| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
J. Biol. Chem., Vol. 280, Issue 41, 34548-34557, October 14, 2005
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
1
From the
Department of Pathology, Brigham and Women's Hospital, and Harvard Medical School, Boston, Massachusetts 02115 and the Department of Pathology and Immunology, Washington University School of Medicine, Saint Louis, Missouri 63110
Received for publication, July 6, 2005
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
-catenin, Cdc42, Rac1, and CLIP-170. IQGAP1 regulates the cytoskeleton, promotes cell motility, and modulates E-cadherin-mediated cell-cell adhesion. However, how IQGAP1 exerts its functions in vivo is still unclear. In this study we investigate the self-association of IQGAP1 and its role in IQGAP1 function. Endogenous IQGAP1 co-immunoprecipitated from MCF-7 cells with IQGAP1 tagged with enhanced green fluorescent protein, indicating that IQGAP1 self-associates in cells. In vitro assays confirmed that IQGAP1 can self-associate and that this effect is mediated by the N-terminal half of the protein. Gel filtration analysis suggested that full-length IQGAP1 exists as a combination of monomers, dimers, and larger oligomers. Analysis performed with multiple fragments of IQGAP1 narrowed the self-association region to amino acids 763–863. In support of this observation, a peptide comprising residues 763–863 disrupted self-association of full-length IQGAP1 in a dose-dependent manner. Similarly, deleting this sequence from IQGAP1 abolished binding to full-length IQGAP1. In addition, the ability of IQGAP1 to increase the amount of active Cdc42 in cells is abrogated upon removal of this region. Consistent with these findings, transfection into cells of a peptide containing the self-association domain significantly reduced the amount of active Cdc42 in cell lysates. These observations define a sequence of IQGAP1 that is necessary for its oligomerization and demonstrate that self-association is required for the normal cellular function of IQGAP1. |
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
-catenin (10, 12), and CLIP-170 (13). Through interactions with these and other proteins, IQGAP1 participates in multiple fundamental cellular activities, including transcription, cell-cell attachment, and regulation of the cytoskeleton (2).
Although it is involved in many important activities, how IQGAP1 exerts its function in vivo is unclear. Accumulating evidence supports the hypothesis first proposed by Ho et al. (5) that IQGAP1 functions as a scaffold. Proteins in biological systems rarely act in isolation but bind other biomolecules to elicit specific cellular responses. The self-association of proteins to form dimers and higher oligomers is a very common phenomenon (for a review see Ref. 14). Recent structural and biophysical studies show that dimerization or oligomerization is a key factor in the regulation of proteins such as enzymes, ion channels, receptors, and transcription factors (14). There is some indirect evidence implying that IQGAP1 self-associates. For example, sedimentation equilibrium of IQGAP1 obtained from bovine adrenal tissue suggested an estimated molecular weight of 358,653 to 401,864 (15). Another study inferred from gel filtration analysis that IQGAP1 may form oligomers (16). However, no data have been published to unequivocally document that IQGAP1 can self-associate. Moreover, although it is commonly stated that IQGAP1 dimerizes via multiple putative coiled-coil repeats in the N-terminal half of the molecule (4, 6), direct evidence to bolster this claim is lacking. For these reasons, we set out in this study to establish whether IQGAP1 self-associates and, if so, which region is responsible. Our data reveal that IQGAP1 self-associates both in vitro and in cells. Most unexpectedly, the oligomerization site was not via the coiled-coil repeats, but localized to amino acids 763–863. Deletion of this sequence disrupted self-association of IQGAP1. Moreover, a mutant IQGAP1 construct lacking this sequence failed to increase active Cdc42 in the cells, suggesting that oligomerization is important for IQGAP1 function.
|
EXPERIMENTAL PROCEDURES |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
-amylase, alcohol dehydrogenase, and albumin markers for gel filtration analysis were purchased from Sigma. Anti-Myc monoclonal antibody (9E10.2) was manufactured by Maine Biotechnology. The anti-GFP and anti-Cdc42 antibodies were purchased from Zymed Laboratories Inc. and Gene Transduction, respectively. The anti-IQGAP1 polyclonal antibody has been characterized previously (5). The anti-IQGAP1 monoclonal antibody was generously provided by Andre Bernards (Massachusetts General Hospital, Boston). Secondary antibodies for enhanced chemiluminescence (ECL) detection were from Amersham Biosciences. All other reagents were of standard analytical grade. Cells Lines—MCF-7 human breast epithelial cells that stably overexpress pcDNA3-myc-IQGAP1 (termed MCF/I cells) have been described previously (12, 17). MCF/I cells have 3-fold more IQGAP1 than the parent MCF-7 line (17). HEK-293T and MCF-7 cells were purchased from Invitrogen and American Type Culture Collection, respectively.
