IQGAP1 Integrates Ca2+/Calmodulin and B-Raf Signaling*

Ca2+ and calmodulin modulate numerous cellular functions, ranging from muscle contraction to the cell cycle. Accumulating evidence indicates that Ca2+ and calmodulin regulate the MAPK signaling pathway at multiple positions in the cascade, but the molecular mechanism underlying these observations is poorly defined. We previously documented that IQGAP1 is a scaffold in the MAPK cascade. IQGAP1 binds to and regulates the activities of ERK, MEK, and B-Raf. Here we demonstrate that IQGAP1 integrates Ca2+ and calmodulin with B-Raf signaling. In vitro analysis reveals that Ca2+ promotes the direct binding of IQGAP1 to B-Raf. This interaction is inhibited by calmodulin in a Ca2+-regulated manner. Epidermal growth factor (EGF) is unable to stimulate B-Raf activity in fibroblasts treated with the Ca2+ ionophore A23187. In contrast, chelation of intracellular free Ca2+ concentrations ([Ca2+]i) significantly enhances EGF-stimulated B-Raf activity, an effect that is dependent on IQGAP1. Incubation of cells with EGF augments the association of B-Raf with IQGAP1. Moreover, Ca2+ regulates the association of B-Raf with IQGAP1 in cells. Increasing [Ca2+]i with Ca2+ ionophores significantly reduces co-immunoprecipitation of B-Raf and IQGAP1, whereas chelation of Ca2+ enhances the interaction. Consistent with these findings, increasing and decreasing [Ca2+]i increase and decrease, respectively, co-immunoprecipitation of calmodulin with IQGAP1. Collectively, our data identify a previously unrecognized mechanism in which the scaffold protein IQGAP1 couples Ca2+ and calmodulin signaling to B-Raf function.

isoforms of the small GTP-binding protein Ras (2,3). By inducing the exchange of GDP for GTP on Ras, the guanine nucleotide exchange factor Sos activates Ras, which mediates Raf activation (4). The Raf family of serine/threonine protein kinases comprises three isoforms: A-Raf, B-Raf, and C-Raf (also known as Raf-1) (5,6). Although both Ras binding and phosphorylation are recognized as important components of Raf activation (5), the specific mechanism by which Raf is activated remains unknown. Nevertheless, active Raf induces a cascade, with sequential phosphorylation of MEK and ERK, which results in modulation of cellular function (2).
It is now widely accepted that cytoplasmic signaling proteins form networks of interactions rather than simple linear pathways (7,8). Individual enzymes receive input from several pathways, where cross-talk from one signaling pathway influences the activity of another (9). Interconnections between signaling cascades result in networks whereby a single signaling component receives information from multiple sources (8). Analogous to other signaling cascades, MAPK function is influenced by several pathways, including Ca 2ϩ . A diverse array of extracellular regulators influence cellular behavior via the second messenger molecule Ca 2ϩ (10). Ca 2ϩ signaling controls numerous cellular functions, ranging from muscle contraction and gene transcription to cell cycle progression and memory (10,11). Calmodulin is a ubiquitous Ca 2ϩ trigger protein that translates Ca 2ϩ signals to changes in the cell (12). Accumulating evidence indicates that Ca 2ϩ and calmodulin regulate MAPK signaling (for review, see Ref. 13). For example, manipulation of intracellular free Ca 2ϩ concentrations ([Ca 2ϩ ] i ) alters Raf/MEK/ERK function. Moreover, inactivation of calmodulin induces activation of MEK/ERK (14). Although the cumulative data are convincing, much of the evidence implicating Ca 2ϩ and calmodulin in MAPK signaling was derived using chemical antagonists, and the molecular mechanisms underlying many of these observations have not been identified.
In addition, IQGAP1 binds both Ca 2ϩ /calmodulin and Ca 2ϩ -free (apo-) calmodulin (21,24,25) in a complex interaction that occurs predominantly via the four IQ motifs of IQGAP1 (34). Published studies from several groups, including ours, have shown that Ca 2ϩ /calmodulin attenuates the interaction of IQGAP1 with all binding partners studied to date. These include Cdc42 (21,24), ␤-catenin (28), E-cadherin (26), actin (25), S100B (35), and Rap1 (22). Via its interaction with IQGAP1, Ca 2ϩ /calmodulin is able to modulate the activity of these IQGAP1 binding partners. These data, combined with the observations that Ca 2ϩ /calmodulin regulates MAPK signaling (13), led us to investigate whether IQGAP1 integrates Ca 2ϩ / calmodulin signaling with the MAPK cascade. Analysis was performed in vitro with pure proteins and in intact cells in which [Ca 2ϩ ] i and IQGAP1 concentrations were manipulated.

