Ankyrin repeat domain 28 (ANKRD28), a novel binding partner of DOCK180, promotes cell migration by regulating focal adhesion formation

E XP ER I ME NTAL C EL L R ES E ARC H 315 (2 0 0 9 ) 8 63 – 8 76

a v a i l a b l e a t w w w. s c i e n c e d i r e c t . c o m

w w w. e l s e v i e r. c o m / l o c a t e / y e x c r

Research Article

Ankyrin repeat domain 28 (ANKRD28), a novel binding partner of DOCK180, promotes cell migration by regulating focal adhesion formation Mitsuhiro Tachibana a,b , Etsuko Kiyokawa a,⁎, Shigeo Hara a,d , Shun-ichiro Iemura e , Tohru Natsume e , Toshiaki Manabe b , Michiyuki Matsuda a,c,d a

Department of Pathology and Biology of Diseases, Graduate School of Medicine, Kyoto University, Yoshida Konoe-cho, Sakyo-ku, Kyoto 606-8501, Japan b Department of Diagnostic Pathology, Graduate School of Medicine, Kyoto University, Kawahara-cho, Shogoin, Sakyo-ku, Kyoto 606-8507, Japan c Laboratory of Bioimaging and Cell Signaling, Graduate School of Biostudies, Kyoto University, Yoshida Konoe-cho, Sakyo-ku, Kyoto 606-8501, Japan d Department of Signal Transduction, Research Institute for Microbial Diseases, Osaka University, Suita, Osaka 565-0871, Japan e National Institute of Advanced Industrial Science and Technology (AIST), Biological Information Research Center (JBIC), Kohtoh-ku, Tokyo 135-0064, Japan

A R T I C L E I N F O R M AT I O N

AB ST R AC T

Article Chronology:

DOCK180 is a guanine exchange factor of Rac1 originally identified as a protein bound to an SH3

Received 30 July 2008

domain of the Crk adaptor protein. DOCK180 induces tyrosine phosphorylation of p130Cas, and

Revised version received

recruits the Crk-p130Cas complex to focal adhesions. To understand the role of DOCK180 in cell

1 December 2008

adhesion and migration, we searched for DOCK180-binding proteins with a nano-LC/MS/MS

Accepted 3 December 2008

system, and identified ANKRD28, a protein that contains twenty-six ankyrin domain repeats.

Available online 24 December 2008

Knockdown of ANKRD28 by RNA interference reduced the velocity of migration of HeLa cells, suggesting that this protein plays a physiologic role in the DOCK180-Rac1 signaling pathway.

Keywords:

Furthermore, knockdown of ANKRD28 was found to alter the distribution of focal adhesion

Ankyrin

proteins such as Crk, paxillin, and p130Cas. On the other hand, expression of ANKRD28, p130Cas, Crk,

Rac

and DOCK180 induced hyper-phosphorylation of p130Cas, and impaired detachment of the cell Cas

p130

membrane during migration. Consequently, cells expressing ANKRD28 exhibited multiple long

DOCK180

cellular processes. ANKRD28 associated with DOCK180 in an SH3-dependent manner and

Paxillin

competed with ELMO, another protein bound to the SH3 domain of DOCK180. In striking

Crk

contrast to ANKRD28, overexpression of ELMO induced extensive lamellipodial protrusion around

Migration

the entire circumference. These data suggest that ANKRD28 specifies the localization and the

Focal adhesion

activity of the DOCK180-Rac1 pathway. © 2008 Elsevier Inc. All rights reserved.

Introduction Cell migration is an important event during early development, inflammatory responses to infection, and wound healing, and an important pathological event during tumor invasion and metas⁎ Corresponding author. Fax: +81 75 753 4698. E-mail address: [email protected] (E. Kiyokawa). 0014-4827/$ – see front matter © 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.yexcr.2008.12.005

tasis. The family of Rho GTPases regulates this process through remodeling of the actin cytoskeletons, and by generating focal complexes and focal adhesions [1]. DOCK180 was originally identified as one of two major binding proteins of the adaptor protein Crk [2]. Later, genetic and

864

E XP ER I ME NTAL CE L L R ES E ARC H 315 (2 0 0 9 ) 8 63 – 8 76

biochemical studies revealed that DOCK180 functions as a guanine-nucleotide exchange factor (GEF) for a small GTPase Rac [3,4]. Orthologs of DOCK180 in C. elegans and D. melanogaster have been identified as Ced-5 (cell death abnormal 5) and Mbc (Myoblast city), respectively, and together with DOCK180 itself comprise an evolutionarily conserved protein group called the CDM (CED-5, DOCK180, MBC) family [5]. In addition, studies in C. elegans identified orthologs of Crk (Ced-2) and Rac (Ced-10) to show that the Crk-DOCK180-Rac signaling pathway is evolutionally conserved [6]. The localization and function of DOCK180 are regulated through interaction with binding partners. DOCK180 contains N-terminal SH3 and C-terminal proline-rich domains, which are required for ELMO and CrkII interaction, respectively [7,8]. Subsequent studies identified two other regions, designated DHR1 and DHR2 (or CZH1 and CZH2/DOCKER, respectively) [2,9], both of which showed high sequence homology among the 11 human DOCK180 superfamily proteins. Numerous recent findings have demonstrated that DHR2 functions as a GEF for Rho-family proteins [8,10–12]. In vitro and in vivo data indicated that phosphatidylinositol 3,4,5-trisphosphate (PIP3) binds to and recruits DOCK180 through DHR1 at the site of lamellipodia in response to platelet-derived growth factor [13]. Coexpression of CrkII and its binding protein p130Cas induces the accumulation of CrkII, p130Cas, and DOCK180 at the focal adhesions in NIH3T3 cells [14], and the enhancement of Rac GEF activity of DOCK180 toward Rac1 [3]. Ced-12 in C. elegans has been identified as a required gene for engulfment of apoptotic cells [7]. Its mammalian homolog ELMO binds to DOCK180 directly and cooperates with CrkII and DOCK180 to promote cell shape changes through enhancement of DOCK180-induced Rac1 activation [15]. It has been suggested that the SH3 domain of DOCK180 binds directly to the DHR2 of DOCK180 to inhibit the activity of the protein in the steady state [16], and that these binding proteins and lipids induce the conformational changes of DOCK180 upon stimulation. Here we identify ANKRD28 as a novel binding partner of DOCK180, and show a potential role of ANKRD28 in regulating focal adhesion for cell migration, of which mechanism is different from that of the other binding partner ELMO.

three contiguous residues at positions 1487 to 89, Ile-Ser-Pro, were replaced by Ala-Ala-Ala, as described previously [15] with accidental mutation by PCR at amino acid 1456 (Asn to Asp). The plasmids pEBB-mELMO1/2-wt-FLAG and pEBB-mELMO1/2-wtGFP were gifts from Dr. K.S. Ravichandran [7]. For vector-based knockdown, the short hairpin RNA (shRNA) sequences targeting human DOCK180 (DK1-#4: 5′-GTTTCTTCAGGACACGTTG-3′; DK1-#7: 5′-GTACGGAGATATGAGGAGA-3′), human ANKRD28 (ANK-#2: 5′-GTACCTTCTAGATCTTGGA-3′; ANK-#5: 5′-GGTGCTGCTGAGATGTTAA-3′), and firefly luciferase (Luc) (5′-GATTATGTCCGGTTATGTA-3′) were cloned into pSuper. retro.puro vector (OligoEngine, Seattle, WA) or pSuper-DsRed2, in which the puromycin-resistant sequence of pSuper.retro.puro was replaced by that for DsRed. For RNAi rescue experiments, shRNA (ANK-#5 and DK1-#4)-resistant cDNAs were created by introducing seven silent point mutations in the target sequences (GAGCGGCGGAAATGTTGATAGATA and TTCCTGCAGGATACCTTAGACGCT), generating pCAGGS-EGFP-ANKRD28R and DOCKR, respectively [19]. For retroviral infection, Moloney murine leukemia virus (MuLV)-based retroviral vector plasmids, pMSCVpac-3HA-ANKRD28, -ELMO2, pCX4bsr-EGFP-Cas, and pCX4bsr-mCherry-paxillin, were constructed [20].

