Reduction of cortactin expression by siRNA or over-expression of its isolated SH3 domain, polyproline region or its α-helical region resulted in a drastic decrease in actin-pedestal formation during infection with EPEC 23. However the role of cortactin's Arp2/3 binding and activating region has not been addressed 23. Therefore, we investigated its contribution to actin assembly on pedestals using EPEC to infect HeLa cells transiently transfected with GFP-cortactin. Pedestals were visualized by immunofluorescent staining of actin using fluorescent phalloidin and bacteria with DAPI. As previously reported 23, no differences on the number of attached bacteria were observed for the transfectants used (data not shown).

The cortactin NTA domain carries a 20DDW22 motif that binds and activates the Arp2/3 complex. Mutation of this motif to 20DDA22, hereafter referred to as W22A, abolished this activity 11. To determine whether this motif is necessary for pedestal formation we transfected HeLa cells with GFP-W22A. We used wild-type (WT) cortactin (GFP-FL) and GFP alone as controls. As shown in Fig. 1, over-expression of GFP-FL cortactin allowed pedestal formation to levels similar to those in cells expressing GFP. Fig. 1C (black bars) shows normalized percentages and standard deviations for GFP-FL. Results of three independent experiments were considered statistically significant (p < 0.01 by Student's t-test). Since the constructs bear a GFP-tag we were able to simultaneously assess the localization of different cortactin forms (Fig. 1A, B and 1C). We observed GFP-FL cortactin to localize in 70% of pedestals, compared to 4% for GFP-transfected cells (open bars). Importantly, the number of pedestals in cells expressing GFP-W22A mutant was significantly lower than in GFP-FL transfected cells (51% vs 83%). This result indicates that cortactin W22A exerts a dominant negative effect, which may mean that cortactin binding and activation of the Arp2/3 complex is necessary for pedestal formation.

Cortactin has a C-terminal SH3 domain that binds several proteins. Mutation of a critical amino acid (W525K) abolishes its binding to known targets 29 such as N-WASP 14. We used this mutant to assess the contribution of the cortactin SH3 domain to pedestal formation; we found that its expression inhibits pedestal formation to an even greater extent than the W22A mutant (31% vs 51%). This indicates that cortactin W525K mutant exerts a dominant-negative effect, corroborating previous results 23. In previous work, we described that the cortactin SH3 domain is able to activate N-WASP and we proposed a model for the regulation of N-WASP activation by cortactin, in which cortactin is switched on by Erk phosphorylation of serines 405 and 418, while it is switched off by Src phosphorylation of tyrosines 421, 466 and 482 14. Next we repeated the pedestal formation assay with cells expressing the cortactin S405,418D double mutant, which mimics Erk phosphorylation and activates N-WASP in vitro, as well as its non-phosphorylatable counterpart (S405,418A). The S405,418D mutant allowed pedestal formation to a similar extent as the WT cortactin (90%) and to a greater extent, although not significantly greater, than the GFP negative control (83%). The phosphoserine-mimicking cortactin mutant accumulated in only 21% of pedestals and showed a weak, diffuse pattern of localization in the cytoplasm and pronounced staining in the nucleus. In contrast, the mutant that abolished Erk phosphorylation (S405,418A) impaired pedestal formation (34%) and its own translocation to them (3%). These results suggest that Erk phosphorylation of cortactin contributes to pedestals formation.

Similarly, we wanted to address the role of Src-mediated phosphorylation of cortactin. We therefore used the phosphotyrosine-mimicking mutant (Y421,466,482D) and the phosphotyrosine deficient mutant (Y421,466,482F). In both cases pedestal formation and location of these constructs on them were impaired (33%/7%; 28%/14%, Fig. 1B and 1C). These results indicate that Src-mediated phoshorylation of cortactin seems to inhibit pedestal formation and that a dynamic phosphorylation of these tyrosine residues play a role in the formation of pedestals.