EXPERIMENTAL PROCEDURES
Materials-Lipofectamine 2000, tissue culture reagents, and Pfx polymerase were purchased from Invitrogen. All of the restriction enzymes were obtained from New England Biolabs. BAPTA-AM, A23187, and ionomycin were purchased from Sigma. [ 35 S]Methionine and 45 CaCl 2 were from New Life Science Products. Glutathione-Sepharose, protein G-Sepharose, and protein A-Sepharose beads were from Amersham Biosciences. Ni 2ϩ -nitrilotriacetic acid-agarose beads and the QIAprep kit were from Qiagen. Fetal bovine serum and EGF were obtained from Invitrogen. Annexin V was from Sigma. The transcription and translation T7 quick-coupled transcription/ translation system was obtained from Promega. Anti-Myc monoclonal antibody (9E10.2) was manufactured by Maine Biotechnology. The anti-B-Raf monoclonal antibody was purchased from Santa Cruz Biotechnology. The anti-IQGAP1 polyclonal antibody and anti-calmodulin monoclonal antibody have been previously characterized (24,36). The anti-IQGAP1 and anti-myoglobin monoclonal antibodies were generously provided by Andre Bernards (Massachusetts General Hospital, Boston, MA) and Jack Ladenson (Washington University School of Medicine, St. Louis, MO), respectively. Secondary antibodies for ECL detection were from Amersham Biosciences. All other reagents were of standard analytical grade.
Cell Culture and Transfection-Mouse embryonic fibroblast (MEF) cells were isolated from embryonic day 14 embryos of IQGAP1 Ϫ/Ϫ mice and normal littermate controls and immortalized as described (33). HEK-293H and MEF cells were maintained in Dulbecco's modified Eagle medium supplemented with 10% (v/v) fetal bovine serum, 100 units penicillin, and 100 g/ml streptomycin and grown at 37°C and 5% CO 2 . The cells were transfected using Lipofectamine 2000 essentially as described (33,37).
Preparation of Fusion Proteins-Glutathione S-transferase (GST)-B-Raf was constructed as described (33). GST fusion proteins were expressed in Escherichia coli and isolated with glutathione-Sepharose as previously described (24,33). Histagged IQGAP1 was expressed and purified from Sf9 insect cells using Ni 2ϩ -nitrilotriacetic acid-agarose beads as previously described (38). All of the fusion proteins were at least 90% pure. 45 Ca 2ϩ Overlay-Ca 2ϩ binding was performed essentially as described (24). Briefly 6 g of GST, 6 g of GST-Rac1, 2 g of calmodulin, and 8 g of GST-B-Raf were adsorbed onto PVDF membrane. The membrane was incubated with 45 CaCl 2 (2 Ci/ ml) in buffer containing 60 mM KCl, 5 mM MgCl 2 , and 10 mM imidazole (pH 6.8) for 10 min at 22°C. After washing in distilled H 2 O, the membrane was air-dried and exposed to x-ray film. To confirm that protein bound to the PVDF membrane, the dot blot was stained with 0.1% (w/v) Amido Black in 45% (v/v) methanol and 10% (v/v) acetic acid for 5 min at 22°C, followed by destaining for 5 min at 22°C in 90% (v/v) methanol and 2% (v/v) acetic acid.
TNT Product Production-[ 35 S]Methionine-labeled TNT products were produced with the TNT quick coupled transcription/translation system (Promega, Madison, WI) as described (33). Briefly, 2 g of IQGAP1 plasmid (wild type and IQGAP1⌬CHD) was incubated with 40 l of TNT Quick Master Mix (Promega) and 2 Ci (1 Ci ϭ 37 GBq) of [ 35 S]methionine at 30°C for 90 min. The products were identified by SDS/ PAGE and autoradiography. All of the products migrated to the expected position on SDS-PAGE.
In Vitro Binding Assays-Equal amounts of [ 35 S]methioninelabeled IQGAP1 constructs or pure His-tagged IQGAP1 protein from insect cells were incubated for 3 h at 4°C with 5 g GST-B-Raf in 1 ml of buffer A (50 mM Tris-HCl, pH 7.4, 150 nM NaCl, 1 mM EGTA, 1% Triton X-100) containing 1 mM phenylmethylsulfonyl fluoride, protease inhibitor mixture, and 1 mM EGTA or 1 mM CaCl 2 . GST (2 g) alone was used as a control. For competition assays, 5 g of calmodulin, 15 g of BSA, or 4 g of annexin V was preincubated with equal amounts of [ 35 S]methionine-labeled IQGAP1 for 60 min, and then equal amounts of GST-B-Raf were added. Complexes were isolated with glutathione-Sepharose, resolved by SDS/PAGE, and processed by autoradiography (for radiolabeled IQGAP1) or Western blotting (for His-tagged IQGAP1).
