HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 277, Issue 42, 39840-39849, October 18, 2002

This Article Abstract Full Text (PDF) All Versions of this Article: 277/42/39840    most recent M207052200v1 Submit a Letter to Editor Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Sklyarova, T. Articles by Gettemans, J. Search for Related Content PubMed PubMed Citation Articles by Sklyarova, T. Articles by Gettemans, J. Social Bookmarking                      What's this?

Fragmin60 Encodes an Actin-binding Protein with a C2 Domain and Controls Actin Thr-203 Phosphorylation in Physarum Plasmodia and Sclerotia^*

Tatyana Sklyarova, Veerle De Corte^§, Kris Meerschaert, Liesbeth Devriendt, Berlinda Vanloo, Juliet Bailey^¶**, Lynnette J. Cook^¶, Mark Goethals, Jozef Van Damme, Magda Puype, Joël Vandekerckhove, and Jan Gettemans^§§§

From the  Department of Biochemistry, Faculty of Medicine and Health Sciences, Ghent University, Rommelaere Institute, Albert Baertsoenkaai 3, B-9000 Ghent, Belgium and the ^¶ University of Leicester, Department of Genetics, University Road, Leicester LE1 7RH, United Kingdom

Received for publication, July 15, 2002, and in revised form, August 1, 2002

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

We report the isolation of a cDNA clone encoding a 60-kDa protein termed fragmin60 that cross-reacts with fragmin antibodies. Unlike other gelsolin-related proteins, fragmin60 contains a unique N-terminal domain that shows similarity with C2 domains of aczonin, protein kinase C, and synaptotagmins. The fragmin60 C2 domain binds three calcium ions, one with nanomolar affinity and two with micromolar affinity. Actin binding by fragmin60 requires higher calcium concentrations than does binding of actin by a fragmin60 mutant lacking the C2 domain, suggesting that the C2 domain secures the actin binding moiety in a conformation preventing actin binding at low calcium concentrations. The fragmin60 C2 domain does not bind phospholipids but interacts with the endogenous homologue of Saccharomyces cerevisiae S-phase kinase-associated protein (Skp1), as shown by pull-down assays and co-expression in mammalian cells. Recombinant fragmin60 promotes in vitro phosphorylation of actin Thr-203 by the actin-fragmin kinase. We further show that in vivo phosphorylation of actin in the fragmin60-actin complex occurs in sclerotia, a dormant stage of Physarum development, as well as in plasmodia. Our findings indicate that we have cloned a novel type of gelsolin-related actin-binding protein that is involved in controlling regulation of actin phosphorylation in vivo.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Many basic cellular functions are dependent on rapid and localized reorganization of actin filaments in response to external or internal signals, and Ca²⁺-sensitive proteins often play a key role in the regulation of these processes (1-3). Actin-binding proteins of the gelsolin family represent an important class of cellular effectors and are involved in various events, including cell motility, secretion, alterations in cell morphology, and apoptosis (reviewed in Refs. 4 and 5). These proteins disassemble actin filaments in the presence of micromolar calcium concentrations through their F-actin severing activity. Polyphosphoinositides such as phosphatidylinositol 4,5-bisphosphate (PIP₂)¹ (6), but also lysophosphatidic acid (7), counteract this activity, and uncap gelsolin from barbed actin filaments, making the filaments ends free for polymerization. Gelsolin-related proteins are characterized by repeats of 125-150 amino acid segments. These segments are organized in triplicate, presumably the result of gene multiplication of a prototypical "gelsolin domain" (8, 9). Members of this family typically exist as either three-domain proteins, such as severin from Dictyostelium discoideum (10, 11), fragmins P and A from Physarum (12-16), and CapG (17); or six-domain proteins, such as adseverin (18) and gelsolin (19).