Antibodies The polyclonal antibodies against DOCK180 and GFP were developed in our laboratory and described previously [2,21].The antibody against ANKRD28 was raised in a rabbit using a human ANKRD28 C-terminus peptide (corresponding to amino acids (aa) 1026–1040 CSFNNIGGEQEYLYT) (Covalab, Lyon, France). Other antibodies used in this study were purchased from the following suppliers: Rat anti-HA antibody (Roche, Basel, Switzerland); mouse anti-c-Raf-1 antibody, mouse anti-p130Cas antibody, mouse anti-phosphotyrosine (pTyr) antibody, mouse anti-Crk antibody, and mouse anti-Rac1 antibody (BD Biosciences, San Jose, CA); Mouse anti-FLAG monoclonal antibody (Sigma-Aldrich, St. Louis, MO); Mouse anti-α-tubulin antibody (EMD Chemical Inc., San Diego, CA); goat anti-ELMO2 antibody (IMGENEX, San Diego, CA); and Goat Alexa Fluor-conjugated anti-mouse and rabbit IgG (Invitrogen, Paisley, United Kingdom).

Cell culture, transfection, and retroviral infection

Materials and methods Plasmids cDNAs encoding the full-length ANKRD28 (KIAA0379) were obtained from Kazusa DNA Research Institute. To construct expression vectors, cDNA were amplified by PCR with the primer set GTCGACATGGCGTTCCTCAAACTCCGT (forward) and GCGGCCGCTCAGTAGGTCTCAGAATCGGA (reverse) and subcloned into pDrive vectors (QIAGEN, Hilden, Germany). After confirming the nucleotide sequences, the cDNA was digested with SalI and NotI and cloned into expression vectors to obtain pCAP-FLAGANKRD28 and pCAGGS-EGFP-ANKRD28. pCXN2-FLAG-DOCK180 and its deletion mutants (PS and dSH3) pCAGGS-EGFP-DOCK180, pSSRp130Cas, and pCAGGS-Myc-CrkII were described previously [14,17,18]. pCAGGS-DOCK180-1-357 and pCNX2-DOCK180-72-520 encode amonoacid residues 1–357 and 72–520 of DOCK180, respectively. pCXN2-FLAG DOCK180-ISP is a mutant in which

HeLa, HEK 293T, and BOSC23 cells were grown at 37 °C in Dulbecco's modified Eagle's medium (DMEM) (Sigma-Aldrich) supplemented with 1 μg/ml of penicillin-streptomycin and 10% fetal bovine serum. 293F cells were maintained at 37 °C in FreeStyle 293 expression medium (Invitrogen), and transfected with various plasmids using 293 fectin reagent (Invitrogen) according to the manufacturer's instructions. In some experiments, a MuLV-based retroviral system was utilized to obtain cells expressing the proteins of interest [20]. Briefly, the murine ecotropic retrovirus receptor EcoVR was first introduced by using viruses, which were produced from BOSC23 cells by transfection of pCX4hyg-EcoVR together with the plasmids pGP (the packaging plasmid) and pVSV-G (the envelope) [20,22]. After selection by hygromycin, the cells expressing EcoVR were infected with retrovirus for various tagged proteins. cDNA of the protein of interest was cloned into pMSCVpac- or pCX4bsr-based vectors, which are resistant to puromycin and blasticidin, respectively, and

E XP ER I ME NTAL C EL L R ES E ARC H 315 (2 0 0 9 ) 8 63 – 8 76

viruses were produced in BOSC23 cells. After treatment with the corresponding drugs, the cells were utilized for experiments.

Identification of interacting proteins 293T cells were transfected with pCXN2-FLAG-DOCK180 and binding proteins were analyzed by direct nano-flow LC (DNLC)/ MS/MS as described previously [23]. Briefly, samples were processed under stable flow using DNLC. The DNLC system consisted of the following components: a high-pressure syringe pump (KYA Technologies, Tokyo, Japan), a gradient device consisting of 10 channel solvent reservoirs (Valco, Houston, TX), a high-pressure pump-mixing module (Hewlett-Packard, Palo Alto, CA) and an ESI column packed with 1 μm reversedphase beads (Kanto Chemical, Tokyo, Japan). Samples processed using the DNLC system were injected at flow rates of 25–500 nl/ min and tandem mass spectra (MS/MS) were obtained.

Immunoprecipitation, SDS-PAGE, and Western blotting Twenty-four or 48 h after transfection, 293F cells were washed with Tris-buffered saline (TBS), and lysed with ice-cold lysis buffer [10 mM Tris–HCl pH 7.5, 5 mM EDTA, 1% Triton X-100, 10 mM NaF, 1 mM Na3VO4, 2 mM phenylmethylsulfunyl fluoride (PMSF), 10 μg/ml aprotinin]. The cleared lysates were incubated with anti-GFP antibody at 4 °C for 60 min. Proteins collected using Protein A-Sepharose beads (GE Healthcare, Little Chalfont, United Kingdom) were washed three times with ice-cold lysis buffer. For immunoblotting, proteins were separated by SDSPAGE and transferred to a polyvinylidene difluoride membrane, followed by detection with specific antibodies. The bound antibodies were detected by an enhanced chemiluminescence (ECL) detection system (GE Healthcare) and LAS-1000 image analyzer (Fuji-Film, Tokyo, Japan). In some experiments, a LI-COR Odyssey Infrared Imaging System scanner (LI-COR Biotechnology, Lincoln, NE) was utilized.

RNA Interference The stealth siRNA oligonucleotides against human ANKRD28 and ELMO2 were designed by and purchased from Invitrogen. The sequences were as follows: for ANKRD28-1: sense, 5′-UUAAUAAGUAUUCCAGGCACUUGCC-3′; for ANKRD28-2: sense, 5′AUGAUAGGCAGCCAAGUGUAAAGGG-3′; for ANKRD28-3: sense, 5′-UUCAAUGUGACCCAUAUAUGCUGCC-3′; for ELMO2-1: sense, 5′-AUAACGGAGGGUAUAAUACUCUGGG-3′; for ELMO2-2: sense, 5′-AAUGCCAGCAUCUCACUGUAGUGGG-3′; and for ELMO2-3: sense, 5′-AAUGGCGUAGGUCUGAAUCUCCUGG-3′. Stealth TM RNAi Negative Control Duplexes (Invitrogen) were utilized as a negative control. Cells were seeded at a density of 2 × 105 cells per well on collagen-coated 6 well plates (Iwaki Glass, Tokyo, Japan) the day before plasmid transfection, and were transfected with siRNA oligonucleotides using RNAi MAX (Invitrogen) according to the manufacturer's protocol.

Immunofluorescence microscopy Fluorescent images were obtained using an inverted microscope (Carl Zeiss, Jena, Germany), or by inverted epifluorescence microscopy using IX71 and IX81 microscopes (Olympus, Tokyo,

865

Japan). All three microscopes were equipped with an Orca II cooled charge coupled device (CCD) camera (Hamamatsu Photonics, Hamamatsu, Japan), and images were processed and analyzed by MetaMorph software (Molecular Devices, Sunnyvale, CA). In some experiments, cells were imaged with an FV1000 confocal microscope equipped with an Argon and HeNe laser (Olympus) and a 60× oil-immersion objective lens (UPlanSApo).

Wound healing assay Wound healing assays were performed essentially as described previously [24]. HeLa cells were seeded at 1 × 105 cells on collagen-coated 35-mm glass bottom dishes (Iwaki Glass), and transfected with pCA-EGFP together with pSuper-DOCK180, -ANKRD28 and -Luciferase by 293fectin. Forty-eight hours after transfection, they were wounded by scraping with a P2 tip, rinsed with phosphate-buffered saline (PBS), and fed with fresh medium containing 10% serum. The medium was covered with mineral oil (Sigma-Aldrich) to preclude evaporation. Beginning at 2 h after wounding, cells were imaged every 15 min for 16 h at the same spots using a Zeiss inverted microscope, or an Olympus IX71 or IX81 inverted microscope. All microscopes were equipped with an objective lens (×10, NA 0.3) and a CCD camera. By tracking the GFP-expressing cells, migrated distance was measured by MetaMorph software. In some experiments, siRNA duplexes (20 nM) were used instead of transfection of shRNA vectors, and the migrated area was measured by MetaMorph software.

