration of the 3FRET donor-sensitized acceptor fluorescence assay (see Materials and Methods for details). (D) Jurkat E6.1 cells stably expressing CFP-WASp and YFP-WIP were transfected with plasmid encoding PKCθ-mCherry. Cells were plated over stimulatory antibody-coated coverslips for 1 min (n = 40 cells), 3 min (n = 41 cells), or 5 min (n = 45 cells), and 3FRET analysis was performed. (E) Graph summarizing the percentage FRET efficiency in each of the three indicated channels in cells plated over activating coverslips for the indicated times. Data are means ± SEM from at least four independent experiments.

Next, we followed the dynamics of the molecular complex consisting of WIP, WASp, and PKCθ throughout TCR activation with 3FRET technology (34), which involves energy transfer between three fluorophores, and thus enables the assessment of conformational changes within and between signaling molecules (Fig. 4C). Cells stably expressing CFP-WASp and YFP-WIP were transfected with plasmid encoding PKCθ tagged with the mCherry fluorophore (PKCθ-mCherry, fig. S3A). Thus, the CFP-YFP, YFP-mCherry, and CFP-mCherry FRET efficiencies reflect the interactions between WASp-WIP, WIP-PKCθ, and WASp-PKCθ, respectively. Throughout T cell activation, there was reduction in the FRET efficiencies between WASp-WIP and WIP-PKCθ (Fig. 4, D and E), in agreement with the previously described double-FRET experiments. These data suggest that WIP interacts with both WASp and PKCθ within the first minute of activation and later dissociates from PKCθ. In contrast, the FRET efficiency between WASp and PKCθ remained high (47.6 ± 4.6%, 40.3 ± 2.4%, and 41.5 ± 2.1% after 1, 3, and 5 min of activation, respectively), indicating that these proteins remained in close proximity throughout the activation process. These results suggest that the interaction between PKCθ and its target WIP is transient and occurs only at the initial stages of TCR activation.

The VCA domain of WASp and the N-terminal region of WIP form a second interaction site in the WIP-WASp complex

Our earlier findings led us to consider the possibility that WIP in its autoinhibitory conformation associates with WASp at a second, phosphorylation-independent binding site, and that phosphorylation of Ser488 of WIP causes only partial dissociation of WIP from WASp. Plausible partners for this second binding site are the N-terminal region of WIP (amino acid residues 1 to 112, WIP-N) and the C-terminal domain (also known as the VCA region) of WASp. Actin binds to both the VCA (40) and N-terminal WH2 domains of WIP (7), and thus may serve as a molecular bridge between WASp and WIP. To investigate this possibility, we compared the interactions of a mutant WASp lacking the VCA domain (Δ430–502, WASp ΔVCA) with actin or WIP to those of wild-type WASp (Fig. 5A and fig. S3, B and C). As expected, less actin precipitated with WASp ΔVCA than with wild-type WASp under stimulatory conditions (Fig. 5A), although actin may bind to other proteins that are present in the WASp complex. Less actin coimmunoprecipitated with wild-type WASp in unstimulated cells than with wild-type WASp in cells stimulated with anti-CD3 antibody (Fig. 5A). This is probably a result of the folded conformation of nonactivated WASp, which makes its VCA domain less accessible to actin. We also observed a substantial reduction in the amount of precipitated WIP–WASp ΔVCA compared to that of WIP–WASp wild type in both stimulated and unstimulated cells (Fig. 5A), suggesting that the VCA domain is involved in the interaction of WASp with WIP.

Fig. 5 The VCA domain of WASp is crucial for its interaction with the WH2 domains of WIP.

(A) Jurkat E6.1 cells were transfected with plasmids encoding CFP–WASp WT or CFP–WASp ΔVCA. Forty-eight hours later, cells were left unstimulated or were stimulated with anti-CD3 antibody before being subjected to immunoprecipitation with an anti–green fluorescent protein (GFP) antibody. Samples were then analyzed by Western blotting with the indicated antibodies. Western blots are representative of three independent experiments. (B) Jurkat E6.1 cells expressing YFP-WIP and either CFP–WASp WT or CFP–WASP ΔVCA were plated over activating coverslips and were stimulated for 2 min (n = 20 cells for WASp WT; n = 44 cells for WASp ΔVCA) or 5 min (n = 31 cells for WASp WT; n = 40 cells for WASp ΔVCA), and FRET efficiency was determined as described earlier. (C) Summary of FRET analysis between different forms of WIP and WASp. FRET efficiency was measured between YFP-WIP and CFP–WASp ΔVCA after 2 min of activation and compared to that between YFP-WIP and CFP–WASp WT. Cells expressing CFP-WASp were transfected with plasmid encoding YFP–WIP Δ1–112 or YFP–WIP Δ451–485. Cells were plated on a stimulatory coverslip for the indicated times and then fixed. FRET efficiency between CFP-WASp and YFP–WIP Δ1–112 (n = 34 cells for 2 min, and n = 37 cells for 5 min) was compared to that between CFP-WASp and YFP–WIP WT. FRET efficiency between CFP-WASp and YFP–WIP Δ451–485 (n = 44 cells for 2 min, and n = 40 cells for 5 min) was compared to that between CFP-WASp and YFP–WIP WT (n = 20 cells for 2 min, and n = 31 cells for 5 min). Results of three independent experiments are summarized.