Some gelsolin homologues possess additional domains, which enable the integration of severing, capping, and nucleation with other activities or interactions. Flightless I from Drosophila, for example, contains, in addition to the six gelsolin-like segments, 16 Leu-rich repeats at its N terminus (20), that interact with TRIP, a double-stranded RNA-binding protein (21). Advillin (p92) is most closely related to villin and includes a headpiece domain (22). Supervillin (205 kDa), a membrane-interacting protein, contains an N-terminal domain with nuclear targeting signals in addition to a C-terminal half with weak similarity to villin (23). GRP125 from Dictyostelium has five gelsolin segments, in addition to a number of unique domains (24).

In this study we describe a new fragmin-related protein from Physarum, fragmin60, which harbors an N-terminal C2 domain in addition to three segments of the gelsolin-like core. C2 domains are independently folding modules of ~130 amino acid residues that are present as single or multiple copies in proteins. C2 modules occur in a variety of proteins, most of which have a role in cell signaling such as phospholipases, protein kinases, activators of GTPases, and proteins involved in membrane trafficking such as synaptotagmins (reviewed in Refs. 25-28). They have diverged evolutionarily into Ca²⁺-dependent and Ca²⁺-independent forms that interact with multiple targets. Most studied C2 domains bind lipids in a Ca²⁺-dependent way, but can also associate with protein partners in a Ca²⁺-dependent or -independent manner. Some representatives are able to bind inositol polyphosphates. Here we show that the C2 domain of fragmin60, although it binds three calcium ions with high affinity, does not interact with phospholipids but associates with the Skp1 homologue from Physarum in a calcium-independent manner. Previously, we showed that fragminP and fragminA both possess a unique quality by promoting phosphorylation of actin in vitro at Thr-203 by the actin-fragmin kinase (AFK), a protein kinase with a kelch domain and a catalytic domain that contains barely 5 conserved amino acids encountered in regular protein kinases (16, 29, 30). We demonstrate here that fragmin60 is involved in controlling actin phosphorylation in plasmodia and particularly in sclerotia, which contain exceptionally high levels of phospho-actin (31).

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

[-³²P]CTP (3000 Ci/mmol) was from ICN (Costa Mesa, CA). DEAE-cellulose was purchased from Whatman (Maidstone, UK). Phospholipids and protease inhibitors were from Sigma. Hydroxyapatite (Bio-Gel^R HT) and molecular weight markers for SDS-PAGE were from Bio-Rad. Nitrocellulose membranes (Hybond C) were from Amersham Biosciences (Buckinghamshire, UK). CaCl₂ stock solution (100 mM) and anti-rabbit IgG-alkaline phosphatase-conjugated antibodies were from Fluka (Buchs, Switzerland). Nitro blue tetrazolium chloride and 5-bromo-4-chloro-3-indolyl phosphate were purchased from Duchefa (Haarlem, The Netherlands). Phenylmethylsulfonyl fluoride (PMSF) was from Serva (Heidelberg, Germany). Trypsin and endo Lys-C (sequencing grade) were from Roche Molecular Biochemicals (Mannheim, Germany). Taq and Pfx polymerases were from Invitrogen. Restriction enzymes were from New England Biolabs, Inc. (Beverly, MA).