Analysis of focal adhesion proteins HeLa cells were transfected with various pSUPER plasmids, treated with puromycin to select transfectants, and further transfected with pCA-RFPN-CrkII. Forty-eight hours after the second transfection, cells were replated onto collagen-coated glass-bottom dishes. Ninety minutes after replating, cells were fixed in 3% paraformaldehyde (PFA)/PBS for 10 min at room temperature. Images were collected using an IX71 Olympus inverted microscope equipped with a 60× objective lens (NA 1.4) and a CCD camera, and analyzed using MetaMorph software. In some experiments, cells were treated with various siRNA duplexes in place of pSUPER transfection and puromycin selection. To analyze p130Cas and paxillin, HeLa cells expressing EcoV receptors were infected with retroviruses for EGFP-p130Cas or mCherry-paxillin, and further treated with various siRNAs and processed similarly.

Northern blotting analysis The filter blotted with mRNA from adult tissue was purchased from Clontech (Mountain View, CA). The two cDNAs of the probe of ANKRD28, corresponding to base pairs 1–1502 and 1502– 3169, were purified and labeled with 32Pi using a Prime It random primer labeling kit (Stratagene, La Jolla, CA) according to the manufacturer's protocol. The filter was incubated with the probes in hybridization solution (Clontech) for 2 h at 55 °C and 2 h at 68 °C, and then washed in buffer A (2× SCC and 0.05% SDS) at room temperature, followed by washing in solution B (0.1× SSC and 0.1% SDS) at 50 °C for 40 min. The filter was exposed and analyzed with a FUJI BAS2000 image analyzer. After

866

E XP ER I ME NTAL CE L L R ES E ARC H 315 (2 0 0 9 ) 8 63 – 8 76

Table 1 – Proteins identified as potential binding partner of DOCK180 Protein name ANKRD28 CEBPE ⁎ ELMO2 LOC129138 ⁎⁎

Results a 2/2 2/2 1/2 1/2

a Positive results/total experiments. ⁎ CCAAT/enhancer binding protein (C/EBP), epsilon. ⁎⁎ Hypothetical protein BC014641.

extensive washing, the same filter was similarly processed using the probes of β-actin.

which of them is the predominant form. In this manuscript, we refer to the protein sequence shown in Supplemental Fig. S1A (NCBI-GeneID 23243, NP_056014, GenBank accession no AB002377) as the ANKRD28 protein. ANKRD28 is mostly comprised of 26 ankyrin repeat domains and intercalating linker domains. The ankyrin repeat, a 33-residue sequence motif, folds the canonical helix-loop-helix-β-hairpin/loop and mediates protein-protein interactions [27]. Although the ankyrin repeat domains of ANKRD28 showed 45% homology and 30% identity to ankyrin 2 (NCBI-GeneID 287, GenBank accession no AC004057), ANKRD28 lacks the C-terminal regulatory domains of ankyrin 2, including the spectrin-binding site. ANKRD28 mRNA was expressed ubiquitously in various human tissues when examined by Northern blotting (Supplemental Fig. S1B).

Interaction of DOCK180 and ANKRD28

Results ANKRD28 is a novel binding partner of DOCK180 To better understand the function of DOCK180, we searched for cellular proteins that physically associate with DOCK180 in mammalian cells. For this purpose, FLAG-DOCK180 was expressed in 293T cells and immunoprecipitated from cell lysates with an anti-FLAG antibody. The immunoprecipitates were eluted with FLAG peptides, digested with Lys-C endopeptidase, and analyzed using a highly sensitive direct nano-flow liquid chromatography/ tandem mass spectrometer (LC-MS/MS) as described previously [23]. We identified four proteins that interacted with DOCK180, as shown in Table 1. Among the listed proteins, Ankyrin repeat protein 28 (ANKRD28) was found to associate with DOCK180 repeatedly, validating further analyses. Recently, ANKRD28 was identified as a PP1-targeting subunit that modulates the phosphorylation of the transcriptional regulator hnRNP K, and was therefore alternatively designated Phosphatase Interactor Targeting K protein (PITK) [25]. In another study, ANKRD28 was identified as one of the PP6-binding partners, and was referred to as PP6-ARS-A [26]. However, as discussed in the supplemental materials (Supplemental Fig. S1A), the differences among the reported amino acid sequences of ANKRD28, PITK, and PP6-ARS-A are limited to their N-termini, and it is not clear at this moment

To confirm the binding of DOCK180 to ANKRD28, 293F cells expressing HA-DOCK180 and GFP-ANKRD28 were subjected to coimmunoprecipitation analysis with anti-GFP antibody (Fig. 1A). HA-DOCK180 was found to associate with GFP-ANKRD28, while HA-Sos1 used as a negative control did not, indicating the specificity of the interaction between DOCK180 and ANKRD28. Other proteins containing ankyrin repeats, such as ACAP1 and GIT1/2, did not bind to DOCK180 (data not shown). We next examined the interaction between DOCK180 and ANKRD28 at the physiological expression level. Before this analysis, we generated specific antisera against ANKRD28 (Supplemental Fig. S2) and quantified the molecule numbers of the endogenous DOCK180, ANKRD28, and ELMO2 in HeLa cells, essentially as described previously [19,31] (Supplemental Fig. S3 and Table 2). DOCK180 was approximately ten times fewer than ANKRD28 and ELMO2. Because the ANKRD28 antibody prepared in our laboratory could precipitate only ca. 10% of total ANKRD28, the amount of co-precipitated DOCK180 was below the detection limit (data not shown). To overcome this problem, we stably expressed GFP-tagged ANKRD28 in HeLa cells (Fig. 1B). In this cell line, the sum of endogenous and exogenous ANKRD28 proteins in the GFP-ANKRD28-expressing cells was twice as much as the endogenous ANKRD28 in the parental cells. Using these cells, we could show the association of the endogenous

Fig. 1 – ANKRD28 binds to DOCK180. (A) 293F cells were transfected with various combinations of plasmids of pCA-3HA-DOCK180, pCA-3HA-FKBP-Sos1, pCA-EGFP-ANKRD28, and pCA-EGFP, which encode proteins denoted as HA-DOCK180, HA-Sos1, GFP-ANKRD28, and GFP, respectively, as indicated at the top of the figure. After 24 h, cells were lysed and supernatants were incubated with anti-GFP antibody, and collected with protein A-Sepharose. After washing beads with ice-cold lysis buffer, bound proteins (denoted as IP: immunoprecipitation) as well as total lysate (denoted as TCL: total cell lysate) were separated by SDS-PAGE and detected by Western blotting with the antibodies against HA and GFP indicated at right. At least two independent experiments demonstrated similar results. (B) HeLa cells were stably infected with a recombinant retrovirus encoding GFP-ANKRD28. Lysate from the parental (left lane) and GFP-ANKRD28 expressing (right lane) HeLa cells were subjected to SDS-PAGE and Western blotting with the indicated antibodies. The closed and open arrow heads denote the endogenous and GFP-tagged ANKRD28, respectively. (C) GFP-ANKRD28 expressing cells were lysed and supernatants were incubated with preimmune or anti-GFP antibody conjugated to Protein A sepharose beads. Proteins bound to the beads were washed and separated by SDS-PAGE, and detected with indicated antibodies. (D–F) 293F cells were transfected with plasmids coding FLAG-tagged DOCK180 (denoted as WT) and its mutants (PS, dSH3, 1–357, and 72–520) together with GFP-tagged ANKRD28. Twenty-four or forty-eight hours later, cells were lysed and processed similarly as in A. In (E), to visualize dSH3 protein, the corresponding plasmid was increased for transfection, and processed similarly as in (D). TCL, total cell lysates; IP, immunoprecipitates. (G, H) 293F cells were transfected with various plasmids for the proteins indicated at the top of the figure and analyzed as described for A. The amount of plasmid for FLAG-ELMO1 (F) or -ANKRD28 (G) was increased from 0.5 to 2.0 μg as shown. TCL, total cell lysates; IP, immunoprecipitates.