Although no appreciable changes in the cellular architecture were observed, we wanted to exclude the possibility that over-expression of cortactin mutants induces a general alteration of the actin cytoskeleton. We therefore used flow cytometry to assess the total basal F-actin content of the different transfectants. Fig. 1C (grey bars) and 1D show the mean fluorescence of GFP transfectants, which did not significantly differ based on Student's t-test. As a control, transfected cell were pretreated with Cytochalasin D, a drug known to inhibit actin polymerization (Fig. 1D, bottom right panel). In addition, this experiment allowed us to calculate the transfection efficiency, which was estimated as 60–70%, based on the analysis of the expression of GFP constructs (data not shown).

Contrary to the transient phosphorylation induced by EHEC, EPEC infection of CH7 mouse fibroblasts induces tyrosine phosphorylation of cortactin 26. Src has been shown to phosphorylate cortactin on tyrosines Y421, 466 and 482 30, which decreases cortactin affinity for N-WASP in vitro 14. In addition, N-WASP-deficient cells do not form pedestals 31. These observations prompted us to examine the phosphorylation status of cortactin in WT mouse embryonic fibroblasts (MEFs), MEFs deficient in N-WASP and MEFs deficient in N-WASP in which the protein was later restored through retroviral transduction (rescued, R). First, we performed Western blotting control experiments to assess the expression of N-WASP, cortactin and actin (Fig. 2A).

Fig. 2A shows that EPEC induces phosphorylation of tyrosine 466 at 3 hours of infection in WT MEFs, as detected using an antibody against phospho-Y466-cortactin. This result was corroborated using a second phospho-specific antibody (pY421, data not shown). Unexpectedly, phosphorylation of tyrosine residue 466 was not induced in N-WASP-deficient cells. This result suggests that tyrosine phosphorylation of cortactin during EPEC infection depends on the presence of N-WASP. To verify this, we infected R cells with EPEC and examined levels of phosphoY466-cortactin. Fig. 2A shows that N-WASP re-expression partially restored cortactin tyrosine phosphorylation levels. In three independent experiments the normalized average induction was 1 ± 0.2 for WT cells, 0 for N-WASP-deficient cells and 0.5 ± 0.1 for R cells. This supports the idea that EPEC-induced tyrosine phosphorylation of cortactin in cells requires N-WASP.

Given the absence of cortactin tyrosine phosphorylation in EPEC-infected N-WASP-deficient cells, we then checked Src activation, using a commercially available phospho-active Src antibody (pY416). Fig. 2B demonstrated that equal activation of Src was achieved during EPEC infection in all cell types studied, while, as expected, the levels of total Src remained constant during infection. This result showed that the lack of cortactin phosphorylation in N-WASP-deficient cells was not due to a block in Src activation. As a further control, we treated the cells with pervanadate and observed robust phosphorylation of cortactin tyrosine 466 (Fig. 2C, upper panel).

Similarly we sought to establish the activation status of Erk in EPEC-infected cells. We used a phospho-specific monoclonal antibody that detects the activated form of Erk1/2 (anti pThr202-pTyr204). EPEC induced the activation of Erk on WT MEFs (Fig. 2D first lane), in agreement with a previous report on T84 epithelial cells 32. However, infection of N-WASP-deficient cells showed reduced activation of Erk which was recovered in R cells. This result implies that Erk is activated by EPEC and may phosphorylate cortactin in vivo. More importantly, N-WASP is absolutely required for the induction of Erk activation at 3 hours of infection. However, WT MEFs treated with ERK inhibitors PD98056 or U0126 showed no difference in the number of pedestals formed (23 and data not shown).

The bacterial protein called Tir initiates what is considered to be the principal signaling cascade, which consists of Tir clustering and concomitant phosphorylation on its tyrosine 474, which then recruits Nck. The latter presumably binds N-WASP to initiate Arp2/3 complex-mediated actin polymerization 1. We wanted to gain insights into how cortactin functions in pedestal signaling. Our initial hypothesis was that cortactin and Tir interact directly. Therefore we used the Scansite database 33 to search for motifs in the Tir sequence to which cortactin SH3 domain could bind. We found a consensus motif (NNSIPPAPPLPSOTD) centered on proline 20 of Tir.