Purification and Peptide Amino Acid Sequencing of Fragmin60

Fresh plasmodia (300 g) were homogenized in a Waring blender in two volumes of extraction buffer (50 mM Tris-HCl, pH 7.5, 10 mM EGTA, 1 mM PMSF, 1 mM DTT, 15% sucrose). After centrifugation at 40,000 × g for 60 min, the supernatant was loaded onto a DEAE-cellulose column (5 cm × 20 cm) equilibrated in TEDA buffer (10 mM Tris-HCl, pH 7.5, 1 mM EGTA, 1 mM DTT, 0.02% NaN₃). Bound proteins were eluted with a 2-liter linear gradient of NaCl (0-250 mM in TEDA buffer). Fractions of 15 ml were collected and analyzed for the presence of fragmin60 following SDS-PAGE on mini slab gels (32) and Western blotting (33) using rabbit anti-fragmin polyclonal antibodies (34) that cross-react with fragmin60. Immunopositive fractions were collected, dialyzed against buffer for hydroxyapatite chromatography (20 mM Tris-HCl, pH 7.5, 10 mM KH2PO₄, 1 mM DTT, 1 mM PMSF), and loaded onto a hydroxyapatite column (2 cm × 15 cm). Proteins were eluted with a gradient of KH₂PO₄ (10-250 mM) in hydroxyapatite buffer. Positive fractions were collected and precipitated overnight with 2 volumes of ethanol. The ethanol precipitate was air-dried and dissolved in 9.5 M urea, 2% Nonidet P-40, 1.6% ampholines, pH 5-7, 0.4% ampholines, pH 3.5-10, and 100 mM DTT. Two-dimensional gel electrophoresis was performed according to Ref. 35. The position of fragmin60 on two-dimensional gels was determined by Western blotting using anti-fragminP antibodies. Protein spots corresponding to fragmin60 were excised from multiple Coomassie-stained gels and equilibrated in SDS-containing buffer, and the protein was electro-eluted and concentrated according to Ref. 36. Approximately 10 µg of protein were digested in gel using endolysin-C or trypsin (sequencing grade). Extracted peptides were separated on a narrow-bore C₄ reversed-phase column as described in Ref. 37. Amino acid sequence analysis was performed on a model 477A pulsed liquid-phase sequenator equipped with a model 120A phenylthiohydantoin amino acid analyzer (Applied Biosystems Inc.).

Cloning of Fragmin60 cDNA; Construction of a Physarum Polycephalum Plasmodial cDNA Library

A cDNA library was constructed from Physarum plasmodial poly(A)⁺-enriched RNA by priming reverse transcription with oligo(dT) and a mixture of random hexanucleotides. Complementary DNA was synthesized by using the SuperScript^TM Choice System (Invitrogen) according to the instructions from the manufacturer. First round PCR was done using a pair of degenerate primers, designed on two peptides that are homologous to the N-terminal part of fragminP (ARHEAAW, 5'-GCNMGICAYGARGCNGCNTGG-3' as forward (+) primer, and DDFLGGA, 5'-NGCICCNCCIARRAARTCRTC-3' as reverse () primer). PCR reactions were done using 2.5 units of Taq DNA polymerase, 1 µM of both primers, and ~1 ng of double-stranded DNA. 35 cycles of PCR were performed in a thermocycler (Biometra) under the following conditions: 95 °C, 1 min; 55 °C, 2 min, 72 °C, 30 s, followed by a final extension of 10 min at 72 °C. The resulting PCR product of 280 bp was cloned into the pCR^®2.1 vector (Invitrogen, Groningen, The Netherlands) and sequenced. Based on the sequence of the PCR product, non-degenerate forward and reverse primers were designed. A 27-mer fragmin60-specific oligonucleotide: 5'-GTATGCGGCAGTTCCGGCTTCATCTTG-3' was used as reverse primer in combination with the 5' LD Amplimer (5'-CTCGGGAAGCGCGCCATTGTGTTGGT-3') from the LD-Insert Screening Amplimer Sets (Clontech) for walking in the 5' direction of the cDNA. Reactions were carried out using TaqPCR Core Kit (Qiagen, Leusden, the Netherlands) using 2.5 pmol of both primers, 1 ng of template DNA, and cycle parameters recommended by the instructions from the manufacturer. A PCR product of 900 bp was obtained and subcloned in the pCR2.1 shuttle vector. Finally, a non-degenerate primer was designed (5'-TTCGTGAATGGAGCCCAAGCAAGGATAC-3') for walking in the 3' direction in combination with a 3' LD Amplimer in the same conditions as for the previous reaction. Both strands of two independent fragmin60 clones were sequenced entirely.