FRET measurements confirmed the biochemical analysis, as well as demonstrated a substantial reduction in the binding between YFP-WIP and CFP–WASp ΔVCA (FRET efficiency of 3.6 ± 1.7% after 2 min of activation) compared to that between YFP-WIP and CFP–WASp wild type in unstimulated cells (33.6 ± 6.5%, P ≤ 0.0001; Fig. 5, B and C). The interaction between WASp and a YFP-tagged WIP mutant from which the C-terminal amino acid residues 451 to 485, which contain the binding site for WASp (33), were deleted (YFP–WIP Δ451–485) was also examined by FRET. We detected lower FRET efficiency between CFP-WASp and YFP–WIP Δ451–485 than between CFP-WASp and YFP–WIP wild type (FRET efficiency of 7.9 ± 3.6% after 2 min, P ≤ 0.002; 6.2 ± 3.0% after 5 min for WIP Δ451–485) (Fig. 5C). This reduced FRET efficiency was comparable with the reduction in FRET efficiency between YFP-WIP and CFP–WASp ΔVCA (P > 0.36, Fig. 5C), suggesting that the WASp VCA domain and the WIP C-terminus make similar contributions to the WIP-WASp interaction. However, even in the cells that coexpressed YFP–WIP Δ451–485 and CFP-WASp, FRET efficiency between WIP and WASp was decreased but not completely abolished, which is suggestive of a residual interaction between WIP and WASp. Western blotting analysis showed that the amount of WASp ΔVCA protein was less than that of wild-type WASp in stimulated cells (fig. S3C). The low abundance of WASp ΔVCA in stimulated cells, together with the substantial decrease in the extent of the interaction between WIP and WASp, suggests that VCA-deficient WASp protein was less protected by WIP than was wild-type WASp, and thus was more susceptible to degradation. Thus, these results suggest the VCA domain as a second potential WIP-interacting domain of WASp, an interaction that stabilizes WASp and might be mediated by actin.

The N terminus of WIP contains two actin-binding WASp-homology-2 (WH2) domains (amino acid residues 32 to 59 and 96 to 112), which contain the putative actin-binding motifs KLKK and KLRS, respectively (6, 7). We next generated cell lines expressing YFP-tagged fusion proteins of the WIP deletion mutants WIP Δ1–61, WIP Δ61–112, and WIP Δ1–112 (fig. S3, D to F), and we performed immunoprecipitation analysis to compare the relative extents of their association with WASp. Only removal of both WH2 domains of WIP reduced the extent of the binding of WIP to WASp (fig. S3, G to I), whereas deletion of either WH2 domain had no effect on the stability of the WIP-WASp complex (fig. S3, G and H). The effect of deletion of both WH2 domains (in the WIP Δ1–112 mutant) was echoed in an analysis of FRET efficiencies between CFP-WASp and the deletion mutants YFP–WIP Δ1–61, YFP–WIP Δ61–112, and YFP–WIP Δ1–112. Only the YFP–WIP Δ1–112 mutant demonstrated a statistically significant reduction in FRET efficiency (5.8 ± 2.8%) compared to that of wild-type WIP (33.6 ± 6.5%, P ≤ 0.0008; Fig. 5C and fig. S3J). The finding that both actin-binding domains in the WIP N-terminus needed to be removed to abolish its interaction with WASp suggests that actin stabilizes the WIP-WASp interaction by binding to both WH2 domains of WIP.

We further examined the potential contribution of actin to the WIP-WASp complex by pretreating cells coexpressing CFP-WASp and YFP-WIP with the actin-depolymerizing agent latrunculin A. FRET analysis showed a marked decrease in the binding between YFP-WIP and CFP-WASp, after 2 min of activation, in latrunculin A–treated cells compared to that in control cells (FRET efficiencies of 8.8 ± 3.1% and 39.09 ± 1.6%, respectively; P ≤ 0.00001, fig. S4A). A similar FRET analysis was performed to examine whether this change in the interaction between WIP and WASp resulted from the reduced direct binding of actin to either WIP or WASp. Accordingly, treatment of cells coexpressing CFP-WASp and YFP-actin with latrunculin A resulted in a significant decrease in the extent of binding between WASp and actin compared to that in untreated cells (FRET efficiencies of 8.3 ± 0.9% and 43.3 ± 3.6%, respectively; P ≤ 0.00001, fig. S4B, left panel). The same effect was obtained by treating cells coexpressing YFP-WIP and CFP-actin with latrunculin A, which resulted in a substantial reduction in the interaction between WIP and actin compared to that in control cells (FRET efficiencies of 5.2 ± 2.9% and 35.9 ± 3.2%, respectively; P ≤ 0.00001, fig. S4B, right panel). Thus, reducing the concentration of actin monomers with latrunculin A decreased the extent of binding of actin to both WASp and WIP, thereby reducing the amount of WIP that interacted with WASp.

