THE RECRUITMENT OF THE INTERLEUIN-1 RECEPTOR-ASSOCIATED KINASE (IRAK) INTO FOCAL ADHESION COMPLEXES IS REQUIRED FOR IL-1 β -INDUCED ERK ACTIVATION

The colocalization of IL-1 receptors with focal adhesions has been implicated in the restriction of IL-1 signal transduction to ERK, however the mechanism of this restriction and the requirement of IL-1 receptor-associated proteins have not been characterized. We determined if theinterleukin-1 receptor-associated kinase (IRAK) colocalizes with focal adhesions and is required for IL-1-dependent ERK activation. Human gingival fibroblasts were incubated with collagen-coated beads to induce the assembly of focal adhesions at sites of cell-bead contact. Phosphorylated IRAK was transiently detected in focal adhesion complexes isolated from fibroblasts stimulated with IL-1 β . Cells forming focal adhesions showed IL-1-induced phosphorylation of ERK, JNK and p38; in contrast, cells plated on poly-L-lysine to prevent focal adhesion formation showed activation only of JNK and p38. ERK activation was partially restored by incubating cells plated on poly-L-lysine with collagen-coated beads prior to IL-1 stimulation. Cells treated with Swinholide A to induce actin filament depolymerization, showed elimination of IL-1-induced ERK activation. Fibroblasts electroinjected with an anti-IRAK antibody to block the recruitment of IRAK into FACs failed to activate ERK after IL-1 treatment, indicating that FAC-associated IRAK is required for ERK activation. These data suggest that the integrity of actin filament arrays and the recruitment of IRAK into focal adhesions are involved in the restriction of IL-1 signaling to ERK.


INTRODUCTION
Interleukin-1 (IL-1) is a potent, multifunctional cytokine that is involved in a host of immune and pro-inflammatory responses (1).The broad spectrum of biological effects attributed to IL-1 results from its ability to induce a wide range of factors which contribute to the inflammatory response.These factors include matrix metalloproteinases (2,3), nitric oxide synthetase (4), prostaglandin E (5) as well as other cytokines (6,7).Consequently, IL-1 is able to mediate significant cellular and tissue damage when its expression is upregulated, as seen in the pathogenesis of chronic inflammatory diseases such as rheumatoid arthritis or periodontal diseases (8)(9)(10)(11).
Despite extensive research, much remains to be elucidated about the regulation and restriction of IL-1 signals which lead to the multiple biological responses attributed to this cytokine.There are two known membrane-bound IL-1 receptors, IL-I receptor type I and type II (IL-1R I and IL-1R II ) (12).IL-1R I alone is capable of generating a signal while IL-1R II acts as a decoy receptor (13)(14)(15)(16).The current model for IL-1 signaling suggests that following ligand binding, the IL-1 receptor associated protein (IL-1RAcP) (17) is recruited to IL-1R 1 (18), subsequently increasing the avidity of the receptor for its ligand (19).The IL-1 associated kinases (IRAK-1,2) (20,21) are also recruited to the signaling complex by the adapter protein MyD88 (22,23) within seconds of IL-1 binding (24,25).Several studies have indicated that IRAK-1 selectively associates with IL-1RAcP while IRAK-2 associates with IL-1R 1 , although the biological significance of this difference in affinities is unknown (24).Interestingly, while both kinases are rapidly phosphorylated, the phosphorylation step is not a requirement for signal transduction but, rather, appears to direct degradation through proteolysis (26).In addition, the overexpression of IRAK in the absence of IL-1 leads to its phosphorylation and degradation thereby providing a possible negative feedback mechanism for regulating the IL-1 signaling pathway (27).Following phosphorylation, IRAK dissociates from the complex in order to initiate downstream signaling events (28).Currently, the only downstream IRAK binding partner elucidated has been TRAF-6, a member of the TNF receptor-associated family.
TRAF-6 is required to mediate the IL-1 dependent activation of the NFκB and JNK/SAPK signal transduction pathways (29).