Cloning of Physarum Skp1 cDNA

Degenerate primers were designed based on peptide sequences obtained by quadrupole time-of-flight mass spectrometry (Q-TOF MS). A 217-bp cDNA fragment was generated by PCR on the cDNA library using primers 5'-DATRTTRAANGTYTTNCRDATYTCYTC-3' (EEIRKTFNI) and 5'-CCNACNCCIGAYGARAARGAYG (PTPDEKDE). The primers for walking in the 5' direction were 5'-CTGTGCGCTTCTCGTCTTTCT-3' and 5'-AATTCGCGGCCGCGTCGAC-3' (adaptor primer). Primers for walking in the 3' direction were 5'-ACTCTTCGAGCTCATCTTGGCTGCTAAC-3' and 5'-CGCGGCCGCGTCGAC-3' (adaptor primer). The full-length clone was obtained using primers 5'-AATGGGCCGCCGAAGCATCTGCCGGTAC-3' (forward) and 5'-CTTCTCTTCGCACCACTCATTCTCCTT-3' (reverse). Both strands were sequenced entirely.

Northern Blotting

All procedures concerning RNA isolation, Northern blotting, synthesis of probes and hybridization conditions are described in Refs. 13, 14, and 39, and references therein.

Prokaryotic Expression and Purification of Recombinant Fragmin60, Its Deletion Mutants, and Skp1

All GST fusion proteins were engineered by PCR using fragmin60 or Skp1 cDNA as template. Briefly, PCR was carried out with a (+) strand primer with a BamHI recognition site and a () strand primer with an XbaI recognition site, corresponding to fragmin60, C2-"hinge" region, fragmin60C2, and C2"hinge", respectively. PCR products were ligated into the pGEX-5X-1 vector (Amersham Biosciences). All cDNA constructs were confirmed by DNA sequencing. Skp1 contained an additional C-terminal His₆ tag added by PCR. Plasmids were transformed into Escherichia coli BL21(DE3)pLysS-competent cells (Stratagene). Expression was induced by the addition of 0.5 mM isopropyl -D-thiogalactopyranoside. After 5 h of growth at 30 °C, cells were harvested by centrifugation, resuspended in TBS buffer (20 mM Tris-HCl, pH 7.5, 150 mM NaCl) with protease inhibitors and lysed by passing the cells twice through a French press. Soluble recombinant proteins were purified using glutathione-Sepharose 4B according to the instructions from the manufacturer and stored frozen at 20 °C. The GST tag was removed from Skp1 with factor X_a.

Protein Binding to Small Unilamellar Vesicles

10 mg of a PS:PC mixture (molar ratio 30:70) in chloroform was dried under a stream of nitrogen, degassed in vacuum overnight, and resuspended in 10 ml of buffer A (20 mM Tris, pH 7.5, 150 mM NaCl, 2 mM EGTA) by vigorous vortexing and subsequent sonication (38). Five µg of various GST-tagged C2 domains in 200 µl of buffer A were mixed with 100 µl of vesicles either in the presence or absence of 3 mM CaCl₂ (final concentration free calcium is 1 mM) in a total volume of 300 µl of buffer A. After 5 min of incubation with agitation at 25 °C, the mixture was centrifuged at 15,000 × g for 15 min at 4 °C. The supernatants were carefully withdrawn and precipitated with trichloroacetic acid, and these pellets as well as lipid-bound complexes were dissolved in 60 µl of SDS sample buffer. A 10-µl sample of both fractions was analyzed by SDS-PAGE on 15% polyacrylamide gels.