The involvement of actin in the WIP-WASp interaction was further confirmed by specifically blocking the interaction between WIP and WASp in experiments with mutants that abrogate actin binding to the WH2 domains of WIP (K47,48E and R112E,S113E). Analysis of FRET between CFP-WASp and YFP-tagged WIP proteins with mutations at either one of the actin-binding motifs demonstrated a moderate decrease in the FRET efficiency (K47,48E: 21.6 ± 5.3%, P ≥ 0.185; R112E,S113E: 22.2 ± 4.1%, P ≥ 0.16; fig. S4C), compared to that between CFP-WASp and YFP–WIP wild type (33.6 ± 6.5%). Point mutations at both actin-binding motifs further reduced the interaction of WASp with WIP, as indicated by the marked reduction in FRET efficiency (14.2 ± 4%, P ≤ 0.01; fig. S4C). Thus, our data suggest that the inability of WIP to bind to actin had a negative effect on the interaction between WIP and WASp (fig. S4C).

To confirm that actin mediated the interaction between WASp and WIP, we specifically measured the FRET efficiency between CFP-tagged actin and YFP-tagged WIP proteins mutated at their actin-binding motifs. These findings agreed with those from earlier experiments in which the WH2 domains were deleted. Mutations at single actin-binding motifs were comparable to deletions of single WH2 domains and had only a mild effect on the WIP-WASp interaction, as determined by comparison of the FRET efficiencies of the mutant WIP proteins (K47,48E: 14.7 ± 4.4%, P ≤ 0.0003; R112E,S113E: 19.4 ± 5.5%, P ≤ 0.008; double mutant: 9.66 ± 2.8%, P ≤ 0.00001; fig. S4D) with that of wild-type WIP (35.2 ± 3.2%). Together, these data suggest that actin mediates the WIP-WASp interaction, possibly by binding to the WH2 domains of WIP and the VCA domain of WASp.

3FRET enables monitoring of the conformational changes in WIP during TCR activation

The data described so far suggest that (i) the WIP-WASp interaction is modified after stimulation of the TCR; (ii) the phosphorylation of WIP on Ser488 induces its partial dissociation from WASp and reduces WASp stability, which is required for actin polymerization; and (iii) WIP and WASp physically interact with each other through two distinct binding sites. The first site is the region at which the N terminus of WASp and the C terminus of WIP interact, whereas the second site is where the N-terminal WH2 domains of WIP and the C-terminal VCA domain of WASp interact, and this interaction is mediated by actin. However, can these findings be combined to describe the inhibitory role of WIP? To address this question, we used 3FRET to monitor and compare WIP-WASp complex assembly in resting and activated T cells, by focusing on the relative positioning of the two interaction sites. T cells stably expressing CFP-WASp (by N-terminal tagging) were transfected with plasmid encoding a doubly tagged WIP, with YFP and mCherry tethered to its N and C termini, respectively (YFP-WIP-mCherry). Thus, we could monitor the three-way complex of WASp with the two ends of WIP, which enabled simultaneous quantification of changes in the conformation of both WIP and the WIP-WASp complex after TCR activation (Fig. 6A and fig. S5A). Thus, analysis of the CFP-YFP and CFP-mCherry FRET efficiencies reported on the intermolecular contacts in the WIP-WASp complex, whereas analysis of the YFP-mCherry FRET efficiency reported on intramolecular proximities within WIP.

Fig. 6 Release of WIP autoinhibition is followed by a conformational change in the WIP-WASp complex.

(A) Schematic representation of the tagged WIP and WASp proteins used for 3FRET analysis. (B and C) Jurkat E6.1 cells expressing the indicated tagged proteins were plated over (B) a nonstimulatory coverslip coated with anti-CD43 antibody for 2 min (n = 50 cells) or 5 min (n = 45 cells) or (C) a stimulatory coverslip coated with anti-CD3 antibody for 2 min (n = 47 cells) or 5 min (n = 54 cells). After the activated cells had spread, they were fixed and FRET efficiency was determined as described earlier. FRET was detected between N-terminal tagged WASp and N-terminal or C-terminal tagged WIP, as well as between the two molecular ends of WIP. (D) Graph summarizing the percentage FRET efficiency in all three indicated channels in cells plated over activating (CD3) or nonactivating (CD43) coverslips for the indicated times. Data are means ± SEM from at least three independent experiments (see table S1 for details). (E) In separate experiments, Jurkat E6.1 cells were transfected with plasmids encoding YFP-WIP-mCherry and WASp-CFP (C-terminal tag). Cells were seeded over a stimulatory coverslip coated with anti-CD3 antibody and fixed after 2 min (n = 38 cells) or 5 min (n = 37 cells) of activation. FRET efficiencies in the three indicated channels were determined. Graph summarizing the percentage FRET efficiency in cells plated over activating coverslips for the indicated times. Data are means ± SEM from at least three independent experiments (see table S2 for details).