Cell Culture and Transfection—HEK-293T, MCF-7, and MCF/I cells were maintained in Dulbecco's modified Eagle's medium supplemented with 10% (v/v) fetal bovine serum, 100 units penicillin, and 100 µg/ml streptomycin and grown at 37 °C in 5% CO2. MCF-7 and MCF/I cells were transfected using FuGENE 6, whereas Lipofectamine 2000 was used to transfect HEK-293T cells (17).
Plasmid Construction—Myc-tagged human IQGAP1, IQGAP1-N (amino acids 2–863) or IQGAP1-C (amino acids 864–1657), in pcDNA3 vector (5) was used. IQGAP1 lacking the CHD domain (IQGAP1
CHD, amino acids 35–365 deleted), IQGAP1WW (amino acids 643–744 deleted), IQGAP1GRD (amino acids 1122–1324 deleted), IQGAP1C (amino acids 1502–1657 deleted), IQGAP1-N1 (amino acids 2–431), and IQGAP1-N2 (amino acids 432–863) were described previously (7, 17, 18). IQGAP1-(746–860) (amino acid 746–860 deleted) was constructed using Myc-tagged wild type IQGAP1 in pcDNA3 vector. To create a PacI site, the pBluescript-KS– plasmid was digested with SspI, and then a short linker containing a PacI site was inserted. This plasmid, named pBluescript-KS–-P, was cut with PacI and ClaI, and a PacI-ClaI fragment from pcDNA3-IQGAP1 was inserted. The resultant plasmid was named pBluescript-IQ. pBluescript-IQGAP1-(746–860) was created by PCR-directed deletion mutagenesis using pBluescript-IQ as a template and Pfx polymerase. Phosphorylated primers complementary to bp 2682–2702 (antisense, 5'-ATTGGCCAGCCACAGCTGTTC-3') and bp 3048–3069 (sense, 5'-GCTGAGGATCCTCCTATGGTTG-3') of IQGAP1 were used for amplification, and the resultant PCR product was gel-purified and self-ligated, producing pBluescript-IQGAP1-(746–860). A PacI-ClaI fragment from pBluescript-IQGAP1-(746–860) was inserted into pcDNA-IQGAP1 cut with PacI and ClaI to generate pcDNA3-IQGAP1-(746–860).
Full-length IQGAP1 was tagged at the N terminus with tandem enhanced green fluorescent protein (EGFP) by digesting pcDNA3-IQGAP1 with XbaI. After making a blunt end with T4 DNA polymerase, partial digestion was performed with BamHI to give a 5577-bp fragment. This fragment was inserted into pEGFP-C1 (Clontech) at a BglII and SmalI site, producing the plasmid pEGFP-IQGAP1. To enhance fluorescence, a second EGFP was added. pEGFP-IQGAP1 was digested with NheI and PacI to give a 2380-bp fragment. The NheI cleavage site was filled with T4 DNA polymerase. The 2380-bp fragment was then inserted into pEGFP-IQGAP1 at a BsrGI-PacI site, in which BsrGI cleavage was also filled with T4 DNA polymerase, producing pEGFP-EGFP-IQGAP1 (named pEGFP2-IQGAP1).
Peptides comprising selected regions of IQGAP1 were produced by PCR. To generate Myc-IQGAP1-(763–964) (the numbers indicate the amino acid residues of IQGAP1 that are included), PCR was performed on full-length IQGAP1 using primers flanking nucleotides 2754 and 3359. The forward primer 5'-CGGGATCCCAGGAATTCCGATCCAGG-3' and the reverse primer 5'-GCTCTAGATCACTCCTTGCTCAAAGCCTTG-3' were designed to generate a 606-bp DNA product containing a BamHI site at the 5' end and a stop codon and a XbaI site at the 3' end. Digestion with BamHI and XbaI generated a fragment that was subcloned into the BamHI-XbaI site of pcDNA3-myc yielding a plasmid containing the cDNA for the 202-amino acid residues of IQGAP1 fused in-frame at its N terminus to a Myc epitope tag. A similar cloning strategy was employed to construct myc-IQGAP1-(216–683) (containing