IL-1-induced signal transduction is mediated by a number of protein families.One such family is the mitogen activated protein kinase (MAPK) family of threonine-tyrosine phosphorylated signaling molecules (30)(31)(32)(33).There are three members of the MAPK family c-Jun NH2-terminal kinases / stress activated protein kinases (JNKs/SAPKs), extracellular signalregulated kinases (ERKs) and p38 MAPK .Many of the upstream and downstream signal transducers are unique to each MAPK, resulting in a cell type-restricted repertoire of responses for JNKs, ERKs and p38 (34)(35)(36)56).The IL-1 dependent activation of these kinases has been implicated in the tissue destruction characteristic of chronic inflammatory diseases, although the kinetics and degree of phosphorylation varies greatly.IL-1 induced phosphorylation of ERK and p38 occurs within 5 minutes while phosphorylated JNK appears after 15 minutes in responsive cell types (34).In HepG2 cells, IL-1β stimulates a 25-fold increase in phospho-p38, a 20-fold increase in phospho-JNK but only a 3-fold increase in phospho-ERK-1,2 (37).Understanding the mechanism of signal restriction that occurs proximal to the IL-1 receptor complex could be a significant step in determining how these MAPK cascades are differentially regulated.Insight into these events could be important in limiting pharmacologically the action of such a potent pro-inflammatory cytokine.
The cytoskeleton is an important mediator and restriction element in many types of intracellular signaling (38) as it provides a structural framework for the physical association of signaling molecules, including those involved in IL-1 signal transduction.In particular, focal adhesion complexes (FACs) are membrane-associated cytoskeletal structures which have been implicated in many signaling cascades (39,40).Immunohistochemistry and I 125 labeling experiments with human gingival fibroblasts have established a tight spatial relationship between IL-1 receptor density and FACs, suggesting a potential role for these structures in IL-1 signaling (41)(42)(43).When examined in the context of IL-1-induced MAPK phosphorylation, FACs are necessary for IL-1-dependent ERK activation, however, JNK and p38 are not similarly restricted (44).
Although immunoprecipitation studies have shown that IRAK and the IL-1R I associate in an IL-1 dependent manner (45), IRAK has not been linked to FACs nor has its involvement in ERK activation been investigated.Indeed little is known about the mechanism of IL-1 signal restriction by FACs, specifically with respect to IL-1-dependent ERK activation.The purpose of this study was to examine the role of IRAK, FACs and by extension the cytoskeletal network, in IL-1 signal restriction.We determined: 1) if IRAK associates with the focal adhesion complex; 2) if this association is required for IL-1 induced ERK activation; and 3) whether or not this association is an IL-1 dependent phenomenon.Human gingival fibroblasts were used as a model to study IRAK-FAC association and ERK activation in response to IL-1 stimulation as they constitutively express high levels of the IL-1R 1 (43,44).

CELL CULTURE
Human gingival fibroblasts were grown in minimal essential medium (α-MEM) containing 10% fetal bovine serum and antibiotics (0.17% penicillin V, 0.1% gentamycin sulphate and 0.01% amphotericin) in a humidified atmosphere of 5% CO 2 in air.Cells between the 5th and 12th passages were used for all experiments.

COLLAGEN-COATED BEAD PREPARATION
Magnetite beads were added to soluble collagen (100 µg/ml) and vortexed.NaOH was added to a final concentration of 0.1 mM to equilibrate pH to 7.4 and facilitate collagen fibril assembly on the beads.The suspension was incubated at 37°C for 20 minutes.The beads were then washed several times and resuspended in PBS and sonicated for 10 seconds (output setting 3, power 15%).

ISOLATION OF FOCAL ADHESIONS
Cells which had reached 80-90% confluence on 60 mm tissue culture dishes in normal growth medium were used.For each experiment 5 dishes of cultured cells were cooled to 4°C.
Collagen-coated magnetite beads were added to each dish.Experiments were conducted in normal growth medium.FACs were isolated from dishes 1 and 2 after 5 and 10 min respectively.