E XP ER I ME NTAL C EL L R ES E ARC H 315 (2 0 0 9 ) 8 63 – 8 76

DOCK180 with GFP-ANKRD28 protein immunoprecipitated with anti-GFP antiserum (Fig. 1C), indicating that the expression levels of endogenous ANKRD28 and DOCK180 are sufficient to form the ANKRD28-DOCK180 complex.

867

The SH3 domain of DOCK180 binds to ANKRD28 DOCK180 contains the N-terminal SH3 domain (amino acids 1– 71) and the C-terminal proline-rich region (1799–1843) [2]. To

868

E XP ER I ME NTAL CE L L R ES E ARC H 315 (2 0 0 9 ) 8 63 – 8 76

Table 2 – Molecule numbers of DOCK180, ANKRD28 and ELMO2 in HeLa cell Number (/cell) DOCK180 ⁎ ANKRD28 ELMO2

5

3.9 × 10 1.7 × 106 1.4 × 106

HeLa cells were transfected with plasmids encoding 3HA-ANKRD28 and GFP-ELMO2, and, one day later, lysed in SDS-lysis buffer. Proteins were separated by SDS-PAGE and quantified by Western blotting with a standard protein, GST-5myc-FLAG-3HA-YFP, as described previously [31]. The tagged proteins expressed in HeLa cells were then used as the standard proteins to quantify DOCK180, ANKRD28, and ELMO2 by Western blotting with antibodies raised against each protein. The actual blotting data are shown in Supplemental Figure S3. Results shown here are averages of three independent experiments. ⁎ obtained from [19].

investigate the domain responsible for binding of DOCK180 to ANKRD28, the truncated mutants of DOCK180, named PS and dSH3, which encoded aa 1–1472 and aa 73–1865, respectively, were expressed with GFP or GFP-ANKRD28, followed by coimmunoprecipitaion with anti-GFP antibody (Fig. 1D). The expression level of the dSH3 mutant was significantly lower than that of the wild type; therefore, we increased the amount of plasmid for dSH3, and re-examined its binding to ANKRD28 protein (Fig. 1E). Even under this condition, the dSH3 protein was not detected in the complex with ANKRD28, while the other mutant PS was found to interact with ANKRD28 very efficiently. These data indicated that the SH3 domain of DOCK180 was required for the association with ANKRD28. To examine whether the N-terminus of DOCK180 was sufficient for ANKRD28 interaction, we generated two deletion mutants of DOCK180, which encoded aa 1–357 and 72–520, respectively. As shown in Fig. 1F, the mutant 1–357, but not 72–520, was co-immunoprecipitated with GFP-ANKRD28, indicating that N-terminus of DOCK180 comprising of the SH3 domain was sufficient for the binding to ANKRD28.

ELMO competes with ANKRD28 for DOCK180 binding It has been reported that aa 1–357 of DOCK180, which include the SH3 domain, are required for interaction with ELMO [15] and that the SH3 domain of DOCK180 also binds intramolecularly to the catalytic DHR2 domain [16]. We next examined whether ELMO1 competed with ANKRD28 for DOCK180 binding. For this, FLAGDOCK180 and GFP-ANKRD28 were co-expressed in 293F in the presence of increasing amounts of FLAG-ELMO1. We found that ELMO1 expression decreased the amount of DOCK180 associated with ANKRD28 (Fig. 1G). The same immune complex did not contain any ELMO1 (data not shown). The reverse experiment with increasing amounts of FLAG-ANKRD28 indicated that ANKRD28 expression also decreased the amount of DOCK180 bound to ELMO1, albeit to a lesser extent (Fig. 1H). Therefore, we concluded that ANKRD28 and ELMO utilized an overlapping region of DOCK180 for their association.

ANKRD28 is required for cell migration Depletion of DOCK180 in C. elegans reduces distal tip cell migration [5]. On the other hand, overexpression of DOCK180 together with CrkII and ELMO enhances haptotaxis and cell spreading in mammalian cells [13,28]. We therefore examined whether ANKRD28 regulates cell migration, as does DOCK180. To knockdown ANKRD28 and DOCK180 proteins by the vectorbased expression of short hairpin RNA (shRNA), we prepared pSUPER-DOCK180 and pSUPER-ANKRD28. HeLa cells were transfected with pSUPER vectors and selected with puromycin (Fig. 2A). Efficient knockdown of the targeted proteins was confirmed by immunoblotting with antiserum against ANKRD28 that was prepared as described in the Materials and methods section (Supplemental Fig. S3). Next, we examined the effect of knockdown of DOCK180 and ANKRD28 in the wound healing assay. As shown in Fig. 2B, the negative control cells transfected with pSUPER-Luc migrated remarkably faster than did the cells transfected with pSUPER-DOCK180 or pSUPER-ANKRD28. To

Fig. 2 – Suppression of ANKRD28 protein decreases wound closure. (A) HeLa cells were transfected with shRNA targeting luciferase (shown as Luc), DOCK180 (DK1-#4 and -#7), and ANKRD28 (ANK-#2 and -#5) as indicated above. After selecting the transfected cells with puromycin for 2–3 days, cells were lysed and analyzed by SDS-PAGE, and Western-blotting with the antibodies indicated at right. (B) HeLa cells were transfected with pCA-EGFP, together with shRNA plasmids for luciferase, DOCK180, and ANKRD28 (denoted as pSUPER-Luc, pSUPER-DK1-#4, and pSUPER-ANK28-#5, respectively) at the ratio 1:6. HeLa cell monolayers were wounded by scraping and observed by epifluorescent microscopy 2 h later. Images for GFP (denoted as GFP), phase contrast (denoted as PC), or differential interference contrast (denoted as DIC) were obtained every 15 min for 16 h. The images shown here are those at time points 0 and 16 h. The white arrows indicate GFP-positive cells at the initial position of the wound edge. Scale bar = 300 μm. (C) Quantification of velocity. From the images obtained in (B), the migration distance was calculated by tracking the GFP-expressing cells. The numbers of cells examined were as follows: Luc (control) (n = 39), DK1-#4 (n = 14), ANK-#5 (n = 22). Bars = standard error of the mean (s.e.m.). ⁎p < 5 × 10− 6 (Student's t-test); ⁎⁎P < 3 × 10− 10 (Welch's t-test). (D) 293T cells were transiently transfected with plasmids for GFP-tagged ANKRD28 (left column; EGFP-ANKRD28) or shRNA-resistant ANKRD28 (right column; EGFP-ANKRD28R), together with pCAGGS [denoted as (-)], pSUPER-Luciferase (Luc) or -ANKRD28 (ANK). Forty-eight hours after transfection, cells were lysed, separated by SDS-PAGE, and analyzed by Western blotting with the antibodies denoted at right. (E) HeLa cells were transfected with the plasmid pSUPER-Ds-Red2-ANKRD28 #5 (denoted as -ANK-#5) or -Luciferase (Luc), together with pCA-EGFP-ANKRD28R (GFP-ANKR) and EGFP (GFP). After transfection, cells were processed similarly as in Fig. 2B. The white arrows indicate the green and red fluorescent protein-expressing cells. Scale bar = 300 μm. (F) Quantification of velocity. From the images obtained in (E), the migration distance was calculated by tracking the GFP-expressing cells. The cell numbers examined were as follows: GFP-ANKR + pSUPER-DsRed2-ANK (n = 9), GFP-ANKR + pSUPER-DsRed2-Luc (n = 21), and GFP + pSUPER-DsRed2-ANK (n = 21). ⁎p < 8 × 10− 6 (Student's t-test); ⁎⁎p = 0.098 (Welch's t-test).

E XP ER I ME NTAL C EL L R ES E ARC H 315 (2 0 0 9 ) 8 63 – 8 76

quantify the wound healing assay, the length of the migration path of each GFP-positive cell was measured and analyzed with Metamorph software. As shown in Fig. 2C, the migration distance of control cells was 222 ± 91 μm/16 h, while those of DOCK180and ANKRD28-knockdown cells were 88 ± 60 μm and 65 ± 41 μm,

869

respectively, indicating that DOCK180 and ANKRD28 were required for cell migration during wound healing. To confirm the specificity of ANKRD28 knockdown, we prepared pCAGGSEGFP-ANKRD28R, an expression vector for the ANKRD28 mRNAs that are resistant to the aforementioned shRNA (Fig. 2D). Using

870

E XP ER I ME NTAL CE L L R ES E ARC H 315 (2 0 0 9 ) 8 63 – 8 76

this plasmid, we examined whether or not the shRNA-resistant ANKRD28 protein could antagonize the effect of shRNA for ANKRD28 (Fig. 2E). As shown in Fig. 2F, expression of GFPANKRD28R, but not GFP, counteracted the effect of ANKRD28 knockdown, confirming that ANKRD28 is required for efficient cell migration.