Pull-down Assays with GST Fusion Proteins

GST fusion proteins used in this study were as follows: GST-C2hinge (coding residues 1-165), GST-C2 (residues 1-141), GST-fragmin60C2 (residues 142-536), and GST-fragmin60. GST fusions of C2 domains from other proteins included synaptotagmin I (C2A), II (C2A), VII (C2A), Doc2C2A, Doc2C2A, and NEDD4. These were generously provided by Dr. B. Davletov (Medical Research Council Laboratory of Molecular Biology, Cambridge, UK; SytIIC2A), Dr. D. Rotin, (Departments of Cell/Lung Biology, Toronto, Ontario, Canada; NEDD4C2), Dr. T. Südhof (Department of Molecular Genetics, and Howard Hughes Medical Institute, University of Texas Southwestern Medical Center, Dallas, TX; SytI and VIIC2A), and Dr. M. Fukuda (Fukuda Initiative Research Unit, RIKEN, Saitama, Japan; Doc and Doc C2A). They were expressed and purified in the same manner as GST-fragmin60 and mutants thereof. Lysates for pull-down assays were prepared from Physarum microplasmodia by disruption in a glass homogenizer in two volumes of buffer B (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 1.5 mM MgCl₂, 1% Triton X-100, 1 mM DTT, 10% glycerol, protease inhibitor mixture (Roche), and 1 mM PMSF), sonication, and centrifugation at 100,000 × g for 1 h. The supernatant was used for the assay. Glutathione-Sepharose beads were preloaded with equivalent amounts (150 µg) of fusion proteins or GST, and mixed with 2 mg of lysate followed by incubation overnight at 4 °C. Beads were washed three times with 0.6 ml (120 volumes of raisin volume) of lysis buffer, once with TBS buffer, and bound proteins were eluted with SDS sample buffer and resolved by SDS-PAGE. Gels were stained with silver or Coomassie or transferred to nitrocellulose membranes. Proteins were detected by ECL using anti-Physarum Skp1 polyclonal antibody or monoclonal anti-His₆ and HRP-conjugated goat anti-rabbit or HRP-conjugated goat anti-mouse antibody, respectively. The interaction between C2 domains and Skp1 was analyzed as follows: GST-C2 fusion protein (20 µg), immobilized on glutathione-Sepharose 4B, and varying amounts of recombinant Skp1 were incubated overnight at 4 °C in TBS buffer, 0.5% Triton X-100, 1% bovine serum albumin. Following several washes with the same buffer the samples were boiled in 5× Laemmli buffer and separated by SDS-PAGE. Skp1 was revealed by Western blotting using monoclonal penta-His antibody (Qiagen).

Baculovirus Cloning and Expression

The fragmin60 C2 domain (amino acids 1-165, including the hinge region) and fragmin60 were cloned as BamHI/NotI fragments into the pFastbac donor plasmid, containing an N-terminal His₆ tag. Following transformation of the respective plasmids into competent DH10Bac E. coli cells, recombinant bacmid DNA was purified and used to transfect Sf21 insect cells. Recombinant baculovirus particles were harvested and used for infection of HI5 insect cells. Cell lysates were prepared 3 days after infection, and the recombinant proteins were purified using nickel-nitrilotriacetic acid resin according to the instructions from the manufacturer.

Co-expression of Fragmin60 C2 and Skp1 in Mammalian Cells

HEK293T cells were maintained in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal calf serum. The fragmin60 C2 domain was cloned into the pEGFP-N1 vector (Clontech) at SalI and XhoI sites and PpSkp1 was cloned into pcDNA6Myc/His (Invitrogen) at the EcoRI and XbaI sites, in frame with C-terminal Myc epitope. HEK293T cells were transiently transfected with both plasmids (or pEGFPN1-C2 alone) using calcium phosphate. Next day the medium was renewed, and 48 h after transfection the cells were washed with PBS buffer and treated with 1 mM dithiobis(succinimidylpropionate) (Pierce) for 30 min at room temperature. The reaction was stopped by adding 20 mM Tris-HCl, pH 7.5, buffer. After 15 min the cells were lysed in TBS buffer containing 1% Triton X-100 and a protease inhibitor mixture mix (Roche Diagnostics, Mannheim, Germany). The lysate was sonicated and insoluble material was removed by centrifugation (20,000 × g, 10 min at 4 °C). 200 µg of lysate were incubated with either 9E10 monoclonal anti-Myc antibody (kindly provided by Dr. P. Zimmermann, Department of Glycobiology, University of Leuven, Leuven, Belgium) or mouse IgG₁. After 1 h, protein G-Sepharose was added. The pellets were washed with lysis buffer and finally analyzed by SDS-PAGE and Western blotting with either anti-C2 or anti-PpSkp1 antibodies.