amino acids 216–683), myc-IQGAP1-(717–916) (containing amino acids 717–916), myc-IQGAP1-(763–863) (containing amino acids 763–863), and myc-IQGAP1-(864–964) using primers flanking nucleotides 1111 and 2514, 2614 and 3213, 2752 and 3054, and 3055 and 3357, of IQGAP1 respectively. The forward primers 5'-CGGGATCCGCATTCATGCTGCTGTTATTGC-3', 5'-CGGGATCCTCTATGCAGCTTTCTCGGGAGG-3', 5'-CGGGATCCCAGGAATTCCGATCCAGG-3', and 5'-CGGGATCCCCTCCTATGGTTGTGGTCCG-3' and the reverse primers 5'-GCTCTAGACCCGGGTCAGCTGTTATTATCTCCTACTGCC-3', 5'-GCTCTAGACCCGGGTCATTTGATATCCATGAGATTGATT-3',5'-GCTCTAGATCAATCCTCAGCATTGATGAGAG-3', and 5'-GCTCTAGATCACTCCTTGCTCAAAGCCTTG-3' were used to generate the IQGAP1-(216–683), IQGAP1-(717–916), IQGAP1-(763–863), and IQGAP1-(864–964), respectively. Each reverse primer included a stop codon directly following the codon corresponding to the last amino acid of the appropriate IQGAP1 sequence. For the construction of IQGAP1-(2–746), PCR was performed on full-length IQGAP1 using forward primer 5'-GCTGGCTACCCTGCAGCG-3' and reverse primer 5'-GCTCTAGATCAATTGGCCAGCCACAGCTGTTC-3' flanking nucleotides 1763 and 2702, which contains a PacI site. The reverse primer has an XbaI site at the 3' end. PCR products were digested with PacI and XbaI, and the resultant fragment were inserted into pcDNA3-IQGAP1 cut with PacI-XbaI. The sequence of all constructs was confirmed by both restriction mapping and DNA sequencing. All peptides are Myc-tagged. Plasmids were purified with a QIAprep Spin miniprep kit according to the manufacturer's instructions.
Preparation of Fusion Proteins—The construction of glutathione S-transferase (GST)-IQGAP1, GST-IQGAP1-N (N-terminal half, amino acids 2–863), GST-IQGAP1-C (C-terminal half, amino acids 864–1657), and GST-WASP-GBD has been described previously (5, 19, 20).
A TEV cleavage site was inserted between GST and IQGAP1 to allow removal of the GST tag. A linker containing the TEV protease recognition site (TEV1, 5'-GAAGATCTGATTACGATATCCCAACGACCGAAAACCTGTATTTTCAGGGCGCCGGATCCCG-3', and TEV2, 5'-CGGGATCCGGCGCCCTGAAAATACAGGTTTTCGGTCGTTGGGATATCGTAATCAGATCTTC-3') was inserted into pGEX4T1 to generate pGEX4T1-TEV. pcDNA3-IQGAP1 was digested with XbaI, and a blunt end was made with T4 DNA polymerase. Following partial digestion with BamHI, the resulting fragment of IQGAP1 was inserted into pGEX4T1-TEV, which had been digested with BamHI-SmaI. The resultant pEGX4T1-TEV-IQGAP1 was digested with PvuI, and the fragment obtained was inserted into pACYC184 (which had been digested by AvaI-HindIII) to generate the pACYC184-TEV-IQGAP1 plasmid. (Note that both the PvuI and HindIII sites were made into a blunt end with T4 DNA polymerase.)
To obtain GST-IQGAP1-(216–683), PCR was performed on full-length IQGAP1 using the forward primer 5'-CGGGATCCGCATTACATGCTGCTGTTATTGC-3' (which included a BamHI site) and the reverse primer 5'-GCTCTAGACCCGGGTCAGCTGTTATTATCTCCTACTGCC-3' (an XmaI site and a stop codon were included) flanking nucleotides 1111 and 2514 of IQGAP1. The resulting 1404-bp product was gel-purified, cut with BamHI and XmaI, and subcloned into the BamHI-XmaI site of pGEX-2T. A similar method was used to construct GST-IQGAP1-(717–916), using the primers 5'-CGGGATCCTCTATGCAGCTTTCTCGGGAGG-3' (forward, BamHI site included) and 5'-GCTCTAGACCCGGGTCATTTGATATCCATGAGATTGATT-3' (reverse, XmaI site and a stop codon included).