In unstimulated T cells, we detected high FRET efficiencies at 2 and 5 min after contact with the plates coated with anti-CD43 antibody, both for the intramolecular interaction between the N terminus and the C terminus of WIP, and for the intermolecular interaction between WIP and WASp (Fig. 6B and table S1). In stimulated cells, however, we measured a statistically significant reduction in FRET efficiencies for all energy transfers after 5 min of activation (Fig. 6C and table S1). These data suggest the possible partial dissociation of the WIP-WASp complex because the CFP-YFP and CFP-mCherry FRET efficiencies were reduced 5 min after stimulation (19.8 ± 3.3% and 23.7 ± 4.1%, respectively) in comparison to those measured 2 min after activation (40.8 ± 2.4%, P ≤ 0.0001, and 51.6 ± 1.0%, P ≤ 0.0001, respectively) or to those in unstimulated cells (42.1 ± 1.6%, P ≤ 0.0006, and 56.1 ± 1.3%, P ≤ 0.0001, respectively). These data agree with our earlier findings from experiments with the standard double-FRET system. The reduction in the YFP-mCherry FRET efficiency (corresponding to WIP intramolecular interactions) between 2 and 5 min after stimulation (24.1 ± 3.7% and 8.9 ± 2.6%, respectively; P ≤ 0.0017, Fig. 6D and table S1) suggested a structural alteration within WIP over the course of T cell activation.

To further elucidate the conformational change in the WIP-WASp complex, we generated a fusion protein of WASp with a C-terminal CFP tag (WASp-CFP), which was then expressed together with YFP-WIP-mCherry in Jurkat cells for 3FRET experiments (Fig. 6E, fig. S5B, and table S2). The YFP-mCherry and CFP-mCherry FRET efficiencies exhibited the expected reduction to 7.1 ± 1.9% and 20.6 ± 7.3% after 5 min of activation, which was in comparison to the FRET efficiencies of 17.2 ± 1.9% and 47.4 ± 2.5%, respectively, that were measured 2 min after activation (P ≤ 0.0015 and P ≤ 0.0008, respectively; table S2). However, the CFP-YFP FRET efficiency, which reports on the interaction between the VCA and WIP-N domains, remained high throughout the course of activation (19.2 ± 4.7% and 19.3 ± 4.3% after 2 and 5 min, respectively; P ≥ 0.9796, Fig. 6E and table S2). This contrast between proteins, which was distinguished only by the location of the fluorescent tag, highlights the difference between the two interaction sites of WIP and WASp. These data suggest that after cellular activation, the phosphorylation-modulated interaction site (site I) involving the C terminus of WIP and the WH1 domain of WASp undergoes dissociation, whereas the phosphorylation-independent interaction site (site II) involving the N-terminal region of WIP and the VCA domain of WASp, which is also bound by actin, remains intact, and that it is the latter interaction that is responsible for the observed constitutive association between both proteins (Fig. 7).

Fig. 7 Suggested molecular model of WIP autoinhibition and the conformational changes that occur in WIP-WASp after TCR activation.

In naïve T cells, WIP and WASp associate through two distinct molecular interfaces: a phosphorylation-dependent site involving the WASp-WH1 domain and the WIP C-terminal domain (site I), and a site involving the WASp-VCA domain and the WIP-N-terminal domain, which are also bound by actin (site II). TCR activation leads to the phosphorylation of WIP on Ser488 by PKCθ, which stimulates the partial dissociation of the WIP C-terminus from WASp-WH1 domains (site I) while leaving the other binding site intact (site II). WASp was tagged at either the N terminus or C terminus, as indicated by the dashed circles.

DISCUSSION

Correct formation of the WIP-WASp complex is crucial for the stability of the actin regulator WASp, for actin-dependent processes, and for the overall immune response (23, 41). Indeed, mutations that impede the formation of this complex are manifested as the symptoms characteristic of the immunodeficiencies WAS and XLT (42). Furthermore, WIP is not exclusively found in immune cells, and it is also a key player in actin-dependent cellular processes, such as migration, adhesion, spreading, and invasion in nonhematopoietic cells (69); thus, regulation of the WASp-WIP interaction has implications beyond the immune response. The molecular nature of the WIP-WASp interaction and the manner by which WIP shields WASp from degradation remain unclear (2831). Specifically, the phenomenon of activation-induced dissociation of the WIP-WASp complex has not been satisfactorily explained.

The process of PKCθ-mediated phosphorylation is involved in T cell activation (31, 43). Our findings establish a correlation between the phosphorylation of WIP and the partial dissociation of the WIP-WASp complex. This was shown by monitoring FRET efficiencies between WASp and mutants of WIP that mimicked its phosphorylated and nonphosphorylated states. In addition, knockdown of PKCθ resulted in a persistent interaction between WIP and WASp. Using 3FRET to simultaneously monitor the interplay between WIP, WASp, and PKCθ, we demonstrated that PKCθ physically interacted with WIP only transiently, with a time dependence similar to that of the interaction between WIP and WASp. Colocalization of WASp and PKCθ appeared to persist during cellular activation, which was surprising given that these proteins exert opposing effects on the stability of the immunological synapse: PKCθ disrupts the symmetry of the immunological synapse and facilitates T cell motility, whereas WASp helps to reestablish the symmetric immunological synapse, resulting in synapse stabilization and T cell arrest (44). It is possible that in T cells, the constitutive interaction between WASp and PKCθ may facilitate a local and rapid transition between a destabilized and stabilized immunological synapse; however, the mechanisms by which WASp and PKCθ exert their effects on the stability of the immunological synapse are not known. In natural killer cells, WASp antagonizes the PKCθ-mediated formation of a complex consisting of WIP, actin, and myosin IIA (30). Because myosin IIA contributes to the generation of contractile forces that destabilize the immunological synapse in T cells (45), WASp might promote T cell arrest by perturbing the WIP–myosin IIA complex. Therefore, the interaction between WASp and PKCθ merits further investigation.