Isothermal Titration Calorimetry

Microcalorimetric titration measurements were performed in a Microcal Omega isothermal titration calorimeter (Microcal Inc., Northampton, MA). All solutions were degassed under vacuum prior to use. In a typical experiment, 1.33 ml of ~ 25 µM GST-C2 in 10 mM Tris-HCl, 0.2 mM DTT, pH 7.5, was titrated in 28 steps with 300 µl of 600 µM CaCl₂ (stock solution obtained from Fluka). During titration, the injection syringe was rotated at 400 rpm. Time between injections was set at 3 min. In a blank experiment, heat evolving from dilution was measured by injecting the CaCl₂ solution into the sample cell filled with ~25 µM GST in 10 mM Tris-HCl, 0.2 mM DTT, pH 7.5. This heat of dilution was subtracted from the corresponding Ca²⁺ binding data of the GST-C2 protein. Data were integrated and fitted to an appropriate binding model using ORIGIN software supplied by Microcal Inc.

Severing of Actin Filaments by Fragmin60

This was performed essentially as described (34). Briefly, actin (8 µM, 25% pyrene-labeled) was polymerized in F-buffer for 15 min. Subsequently the filaments were pre-capped with the actin-fragmin complex (40 nM final concentration). The mixture was stored overnight on ice. This solution was then diluted in G-buffer to a final concentration of 400 nM, which is below the critical monomer concentration of the pointed ends. Addition of a severing protein will create new pointed ends, and the fluorescence will decrease proportionally to the number of free () ends. The effect of phospholipids on the severing activity of fragmin60 was investigated by pre-incubating fragmin60 with the indicated phospholipid concentrations for 5 min at room temperature. Subsequently the mixture was added to the pyrene-labeled F-actin solution, and the decrease in fluorescence was measured over time.

Stoichiometry Measurement of Actin-Fragmin60 Complex Formation

Pyrenyl G-actin (100% labeled, 140 nM final concentration) was mixed for 10 min with increasing concentrations of fragmin60 (between 0- and 2-fold molar excess over actin) under non-polymerizing conditions. The relative increase in fluorescence was measured for each mixture. These experiments were carried out in the presence of either 0.2 mM Ca²⁺ or 2 mM EGTA. Mg²⁺-actin was used for experiments involving EGTA.

Generation of Polyclonal Anti-C2 and Anti-PpSkp1 Antibodies

The GST moiety of purified proteins was removed with factor X_a, and subsequently the proteins were dialyzed overnight against 50 mM NaHCO₃ at 4 °C. Samples of 120 µg were concentrated to dryness in a SpeedVac concentrator (Savant Instruments, Farmingdale, NY), and rabbits were immunized (Centre d'Economie Rurale, Laboratoire d'Hormonologie, Marloie, Belgium). Polyclonal antibodies against the C2 domain were affinity-purified on a CNBr-Sepharose-coupled C2 affinity column and eluted with ethanolamine as described in Ref. 34.

Immunoprecipitation Experiments

Nitrogen-frozen pellets of plasmodia or sclerotia (2 mg, wet weight) were lysed by mechanical disruption and extracted in 20 mM Tris-HCl, pH 7.5, 1 mM DTT, 1 mM EDTA, 1 mM EGTA, 50 mM NaF, 1 µM okadaic acid, 1 µM staurosporine, and protease inhibitors. Extract (750 µl) was incubated with 20 µl of anti-C2 antibodies or 10 µl of anti-fragminP antibodies (both affinity-purified) overnight at 4 °C in the presence of protein G-Sepharose. Pellets were washed three times with TBS containing 1% Triton X-100 and 0.5% Nonidet P-40, followed by three washes with TBS. Pellets were boiled 5 min in Laemmli sample buffer (40), and the supernatant was analyzed by SDS-PAGE. 10 µl of the supernatants after immunoprecipitation was also electrophoresed on an SDS gel.

Miscellaneous Procedures

Protein concentration was measured according to Ref. 41 using bovine serum albumin as standard. Calculation of free calcium concentrations was done using the MaxChelator program located at www.stanford.edu/~cpatton/maxc.html. Antibodies were cross-linked to protein G-Sepharose using the Seize^TM X protein G immunoprecipitation kit (Pierce), used according to instructions from the manufacturer.