ANKRD28 is required for CrkII localization to the focal adhesion It has been shown that DOCK180 is recruited to the focal adhesion in the presence of CrkII and p130Cas [14]. We therefore examined the distribution of RFP-CrkII in DOCK180- or ANKRD28-depleted cells to investigate their role in the formation of focal adhesions (Fig. 3A). In the control cells, a substantial fraction of RFP-CrkII was concentrated at needle-like structures, probably the focal adhesion. This patchy accumulation of RFP-CrkII was observed mostly at the cell periphery, and thus we will refer to this distribution as a “peripheral pattern” hereafter (Fig. 3B, left panel). In contrast, in cells depleted of DOCK180 or ANKRD28 proteins, RFP-CrkII showed two additional distribution patterns. The first of these was a dispersed pattern, in which RFP-CrkII was observed in patchy structures appearing not only in the periphery but also in the central area of the cells (Fig. 3B, second and fourth panels). The second was a diffuse pattern, in which RFP-CrkII was located mostly in the cytoplasm without showing any patchy structure (Fig. 3B, third and fifth panels). The cells were classified into one of these three categories and the percentage of each phenotype is summarized in Fig. 3C. When DOCK180 or ANKRD28 was depleted from the cells, the distribution of RFP-CrkII was shifted from the peripheral pattern to the dispersed and diffuse patterns, indicating that both DOCK180 and ANKRD28 were required for the proper delivery of RFP-CrkII to the peripheral region of the cells. Because ELMO competes with ANKRD28 in the DOCK180 binding, we also examined the distribution of RFP-CrkII in cells depleted of ELMO2, which is a dominant form of ELMO proteins in HeLa cells. Since the preparation of shRNA for ELMO2 was unsuccessful, the small interfering RNA (siRNA)-mediated knockdown technique was used to deplete ELMO2 and ANKRD28 in this experiment (Fig. 3D). siRNA-mediated knockdown of ANKRD28 reproduced the effect of shRNA in the distribution of RFP-CrkII (Fig. 3E). Furthermore, we observed that ELMO2 knockdown also increased the fraction of cells showing the dispersed pattern of RFP-CrkII distribution, albeit to a lesser extent. In agreement with this observation, the siRNA for ELMO2 significantly inhibited cell migration in the wound healing assay (Supplemental Fig. S4). It has been reported that ELMO depletion induces DOCK180 ubiquitylation and thereby its degradation [29]. In fact, we confirmed that depletion of ELMO2 protein reduced DOCK180 protein, as reported previously, whereas depletion of ANKRD28 did not alter the amount of DOCK180 (Supplemental Fig. S5). Therefore, we concluded that the observed effect of ELMO2 depletion on the CrkII distribution was caused by the reduction in DOCK180 protein.

ANKRD28 is required for p130Cas and paxillin localization to the focal adhesion CrkII is recruited to the focal adhesion by the interaction with phosphorylated tyrosine residues on p130Cas [30], which urged us

to examine whether depletion of ANKRD28 or ELMO2 perturbs p130Cas recruitment to focal adhesions. For this, we used HeLa cells expressing EGFP-p130Cas (Fig. 4A). In control cells, the distribution of EGFP-p130Cas was similar to that of CrkII, which was categorized as the peripheral pattern (Fig. 4B). In clear contrast, the distribution of EGFP-p130Cas was altered in cells depleted of ELMO2 or ANKRD28. As shown in Fig. 4C, ca. 40% of cells depleted of ELMO2 or ANKRD28 were classified into the dispersed pattern. Similar experiments were conducted with cells expressing mCherry-paxillin (Figs. 4D-F). Again, mCherry-paxillin was localized irregularly in cells depleted of ELMO2 or ANKRD28. To avoid examiner's variance, we performed a blind test in which two experienced pathologists classified the mCherry-paxillin-expressing cells into either the peripheral or the dispersed pattern without any information (Supplemental Fig. S6). Although the absolute numbers varied between the two examiners, they reached the same conclusion, i.e., the mCherry-paxillin cells classified as having the irregular pattern appeared more frequently in ANKRD28-depleted cells than control cells. To confirm that ANKRD28 depletion caused the mislocalization of paxillin, we expressed shRNA-resistant ANKRD28 mRNA in the presence of shRNA. The effect of shRNA-resistant DOCK180 was also examined in parallel (Supplemental Figs. 7A and B). As expected, co-expression of the shRNA-resistant mRNAs restored the original localization of paxillin (Supplemental Fig. 7C). These observations strongly suggested that ANKRD28 cooperated with DOCK180 to localize the focal adhesion proteins, CrkII, p130Cas, and paxillin, to the peripheral region of the cells.

Expression of ANKRD28 induces hyperphosphorylation of p130Cas We previously showed that expression of DOCK180 enhances p130Cas-CrkII complex formation and the tyrosine phosphorylation of p130Cas on multiple tyrosine residues [14]. This observation urged us to examine the effect of ANKRD28 expression on the p130Cas phosphorylation. Twenty-four hours after transfection, cells were either kept in suspension, or plated onto the collagencoated dish and cultured for another 24 h. The electro-mobility of p130Cas in SDS-polyacrylamide gels was reduced in the presence of CrkII and DOCK180 as described previously [14] (Fig. 5). Importantly, this reduced electro-mobility of p130Cas was observed only when cells were attached to the collagen-coated dish (compare the 5th lanes of suspension and adherence conditions), suggesting that integrin stimulation was required for p130Cas phosphorylation. Furthermore, co-expression of ANKRD28 protein increased the fraction of the slowly migrating form of p130Cas. We confirmed that the slowly migrating form of p130Cas in ANKRD28-expressing cells was the tyrosine-phosphorylated form by probing with an antiphosphotyrosine antibody (the second top panel). Notably, expression of ANKRD28 in the absence of DOCK180 did not induce p130Cas hyperphosphorylation.

The SH3 domain and Rac GEF activity of DOCK180 are required for p130Cas hyperphosphorylation To understand the mechanism by which DOCK180 and ANKRD28 induce p130Cas hyperphosphorylation, we expressed a series of DOCK180 mutants in the presence of CrkII and ANKRD28. The DOCK180 mutants without GEF function for Rac, ISP and PS [15]

E XP ER I ME NTAL C EL L R ES E ARC H 315 (2 0 0 9 ) 8 63 – 8 76

failed to induce p130Cas hyperphosphorylation, implying that Rac1 activation may be involved in this process. A mutant without the SH3 domain, dSH3, also did not induce p130Cas hyperphosphorylation, suggesting that an SH3 domain-binding protein(s), possibly ANKRD28, is also necessary for this phenomenon. To examine whether ELMO induces p130Cas hyperphosphorylation, as does ANKRD28, GFP-ELMO was expressed with DOCK180 and CrkII and plated onto collagen-coated dishes. In contrast to

871

cells expressing ANKRD28, the hyperphosphorylation of p130Cas was not observed in ELMO-expressing cells (Fig. 5C).