Immunoprecipitation-Immunoprecipitation was performed as described previously (39) with minor modifications. Briefly, subconfluent 293H cells were co-transfected with Myc-IQGAP1 and HA-B-Raf. pcDNA3 vector was added to ensure that the total amount of plasmid was the same in each sample. After 24 h, the medium was replaced with serum-free Dulbecco's modified Eagle medium, and 12 h later the cells were incubated with pertinent reagents. To examine the effects of manipulating [Ca 2ϩ ] i , 30 M BAPTA-AM, 5 M A23187, or 5 M ionomycin was added for 30, 20, or 5 min, respectively (see Figs. 7 and 8). (Analysis with trypan blue revealed that none of the agents significantly reduced cell viability under the assay conditions used.) An equal volume of Me 2 SO was added as vehicle. The cells were washed twice with ice-cold phosphate-buffered saline, followed by incubation with 100 mM dithiobis(succinimidyl propionate) for 10 min (for cross-linking), which was neutralized by adding 500 l of 1 M Tris-HCl (pH 7.4) to the medium (final concentration is 50 mM Tris-HCl). The cells were washed twice with ice-cold phosphate-buffered saline and lysed in buffer B (20 mM Tris-HCl, pH 8.0, 137 mM NaCl, 10% glycerol, and 0.2% Triton X-100) containing 1 mM phenylmethylsulfonyl fluoride and a protease inhibitor mixture and subjected to centrifugation at 15,000 ϫ g for 10 min at 4°C to remove the debris. Supernatants were precleared for 30 min at 4°C with protein G-Sepharose beads. Anti-Myc monoclonal or anti-IQGAP1 polyclonal antibodies were incubated with protein G-or protein A-Sepharose beads, respectively, for 2 h at 4°C, washed four times with buffer B and incubated for 3 h with at 4°C with equal amounts of precleared protein lysate. Antimyoglobin monoclonal antibody or nonimmune rabbit serum were used as controls for immunoprecipitation with monoclonal or polyclonal antibodies, respectively. Complexes were sedimented by centrifugation, washed five times with buffer B, and heated for 5 min at 100°C in solubilization buffer. The samples were resolved by SDS-PAGE and transferred to PVDF, and the blots were probed with anti-IQGAP1, anti-B-Raf, or anti-calmodulin monoclonal antibodies as indicated in the relevant figure legends. Antigen-antibody complexes were visualized with horseradish peroxidase-conjugated goat anti-mouse secondary antibody and developed by ECL.
In Vitro B-Raf Kinase Assays-B-Raf kinase activity was quantified using inactive GST-MEK (kindly provided by Michael White, University of Texas Southwestern Medical Center) as substrate as described (33). Briefly, the cells were serum-starved and then incubated with vehicle, 30 M BAPTA-AM, or 5 M A23187 for 20 min at 37°C. After washing in phosphate-buffered saline, the cells were treated with or without 100 ng/ml EGF for 10 min at 22°C and lysed in radioimmune precipitation assay buffer (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1% Nonidet P-40, 1 mM EDTA, 0.25% deoxycholic acid) containing protease and phosphatase inhibitor cocktails. The lysates were immunoprecipitated with anti-B-Raf monoclonal antibody. Immunoprecipitates were washed three times and incubated with 500 M ATP and 1 g of inactive GST-MEK in a kinase assay buffer containing 20 mM MOPS, pH 7.2, 25 mM ␤-glycerol phosphate, 5 mM EGTA, 1 mM sodium orthovanadate, and 1 mM dithiothreitol. After incubating at 30°C for 30 min, GST-MEK phosphorylation was quantified by probing Western blots with anti-phospho-MEK1/2 antibody (Upstate/ Millipore). The amount of B-Raf immunoprecipitated was assessed by Western blotting with anti-B-Raf monoclonal antibody.
Miscellaneous-Protein assays were performed using the DC protein assay from Bio-Rad. Densitometry of ECL signals were analyzed with Un-scan-it software (Silk Scientific Corp.) Statistical analysis was performed by one way analysis of variance or Student's t test with GraphPad Prism 4 (GraphPad software).

Ca 2ϩ
Regulates the Binding of IQGAP1 to B-Raf in Vitro-We previously documented a direct interaction between IQGAP1 and B-Raf (33). In vitro analysis was conducted to explore the effect of Ca 2ϩ on this interaction. Specific binding of IQGAP1, both pure protein (Fig. 1A) and IQGAP1 labeled with [ 35 S]methionine using the TNT system (Fig. 1B), to GST-B-Raf is observed. Significantly more IQGAP1 binds to B-Raf in the presence of Ca 2ϩ than when Ca 2ϩ is chelated with EGTA ( Fig.  1). Quantification of four independent analyses reveals that the binding between B-Raf and IQGAP1 in the absence of Ca 2ϩ is ϳ50% of that when Ca 2ϩ is present (Fig. 1C). Coomassie staining validates that equal amounts of GST-B-Raf are present in the Ca 2ϩ -and EGTA-containing samples (Fig. 1B, lower panel). These data demonstrate that Ca 2ϩ regulates the interaction between IQGAP1 and B-Raf.