We first performed pull-down experiments with purified recombinant Tir 34 and cortactin proteins (Fig. 3). We produced WT GST-Tir that was purified using GSH beads and treated with PreScission enzyme, which excised Tir and at the same time removed the GST tag (Fig. 3, Coomassie gel). This Tir protein was used as the input in pull-down experiments with GST-cortactin. The first line of Fig. 3A shows that cortactin binds Tir in vitro.

To map the domains involved in the interaction, we performed pull-down experiments using cortactin mutants as follows: full-length W525K (mutated SH3 domain), the N-terminus (NH2, residues 1–333), and the isolated SH3 domain (SH3, residues 458–546). GST was used as a negative control. In agreement with our initial hypothesis, the isolated SH3 domain of cortactin bound Tir. However, the N-terminal domain of cortactin also bound Tir (Fig. 3A, third line). This unforeseen interaction was confirmed in experiments with cortactin carrying the point mutation, W525K in the SH3 domain (Fig. 3A). We obtained similar results using as input the Tir phospho-mimicking mutant TirY474D (TirD, data not shown). Next we tested the cortactin S405,418D (SD) and Y421,466,482D (3D) mutants which were similar to the WT form in their ability to bind both Tir (Fig. 3B) and TirD (data not shown). These results demonstrate that cortactin and Tir interact directly in vitro, that this interaction involves both the N-terminal part and the SH3 domain, and that it appears to be independent of cortactin phosphorylation.

Given the direct interaction between Tir and cortactin, we wondered whether Tir can activate the ability of cortactin to promote Arp2/3-mediated actin polymerization. We coupled recombinant Tir protein (or TirD, data not shown) to 1 μm beads, and then we washed the beads with Xb buffer and blocked them in Xb buffer containing 1% BSA. Next we incubated them with purified Arp and actin in Xb buffer containing WT and cortactin mutants. Fig. 3C shows that Tir activated WT cortactin and both SD and 3D mutants. Similar results were obtained for TirD (data not shown). The W525K mutant was also activated, although weakly. As expected, W22A cortactin was not activated, indicating that the effect was mediated by cortactin activation of the Arp2/3 complex. As a negative control we used naked beads that showed no activation. Conversely, experiments in which cortactin and its mutants were coupled to GSH beads showed similar results (data not shown). These results indicate that Tir activates the ability of cortactin to promote Arp2/3-mediated actin polymerization in vitro.

Because cortactin binds directly both Tir (Fig. 3) and N-WASP 14, we analyze cortactin-Tir interaction in N-WASP-deficient cells. Since those cells do not form pedestals 35, we wondered if Tir would be present at similar levels to WT cells. To address this question, we used a previously described fractionation protocol that enriches in Tir-containing membranes (36 and Materials and Methods). As shown in Fig. 4A, as expected Tir was enriched in the pellets compared to supernatants, as detected by western-blotting with anti-Tir mAb. We observed that a band with slower electrophoretic motility was the predominant form of Tir in the pellets, which represents fully-modified Tir 37. WT and N-WASP-deficient cells presented detectable amounts of mature Tir that was slightly reduced on R cells.

FL cortactin has a closed conformation 1420. Therefore, we decided to use N-terminal cortactin (NH₂), the SH3 domain (SH3) and GST as a negative control to perform pull-down experiments with lysates of EPEC infected and uninfected WT, N-WASP-deficient and R cells. Western blotting in Fig. 4B shows that NH₂ bound Tir in EPEC infected but not uninfected cells, with no appreciable differences between WT, N-WASP and R cells. Similar results were obtained with total cell lysates although longer exposure times where necessary to detect Tir (data not shown).