Following 10 min at 4°C the temperature of the remaining dishes was increased to 37°C and the FACs were isolated from dishes 3, 4 and 5 after an additional 2, 5 and 10 min respectively.Focal adhesion complexes were isolated according to the methods of Plopper and Ingber (1993) (52).Cells were gently washed 3 times with ice-cold PBS to remove unbound beads and scraped into ice cold cytoskeleton extraction buffer (CSKB; 5% Triton-X 100, 50 mM NaCl, 300 mM sucrose, 3 mM MgCl 2 , 20 µg/ml aprotinin, 1 µg/ml leupeptin, 1 µg/ml pepstatin, 1 mM PMSF, 10 mM PIPES, pH 6.8).The cell bead suspension was sonicated for 10 seconds (output setting 3, power 15% Branson) and the beads were isolated from the lysate using a magnetic separation stand.The beads were resuspended in fresh ice-cold CKSB, homogenized with a Dounce homogenizer (20 strokes) and re-isolated magnetically.The beads were washed thoroughly in CSKB, pelleted with a microcentrifuge, resuspended in Laemmli sample buffer and placed in a boiling water bath for 10 minutes to allow the collagen-associated complexes to dissociate from the beads.The beads were pelleted and the lysate collected for analysis.

IMMUNOBLOT ANALYSIS
The protein concentrations of the cell lysates were determined by a Bradford assay (Biorad, Hercules, CA).Equal amounts of protein were loaded into an SDS-polyacrylamide gel (10% acrylamide), resolved by electrophoresis and transferred to a nitrocellulose membrane.The membrane was incubated overnight at 4°C in a Tris-buffered saline solution with 5% milk to block non-specific binding sites.Membranes were incubated with the primary antibodies for 1-4 hours at room temperature in Tris-buffered saline with 0.1% Tween-20.Horseradish peroxidase secondary antibodies were incubated for 1 hour at room temperature in Tris-buffered saline with 0.1% Tween-20 and 5% milk.Labeled proteins were visualized by chemiluminescence (ECL Chemiluminescent Kit).

IMMUNOFLUORESCENCE STAINING
Chamber slides ( 8-well; Labtek) were coated with fibronectin (10 µg/ml in PBS).Cells were plated and allowed to spread for 24 hours prior to treatment.Following treatment cells were fixed in 3% paraformaldehyde in PBS for ten minutes at room temperature, blocked and permeabilized in PBS with 0.2%Triton-X 100 and 0.2% BSA for 15 minutes at room temperature.Antibodies were diluted in PBS with 0.2% Triton-X 100 and 0.2% BSA.
Immunofluorescence staining for vinculin and IRAK was performed with mouse monoclonal antivinculin or anti-IRAK antibody (1:50 and 1:20 dilution respectively) for 1 hour at room temperature or 3 hours at 37°C.Slides were washed with PBS, incubated with goat anti-mouse FITC-conjugated antibody (1:50 dilution) for 60 minutes at 4°C, washed and coverslipped.

ELECTROPORATION
Cells were harvested by trypsinization, pelleted and resuspended in serum-free α-MEM buffered with 12.5 mM N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid (Hepes).A 30 µl aliquot of cells was placed in a cuvette with 30 µg of mouse monoclonal anti-IRAK antibody in Hepes-buffered α-MEM at 4°C.The cells were electroporated at 100 V/cm and capacitance 960 µF using a Bio-Rad gene pulser with a capacitance expander and gene pulser cuvettes (0.2 cm inter-electrode distance).Cells were incubated at 4°C for 10 minutes and replated in normal growth medium.After 4 hours the medium was aspirated from the cells to remove cellular debris and normal growth medium was added back.

IRAK Recruitment into Nascent FACs Requires IL-1
Focal adhesion complexes (FACs) are actin-rich attachment domains that assemble at sites on the cell membrane where integrin receptors bind to the extracellular matrix (ECM) (46).Magnetite microbeads coated with ECM molecules such as collagen and fibronectin have been used to stimulate the formation as well as to enable the isolation of FACs at the sites of microbead-cell contacts (50).We first confirmed that collagen-coated magnetite microbeads could be used to isolate specific focal adhesion proteins including β-actin, vinculin and talin.