Ca 2ϩ Binds to B-Raf-The mechanism underlying the effect of Ca 2ϩ on the interaction between IQGAP1 and B-Raf was investigated. We previously documented that Ca 2ϩ binds to IQGAP1 (24), suggesting a possible means by which Ca 2ϩ enhances IQGAP1 binding to B-Raf. Nevertheless, it remained possible that our observation may be due to Ca 2ϩ regulating the conformation of B-Raf. To the best of our knowledge, no prior studies have reported a potential direct interaction between Ca 2ϩ and B-Raf. We therefore examined this possibility using 45 Ca 2ϩ overlay (24). Analysis reveals that purified B-Raf binds Ca 2ϩ directly (Fig. 2). Specificity was verified by the absence of Ca 2ϩ binding to GST alone or to GST-Rac1. Calmodulin, a well characterized Ca 2ϩ -binding protein that binds four molecules of Ca 2ϩ (12,40), serves as a positive control (Fig. 2).
The Interaction of Ca 2ϩ with IQGAP1 Regulates Its Binding to B-Raf-To determine whether the binding of Ca 2ϩ to IQGAP1 and/or B-Raf promotes the interaction between the two proteins, we used IQGAP1⌬CHD. This mutant construct lacks the CHD, which is the region on IQGAP1 to which Ca 2ϩ binds (24). In the presence of Ca 2ϩ , significantly less IQGAP1⌬CHD than wild type IQGAP1 binds to GST-B-Raf (Fig. 3). Chelation of Ca 2ϩ does not substantially reduce the interaction between IQGAP1⌬CHD and B-Raf (Fig. 3). These data imply that Ca 2ϩ binding to IQGAP1, not to B-Raf, augments the association between the two proteins.
Calmodulin Attenuates the Binding of B-Raf to IQGAP1 in Vitro-Previously published data from both our group (21,22,24,26,28) and others (25,35) document that calmodulin attenuates the binding of IQGAP1 to several binding partners. Therefore, we investigated the potential effect of calmodulin on the interaction between IQGAP1 and B-Raf. Analogous to its effect on other IQGAP1 binding partners, calmodulin significantly reduces the interaction between B-Raf and IQGAP1 in vitro (Fig. 4). The effect of Ca 2ϩ in the presence of calmodulin is different from that seen without calmodulin. Preincubation of IQGAP1 with Ca 2ϩ /calmodulin prior to adding GST-B-Raf to the assay completely abrogates the binding of IQGAP1 to B-Raf (Fig. 4). In contrast, apocalmodulin (Ca 2ϩ -free calmodulin) reduces the interac-tion between IQGAP1 and B-Raf by 43%. The effect is specific for calmodulin because BSA does not significantly alter the association of IQGAP1 with GST-B-Raf, regardless of whether Ca 2ϩ is present or absent (Fig. 4). As documented in earlier figures, equal amounts of GST-B-Raf were present in each sample, and no IQGAP1 bound to GST alone in this assay (data not shown). Note that, as shown in earlier data, significantly more IQGAP1 binds to B-Raf in the presence than the absence of Ca 2ϩ (Fig. 4). Importantly, calmodulin does not associate directly with B-Raf (41). The specificity of inhibition by calmodulin was also evaluated by examining annexin V, a well characterized Ca 2ϩ -binding protein. Analogous to BSA, annexin V does not significantly impair the binding of IQGAP1 to B-Raf (Fig. 4C). These data reveal that both Ca 2ϩ /calmodulin and apocalmodulin reduce IQGAP1 binding to B-Raf, but the inhibition by Ca 2ϩ /calmodulin is substantially greater than that produced by apocalmodulin.

Manipulation of [Ca 2ϩ ] i Alters EGF-stimulated B-Raf
Kinase Activity-EGF stimulation produces a rapid change in [Ca 2ϩ ] i (42). Several investigators have documented that Ca 2ϩ and calmodulin regulate the Ras/MEK/ERK pathway (13). Although the coupling between Ca 2ϩ and Ras signaling has been well studied (43,44), little is known about the effects of Ca 2ϩ on  B-Raf. Therefore, we manipulated [Ca 2ϩ ] i and measured B-Raf kinase activity using inactive GST-MEK as substrate (33). Incubation with EGF induces a robust increase in B-Raf kinase activity in normal fibroblasts (Fig. 5A, compare lanes 1 and  2). Increasing [Ca 2ϩ ] i with the Ca 2ϩ ionophore A23187 has no effect on B-Raf activity in serum-starved cells (i.e. basal B-Raf activity). An unexpected finding is that EGF is unable to enhance B-Raf kinase activity in cells that are pretreated with A23187 (Fig. 5A, compare lanes 3 and 4). In contrast, chelation of [Ca 2ϩ ] i with BAPTA augments the effect of EGF on B-Raf. Compared with vehicle-treated cells, the magnitude of EGF stimulation of B-Raf kinase is enhanced by 4.2fold when [Ca 2ϩ ] i is chelated (Fig. 5B, compare lanes 2 and  4). Interestingly, BAPTA also significantly increases B-Raf activity in serum-starved cells (Fig. 5B, compare lanes 1 and  3).