In contrast, neither the isolated SH3 domain nor the GST negative control bound Tir in any of the cells types used. In view of these results, we can conclude that in cells, cortactin binds Tir primarily through its N-terminal region. To test whether the SH3 domain of cortactin prefers to bind N-WASP over Tir, we performed pull-downs with clarified total lysates, and we then stripped and reprobed the blots with anti-N-WASP antibody. This approach was necessary because the pellets did not contain easily detectable levels of N-WASP (data not shown). As previously described 14, the SH3 domain of cortactin was able to pull-down N-WASP in WT cells but not N-WASP-deficient cells (data not shown). This argues in favor of the conclusion that the N-terminal region of cortactin is involved in binding Tir, while the SH3 domain is involved in binding N-WASP.

HeLa human epithelial cells were obtained from ATCC and grown in Iscove's Modified Dulbecco's Media (IMDM) supplemented with 10% FBS (fetal bovine serum). N-WASP-deficient and R mouse embryonic fibroblasts (MEFs) were obtained from Dr. Scott Snapper (Massachusetts General Hospital, Boston, USA) and Nck1/2-deficient MEFs from Dr. Tony Pawson (Samuel Lunenfeld Research Institute, Mount Sinai Hospital, Toronto, Canada). Enteropathogenic Escherichia coli (EPEC) E2348/69, as well as monoclonal antibodies against the N- and C-termini Tir (2A5 and 2C3, respectively) were provided by Dr. Brett B. Finlay (University of British Columbia, Vancouver, Canada). Anti-N-WASP antibody was previously described 44. Commercial antibodies used were: anti-cortactin 4F11 monoclonal antibody, anti-Src GD11 monoclonal and polyclonal anti-phosphoY416 (activated Src) antibodies (Millipore), and anti-actin C4 monoclonal antibody (MP Biomedical). Anti-phospho-cortactin Y466 polyclonal antibody was from Abcam (Cambridge, UK). Polyclonal anti-Erk1/2 (p44/42) and monoclonal anti-phospho-Erk (Thr202/Tyr204, clone E10) antibodies were from Cell Signaling. Anti-rabbit and anti-mouse horseradish peroxidase (HRP) antibodies were from Amersham Pharmacia Biotech.

Wild-type cortactin and selected mutants 14 were sub-cloned in frame with GFP at the N-terminus in the plasmid pC2-EGFP (Invitrogen) and verified by sequencing. The constructs used were full-length wild-type cortactin (FL), and the following derivatives: the single point mutants W22A and W525K; the double mutant S405,418D; the triple mutant Y421,466,482D; an N-terminal fragment of cortactin (NH2) containing residues 1–333, and a cortactin fragment (residues 458–546) containing the SH3 domain (SH3) aas. Two new mutants: S405,418A and Y421,466,482F, were generated using PCR and GST-FL as the template with the QuikChange site-directed mutagenesis kit (Stratagene). The Tir Y474D mutant was produced using the QuikChange kit.

Cell transfection was carried out using Lipofectamine 2000 (Invitrogen). Briefly, HeLa cells were grown to 60–70% confluence in 6-well plates. Transfections were incubated for 16 hours in medium containing 10% FBS but no antibiotics. Western blotting was done on cells from a single well by directly adding 300 μl of 2× Laemmli buffer and scraping the cells. Samples were homogenized by six passages through a syringe with a 25-gauge needle, followed by centrifugation at 21,000 g for 15 min at 4°C. Samples were resolved by 10% SDS-PAGE and analyzed by Western blotting and developed with ECL (Amersham). Band densitometry was carried out using NIH ImageJ software. Normalization for each experiment was done by first, normalizing actin and next, the protein. The average difference was calculated from three independent experiments and reported as ± standard deviations.