We also identified the functional consequences of the phosphorylation of Ser488 of WIP. Activation-induced recruitment of WIP to TCR clusters, the translocation of WIP to actin-rich sites, and cellular spreading were all influenced by the phosphorylation state of Ser488. Live-cell imaging analysis showed that phosphorylated WIP, which was weakly associated with WASp, moved and dissipated rapidly upon TCR stimulation, whereas nonphosphorylated mutant forms of WIP bound strongly to WASp and exhibited poor clustering, with slower cluster movement and increased lifetimes of visible clusters. The behavior of clusters of wild-type WIP resembled that of clusters of the constitutively phosphorylated mutant WIP S488D immediately after activation, suggesting a mechanism of transient activation by phosphorylation. Phosphorylation was also required for the interaction between WIP and actin; consequently, actin rearrangement was impaired in cells expressing nonphosphorylated WIP, which suggests that WIP phosphorylation is required for appropriate cytoskeletal behavior in these cells.

It has been suggested that WASp undergoes cycles of phosphorylation and dephosphorylation (46), and that these cycles may be essential for the WASp-dependent chemotaxis of macrophages and dendritic cells (47, 48). Similarly, WIP-dependent actin remodeling may also be regulated by cycles of phosphorylation and dephosphorylation, although these mechanisms require further investigation. Thus, our results indicate the crucial role of WIP phosphorylation in controlling the cellular dynamics of WIP and of homeostatic cytoskeletal behavior. Furthermore, we found that phosphorylation of WIP, which precedes the phosphorylation of WASp (24), was a prerequisite for the ubiquitylation and subsequent degradation of WASp, and that nonphosphorylated WIP efficiently protected WASp from degradation. Several studies found N-WASp to be either protected or inhibited by WIP and showed that N-WASp was released from inhibition by various activators, including Nck, Toca-1, Cdc42, and SNX9 (4951); however, the mechanistic details of these modes of activation are unclear. Thus, our results suggest a possible mechanism by which the interaction between WIP and N-WASp may also be affected by phosphorylation of WIP at Ser488; however, this mode of activation requires further investigation.

In light of evidence of activation-induced partial dissociation of the WIP-WASp complex, it was surprising that our immunoprecipitation experiments did not reflect this dissociation event. These biochemical data echoed findings from a previous study that suggested that a WIP-WASp complex was unaffected by PKCθ-dependent phosphorylation (28), highlighting the potential limitations of relying solely on biochemical methods to study protein-protein interactions. The contradictory results of immunoprecipitation and live-cell imaging methods led us to demonstrate the presence of a previously uncharacterized second (and phosphorylation-independent) interaction between the two proteins, in addition to the well-established binding interface involving the WASp WH1 domain and the WIP C-terminal region. Deletion of either the VCA region of WASp or the N-terminal 112 residues of WIP resulted in dissociation of the WIP-WASp complex upon T cell activation. A VCA-deficient WASp mutant underwent enhanced degradation after activation. These results establish a role for the VCA domain of WASp in maintaining the stability of WASp through its interaction with WIP, which protects it from degradation.

On the basis of our findings, we suggest that the binding of WASp by WIP is mediated by two distinct interaction surfaces, the first involving the N-terminal WH1 domain of WASp and the C-terminal region of WIP (site I) (32), and the second involving the C-terminal VCA domain of WASp and the N-terminal region of WIP, both of which are bound by actin (site II). The presence of the second interaction site is consistent with the observation that a WIP-derived 41-mer peptide spanning the canonical WASp-binding domain exhibits lower affinity for WASp than does full-length WIP (52). The interaction through site I involves direct contact between the two proteins, which is dependent on the PKCθ-mediated phosphorylation of Ser488 in the recognition sequence (residues 485 to 490) located in the C-terminal region of WIP. Dissociation of the proteins at this interface is a defining event in T cell activation, which leads to modified WIP dynamics, actin rearrangement, and, eventually, the proteasomal degradation of WASp. The interaction at site II was abolished only by deletion of the entire tandem actin-binding domain of WIP, which suggests that actin may contribute to the interaction between the two proteins at this site. The use of latrunculin A and, more specifically, point mutations that abrogate the binding of actin to WIP or WASp, or both, suggest that actin plays a role in mediating WIP-WASp complex formation. These results suggest that the interaction between WIP and WASp, two actin-modulating proteins, is mediated and regulated by actin itself. The actin network was previously suggested to feed back to the WASp homolog N-WASp by binding to the WH2 domains of N-WASp and potentially enhancing its recruitment and clustering at the plasma membrane (53, 54). The discovery of an actin-mediated interaction between WIP and WASp further suggests actin as a putative regulator of actin polymerization by binding to WASp, WIP, or both.