To obtain His-IQGAP1-(763–863), PCR was performed on full-length IQGAP1 using primers 5'-CGGGATCCCAGGAATTCCGATCCAGG-3' (forward, BamHI site included) and 5'-GCTCTAGATCAATCCTCAGCATTGATGAGAG-3' (reverse, XbaI site and a stop codon included) flanking nucleotides 2752 and 3054. The resulting product was gel-purified and cut with XbaI. After making a blunt end with T4 DNA polymerase, the resultant product was cut with BamHI to give a 303-bp fragment. The product was subcloned into the BamHI-EcoRI site of pRSET A containing a His tag, in which EcoRI cleavage was also filled with T4 DNA polymerase, producing pRSET-His-IQGAP1-(763–863). Similar strategy was taken to construct His-IQGAP1-(717–916). Briefly, PCR was performed on full-length IQGAP1 flanking nucleotides 2614 and 3213 using primers 5'-CGGGATCCTCTATGCAGCTTTCTCGGGAGG-3' (forward, BamHI site included) and 5'-GCTCTAGACCCGGGTCATTTGATATCCATGAGATTGATT-3' (reverse, SmaI site and a stop codon were included). The resulting 600-bp products were gel-purified and cut with BamHI and SmaI. The product were subcloned into the BamHI-SmaI site of pRSET A containing a His tag, in which EcoRI cleavage was also filled with T4 DNA polymerase producing pRSET-His-IQGAP1-(717–916). pRSET-His-IQGAP1-(763–863) and pRSET-His-IQGAP1-(717–916) were expressed in Escherichia coli BL21(DE3) and purified by nickel affinity chromatography using Ni2+-NTA-agarose (Qiagen). GST fusion proteins were expressed in E. coli and isolated with glutathione-Sepharose as described previously (5).
Immunoprecipitation—Immunoprecipitation was performed as described previously (7) with minor modifications. Briefly, subconfluent MCF-7 cells were transfected with pEGFP2-IQGAP1, IQGAP1-N, or IQGAP1-C. Twenty four hours later, cells were washed twice with ice-cold phosphate-buffered saline, lysed in buffer A containing PMSF and a protease inhibitor mixture, and subjected to centrifugation at 15,000 x g for 10 min at 4 °C to remove the debris. Supernatants were precleared with protein A-Sepharose for 30 min at 4 °C. Anti-IQGAP1 or anti-GFP antibodies were incubated with protein A-Sepharose beads for 2 h at 4°C, washed four times with buffer A, and incubated for 3 h with equal amounts of precleared protein lysate. Complexes were sedimented by centrifugation, washed five times with buffer A, and heated for 5 min at 100 °C in solubilization buffer. Samples were resolved by SDS-PAGE and transferred to PVDF, and blots were probed with anti-IQGAP1 monoclonal antibody (which recognizes the C-terminal half of IQGAP1), anti-IQGAP1 polyclonal antibody (which recognizes the N-terminal half of IQGAP1), anti-Myc, or anti-GFP antibodies. Antigen-antibody complexes were visualized with the appropriate (rabbit or mouse) horseradish peroxidase-conjugated secondary antibody and developed by ECL.
TNT Product Production—[35S]Methionine-labeled TNT products were produced with the TNT Quick-Coupled transcription/translation system (Promega) according to the manufacturer's instructions. Briefly, 1 µg of IQGAP1, IQGAP1CHD, IQGAP1WW, IQGAP1-(746–860), IQGAP1GRD, IQGAP1C, IQGAP1-N, IQGAP1-C, IQGAP1-N1, IQGAP1-N2, IQGAP1-(216–683), IQGAP1-(763–964), IQGAP1-(763–863), IQGAP1-(864–964), or IQGAP1-(717–916) was incubated with 40 µl of TNT Quick master mix (Promega) and 2 µCi of [35S]methionine (PerkinElmer Life Sciences) at 30 °C for 90 min. Products were identified by SDS-PAGE and autoradiography.
In Vitro Binding Assays—Equal amounts of [35S]methionine-labeled IQGAP1 constructs described in the previous paragraph were incubated for 3 h at 4°C with GST-IQGAP1, GST-IQGAP1-N, GST-IQGAP1-(216–683), GST-IQGAP1-(717–916), His-IQGAP1-(763–863), or His-IQGAP1-(717–916) in 1 ml of buffer A (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1 mM EGTA, 1% Triton X-100) containing 1 mM PMSF and protease inhibitors mixture. After centrifugation, samples were washed six times with buffer A and resolved by SDS-PAGE. TNT products were detected by autoradiography of the dried gel. For the competition assays, [35S]methionine-labeled IQGAP1-(763–863) or IQGAP1-(717–916) was preincubated with GST-IQGAP1 for 1 h at 4 °C in buffer A containing PMSF and protease inhibitors, and then equal amounts of [35S]methionine-labeled full-length IQGAP1 was added. After a 2-h incubation, samples were washed six times with buffer A, resolved by SDS-PAGE or Tricine/SDS-PAGE, and followed by autoradiography.