EPEC infections were carried out as follows. Overnight bacterial culture were grown at 37°C with shaking at 200 r.p.m., and 1 μl of culture was added per well of a 6-well-plate. Pedestals were allowed to form for 3 hours in medium containing 10% FBS and no antibiotics at 37°C and 5% CO₂. When indicated, EPEC was preactivated by incubating a 1/100 dilution of the O/N culture for 2 hours in medium containing 10% FBS and no antibiotics at 37°C and 5% CO₂; the amount of bacteria from this preactivated culture that was added to wells was 1/8 that of non-preactivated bacteria. Quantification was done by counting the numbers of pedestals of attached bacteria for a total of 100 cells. Experiments were performed at least three times. Statistical analysis was carried out using Student's t-test in Microsoft Excel.

Cells were fixed with 4% formalin solution (Sigma) at room temperature and permeabilized with 0.1% Triton X-100 for 5 min. After two washes with PBS, cells were blocked with 2% BSA in PBS for 10 min and then sequentially stained with 1 μg/ml tetramethyl rhodamine isotiocianate (TRITC)-phalloidin (Sigma) and DAPI (300 nM). Photographs were taken using a Nikon Eclipse TE 200-U fluorescence microscope equipped with a Hamamatsu camera. Images were processed with Adobe Phothoshop. For determination of total content of F-actin, TRITC-phalloidin was used at 5 μg/ml. As a control, cells were pretreated 15 min with cytochalasin D at 2 μg/ml. Samples were sorted by fluorescence using a FACS Scalibur station. Experiments were performed at least three times. Statistical analysis was carried out using Student's t-test in Microsoft Excel.

GST recombinant proteins were produced, purified and, when necessary, treated with PreScission protease according to the manufacturer's recommendations (GE Healthcare) to remove GST. Carboxilate microspheres (1 μm; Polysciences Inc.) were coupled to Tir/TirD in solution (500 nM) and washed once with Xb buffer (10 mM HEPES pH 7.8, 100 mM KCl, 0.1 mM CaCl₂, 1 mM MgCl₂, 1 mM ATP) and twice with Xb buffer-1% BSA to block nonspecific interactions. Purified actin (2.5 μM) and Arp (300 nM) were used. Cortactin and its mutants (500 nM) were added to a final volume of 25 μl of Xb buffer. After 1 hour TRITC-phalloidin was added to a final concentration of 3.3 μM. The solution containing the beads was placed on a slide and sealed with paraffin. Pictures were acquired while keeping all relevant parameters fixed (gain, exposure time) to allow for fluorescence intensity comparison. Experiments were performed at least three times.

After EPEC infection, MEFs were fractionated as described 36 with some modifications. Briefly, MEFs were grown to 70–80% confluence in 150-mm plates and infected with preactivated EPEC. After 3 hours of infection, cells were washed once with ice-cold PBS and rapidly lysed at 4°C by overlaying the cell monolayer for 10 min with 1 ml of buffer containing imidazole (pH 7.4), 250 mM sucrose, protease-phosphatase inhibitor cocktail (Complete™) and phosphatase inhibitor (PhosStop, Amersham). Then the cells were collected using a cell scraper (Sarstedt) and disrupted by six passages through a syringe with a 25-gauge needle, followed by 15 min centrifugation at 3,000 g to remove cellular debris, bacteria and nuclei. Clarified lysates were centrifuged again for 1 hour at 21,000 g to separate the membrane (pellet) from the cytoplasmic fraction (supernatant). Both of these final fractions were stored at -70°C until further use.

For pull-down experiments pellets were resuspended in 400 μl of lysis buffer containing 0.1% Triton-X100, and a fourth was used for each pull-down. GST, the GST-N terminal (NH2) cortactin fragment and the GST-SH3 cortactin domain were produced in BL21 E. coli, purified and coupled to GSH-beads. Pull-downs were washed three times with 100 μl of lysis buffer diluted 1:10 in PBS containing 0.05% Tween 20.

Pull-down experiments with recombinant proteins were performed as previously described 14. When necessary the GST was removed by Precission enzyme treatment (Amersham).

Pervanadate treatment was carried out by mixing 1 mM of NaVO₄ with 1% H₂O₂ and diluting two-fold with IMDM medium for 30 min at 37°C and 5% CO₂.