To evaluate whether altering [Ca 2ϩ ] i produces a transient stimulation of B-Raf activity, we examined additional time points. MEFs were incubated with A23187 or BAPTA, and B-Raf kinase activity was evaluated at different time intervals. Increasing [Ca 2ϩ ] i with A23187 has no significant effect on basal B-Raf activity at any time point studied (Fig. 6A). The results in normal and IQGAP1-null MEFs are essentially the same. Chelating [Ca 2ϩ ] i increases B-Raf activity as early as 2 min after adding BAPTA to control MEFs (Fig. 6B). As demonstrated in Fig. 5, BAPTA does not significantly alter B-Raf kinase activity in IQGAP1-null MEFs at any time point examined.
IQGAP1 Mediates the Effect of [Ca 2ϩ ] i on B-Raf-IQGAP1 is a scaffold for B-Raf/MEK/ERK signaling (29,30,33). This observation leads to the possibility that IQGAP1 may link Ca 2ϩ signaling to B-Raf function. Therefore, we examined the effects of manipulating [Ca 2ϩ ] i on B-Raf kinase activity in IQGAP1null MEFs (33). EGF is unable to stimulate B-Raf activity in these cells (Fig. 5) (Fig. 5A). Analysis of the effect of chelating [Ca 2ϩ ] i reveals that, in contrast to control MEFs, BAPTA does not significantly alter basal B-Raf kinase activity in IQGAP1-null MEFs (Fig. 5B, compare lanes 5 and 7). Although EGF is able to increase B-Raf kinase activity in IQGAP1-null MEFs treated with BAPTA, the magnitude of the increase is significantly less (p Ͻ 0.01) than that seen in control MEFs and is not statistically significantly different from BAPTA-treated IQGAP1 Ϫ/Ϫ MEFs in the absence of EGF (Fig.  5B, compare lanes 7 and 8). (Note that although EGF is unable to promote B-Raf activity in IQGAP1-null cells, these knockout cells have slightly higher basal B-Raf activity (Fig. 5). The reason for this finding, which we have documented previously (33), is unknown.) Collectively these data suggest that the effects of altering [Ca 2ϩ ] i on B-Raf kinase activity are mediated, at least in part, via IQGAP1.

Effect of Altering [Ca 2ϩ ] i on the Interaction between B-Raf and IQGAP1
-To further explore the mechanism by which IQGAP1 couples Ca 2ϩ to B-Raf, we analyzed the effect of manipulating [Ca 2ϩ ] i on the interaction between IQGAP1 and B-Raf. [Ca 2ϩ ] i was increased with the Ca 2ϩ ionophores A23187 and ionomycin. A23187 transports Ca 2ϩ from the extracellular environment to the cytosol, whereas ionomycin stimulates store-regulated Ca 2ϩ influx (45). Increasing [Ca 2ϩ ] i with A23187 reduces by 70% the amount of B-Raf that co-immunoprecipitates with IQGAP1 (Fig. 7A). Essentially identical results are obtained with the chemically distinct compound ionomycin. The ionophores do not alter the total amounts of IQGAP1 or B-Raf in the cell (Fig. 7A). Consistent with these findings, chelation of [Ca 2ϩ ] i with the cell-permeable compound BAPTA significantly increases the interaction between IQGAP1 and B-Raf in cells (Fig. 7). Like the Ca 2ϩ ionophores, BAPTA does not alter the total amounts of IQGAP1 or B-Raf in the lysate.

Changing [Ca 2ϩ ] i Alters the Interaction between IQGAP1 and Calmodulin-The reduced interaction between IQGAP1 and B-Raf in cells produced by increasing [Ca 2ϩ
] i suggests the following mechanism: increasing [Ca 2ϩ ] i enhances the binding of calmodulin to IQGAP1 (21), which attenuates binding of B-Raf to IQGAP1. To address this possibility, we examined co-immunoprecipitation of calmodulin with IQGAP1. Consistent with our model, A23187 increases by ϳ3-fold the amount of cal-

. Calmodulin attenuates the binding of B-Raf to IQGAP1 in vitro.