ANKRD28, but not ELMO2, induced cell-tale elongation Lastly, we compared the effect of ANKRD28 and ELMO2 on the cell shape. 293F cells were transfected with plasmids for these proteins, together with those for CrkII, p130Cas, and DOCK180

Fig. 3 – Suppression of ANKRD28 and DOCK180 induces mislocalization of CrkII from focal adhesions. (A) HeLa cells were transfected with shRNA targeting luciferase (shown as Luc), DOCK180 (DK1-#4), and ANKRD28 (ANK-#5). Four days after transfection, cells were lysed, separated by SDS-PAGE, and analyzed with antibodies indicated at right. (B) HeLa cells were transfected with one of three shRNAs – pSUPER-Luciferase (Luc), pSUPER-DOCK180-#4 (DOCK180 KD), or pSUPER-ANKRD28-#5 (ANKRD28 KD) – and treated with puromycin to select the transfectants. Forty-eight hours later, cells were further transfected with pCAGGS-RFPN-CrkII (denoted as RFP-CrkII). Forty-eight hours after the second transfection, cells were trypsinized, replated on the collagen-coated glass-bottom dishes for 90 min, and fixed with 3% PFA. Fluorescent images were obtained by epifluorescence microscopy, and representative images for three groups, Peripheral, Dispersed, and Diffuse, are shown. Scale bar = 20 μm. (C) Quantification of RFP-CrkII localization. The HeLa cells shown in B were divided into three groups: Peripheral (black bar), Dispersed (open bar), and Diffuse (gray bar). The cell numbers examined were as follows: pSUPER-Luc transfected cells (denoted as Luciferase; n = 27, 48, and 40), pSUPER-DK1-#4 (denoted as DOCK180, n = 61, 66 and 57), and pSUPER-ANK-#5 (denoted as ANKRD28, n = 97, 85 and 63). The averages ± s. e. m. from three independent experiments are shown. (D) HeLa cells were transfected with control siRNA, siRNA targeting ANKRD28 (ANKRD28-#1, -#2, -#3), and siRNA targeting ELMO2 (ELMO2-#1, -#2, -#3) Seventy-two hours after transfection, the cells were lysed and the proteins were separated by SDS-PAGE and analyzed by Western-blotting using the antibodies denoted at right. (E) Quantification of RFP-CrkII localization. RFP-CrkII-expressing HeLa cells were treated with siRNA and processed as in B. The cell numbers examined were as follows: Control (n = 24, 48, and 32), ANKRD28-#3 (n = 37, 35, and 29), and ELMO2-#3 (n = 20, 26, and 23). Data are the means ± s. e. m. from three independent experiments.

872

E XP ER I ME NTAL CE L L R ES E ARC H 315 (2 0 0 9 ) 8 63 – 8 76

(Fig. 6A). Expression of p130Cas and CrkII did not significantly alter the polygonal shape of 293F cells (left upper panel). However, in cells expressing DOCK180, p130Cas, and CrkII, we observed elongated cellular branches (right upper panel). Additional ANKRD28 expression strengthened this phenotype, changing the elongated branch to a neurite-like extension (left lower panel). In clear contrast, expression of ELMO2 induced marked lamellipodia along the entire circumference (right lower panel), as observed by Gumienny et al. with LR73 cell expressing ELMO1/DOCK180/CrkII

[7]. To investigate the mechanism by which the elongated cellular processes were generated, cells were plated onto collagen-coated dishes and time-lapse imaged (Fig. 6B). In these randomly migrating cells, the elongated branches were generated at the trailing edge of the cells, seemingly because the detachment of the cellular processes was impaired. We did not observe active extension of the cellular processes toward the forward direction. For the quantification of cell shape, the following three indexes were applied: the number of elongated tails, the longest distance

Fig. 4 – Suppression of ANKRD28 and ELMO2 induces mislocalization of both p130Cas and paxillin from focal adhesions. (A, D) GFP-p130Cas (A) and mCherry-paxillin (D)-expressing HeLa cells were transfected without siRNA (-), or with control siRNA (Cont.), those targeting ELMO2-#3 (ELMO), and ANKRD28-#3 (ANK). Seventy-two hours after transfection, the cells were lysed and the proteins were separated by SDS-PAGE and analyzed by Western-blotting using the antibodies denoted at right. Two independent experiments demonstrated similar results. (B, E) Representative confocal images of GFP-p130Cas (B) and mCherry-paxillin (E) expressing HeLa cells transfected with various siRNA duplexes to deplete proteins. Images were obtained with confocal microscopy. Scale bar = 10 μm. (C, F) Quantification of GFP-p130Cas (C) and mCherry (F) localization. The HeLa cells shown in B were divided into two groups: Peripheral (black bars) and Dispersed (open bars). The cell numbers examined in C were as follows: Non-siRNA-treated cells (n = 182 and 174), Control (n = 168 and 203), ELMO2-#3 (n = 223 and 159), and ANKRD28-#3 (n = 160 and 202). The cell numbers examined in F were as follows: Non-siRNA-treated cells (n = 146 and 193), Control (n = 178 and 160), ELMO2-#3 (n = 251 and 105), and ANKRD28-#3 (n = 160 and 147). The averages± deviation from two independent experiments are shown.

E XP ER I ME NTAL C EL L R ES E ARC H 315 (2 0 0 9 ) 8 63 – 8 76

from the tip of extended cellular processes to the center of the nucleus, and the cell area (Fig. 6C). The expression of DOCK180 in cells expressing CrkII and p130Cas increased the number and the length of the elongated cellular processes, which were significantly enhanced by the expression of ANKRD28. ELMO slightly decreased the number of cellular processes, had no effect on their length and, most strikingly, increased the cell area. The last phenotype of ELMO-expressing cells was apparently caused by the marked induction of lamellipodia. Thus, two proteins bound to the SH3 domain of DOCK180 are involved in cell migration by very distinct mechanisms.

Discussion In this study, we have identified ANKRD28 as a novel binding partner of DOCK180 by nano-LC/MS/MS. Knockdown of ANKRD28

873

as well as DOCK180 perturbed the localization of CrkII, p130Cas, and paxillin to the focal adhesions, and inhibited cell migration, which observation strongly argues for the physiological relevance of this binding. It is unlikely that ANKRD28 regulates the CrkII-DOCK180 complex by any known enzymatic mechanisms, considering its amino-acid sequence. Rather, based on the observation that depletion of ANKRD28 mislocalized CrkII, p130Cas, and paxillin in focal adhesions (Figs. 3 and 4), we speculate that ANKRD28 functions as a scaffold protein to assemble components of focal adhesion proteins such as CrkII, DOCK180, p130Cas, and paxillin (summarized in Fig. 7). In support of this hypothesis, we showed that ANKRD28 overexpression enhanced phosphorylation of the p130Cas protein in the presence of CrkII and DOCK180 (Fig. 5A). It has been shown that DOCK180 upregulates the complex formation of CrkII and p130Cas, concomitant with p130Cas hyperphosphorylation [14]. It has also been observed that DOCK180 overexpression induces accumulation of p130Cas, DOCK180, and CrkII at focal adhesions, strongly supporting the idea that the complex of p130Cas-CrkII-DOCK180 regulates cell adhesion at focal adhesions. Using the purified pseudopodia from migrating cells, Klemke and colleagues have shown that the Crk/130Cas complex is formed in the growing pseudopodia to upregulate Rac [32-34]. Examination of the presence of DOCK180, ELMO, and/or ANKRD28 in this purified pseudopodia would also shed light on the differential roles of ELMO and ANKRD28. ANKRD28 contains 26 ankyrin repeats, the functions of which are to mediate protein-protein interactions [27]. Furthermore, tandem ankyrin repeats exhibit tertiary-structure-based elasticity and behave as a liner and fully reversible spring [35]. It should be noted that mechanical strain applied to cells can induce p130Cas extension and tyrosine phosphorylation at the cell periphery [36]. Therefore, it is tempting to speculate that ANKRD28 may cooperate

Fig. 5 – Overexpression of ANKRD28 enhances p130Cas hyper-phosphorylation induced by the CrkII-DOCK180 complex. (A) 293F cells, which were cultured in suspension, were transfected with various combinations of pSSRα-p130Cas (shown as p130Cas), pCAGGS-FLAG-DOCK180 (FLAG-DOCK180), pCAGGS-EGFP-ANKRD28 (GFP-ANK), and pCAGGS-Myc-CrkII (Myc-CrkII) as indicated at the top of the figure. Twenty-four hours later, one half of the total cells were plated onto collagen-coated dishes (the Adherence group), and the other half were maintained in suspension (the Suspension group). Twenty-four hours after plating, cells were lysed, separated by SDS-PAGE, and analyzed by Western-blotting. The corresponding proteins are denoted at left. pTyr-Cas: phosphorylated p130Cas. The results shown are representative of three independent experiments. (B) 293F cells were transfected with plasmids for the wild-type (denoted as WT) and various DOCK180 mutants together with those for p130Cas, ANKRD28, and CrkII. Twenty-four hours later, the cells were plated onto collagen-coated dishes. After 24 h of plating, cells were lysed and processed as described in Fig. 3A. (C) 293F cells were transfected with various combinations of p130Cas, FLAG-DOCK180, GFP-ANKRD28/ELMO2 (shown as A and E, respectively), and CrkII, as indicated at the top of the figure. Transfected cells were plated onto a collagen-coated dish and analyzed as described in A.