Preliminary time-course experiments (at 37°C) revealed that recruitment of these focal adhesion proteins to the sites of cell-bead contact occurred too rapidly to obtain reproducible measurements.Subsequent experiments were conducted at 4°C from 0 to 10 min, followed by an increase in temperature to 37°C from 10 to 20 min.Decreasing the reaction temperature slowed the recruitment of focal adhesion proteins into the bead complex and facilitated accurate protein quantification.Immunoblot analysis of protein bound to an equal number of beads demonstrated the time-dependent recruitment of the focal adhesion proteins β-actin, vinculin and talin into nascent focal adhesions (Fig 1).The amount of bead-bound proteins isolated at 5 min was very low but increased sharply at later incubation times (> 5 min).
Suggested location of Fig. 1 As the collagen coating of the microbeads could have acted as a non-specific trap for cellular proteins, we repeated the experiment with cells which had been pretreated with 1µM Latrunculin B (Lat B) for 30 min.This toxin sequesters actin monomers and promotes the depolymerization of actin filaments (47).We anticipated that cells pretreated with Lat B would not show focal adhesion protein binding to collagen-coated beads if actin filaments were disrupted prior to incubation of cells with beads.As expected, there was no β-actin, vinculin or talin in protein lysates obtained from these beads (Fig 1), indicating that the focal adhesion proteins isolated as described above were recruited into bead-bound complexes and were not an artifact of non-specific protein absorption to the beads.In addition, beads coated with BSA (1 mg/ml) showed very little binding of β-actin, vinculin or talin (data not shown).
Suggested location of Fig. 2 Suggested location of Fig. 3 Cultured human gingival fibroblasts express 11,000 +/-100 IL-1 receptors per cell (42) indicating that this cell type is a useful model system to study IL-1 signaling.Although previous studies have shown that the IL-1 receptors localize to focal contact sites in human fibroblasts and keratinocytes (41)(42)(43), and that IRAK and IL-1R I associate in IL-1-stimulated cells (45), a physical association between IRAK, IL-1R 1 and focal adhesions has not been established.We first determined by immunoblot analysis that IL-1R 1 could be detected in bead-associated complexes at 10 and 20 min following the addition of collagen-coated microbeads to fibroblasts Cells were loaded with anti-IRAK antibody to complex IRAK and prevent its association with other receptor complex signaling molecules.Electroporation was used to create transient pores in the cell membrane (48) and was optimized to facilitate the diffusion of large (approximately 150 kDa) protein molecules across the cell membrane (49).We determined using fluorescence microscopy and 150 kDa FITC-conjugated dextran that a field strength of 100 V/cm and a 960 µF capacitor was required to load 95% of cells with the FITC-dextran, a surrogate for mouse monoclonal anti-IRAK antibody.Electroporated cells were plated in normal growth medium (α-MEM / 10% FBS) overnight and stimulated with IL-1β for 5 min.In cells electroporated with an irrelevant isotype control antibody, immunoblots of SDS-PAGE separated lysates showed an increase of IL-1-induced ERK phosphorylation compared with unstimulated cells (Fig 7-A).
However, we were unable to detect an increase in phosphorylated-ERK in cells that had been electroporated with the anti-IRAK antibody (Fig 7-A).This indicates that IRAK is required to mediate IL-1-induced activation of ERK.To ensure that electroporation did not prevent the almost unchanged, despite the reduction in IRAK, possibly due to the association of multiple IL-1 receptors with a single IRAK molecule (26).In our study, although IRAK could not be detected after 15 min of IL-1β stimulation, it is possible that the number of IRAK molecules present in the isolated FACs were too low to be detected by immunoblotting.
While the precise requirement for FACs in IL-1-induced signaling, and more specifically the functional relationship between FACs and the IL-1 receptor complex remains unclear, we showed previously that FACs are necessary for IL-1-induced ERK activation and calcium responses (44,51).In this paper we show that it was the lack of FACs and not the altered cell morphology or the presence of poly-L-lysine which inhibited ERK activation: we were able to partially restore IL-1 induced ERK activation in rounded fibroblasts plated on poly-L-lysine by inducing the formation of FACs with collagen-coated beads.A FAC requirement for signaling is not limited to IL-1 as recently, FACs were found to restrict G-protein coupled receptor-induced ERK activation in PC12 rat pheochromacytoma cells and this restriction coincided with the expression of the calcium-regulated focal adhesion kinase Pyk2 (57).Thus FACs may provide a more general mechanism for signal restriction to ERK by participating in multiple signaling cascades originating at the cell membrane.As focal adhesions are not formed by all cell types or at all times, alterations in the abundance of focal adhesions may determine the level of IL-1 responsiveness in a given cell.Indeed, we were able to show that FAC restriction of the IL-1 signal is specific to ERK activation while both JNK and p38 were phosphorylated in the absence of FACs.