A, 5 g of calmodulin (CaM) or 15 g of BSA was incubated with equal amounts of [ 35 S]methionine-labeled IQGAP1 for 60 min at 4°C in buffer containing 1 mM EGTA or 1 mM CaCl 2 . Control samples contain radiolabeled IQGAP1 with no added calmodulin or BSA. Equal amounts of GST-B-Raf were added to each sample, and after incubating for 3 h at 4°C, the complexes were isolated with glutathione-Sepharose and processed by SDS-PAGE and autoradiography. B, the amount of IQGAP1 bound to GST-B-Raf was quantified by densitometry. The data, expressed relative to the amount of IQGAP1 in the control sample with Ca 2ϩ , are the means Ϯ S.E. (n ϭ 2). *, p Ͻ 0.05; †, p Ͻ 0.01. C, radiolabeled IQGAP1 was incubated without (Ϫ) or with (ϩ) 4 g of annexin V in buffer containing 1 mM EGTA or 1 mM CaCl 2 . The samples were processed as described for A. modulin that co-immunoprecipitates with IQGAP1 (Fig. 8). Moreover, chelation of [Ca 2ϩ ] i with BAPTA significantly reduces the amount of calmodulin that co-immunoprecipitates with IQGAP1. Nonimmune rabbit serum validates the specificity of the interaction (Fig. 8). These data, which are consistent with our in vitro competition assays, suggest that modulating [Ca 2ϩ ] i regulates the interaction of B-Raf with IQGAP1 via calmodulin.

EGF Increases the Interaction Between IQGAP1 and B-Raf-
We previously documented that IQGAP1 enhances B-Raf kinase in vitro and is necessary for EGF to promote B-Raf activity in cells (33). Therefore, we examined the effect of EGF on the interaction between IQGAP1 and B-Raf in cells. Incubation with EGF significantly enhances the specific co-immunoprecipitation of B-Raf with IQGAP1 by 2.4 Ϯ 0.35-fold (x Ϯ S.E.) (Fig. 9). EGF does not alter the total amounts of IQGAP1 or B-Raf in the cell lysates (Fig. 9).

DISCUSSION
Although the focus of increased attention in the last few years, the molecular mechanism of regulation of B-Raf function remains incompletely understood (6). IQGAP1 was recently Procedures." After lysis, equal amounts of protein were immunoprecipitated (IP) with anti-Myc (the IQGAP1 is Myc-tagged) monoclonal antibody and resolved by SDS-PAGE. Anti-myoglobin (Mb) monoclonal antibody was used as control. Equal aliquots of protein lysate were also resolved by Western blotting (Lysate). Western blots were probed with anti-IQGAP1 and anti-B-Raf antibodies. The data are representative of at least three independent experiments. B, the amount of B-Raf in immunoprecipitates was quantified by densitometry and corrected for the amount of IQGAP1 in the corresponding sample. The data are expressed relative to the amount of B-Raf in vehicle-treated samples and represent the means Ϯ S.E. (n ϭ 3 for A23187 and ionomycin; n ϭ 6 for BAPTA). *, p Ͻ 0.001; **, p Ͻ 0.0001.

FIGURE 8. Changing [Ca 2؉ ] i alters the interaction between IQGAP1 and calmodulin.
A, HEK-293H cells were processed as described in the legend for Fig. 7. Equal amounts of protein were immunoprecipitated (IP) with anti-IQGAP1 polyclonal antibody and resolved by SDS-PAGE. Nonimmune rabbit serum (NIRS) was used as control. Equal aliquots of protein lysate were also resolved by Western blotting (Lysate). Western blots were probed with anti-IQGAP1 and anti-calmodulin (CaM) monoclonal antibodies. The data are representative of at least two independent experiments. B, the amount of calmodulin in immunoprecipitates was quantified by densitometry and corrected for the amount of IQGAP1 in the corresponding sample. The data are expressed relative to the amount of calmodulin in vehicle-treated samples and represent the means Ϯ S.E. (n ϭ 3 for A23187; n ϭ 2 for BAPTA). *, p Ͻ 0.01; **, p Ͻ 0.0001.
shown to function as a scaffold in B-Raf signaling (33). In the present work, we extend these findings by demonstrating that IQGAP1 serves to mediate cross-talk from the Ca 2ϩ and calmodulin signaling pathways to B-Raf signaling.