Gel Filtration Chromatography—GST-IQGAP1 (containing the TEV protease recognition site) was immobilized on glutathione-Sepharose and incubated with 5 µl of TEV protease at 4 °C for 16 h. The GST was pelleted by centrifugation, and the supernatant was passed through a 0.45-µm filter. Fast protein liquid chromatography (FPLC) separation was performed on an KTATMFPLC system (Amersham Biosciences) equipped with a UPC-900 monitor and a P-920 pump. The system was monitored and controlled by methods run by the UNICORN control system (5.01 version). A Superose 6 10/300 GL column (Amersham Biosciences) was pre-equilibrated with 4-column volumes of 50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1 mM EGTA, and 0.1% Triton X-100 (equilibration buffer). After loading the sample (30 µg of IQGAP1 in 500 µl of equilibration buffer containing 1% Triton X-100) onto the column at a flow rate of 0.5 ml/min, the equilibration buffer was passed through the column, and 0.5-ml fractions were collected. An aliquot (450 µl) of each fraction was precipitated with trichloroacetic acid, solubilized, and subjected to SDS-PAGE followed by Western blotting. The column was calibrated using blue dextran (2,000 kDa), apoferritin (443 kDa), -amylase (200 kDa), alcohol dehydrogenase (150 kDa), and albumin (67 kDa) as standards. IQGAP1 was also chromatographed on a TosoHaas (Montgomeryville, PA) G4000SW (600 x 7.5 mm inner diameter) column operated at 0.5 ml/min with 50 mM sodium phosphate, 150 mM NaCl, pH 6.5. Fractions were collected at 0.5-min intervals and analyzed as described above.
Measurement of Activated Cdc42—Active Cdc42 was measured with a GST fusion construct of the GTPase-binding domain (GBD) of Wiskott-Aldrich syndrome protein (WASP) as described previously (20). The GST-WASP-GBD binds only GTP-Cdc42 (20). Briefly, HEK-293T cells were transfected with selected constructs (indicated in the legend to Fig. 11), followed by lysis with buffer B (20 mM HEPES, pH 7.4, 150 mM NaCl, 1% Nonidet P-40, 20 mM NaF, 1 mM MgCl2) containing 20 µM GTP and protease inhibitor mixture. Equal amounts of protein lysate, precleared with glutathione-Sepharose beads, were incubated with 40 µg of GST-WASP-GBD for 2 h at 4°C. Complexes were collected with glutathione-Sepharose, washed, and resolved by SDS-PAGE. The resultant Western blots were probed with anti-Cdc42 antibody. Equal amounts of protein lysate were also subjected to SDS-PAGE, and blots were probed with anti--tubulin, anti-IQGAP1, and anti-Myc antibodies.
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 assessed by Student's t test with Excel (Micro software).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
|
|
|
IQGAP1-N or IQGAP1-C was transfected into MCF-7 cells, and lysates were immunoprecipitated with anti-IQGAP1 monoclonal or polyclonal antibodies. Both proteins were expressed in the cells (Fig. 4A, Lysate). Probing immunoprecipitates with anti-Myc antibody revealed that IQGAP1-N, but not IQGAP1-C, co-immunoprecipitated with endogenous IQGAP1 (Fig. 4A, filled arrowhead). Probing the blot with anti-IQGAP1 polyclonal antibody confirmed the co-immunoprecipitation of the N-terminal half with endogenous IQGAP1 (Fig. 4C). The bands migrating in the region 
95 kDa, close to the migration of IQGAP1-C, in the samples immunoprecipitated with polyclonal anti-IQGAP1 antibody (Fig. 4, B, lower panel, and C) are degradation products of endogenous IQGAP1. Note that they are not recognized by the anti-Myc antibody (Fig. 4A). These bands are detected only in the samples immunoprecipitated with polyclonal antibody because this antibody immunoprecipitates significantly more IQGAP1 than the monoclonal antibody (Fig. 4B, upper panel). Probing the top half of the blot with anti-IQGAP1 monoclonal antibody revealed that equal amounts of endogenous IQGAP1 was present in lysates (Fig. 4B, upper panel).
Amino Acids 763–863 of IQGAP1 Are Necessary for Self-Association—Further analysis was performed to narrow the self-association domain of IQGAP1. Because the N-terminal half of IQGAP1 is responsible for self-association, it was divided into two equal portions, termed N1 and N2 (Fig. 5A). In addition, a fragment comprising amino acids 763–964 was constructed to investigate the possibility that the self-association region may overlap the junction between the N- and C-terminal halves of IQGAP1. Pull down with GST-IQGAP1 of [35S]methionine-labeled fragments revealed binding of N2 and IQGAP1-(763–964) but not of N1 (Fig. 5B). These data indicate that the self-association site is between residues 763 and 863 of IQGAP1. To confirm this site, the IQGAP1-(763–964) peptide was divided into two equal halves, termed IQGAP1-(763–863) and IQGAP1-(864–964) (Fig. 5A). Consistent with the data in Fig. 5B, IQGAP1-(763–863), but not IQGAP1-(864–964), bound to full-length IQGAP1 (Fig. 5C). Similarly, pull down with GST-IQGAP1-N showed specific binding of N2 (Fig. 5B), IQGAP1-(763–964) (Fig. 5, B and C), and IQGAP1-(763–863) (Fig. 5C) to the N-terminal half of IQGAP1 (Fig. 5, B and C). These data reveal that amino acids 763–863 of IQGAP1 mediate self-association.