Evidence supporting this model of the WIP-WASp complex was provided by 3FRET analysis (34) with tags placed on both termini of WIP and on either terminal domain of WASp. Whereas tagging WASp at the N terminus (or site I) resulted in a reduction in all three FRET efficiencies after T cell activation, tagging WASp at the C terminus (site II) resulted in persistent high FRET efficiency with the N terminus of WIP throughout cellular activation. This illustrates the fundamental difference between the two sites in terms of modulation by phosphorylation. Both WIP tags were in close proximity in the inactivated state. We conclude that before cellular activation, WIP adopts a compact, autoinhibitory conformation, which brings together both interaction sites, and that activation releases the interaction between WIP and WASp at site I while preserving the interaction at site II.

In summary, our live-cell imaging approach provided insights into the mechanism by which WIP binds to and exerts its dual effect on WASp. Furthermore, the data presented here provide evidence for the regulated nature of the WIP-WASp interaction that controls WASp stability and actin homeostasis throughout TCR activation. The importance of WASp in the regulation of T cell function is emphasized by the fact that constitutive degradation of WASp and interference with reorganization of the actin cytoskeleton result in an impaired immune response, as manifested by the immunodeficiencies WAS and XLT (24). On the other hand, WASp homologs are primary players in the control of cell motility, adhesion, and invasion, and are increased in abundance in various nonhematologic cancers (5, 55). Thus, our findings about the nature of the WIP-WASp interaction are relevant for ongoing efforts to understand, and possibly control, lymphocyte responses.

MATERIALS AND METHODS

Antibodies and reagents

Antibodies and reagents were obtained from the indicated suppliers. The following antibodies were used for imaging: mouse anti-CD3ε (UCHT or HIT3a), mouse anti-CD28, and mouse anti-CD43 (BD Biosciences). The following primary antibodies were used for immunoprecipitations and Western blotting: mouse anti-GFP (Roche), mouse anti-WASp D1 (Santa Cruz Biotechnology), rabbit anti-WIP (Santa Cruz Biotechnology), rabbit anti-PKCθ (Epitomics), and mouse anti–glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (Biodesign). The horseradish peroxidase–conjugated secondary antibodies used included goat anti-mouse (Sigma-Aldrich) and goat anti-rabbit (Santa Cruz Biotechnology). Pools of three independent RNA duplexes specific for human PKCθ were obtained from Invitrogen and have the following sequences: GAGUCUCCGUUGGAUGAGGUGGAUA, GCAUCCGUUUCUGACGCACAUGUUU, and CCGGCCGAAAGUGAAAUCACCAUUU. Pools of nontargeting (nonspecific) negative control siRNA duplexes were obtained from Dharmacon and have the following sequences: UAGCGACUAAACACAUCAA, UAAGGCUAUGAAGAGAUAC, AUGUAUUGGCCUGUAUUAG, AUGAACGUGAAUUGCUCAA, and UGGUUUACAUGUCGACUAA.

Expression vectors and plasmid construction

The expression vectors pEYFP-C1, pEYFP-N1, pECFP-C1, and pECFP-N1 were obtained from Clontech, and pcDNA3.1+/Hygro was obtained from Invitrogen. Complementary DNA (cDNA) encoding human WASp was provided by D. Nelson (National Cancer Institute, National Institutes of Health). Plasmid encoding FLAG-tagged WIP was purchased from Addgene. cDNAs encoding WIP or WASp were cloned into the expression vectors pECFP or pEYFP to obtain CFP- or YFP-tagged proteins. Aequorea GFP derivatives were rendered monomeric by the A206K substitution, as was previously described (56). The cDNA encoding FLAG-tagged WIP was cloned into pEYFP-C1 with restriction enzymes Hind III and Bgl II. The YFP-WIP-mCherry construct was generated by insertion of the cDNA encoding YFP-WIP into the mCherry plasmid with the restriction enzymes Hind III and Nhe I. WIP point mutations (R485,490K, S488D, K47,48E, and R112E,S113E) and WIP deletion mutants (WIP Δ1–61, WIP Δ61–112, WIP Δ1–112, and WIP Δ451–485) were generated with the QuikChange II XL Site-Directed Mutagenesis Kit (Stratagene) according to the manufacturer’s instructions, and the mutated sequences were introduced into the YFP-WIP construct. WIP deletion mutations were performed with the following primer pairs: WIP Δ1–112 forward: ACAAGTCCGGACTCAGATCTACGGCCAACAGG, WIP Δ1–112 reverse: CCTGTTGGCCGTAGATCTGAGTCCGGACTTGT; WIP Δ1–61 forward: GTCCGGACTCAGATCTAAAGGAGCTGGTGCTG, WIP Δ1–61 reverse: CAGCACCAGCTCCTTTAGATCTGAGTCCGGAC; WIP Δ61–112 forward: GAAGTGCACCAATACTGGACAAAACGGCCAACAGG, WIP Δ61–112 reverse: CCTGTTGGCCGTTTGTCCAGTATTGGTGCACTTC; WIP Δ451–485 forward: CATGTGAAGATGAGTGGAGTGGATCCAACCGAAG, WIP Δ451–485 reverse: CTTCGGTTGGATCCACTCCACTCATCTTCACATG. All constructs were verified by DNA sequencing. Deletion of the C-terminal VCA domain of WASp (ΔVCA, which is deficient in residues 430 to 502) was performed by polymerase chain reaction amplification of cDNA encoding WASp with the following primer pair: forward, TATAATCCGGAAGTGGGGGCCCAATGGGA, which includes a Bsp EI site, and reverse, AATATGAATTCTCAACCCCCACCAGGGGCCAG, which includes an Eco RI site. The amplification product was cloned into pECFP-C1, and the construct was verified by DNA sequencing.