In vitro analysis with pure proteins reveals that the amount of IQGAP1 bound to B-Raf is increased by ϳ2-fold when Ca 2ϩ is present in the assay. A direct interaction of Ca 2ϩ with IQGAP1 has been previously documented (24), but, to our knowledge, no prior examination of the possible binding of Ca 2ϩ to B-Raf has been published. We observed a specific and direct association of Ca 2ϩ with B-Raf. It remained possible, therefore, that the increased interaction between B-Raf and IQGAP1 produced by Ca 2ϩ could be mediated by Ca 2ϩ binding to B-Raf, to IQGAP1, or to both proteins. Although the specific site on B-Raf where Ca 2ϩ binds remains to be determined, Ca 2ϩ was previously shown to bind exclusively to the CHD of IQGAP1 (24). Analysis with IQGAP1⌬CHD, a mutant construct lacking the CHD (24), reveals that the binding of IQGAP1⌬CHD to B-Raf is equivalent to that of wild type IQGAP1 in the absence of Ca 2ϩ . The observation that Ca 2ϩ fails to significantly enhance the interaction between IQGAP1⌬CHD and B-Raf supports the notion that Ca 2ϩ binding to IQGAP1 enhances its association with B-Raf. The biological function of Ca 2ϩ binding to B-Raf is unknown. Analysis of the kinase activity of B-Raf immunoprecipitated from HEK 293H cells failed to demonstrate a significant and reproducible difference when Ca 2ϩ in the assay was chelated with EGTA (data not shown). Additional studies are necessary to identify the Ca 2ϩ -binding domain on B-Raf, which would allow the creation of the appropriate mutant constructs to dissect out the functional sequelae of Ca 2ϩ binding.
Inclusion of calmodulin in the in vitro assay alters the effects of Ca 2ϩ on the association between IQGAP1 and B-Raf. Ca 2ϩ / calmodulin abrogates binding of IQGAP1 to B-Raf. Calmodulin binding reduces the interaction between IQGAP1 and all of its binding partners that have been examined. These include Cdc42 (21), ␤-catenin (28), E-cadherin (26), actin (25), S100B (35), and Rap1 (22). Calmodulin binds in both the apo-and Ca 2ϩ -bound forms to the IQ motifs of IQGAP1 (34). The association is complex, and Ca 2ϩ increases the interaction by ϳ2-fold (21,25,34). The effects of Ca 2ϩ on the regulation of IQGAP1 by calmodulin are intriguing and depend on the specific IQGAP1 binding partner. Ca 2ϩ is required for calmodulin to inhibit IQGAP1 binding to some proteins, such as Cdc42 (21) and actin (25), but is not necessary for calmodulin to prevent binding to IQGAP1 of other proteins, such as Rap1 (22) and ␤-catenin (28). With B-Raf, a different effect is observed. Apocalmodulin reduces the interaction between IQGAP1 and B-Raf by ϳ40%, whereas Ca 2ϩ /calmodulin abrogates binding. These observations imply that B-Raf is unable to access its binding site on IQGAP1 when Ca 2ϩ /calmodulin is bound. (Note that there is no direct interaction between B-Raf and calmodulin (41).) The inhibition may be due to an alteration in IQGAP1 tertiary conformation when bound to Ca 2ϩ /calmodulin, as proposed previously (15,17,21), direct competition between Ca 2ϩ /calmodulin and B-Raf for binding to IQGAP1, or perhaps both. Regardless of the mechanism, these in vitro data form the basis of the hypothesis that Ca 2ϩ /calmodulin might modulate B-Raf function via IQGAP1.
Ca 2ϩ and calmodulin are key modulators of the Ras/Raf/ MEK/ERK pathway, participating in the determination of the fate of cells in response to Ras activation (13). Evidence supports both direct and indirect interactions of calmodulin with several components of this pathway. For example, calmodulin binds directly to EGF receptors, K-Ras, and C-Raf in a Ca 2ϩ -dependent manner (41,46,47). No binding of H-Ras, N-Ras, B-Raf, Grb2, SOS, MEK or ERK to calmodulin was detected (41,47). A calmodulin antagonist augments tyrosine phosphorylation of the EGF receptor, attenuates EGF receptor recycling and degradation and interferes with C-Raf activity (48). An accumulating body of evidence reveals coupling between Ras and Ca 2ϩ signaling (43). In neuronal cells Ca 2ϩ influx via N-methyl-D-aspartate receptors activates Ras by activating the Ca 2ϩ /calmodulin-dependent enzyme p140 Ras-GRF1 (49). Two Ca 2ϩ -regulated Ras GTPase-activating proteins, CAPRI and RASAL, switch off the Ras signal (50). In addition, Ca 2ϩ modulates Ras by inducing its translocation to the Golgi. Ca 2ϩ /calmodulin binds to the tail of K-Ras, releasing it from the plasma membrane (51). This releases active, GTP-K-Ras from the plasma membrane, enabling it to generate sustained activity from the Golgi. Collectively, these data indicate that Ca 2ϩ and calmodulin have an important role in regulating the Ras/Raf/MEK/ERK signaling cascade.