|
-helical coiled-coil structural motif mediates oligomerization of a large number of proteins (for a review see Ref. 21). The IQGAP1 repeat region has six putative coiled-coils located between amino acids 168 and 601 as predicted by the COILS program (www.ch.embnet.org/software/COILS_form.html). However, our data indicate that the segment containing the self-association domain resides in amino acids 763–863, which is outside the coiled-coil region. It is conceivable that the N1 and N2 constructs described above have disrupted the coiled-coil region, thereby impeding binding. To establish whether the coiled-coil also participates in the self-association of IQGAP1, we constructed a peptide encompassing amino acids 2–746, which contains the entire putative coiled-coil region (Fig. 6A). [35S]Methionine-labeled IQGAP1-(2–746) was incubated separately with GST-IQGAP1 and GST-IQGAP1-N on glutathione-Sepharose beads. Constructs that bound to the GST fusion proteins were resolved by SDS-PAGE and identified by autoradiography. No binding of IQGAP1-(2–746) was detected to either fusion construct (Fig. 6B, right panel). Note that both the positive and negative controls (IQGAP1-N2 and IQGAP1-N1, respectively) gave the expected results (Fig. 6B). These findings suggest that the putative coiled-coil region of IQGAP1 is not capable of binding to IQGAP1 and therefore does not participate in self-association.
|
|
|
|
Deletion of Amino Acids 746–860 Abrogated Self-association of IQGAP1—The data in Fig. 7A reveal that amino acid residues 763–863 of IQGAP1 are sufficient for self-association. If this sequence is necessary for binding, one would predict that deletion of this region would eliminate IQGAP1 self-association. It is noteworthy that amino acids 763–863 are located in the IQ domain. Because the IQ domain begins at residue 746, a little proximal to the self-association region, we deleted amino acids 746–860 from IQGAP1. IQGAP1-(746–860) (Fig. 8A) was labeled with [35S]methionine and incubated with GST-IQGAP1 on glutathione-Sepharose. IQGAP1-(746–860) was incapable of binding to full-length IQGAP1 (Fig. 8B, right panel). These data reveal that amino acids 763–863 of IQGAP1 are both necessary and sufficient for self-association of IQGAP1.
A Peptide Comprising Amino Acids 216–683 of IQGAP1 Does Not Self-associate—A prior publication (16) suggested that IQGAP1 forms oligomers through amino acids 216–683. Because these findings differ from our observations, we examined a peptide corresponding to amino acids 216–683 of IQGAP1 in our assay system. The binding of IQGAP1-(216–683) was compared with that of IQGAP1-(717–916) using radiolabeled peptides. Consistent with the data in Fig. 7D that it can self-associate, [35S]methionine-labeled IQGAP1-(717–916) bound specifically to GST fusion proteins of full-length IQGAP1 and IQGAP1-(717–916) (Fig. 9). In contrast, IQGAP1-(216–683) interacted with neither full-length IQGAP1 nor GST-IQGAP1-(717–916). Similarly, GST-IQGAP1-(216–683) bound neither IQGAP1-(216–683) nor IQGAP1-(717–916) (Fig. 9). In addition, GST-IQGAP1-(216–683) did not pull down endogenous IQGAP1 from MCF-7 breast epithelial cell lysates (data not shown). Therefore, we did not detect any binding of IQGAP1-(216–683) to full-length IQGAP1 or to itself under our assay conditions. Note that GST IQGAP1-(216–683) bound to IQGAP1 in bovine brain cytosol (16). Although the reasons for the different results are not known, it is possible that the binding partners of endogenous IQGAP1 in human breast epithelial cells are different to those in bovine brain cytosol.
|
800, 400, and 200 kDa (Fig. 10). These peaks may contain tetramers, dimers, and monomers, respectively, of IQGAP1. Similarly, despite a different buffer and packing material, chromatographic separation of IQGAP1 on a TosoHaas G4000SW column yielded essentially identical results (data not shown). Collectively, these findings support the concept that IQGAP1 can form multimers.