874

E XP ER I ME NTAL CE L L R ES E ARC H 315 (2 0 0 9 ) 8 63 – 8 76

Fig. 6 – Overexpressed ANKRD28, but not ELMO, lengthens the trailing tails. (A) Representative images of cells expressing various combinations of proteins as indicated at the top of the panels. 293F cells were transfected with pCA-EGFP, together with plasmids for p130Cas (denoted as Cas), Myc-CrkII (CrkII), FLAG-DOCK180 (DOCK180), GFP-ANKRD 28 (ANKRD28), and GFP-ELMO2 (ELMO2), plated onto collagen-coated glass-bottom dishes for 12 h, and fixed with 3% PFA. Phase-contrast images of GFP-expressing cells were obtained using an epifluorescence microscope. Scale bar=20 μm. (B) 293F cells were transfected with pCA-EGFP, together with plasmids for p130Cas (denoted as Cas), Myc-CrkII (CrkII), FLAG-DOCK180 (DOCK180), FLAG- ANKRD28 (ANKRD28), and FLAG-ELMO2 (ELMO2) as indicated at the left of each panel. Cells were plated onto collagen-coated glass-bottom dishes for 1.5 h, then imaged with an epifluorescent microscope. Images for phase contrast (denoted as PC) and GFP (denoted as GFP) and were obtained every 1 min for 12 h. Images over a selected 2-h period are shown in this figure, and the time points are denoted at bottom right. The white arrowheads on the upper panels indicate GFP-positive cells and the arrows outlined in black indicate trailing tails. Scale bar=20 μm. (C) Quantification of the number of elongated tails (upper graph), the longest distance from tip to nucleus (middle graph), and the cell area (bottom graph). 293F cells were transfected with plasmids for GFP together with those for p130Cas, FLAG-DOCK180 (shown as DOCK180), GFP-ANKRD28 or ELMO2 (shown as A or E, respectively), and MycCrkII (denoted as CrkII), as indicated at left. Cells were plated onto collagen-coated dishes and observed by epifluorescent microscopy. GFPexpressing cells were designated as transfected cells and analyzed using MetaMorph software. The cell numbers examined were as follows: Cas+CrkII (n=11, 24, and 30), Cas+CrkII+DOCK180 (n=11, 15, and 20), Cas+CrkII+DOCK180+ANKRD28 (n=11, 27, and 30), and Cas+CrkII+ DOCK180+ELMO2 (n=11, 21, and 38). (Upper panel) The number of cells having elongated tails was counted among cells transfected with the various plasmids denoted at the top of the panel. Elongated tails were defined as protrusions two-fold longer than the cell body. The results are the average from three independent experiments±s.e.m. ⁎p<0.08, Student's t-test. (Middle panel) The length of the longest tail from the center of the nucleus was measured. The results are the average from three independent experiments±s.e.m. ⁎p<0.002, Student's t-test. (Lower panel) To measure the cell area, arbitrary thresholds were determined manually in each GFP-expressing cell and the area was obtained. The results are the average from three independent experiments±s.e.m. ⁎p<0.003, Student's t-test.

E XP ER I ME NTAL C EL L R ES E ARC H 315 (2 0 0 9 ) 8 63 – 8 76

875

Acknowledgments We thank the Kazusa DNA Research Institute for providing the cDNA of ANKRD28, and Drs. Tsuyoshi Akagi and Hisataka Sabe for the gift of MuLV vectors and paxillin cDNA, respectively. We are grateful to Ms. Natsuko Yoshida, Noriko Fujimoto, Akiko Abe, Keisho Fukuhara, and Yasuko Kasakawa for their technical assistance. We also thank the staff of the Matsuda Laboratory for their technical advice and helpful input. This work was supported by a Grant-in-Aid for Scientific Research on Priority Areas and a Grant-in-Aid from the Special Coordination Funds for Promoting Science and Technology from the Ministry of Education, Culture, Sports, and Science of Japan and by Sagawa Cancer Research Grant.

Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.yexcr.2008.12.005.

REFERENCES

Fig. 7 – Potential roles of ANKRD28 in focal adhesion maintenance. ANKRD28 forms a complex with p130Cas-Crk-DOCK180 in the trailing edges, while ELMO binds to RhoG and DOCK180 at the leading edges to induce membrane ruffling. Both ANKRD28 and ELMO are required for cell migration.

with p130Cas to sense the strain and to regulate focal adhesion dynamics and cell migration. The observation that ELMO expression inhibited ANKRD28 binding to DOCK180 (Fig. 1F) and vice versa (Fig. 1G) strongly suggested that ELMO and ANKRD28 recognized an overlapping region, probably the SH3 domain, of DOCK180. In fact, we showed that DOCK180 mutants comtaining the SH3 domain, but not those lacking the SH3 domain, were bound to ANKRD28 (Figs. 1D–F). There are two possible explanations for this observation. First, the SH3 domain may be required for proper conformational change of DOCK180 to expose the ANKRD28-binding sites located outside the SH3 domain. Second, both the SH3 domain and the other domain (s) may conform to an ANKRD28-binding interface. In a similar fashion, ELMO has been shown to bind to DOCK180 via the SH3 domain and a region outside of the SH3 [16]. Expression of ANKRD28 and ELMO, both of which bound to the SH3 domain of DOCK180, exhibited remarkably different phenotypes (Fig. 6). ELMO induced membrane ruffling along the entire circumference of the cells, while ANKRD28 inhibited disassembly of focal adhesions, resulting in multiple elongated cellular processes. In both cases, DOCK180 played an essential role. Therefore, ELMO and ANKRD28 may function to determine not only the localization but also the outcome of the activation of the DOCK180-Rac1 pathway. ELMO, which is activated by RhoG at the plasma membrane [37], forms a complex with DOCK180 at the rim of the cells and activates Rac1, culminating in the induction of membrane ruffling. On the other hand, through association with ANKRD28, DOCK180 may regulate stability of focal adhesions, probably via Rac1 (Fig. 7).

[1] A.J. Ridley, M.A. Schwartz, K. Burridge, R.A. Firtel, M.H. Ginsberg, G. Borisy, J.T. Parsons, A.R. Horwitz, Cell migration: integrating signals from front to back, Science 302 (2003) 1704–1709. [2] H. Hasegawa, E. Kiyokawa, S. Tanaka, K. Nagashima, N. Gotoh, M. Shibuya, T. Kurata, M. Matsuda, DOCK180, a major CRK-binding protein, alters cell morphology upon translocation to the cell membrane, Mol. Cell. Biol. 16 (1996) 1770–1776. [3] E. Kiyokawa, Y. Hashimoto, S. Kobayashi, H. Sugimura, T. Kurata, M. Matsuda, Activation of Rac1 by a Crk SH3-binding protein, DOCK180, Genes Dev. 12 (1998) 3331–3336. [4] K.M. Nolan, K. Barrett, Y. Lu, K.Q. Hu, S. Vincent, J. Settleman, Myoblast city, the Drosophila homolog of DOCK180/CED-5, is required in a Rac signaling pathway utilized for multiple developmental processes, Genes Dev. 12 (1998) 3337–3342. [5] Y.C. Wu, H.R. Horvitz, C. elegans phagocytosis and cell-migration protein CED-5 is similar to human DOCK180, Nature 392 (1998) 501–504. [6] P.W. Reddien, H.R. Horvitz, CED-2/CrkII and CED-10/Rac control phagocytosis and cell migration in Caenorhabditis elegans, Nat. Cell. Biol. 2 (2002) 131–136. [7] T.L. Gumienny, E. Brugnera, A.C. Tosello-Trampont, J.M. Kinchen, L.B. Haney, K. Nishiwaki, S.F. Walk, M.E. Nemergut, I.G. Macara, R. Francis, T. Schedl, Y. Qin, L. van Alelst, M. Hengartner, K. Ravichandran, CED-12/ELMO, a novel member of the CrkII/ Dock180/Rac Pathway, is required for phagocytosis and cell migration, Cell 107 (2001) 27–41. [8] N. Meller, M. Irani-Tehrani, W.B. Kiosses, M.A. del Pozo, M.A. Schwartz, Zizimin1, a novel Cdc42 activator, reveals a new GEF domain for Rho proteins, Nat. Cell Biol. 4 (2002) 639–647. [9] N. Meller, S. Merlot, C. Guda, CZH proteins: a new family of Rho-GEFs, J. Cell Sci. 118 (2005) 4937–4946. [10] J.F. Cote, K. Vuori, Identification of an evolutionarily conserved superfamily of DOCK180-related proteins with guanine nucleotide exchange activity, J. Cell Sci. 115 (2002) 4901–4913. [11] Y. Fukui, O. Hashimoto, T. Sanui, T. Oono, H. Koga, M. Abe, A. Inayoshi, M. Noda, M. Oike, T. Shirai, T. Sasazuki, Haematopoietic cell-specific CDM family protein DOCK2 is essential for lymphocyte migration, Nature 412 (2001) 826–831. [12] M. Watabe-Uchida, K.A. John, J.A. Janas, S.E. Newey, L. van Aelst, The Rac activator DOCK7 regulates neuronal polarity through