Transfections and flow cytometric analysis

Cells were transfected with an Amaxa electroporator and appropriate Amaxa solutions. Both transiently transfected T cell cultures and stable clones were used in this study, as indicated in the figure legends. Transiently transfected cells were used 48 hours after transfection. Stable clones were derived from transiently transfected cells with a combination of drug selection. Cells were selected in either neomycin or hygromycin, which was followed by cell sorting. Fluorescence analysis and cell sorting were performed with a FACSAria flow cytometer (Becton Dickinson Biosciences) and FlowJo software.

Cell culture

Jurkat E6.1 cells were cultured as described previously (35, 57). Where indicated, Jurkat E6.1 cells were preincubated with 10 μM latrunculin A (Cayman Chemicals) for 10 min at 37°C before the cells were activated. Human primary PBLs were isolated from whole blood of healthy donors, as previously described (35). The cells were activated with anti-CD3ε antibody (OKT3, 10 μg/ml) and anti-CD28 antibody (10 μg/ml) for 30 min on ice. The cells were then warmed to 37°C for 10 min and stimulated with anti-mouse immunoglobulin G (50 μg/ml) for 2 min.

Immunoprecipitation and Western blotting

Immunoprecipitations and Western blotting analysis were performed as previously described (24). Densitometric analysis of band intensities was performed with ImageJ software, with final results normalized with GFP or GAPDH as loading controls for immunoprecipitated samples and whole-cell lysates, respectively. Relative protein abundance or relative extent of coimmunoprecipitated proteins was compared with the relevant controls.

Cellular imaging by confocal microscopy

Dynamic fluorescent and interference reflection microscopy images were collected on a Zeiss 510 Meta confocal microscope. Image stacks were collected over time with an autofocusing algorithm based on reflection images obtained by imaging the plane of contact between the T cell and the coverslip. All images were collected with a 63× Plan-Apochromat objective lens (Carl Zeiss). For live-cell imaging, a hot-air blower (Nevetec) was used to maintain the sample at 37°C. Fine adjustments were made with a digital temperature probe to monitor the buffer temperature in the chamber. Image processing and measurements were performed with IP Lab software, version 3.9.

Spreading assay

Spreading assays were performed as previously described (35). Briefly, T cells (2 × 106 cells/ml) were seeded on the bottom of chambered cover glasses (Lab-Tek) that were precoated with anti-CD3 stimulatory monoclonal antibodies (10 μg/ml). The cells were incubated in imaging buffer (RPMI without phenol red and containing 10% fetal calf serum and 25 mM Hepes) at 37°C, 5% CO2, for the times indicated in the figure legends. Cells were fixed for 25 min with 2.5% paraformaldehyde in phosphate-buffered saline (PBS) and then were washed three times with PBS.

Image processing and quantitation

The acquired images were extracted with the LSM browser (Carl Zeiss), cropped, and composed into figures within Adobe Photoshop. Movies were prepared from z-stacks by preparing a maximum intensity projection of a given time point and then arranging a sequence of all of the projections. To quantify cluster movement data in the obtained images and to generate averaged tracks, we used algorithms based on previously published methods (58, 59) and specifically optimized in our system. Kymographs were made by drawing regions of interest around moving clusters, and cluster traces were prepared with the kymographs and analyzed with IP Lab software version 3.9.