Mass Spectrometry

Samples (20 µl) were loaded on a trapping column (inner diameter, 0.8 mm × 2 mm - P/N MGU-80-C18PM, LC-packings) run with solvent A (0.05% (v/v) formic acid in water) at a flow of 10 µl/min. At 5 min after injection, the trapping column was connected in back flush to the separation column (a reversed phase nano-column, inner diameter 0.075 mm × 150 mm; P/N NAN75-15-03-C18PM-MS, LC-Packings). A linear gradient was developed as follows: 5-100% solvent B in 55 min at a flow rate of 60 µl/min (solvent B: 0.04% (v/v) formic acid in acetonitrile:water (v/v 70:30)). By means of a flow splitter (split ratio 300:1), a flow of about 0.2 µl/min was directed to the separation column. The outlet of the separation column was connected on-line to the electrospray ion source of the mass spectrometer (Z-spray nanosprayer and Q-TOF mass spectrometer, Micromass). The mass spectrometer was controlled with a software package from the manufacturer (Masslynx, version 3.2 b004). Data were acquired in data-dependent mode. In the survey mode, spectra were acquired in the 300-500 m/z range with an integrated scan time of 1 s. In the MS-MS mode spectra were obtained in the 50-2200 m/z range with an integrated scan time of 1 s. De novo sequence deduction was done by aid of PepSeq (PepSeq is a part of the Masslynx package).

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Identification of a Physarum 60-kDa Protein as a Fragmin-related Actin-binding Protein

In an effort to identify and study new regulatory microfilament components, we focused on a 60-kDa protein detected on blots of total plasmodia lysates probed with anti-fragmin polyclonal antibodies (Fig. 1A). To obtain amino acid sequences, we partially purified this protein by column chromatography. The cytosol of plasmodia was fractionated by DEAE-ion exchange and hydroxyapatite chromatography, and the presence of the 60-kDa protein was followed by SDS-PAGE and Western blotting with anti-fragminP antibodies. Two-dimensional gel electrophoresis revealed several isoforms of the 60-kDa protein (Fig. 1B). Edman sequencing of peptides showed strong similarity with plasmodial fragmin; others, however, showed no similarity with proteins in the databases. A cDNA clone encoding the entire 60-kDa protein was obtained from a Physarum cDNA library. The coding region covers 1608 nucleotides encoding 536 amino acids, resulting in a polypeptide with a predicted molecular mass of 59,773 daltons and a pI of 5.84. An alignment between the 60-kDa protein and the two Physarum fragmin isoforms (frgP (Ref. 13) and frgA (Ref. 14)) is represented in Fig. 2, and shows that the 60-kDa protein contains a C-terminal region of 371 amino acids that is very similar to both frgP and frgA. The amino acid sequence identity/similarity with fragmin60 in this region amounts to 55%/70% for frgP, 54%/69% for frgA, and 50%/66% for Dictyostelium severin. We therefore conclude that we have cloned a third member of the fragmin gene family and a novel member of the fragmin/gelsolin family of actin-binding proteins. Given the high degree of similarity with both fragmin isoforms, and its molecular weight, we termed the protein fragmin60 (frg60).

View larger version (23K): [in this window] [in a new window]   Fig. 1.   Identification of a 60-kDa protein in Physarum extracts that cross-reacts with fragminP antibodies. A, one-dimensional SDS-PAGE and Western blot of 10 µg of cytosolic proteins, electrophoresed on a 10% gel and probed with anti-fragminP antiserum. FragminP (arrow) and the 60-kDa protein (arrowhead) are indicated. B, two-dimensional gel electrophoresis of the 60-kDa protein (upper panel) and fragminP (lower panel) showing different isoforms. Only the part of the two-dimensional gel is shown where the respective proteins migrate. The 60-kDa protein was visualized with fragminP antibodies; fragminP was visualized by Coomassie staining. IEF, isoelectric focusing.