876

[13]

[14]

[15]

[16]

[17]

[18]

[19]

[20]

[21]

[22]

[23]

[24]

E XP ER I ME NTAL CE L L R ES E ARC H 315 (2 0 0 9 ) 8 63 – 8 76

local phosphorylation of stathmin/Op18, Neuron 51 (2006) 727–739. J.F. Cote, A.B. Motoyama, J.A. Bush, K. Vuori, A novel and evolutionarily conserved PtdIns(3,4,5)P(3)-binding domain is necessary for DOCK180 signalling, Nat. Cell Biol. (2005) 797–807. E. Kiyokawa, Y. Hashimoto, T. Kurata, H. Sugimura, M. Matsuda, Evidence that DOCK180 up-regulates signals from the CrkII-p130Cas complex, J. Biol. Chem. 273 (1998) 24479–24484. E. Brugnera, L. Haney, C. Grimsley, M. Lu, S.F. Walk, A.C. Tosello-Trampont, I.G. Macara, H. Madhani, G.R. Fink, K.S. Ravichandran, Unconventional Rac-GEF activity is mediated through the Dock180-ELMO complex, Nat. Cell Biol. 4 (2002) 574–582. X. Grimsley, M. Hall, J. Sondek, M.O. Hengartner, V. Yajnik, K.S. Ravichandran, A steric-inhibition model for regulation of nucleotide exchange via the Dock180 family of GEFs, Curr. Biol. 15 (2005) 371–377. S. Kobayashi, T. Shirai, E. Kiyokawa, N. Mochizuki, M. Matsuda, Y. Fukui, Membrane recruitment of DOCK180 by binding to PtdIns (3,4,5)P3, Biochem. J. 354 (2001) 73–78. R. Sakai, A. Iwamatsu, N. Hirano, S. Ogawa, T. Tanaka, H. Mano, Y. Yazaki, H. Hirai, A novel signaling molecule, p130, forms stable complexes in vivo with v-Crk and v-Src in a tyrosine phosphorylation-dependent manner, EMBO J. 13 (1994) 3748–3756. S. Hara, E. Kiyokawa, S. Iemura, T. Natsume, T. Wassmer, P.J. Cullen, H. Hiai, M. Matsuda, The DHR1 domain of DOCK180 binds to SNX5 and regulates cation-independent mannose 6-phosphate receptor transport, Mol. Biol. Cell 19 (2008) 3823–3835. T. Akagi, T. Shishido, K. Murata, H. Hanafusa, v-Crk activates the phosphoinositide 3-kinase/AKT pathway in transformation, Proc. Natl. Acad. Sci. 97 (2002) 7290–7295. K. Kurokawa, N. Mochizuki, Y. Ohba, H. Mizuno, A. Miyawaki, M. Matsuda, A pair of fluorescent resonance energy transfer-based probes for tyrosine phosphorylation of the CrkII adaptor protein in vivo, J. Biol. Chem. 276 (2001) 31305–31310. T. Akagi, K. Sasai, H. Hanafusa, Refractory nature of normal human diploid fibroblasts with respect to oncogene-mediated transformation, Proc. Natl. Acad. Sci. 100 (2003) 13567–13572. T. Natsume, Y. Yamauchi, H. Nakayama, T. Shinkawa, M. Yanagida, N. Takahashi, T. Isobe, A direct nanoflow liquid chromatography-tandem mass spectrometry system for interaction proteomics, Anal. Chem. 74 (2002) 4725–4733. H. Yano, Y. Mazaki, K. Kurokawa, S.K. Hanks, M. Matsuda, H. Sabe, Roles played by a subset of integrin signaling molecules in cadherin-based cell-cell adhesion, J. Cell Biol. 166 (2004) 283–295.

[25] N.C. Kwiek, D.F. Thacker, M.B. Datto, H.B. Megosh, T.A.J. Haystead, PITK, a PP1 targeting subunit that modulates the phosphorylation of the transcriptional regulator hnRNP K, Cell. Signal. 18 (2006) 1769–1778. [26] B. Stefansson, T. Ohama, A.E. Daugherty, D.L. Brautigan, Protein phosphatase 6 Regulatory subunits composed of ankyrin repeat domains, Biochemistry 47 (2008) 1442–1451. [27] L.K. Mosavi, T.J. Cammett, D.C. Desrosiers, Z.Y. Peng, The ankyrin repeat as molecular architecture for protein recognition, Protein Sci. 13 (2004) 1435–1448. [28] C.M. Grimsley, J.M. Kinchen, A.C. Tosello-Trampont, E. Brugnera, L.B. Haney, M. Lu, Q. Chen, D. Klingele, M.O. Hengartner, K.S. Ravichandran, Dock180 and ELMO1 proteins cooperate to promote evolutionarily conserved Rac-dependent cell migration, J. Biol. Chem. 279 (2004) 6087–6097. [29] Y. Makino, M. Tsuda, S. Ichihara, T. Watanabe, M. Sakai, H. Sawa, K. Nagashima, S. Hatakeyama, S. Tanaka, Elmo1 inhibits ubiquitylation of Dock180, J. Cell Sci. 119 (2006) 923–932. [30] E. Kiyokawa, N. Mochizuki, T. Kurata, M. Matsuda, Role of Crk oncogene product in physiologic signaling, Crit. Rev. Oncog. 8 (1997) 329–342. [31] K. Aoki, T. Nakamura, T. Inoue, T. Meyer, M. Matsuda, An essential role for the SHIP2-dependent negative feedback loop in neuritogenesis of nerve growth factor-stimulated PC12 cells J. Cell Biol. 177 (2007) 817–827. [32] R.L. Klemke, J. Leng, R. Molander, P.C. Brooks, K. Vuori, D.A. Cheresh, CAS/Crk coupling serves as a “Molecular Switch” for induction of cell migration, J. Cell Biol. 140 (1998) 961–972. [33] S.Y. Cho, R.L. Klemke, Extracellular-regulated kinase activation and CAS/Crk coupling regulate cell migration and suppress apoptosis during invasion of the extracellular matrix, J. Cell Biol. 149 (2000) 223–236. [34] S.Y. Cho, R.L. Klemke, Purification of pseudopodia from polarized cells reveals redistribution and activation of Rac through assembly of a CAS/Crk scaffold, J. Cell Biol. 156 (2002) 725–736. [35] G. Lee, K. Abdi, Y. Jiang, P. Michaely, V. Bennett, P.E. Marszalek, Nanospring behaviour of ankyrin repeats, Nature 440 (2006) 246–249. [36] Y. Sawada, M. Tamada, B.J. Dubin-Thaler, O. Cherniavskaya, R. Sakai, S. Tanaka, M.P. Sheetz, Force sensing by mechanical extension of the Src family kinase substrate p130Cas, Cell 127 (2006) 1015–1026. [37] H. Katoh, M. Negishi, RhoG activates Rac1 by direct interaction with the Dock180-binding protein Elmo, Nature 424 (2003) 461–464.