Oligomerization Is Necessary for IQGAP1 to Increase Active Cdc42—The observations that IQGAP1 self-associates both in vitro and in cells leads to the question of whether the formation of multimers is important for the biological function of IQGAP1. Previous work from our laboratory documented that overexpression of full-length IQGAP1 increases active Cdc42 in cells (17). Therefore, active Cdc42 was measured in cell lysates using GST-WASP-GBD (which binds only to active, GTP-Cdc42 (17, 20)) to investigate the possible functional sequelae of disrupting IQGAP1 self-association. Two complementary strategies were used. In the first, cells were transfected with the mutant IQGAP1-(746–860), which cannot self-associate, and compared with overexpression of wild type IQGAP1. Transient transfection of wild type IQGAP1 increased GTP-Cdc42 by 2-fold (Fig. 11), consistent with our prior observations (17). In contrast, IQGAP1-(746–860) had no effect on the amount of active Cdc42 in HEK-293T cell lysates. Note that IQGAP1-(746–860) was expressed at slightly higher levels than wild type IQGAP1 (Fig. 11). The second strategy was to disrupt oligomerization of endogenous IQGAP1 in cells using a peptide. A Myc-tagged peptide corresponding to amino acid residues 717–916 of IQGAP1 was transfected into HEK-293T cells. Probing Western blots with anti-Myc antibody revealed that the peptide was expressed well and migrated to the expected position on SDS-PAGE (Fig. 11A). More importantly, transfection of the peptide reduced by 60% the amount of active Cdc42 in cell lysates (Fig. 11). Collectively, these data reveal that self-association of IQGAP1 is necessary for it to modulate Cdc42.
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
Combining analysis in vitro and in intact cells, we established that IQGAP1 self-associates. To evaluate self-association in cells, we transfected human breast epithelial cells with EGFP2-IQGAP1. This approach permits discrimination of the transfected protein from endogenous IQGAP1 by two methods, namely altered migration on SDS-PAGE (the EGFP tags retard mobility) and recognition by anti-GFP antibodies. Endogenous IQGAP1 co-immunoprecipitated with EGFP-tagged IQGAP1, validating that IQGAP1 can self-associate in cells. (We verified by co-immunoprecipitation that adding an EGFP tag to the N terminus of IQGAP1 did not alter its interaction with several targets, including calmodulin, Cdc42, -catenin, and E-cadherin.3 Others (6) have also reported that an EGFP tag does not alter IQGAP1 function.) In vitro analysis similarly demonstrated that IQGAP1 can self-associate. Fusion proteins of full-length IQGAP1 or the N-terminal half of IQGAP1 were able to self-associate and could also bind one another. These data document that IQGAP1 self-associates.
Two prior studies have stated that IQGAP1 self-associates. In the first, IQGAP1 was estimated to have a molecular weight of 358,653 to 401,864 by sedimentation equilibrium (15). The authors concluded that these findings indicated that IQGAP1 comprises exclusively two subunits. The second study identified a very broad peak on gel filtration, which was interpreted to indicate that IQGAP1 existed as monomers, dimers, and trimers (16). The different methods of analysis, coupled with different sources of IQGAP1 protein (purified or expressed in bacteria), may account for the discrepancies between these two studies. Estimation of molecular weight can be correlated with elution on gel filtration only when the unknown protein has the same shape or conformation as the proteins used to calibrate the column (22). Moreover, neither of the prior studies performed detailed in vitro analysis to confirm dimerization. For these reasons, we conducted the work described here.
Coiled-coil proteins are characterized by a repeating pattern of seven residues, (abcdefg)n, with hydrophobic amino acids predominating at positions a and d of the heptad repeat (21). The coiled-coil is thought to be one of the principal motifs, which mediates subunit oligomerization of a large number of proteins (21). The N-terminal half of IQGAP1 contains six tandem repeats (1). It has been assumed previously, without detailed investigation, that this region of IQGAP1 mediates its dimerization (4, 6). On the basis of the putative coiled-coil domains between residues 168 and 601 of IQGAP1 (predicted by the COILS program), we anticipated that this region of IQGAP1 would mediate self-association. Experimental evidence did not support this hypothesis. The N-terminal half of IQGAP1 (amino acids 2–863), which includes the putative coiled-coil region, bound to full-length IQGAP1. However, deletion of the terminal 115 amino acids (residues 746–860, distal to the coiled-coil region) from this construct abrogated binding. Moreover, a peptide comprising the N-terminal 746 amino acids (which contains all the coiled-coils) failed to bind to either full-length IQGAP1 or the N-terminal half of IQGAP1 in vitro. Consistent with these findings, a peptide comprising residues 763–863 of IQGAP1 both bound to full-length IQGAP1 and was capable of self-association. Finally, excision of this region from IQGAP1 abrogated its ability to bind full-length IQGAP1, indicating that residues 763–863 are both necessary and sufficient to mediate the self-association of IQGAP1.