We have shown here that actin filaments are important for IL-1-induced ERK activation.A tight, reciprocal association exists between IL-1-induced calcium signaling and the expected to prevent the transport of ERK, MEKK1 and other potential binding partners to specific activation sites and block signal transduction.
Analyses of IRAK-deficient mouse fibroblasts have shown that IRAK is required for IL-1-mediated JNK, p38 and NF-kappaB activation (53,63).As we have shown that IL-1-induced ERK activity is clearly differentially regulated from the other MAPKs, and as IL-1 is able to phosphorylate FAC-associated proteins (59), we wished to determine if IRAK was a mediator of IL-1 signaling to ERK.Cells electroinjected with an anti-IRAK antibody were unable to undergo IL-1-induced ERK phosphorylation, demonstrating the requirement for IRAK in IL-1-induced ERK activation.Prior to this study, no IL-1 receptor complex member, other than IL-1R had been spatially associated with FACs (41,43).We determined that IRAK was transiently recruited to the FAC.This recruitment was IL-1 dependent and was necessary for IL-1-induced ERK activation.
Focal adhesions were originally identified simply as actin-dependent cell adhesion structures.It is now thought that these structures are also involved in signaling events that originate at the cell membrane and regulate a variety of cell processes such as proliferation, apoptosis, migration and cell spreading (40,64).The actin cytoskeleton has also been linked to signal transduction, and may be involved in mediating the interaction of specific signaling molecules in a non-random manner.As the organization and remodeling of the cytoskeleton is required for many cellular processes such as wound healing and inflammatory diseases a improved understanding of how FACs and actin regulate and restrict IL-1 signaling will be essential for understanding the action of this cytokine.

( 5 Focal 6 Actin 7 IRAK
Fig 2).To examine the association of IRAK with the focal adhesion complex, collagen bead isolation procedures were carried out on cells which had received either no treatment, IL-1β stimulation alone or Lat B treatment prior to IL-1 stimulation.IRAK is detectable by immunoblotting as an 80 kDa unphosphorylated, inactive form or a 100 kDa phosphorylated, active form(25).We detected bands of approximately 80 kDa and 100 kDa (data not shown) in immunoblots of cell lysates obtained from cells treated with IL-1β (5 min) and probed with a human anti-IRAK antibody.A 100 kDa band which co-migrated with a phosphorylated IRAK standard (Transduction Laboratories) was detected in FACs isolated from cells which had been treated with IL-1β (20 ng/ml) alone (Fig3).The IRAK-FAC association was detected after 5 min of IL-1β stimulation (IL-1β was added 5 min after microbeads) but this association was transient: after 7 min of incubation with IL-1β, IRAK had almost completely dissociated from the FAC.Neither the 80 kDa nor the 100 kDa forms of IRAK were detectable in FACs isolated from untreated fibroblasts, nor from cells which had been treated with Lat B (1 µM) 30 min prior to IL-1β stimulation (Fig3).Suggested location of Fig.4Previously, fibronectin-coated microbeads and fluorescence microscopy have been used to show FAC assembly at sites of microbead-cell contact(50,51).We used a modification of this technique in order to demonstrate the recruitment of IRAK into the FACs of IL-1 stimulated cells.Human gingival fibroblasts were plated on fibronectin-coated (10 µg/ml) glass slides and incubated with collagen-coated latex microbeads for 15 and 30 min at 37°C.Fluorescence microscopy of cells immunostained for the focal adhesion protein vinculin, showed distinct, brightly stained streaks (Fig 4) indicating that gingival fibroblasts were capable of forming typical focal contacts on their ventral surface.In cells incubated for 15 min with beads, only faint vinculin staining was detectable at the periphery of the microbead (Fig 4) while much more intense staining was visible after 30 min (Fig 4), findings that are consistent with the immunoblotting shown in Fig 1.When cells were incubated with collagen-coated microbeads for 30 min there was no IRAK immunostaining