Double-color and triple-color FRET systems

Double-color FRET was performed as described previously (35, 39, 60). Here, we used CFP (excitation wavelength: 468 nm; emission filter wavelength: 465 to 510 nm) as a donor and YFP [excitation wavelength: 514 nm; emission filter wavelength: 530 nm long-pass (LP)] as an acceptor. The 3FRET technique developed in our laboratory was applied as described previously (34). Three fluorophore-exciting lasers were used. For mCFP excitation, we used a laser with a wavelength of 458 nm; for mYFP, we used a laser with a wavelength of 514 nm; and for mCherry, we used a laser with a wavelength of 594 nm. Emission was monitored with a filter at a range of 465 to 510 nm for mCFP, 530 to 600 nm for mYFP, and 615 nm LP for mCherry. Three channels were used to gather images for each FRET pair. One channel was optimized for donor fluorescence (donor excitation, donor emission image), a second channel was optimized for acceptor fluorescence (acceptor excitation, acceptor emission image), and a third channel was optimized for FRET (donor excitation, acceptor emission image). The third image provided raw, uncorrected FRET data, which contained two non-FRET components: the “bleed-through” of the donor emission into the acceptor detection channel, and the cross-excitation of the acceptor by the donor-excitation laser. Energy transfer in the 3FRET system occurs because of overlapping spectral areas and consists of three FRET pairs: mCFP-mYFP (excitation at 458 nm, detection at 530 to 600 nm), mCFP-mCherry (excitation at 458 nm, detection at 615 nm LP), and mYFP-mCherry (excitation at 514 nm, detection at 615 nm LP), with the first fluorophore serving as the donor and the latter serving as the acceptor. The total energy transfer between mCFP and mCherry consists of both direct FRET and indirect transfer through mYFP. Each cell was imaged with six different channel settings: mCFP, mYFP, mCherry, mCFP-mYFP FRET, mCFP-mCherry FRET, and mYFP-mCherry FRET. Three images were used to calculate each FRET pair (mCFP-mYFP, mCFP-mCherry, and mYFP-mCherry): the donor channel, the acceptor channel, and the FRET channel.

FRET correction

The non-FRET components were calculated and removed with calibration curves derived from images of single-labeled cells containing CFP, YFP, or mCherry, as previously described (34). Sets of reference images were obtained with the same acquisition parameters as those used for the FRET experimental images. Bleed-through components were calculated as a function of the intensity of the expressed fluorescent proteins with data gathered from single-labeled cell lines. By using the intensity measured pixel by pixel through the different filter sets, as well as cross-talk elements that were isolated from the control, single-labeled cells, we corrected the measured FRET and determined the actual FRET efficiency at every pixel.

FRET efficiency calculation

The relative FRET efficiency (FRETeff) was calculated on a pixel-by-pixel basis with the following equation (60):FRETeff=FRETcorr/(FRETcorr+donor)×100%where FRETcorr is the pixel intensity in the corrected FRET image, and donor is the intensity of the corresponding pixel in the appropriate donor channel image. To increase the reliability of the calculations and to prevent low-level noise from distorting the calculated ratio, we excluded pixels below 50 intensity units, as well as saturated pixels from the calculations, and set their intensities to zero. These pixels are shown in black in the pseudo-colored FRET efficiency images. To estimate the reliability of the obtained FRET efficiency values and to exclude the possibility of obtaining false-positive FRET, we prepared cells expressing free CFP, free YFP, and free mCherry as negative controls. The FRET efficiency in the negative control system was measured and calculated in the same manner as that in the main experiment. FRET efficiency values obtained from the negative control samples were subtracted from the values obtained in the main experiments.

Actin shape index

A quantitative estimate of the shape changes in CFP-actin was obtained as previously described (39).

Statistical analysis

Standard errors were calculated with Microsoft Excel. Student’s t test was used for unpaired samples.

SUPPLEMENTARY MATERIALS

www.sciencesignaling.org/cgi/content/full/7/331/ra60/DC1

Fig. S1. Analysis of protein abundance and interactions between wild-type and mutant forms of WIP and WASp and actin.

Fig. S2. Phosphorylation of WIP by PKCθ controls WASp ubiquitylation and stability.

Fig. S3. The N terminus of WIP and the VCA domain of WASp are required for the WIP-WASp interaction.

Fig. S4. Involvement of actin in the WIP-WASp interaction.

Fig. S5. Analysis of tagged protein abundance in cells used for 3FRET.

Table S1. FRET efficiency and statistical analysis of the interaction of WASp tagged at its N terminus with the N terminus or C terminus of WIP, as well as intramolecular FRET between the two molecular ends of WIP.

Table S2. FRET efficiency and statistical analysis of the interaction of WASp tagged at its C terminus with the N terminus or C terminus of WIP, as well as intramolecular FRET between the two molecular ends of WIP.

Movie S1. Dynamic movement of CFP-WASp and YFP–WIP wild type in live activated T cells.

Movie S2. Dynamic movement of YFP–WIP wild type in live activated T cells.

Movie S3. Dynamic movement of YFP–WIP R485,490K in live activated T cells.

Movie S4. Dynamic movement of YFP–WIP S488D in live activated T cells.

Movie S5. Dynamic movement of CFP-actin and YFP–WIP wild type in live activated T cells.

REFERENCES AND NOTES

Acknowledgments: We thank N. Joseph for her technical assistance. Funding: This research was funded by the Israeli Ministry of Health through the Office of the Chief Scientist, the Israel Cancer Association through the Estate of the late Alexander Smidoda, and by the Israel Science Foundation Legacy Heritage Fund (grant no. 491/10). Author contributions: M.B.-S. designed the research; S.F., B.R., M.H.P., S.E., O.M., and E.N. performed the experiments; M.B.-S., S.F., B.R., M.H.P., E.N., and J.C. analyzed the data; and M.B.-S. wrote the paper. Competing interests: The authors declare that they have no competing interests.
View Abstract

Stay Connected to Science Signaling

Navigate This Article