Our findings differ from a prior publication. By using GST fusion constructs of four selected portions of IQGAP1, amino acids 216–683 were reported to pull down endogenous IQGAP1 from bovine brain cytosol, whereas a peptide comprising amino acids 521–914 could not bind IQGAP1 in cell lysates (16). Neither in vitro analysis with pure proteins nor detailed characterization of the binding region was performed in that study. In contrast to that work, we were unable to detect binding of a peptide comprising residues 216–683 to full-length IQGAP1. Similarly, the 216–683 peptide did not self-associate. Several methodological differences may account for the differences between the studies. In the prior work, the buffer consisted of 20 mM Tris-HCl, 1 mM EDTA, and 1 mM dithiothreitol. Our assay conditions were more stringent. The buffer we used contained physiological concentrations of salt (150 mM NaCl) to minimize possible nonspecific electrostatic interactions. Next, the methods of analysis were different. We used in vitro translated products or cultured human breast epithelial cell lysate, whereas Fukata et al. (16) used bovine brain cytosol. It is possible that isolation of full-length IQGAP1 with the 216–683 fragment from bovine brain cytosol is mediated via binding to (an)other protein(s) in a ternary complex, rather than direct binding to the GST-tagged peptide. The presence of multiple bands visible on the silver-stained gel derived from the pull down with the 216–683 peptide in their study lends credence to this hypothesis.
It is well recognized that self-association regulates protein function (14, 23–25). In particular, the function of several scaffolding proteins is regulated by self-association (26, 27). Therefore, we analyzed the possible functional sequelae of self-association of IQGAP1 by using its effect on Cdc42 as a readout. IQGAP1 binds directly to Cdc42, maintaining Cdc42 in the active GTP-bound form (5, 28). Moreover, overexpression of full-length IQGAP1 and a dominant negative IQGAP1 construct increase and decrease, respectively, activated Cdc42 in cell lysates (17, 29). In contrast, overexpression of IQGAP1IQ, which lacks amino acids 699–905, had no effect on the amount of active Cdc42 (17). Here we observed that transfection of IQGAP1-(746–860) into cells failed to augment activated Cdc42 in cell lysates. Congruent with these results, a peptide that disrupted IQGAP1 self-association reduced the amount of activated Cdc42 in cell lysates. Collectively, these data suggest that oligomerization is necessary for IQGAP1 to increase activated Cdc42.
Some caveats of the functional data should be borne in mind. Protein self-association domains are often fairly large (30, 31), necessitating excision of several amino acids to prevent self-association. It is possible that removal of amino acid residues 746–860 from IQGAP1 may alter its conformation. For this reason, we complemented the functional analysis of the IQGAP1 deletion mutant with transfection of the peptide corresponding to the self-association region. The peptide, which specifically inhibited in vitro self-association in a dose-dependent manner, significantly reduced the amount of active Cdc42 in cells. Although it is not possible to unequivocally establish that the 717–916 peptide reduced active Cdc42 via IQGAP1, in combination with results of IQGAP1-(746–860), the findings strongly suggest that oligomerization is necessary for IQGAP1 to increase activated Cdc42. A second consideration is that IQGAP1-(746–860) lacks the major calmodulin binding region (5, 19), preventing it from associating effectively with calmodulin. However, binding to calmodulin abrogates the interaction of IQGAP1 with Cdc42 (8). Thus, one would anticipate that IQGAP1-(746–860) would exhibit an enhanced ability to increase active Cdc42, an effect opposite to what we observed.
In conclusion, our work is the first detailed characterization of IQGAP1 self-association. We document both in vitro and in intact cells that IQGAP1 can self-associate. More importantly, this study demonstrates for the first time that oligomerization of IQGAP1 is important for its function. The identification of the self-association domain presented here is likely to augment the ability of future studies to elucidate the molecular mechanisms and myriad biological functions of IQGAP1, and provides a unique motif for the genetic engineering of protein self-association domains.
|
FOOTNOTES |
|---|
1 To whom correspondence should be addressed: Brigham and Women's Hospital, Thorn 530, 75 Francis St., Boston, MA 02115. Tel.: 617-732-6627; Fax: 617-278-6921; E-mail: dsacks{at}rics.bwh.harvard.edu.
2 The abbreviations used are: GAP, GTPase-activating protein; Ni2+-NTA, nickel-nitrilotriacetic acid; GST, glutathione S-transferase; CHD, calponin homology domain; GRD, RasGAP-related domain; GBD, GTPase-binding domain; WASP, Wiskott-Aldrich syndrome protein; EGFP, enhanced green fluorescent protein; TEV, tobacco etch virus; TNT, transcription and translation; PMSF, phenylmethylsulfonyl fluoride; PVDF, polyvinylidene difluoride; FPLC, fast protein liquid chromatography; Tricine, N-[2-hydroxy-1,1-bis(hydroxymethyl)ethyl]glycine.
3 Z. Li and D. B. Sacks, unpublished observations.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