around beads, however, after 5 min of IL-1β stimulation at 37°C IRAK staining was localized to the periphery of the microbeads (Fig 4).After 20 min of IL-1 treatment, the staining had decreased to control levels (Fig 4).Suggested location of Fig. Adhesions are Required for IL-1-Induced ERK Activation As the data above showed that IRAK recruitment into FACs requires IL-1, we next determined if FACs, the actin cytoskeleton and IRAK restrict IL-1-induced MAPK activation.When plated on normal tissue culture plastic, phase contrast microscopy demonstrated the ability of cells to spread and presumably to form focal contacts (Fig 5-A).Indeed, immunofluorescence microscopy of well-spread fibroblasts stained for vinculin revealed discrete, brightly stained sites, indicating the presence of focal adhesions (Fig 5-C).In contrast to cells plated on tissue culture plastic, cells plated on poly-L-lysine cannot form integrin-mediated attachments to the substrate (41) remained rounded and were unable to form FACs (Fig 5-B).Cells plated on poly-L-lysine were unable to activate ERK in response to IL-1β stimulation (Fig 5-G) unlike cells plated on tissue culture plastic which shoed ERK activation (Fig 5-E).However, the presence of focal adhesions did not restrict IL-1 signaling to other MAPK family members: IL-1 stimulation of cells plated on poly-L-lysine resulted in a significant increase in phosphorylated JNK and p38 (Fig 5-G).As plating on poly-L-lysine induced cell rounding and dramatic changes in cell shape, we determined if the apparent focal adhesion restriction of IL-1induced ERK activation was actually attributable to changes in cell shape.Accordingly, IL-1induced ERK activation was partially restored if cells plated on poly-l-lysine were pre-incubated with collagen-coated microbeads (Fig 5-F), indicating that FACs were indeed critical for the IL-1 restricted signaling.Suggested location of Fig. Filaments and IL-1 Signaling to ERK Focal adhesion complexes are in part comprised of actin filaments.Therefore, to test the requirement for actin filaments in IL-1 signal restriction, fibroblasts were treated with two monomeric actin sequestering drugs, Lat B and Swinholide A (SWA).These agents were used to produce different levels of actin disassembly in treated cells.Short-term Lat B treatment causes loss of stress fibers and the reorganization of actin filaments into fine, branched cellular processes (52) which radiate from the cell body.SWA acts by completely severing existing actin filaments in addition to preventing actin polymerization (47).Fibroblasts were plated on fibronectincoated (10 µg/ml) glass slides and treated with Lat B (1 µM) for 30 min or SWA (50 ng/ml) overnight and immunostained for vinculin or stained with rhodamine phalloidin to show actin filaments.In Lat B-treated cells, actin filament-enriched extensions, or runners, formed where the cytoskeleton collapsed around previously formed FACs that remained attached to the fibronectin substrate (Fig 6-B).The vinculin in these cells was no longer visible as discrete points, but instead was visible as branch-like formations in the runners (Fig 6-C).These cells retained the ability to phosphorylate ERK in response to IL-1 treatment, however, the level of phosphorylated ERK was markedly reduced (Fig 6-F).SWA treatment resulted in drastically altered cell morphology including rounding of the cell and complete loss of phalloidin staining for actin filaments (Fig 6-D).Immunostaining revealed the retention of some discrete, but very small, vinculin patches (Fig 6-E).SWA-treated cells did not phosphorylate ERK in response to IL-1 stimulation (Fig 6-G), indicating that in addition to focal adhesion complexes, the retention of actin filaments is necessary for IL-1 signal transduction.Suggested location of Fig.Is Required for IL-1 Dependent ERK Activation Experiments using cells from Irak knockout mice (53, 63) have established that IRAK is a necessary component of IL-1-induced NF-κB and JNK activation.Since the IL-1 signal transduction pathways for ERK and JNK are differentially restricted in the context of focal adhesions, we investigated the effect of the loss of available IRAK on IL-1 signaling to ERK.