The effects produced by Ca 2ϩ and calmodulin on MAPK signaling depend to a large extent on the cells examined. In neurons, a rise in [Ca 2ϩ ] i can activate ERK independent of nerve growth factor (52). Studies conducted with calmodulin antagonists reveal that Ca 2ϩ and calmodulin are necessary for nerve growth factor-stimulated phosphorylation of ERK (13). In contrast, Ca 2ϩ and calmodulin inhibit ERK signaling in fibroblasts and keratinocytes (13). For example, Ca 2ϩ inhibits EGF-induced ERK2 activation in keratinocytes (53). Similarly, an increase in [Ca 2ϩ ] i inactivates ERK, whereas buffering [Ca 2ϩ ] i leads to sustained activation of ERK (54). Furthermore, the calmodulin antagonist W13 activates Ras, C-Raf, MEK, and ERK in serum-starved fibroblasts (14). Not all the data in nonneuronal cells are consistent. W13 is reported to reduce C-Raf, MEK, and ERK activity in serum-starved COS-1 cells (48).
The published effects of Ca 2ϩ on B-Raf are also discrepant. In neuronal cells, depolarization-mediated Ca 2ϩ influx led to activation of B-Raf via Rap1 and protein kinase A (55). In HeLa cells, Ca 2ϩ induces a modest, but significant, activation of B-Raf in a Ras-dependent manner (56). Conversely, Ca 2ϩ restriction increases the levels of B-Raf protein and B-Raf kinase activity in M1 kidney cells (57). Similarly, preincubation of MEFs with BAPTA increases activation of ERK1/2 by EGF (58), but B-Raf activity was not evaluated in this study. The last two findings are congruent with our data. We observed that chelating [Ca 2ϩ ] i in fibroblasts markedly increases B-Raf kinase activity. Both basal and EGF-stimulated B-Raf kinase are enhanced by BAPTA. The reasons for the differences among the studies are not known, but it is likely that the route of Ca 2ϩ entry, the amplitude and frequency of the Ca 2ϩ signals, and differences among cell types are likely to contribute.
In this study we describe a previously unidentified mechanism by which Ca 2ϩ can modulate B-Raf signaling. Our data suggest a model in which IQGAP1 integrates Ca 2ϩ and B-Raf signaling (Fig. 10). A localized increase in [Ca 2ϩ ] i enhances the interaction between calmodulin and IQGAP1, which reduces the binding of B-Raf to IQGAP1. Because IQGAP1 significantly increases B-Raf kinase activity (33), attenuating the interaction will decrease B-Raf signaling. In this scheme, IQGAP1 serves as a regulatory switch to allow changes in [Ca 2ϩ ] i to modulate B-Raf/MEK/ERK signaling. Because EGF receptor activation increases [Ca 2ϩ ] i via phospholipase C␥, it is tempting to speculate that the rise in [Ca 2ϩ ] i feeds back to attenuate B-Raf activation via IQGAP1.
Several scaffold proteins are important in the regulation of MAPK signaling (for reviews, see Refs (59 -61). Whereas a number of these scaffolds bind C-Raf, MEK, and ERK, few proteins have been identified that assemble a complex containing B-Raf, MEK, and ERK. One of the best characterized MEK/ERK scaffolds, KSR (kinase suppressor of Ras), was recently shown to be in a complex with B-Raf (62). Moreover, neither the contribution of individual scaffolds nor the potential communication between different scaffolds is known. Similarly, the possible role of MAPK scaffolds in facilitating cross-talk with other signaling pathways is not understood. For example, very little is known about the possible role of scaffolds in integrating Ca 2ϩ and MAPK signaling. Based on our data, it seems feasible that Ca 2ϩ may regulate the scaffolding function of IQGAP1. A localized increase in [Ca 2ϩ ] i might shift the equilibrium, reducing the interaction of B-Raf (and perhaps MEK/ERK) with IQGAP1. The released B-Raf, MEK, and ERK might bind (an)other scaffold(s), resulting in altered ERK function. This dynamic regulation of IQGAP1 scaffold function may contribute to spatio-temporal modulation of the MAPK cascade.
Prior studies reveal that IQGAP1 is a scaffold in the MAPK pathway (29,30,32,33). IQGAP1 binds directly to B-Raf, MEK, and ERK, modulating their activation by EGF. Consistent with its role as a scaffold, IQGAP1 mediates communication between signaling cascades. For example, IQGAP1 links calmodulin to E-cadherin-mediated cell-cell attachment (26), to Cdc42 signaling (24) and to ␤-catenin-mediated transcriptional co-activation (28). More recent evidence shows that IQGAP1 links N-cadherin to ERK signaling during fear memory formation (32). Here we show that IQGAP1 integrates signaling pathways, enabling Ca 2ϩ to regulate B-Raf activity. The data in this study indicate that manipulation of [Ca 2ϩ ] i modulates B-Raf activity by altering its interaction with the scaffold IQGAP1. These findings uncover an additional mechanism by which Ca 2ϩ couples to the Ras/Raf/MEK/ERK pathway, identifying a previously unknown means of regulating the MAPK cascade.