View larger version (55K): [in this window] [in a new window]   Fig. 2.   Complete amino acid sequence of fragmin60 and alignment with fragminP and fragminA. Fragmin60 contains a unique N-terminal 166-amino acid region. Conserved amino acids in all three isoforms, and in two of the three isoforms, are shown in red and blue, respectively. The C2 domain signature is underlined, whereas the PIP2-binding region is highlighted in italics. The fragmin60 nucleotide sequence is available from GenBankTM under accession no. AF303112.

Unlike any other actin-binding protein of the gelsolin family, fragmin60 contains an N-terminal part of 165 amino acids. Screening of databases using the Blast algorithm with the fragmin60 N-terminal segment (the 141-amino acid N-terminal fragment and the following 25-amino acid region that probably functions as a hinge) revealed strong similarity with C2 domains. The signature sequence characteristic of C2 domains is underlined in Fig. 2. The highest identity/similarity was found with the C2 domains of the profilin-binding protein aczonin (34%/54%), protein kinase C-1A from Hydra vulgaris (32%/52%), and synaptotagmins I-C2A (27%/51%) and V-C2A (29%/51%) from Gallus gallus and Homo sapiens, respectively.

Biochemical Characterization of Fragmin60

Affinity-purified polyclonal antibodies raised against the recombinant fragmin60 C2 domain recognized a 60- and a 42-kDa protein on Western blots using total protein extracts of plasmodia (Fig. 3A, lane 1). The second band probably represents another protein with a C2 domain. Fragmin60 as well as actin were immunoprecipitated by these antibodies (Fig. 3A, lanes 2 and 3).

View larger version (26K): [in this window] [in a new window]   Fig. 3.   Fragmin60 binds actin and severs actin filaments. A, actin co-immunoprecipitates with fragmin60. Lane 1, Western blot analysis of a total lysate from plasmodia probed with affinity-purified anti-C2 domain antibodies. The asterisk indicates a 42-kDa cross-reacting protein. Lane 2, Coomassie-stained SDS-PAGE of proteins immunoprecipitated from Physarum lysate with anti-C2 antibodies. Lane 3, immunoblot of the proteins in lane 2 probed with monoclonal anti-actin antibodies. IgGH, immunoglobulin heavy chain. B, fragmin60 severs actin filaments. Pyrene-F-actin (8 µM) pre-capped with native actin-fragmin was diluted to 0.4 µM in G-buffer in the absence of fragmin60 (triangles; control), or in the presence of 20 nM (squares) or 40 nM fragmin60 (circles), and the decrease in fluorescence was measured over time. Values represent mean of three independent experiments. C, modulation of actin-fragmin60 interaction by calcium, EGTA or PIP2. Pyrene-labeled G-actin was mixed with fragmin60 in G-buffer in ratios as indicated in the x axis in the presence of 0.2 mM Ca2+ (squares). Alternatively, preformed actin-fragmin60 complexes were treated with 2 mM EGTA (triangles), or fragmin60 was pre-incubated with 2 mM EGTA before complex formation with actin (circles), and the resulting changes in fluorescence were measured. Inhibition of actin binding was also observed when fragmin60 was incubated with G-actin in the presence of 14 µM PIP2, but is not shown here for clarity. The results represent mean of triplicate determinations.

In view of its high similarity with actin-severing proteins, we examined the effect of fragmin60 on F-actin depolymerization. Pyrene-labeled F-actin filaments were pre-capped with native actin-fragmin (16), and actin depolymerization was observed by fluorescence decrease over time. Fast actin depolymerization was observed (Fig. 3B) in the presence of fragmin60 (20 or 40 nM) in a concentration-dependent manner. These findings are similar to those reported for fragmin (13, 34), indicating that fragmin60 is an F-actin-severing protein.

F-actin severing requires binding of two actin monomers (42, 43). We investigated the stoichiometry of monomeric actin binding by fragmin60 under non-polymerizing conditions. Binding of pyrene-labeled G-actin by an actin-binding protein results in an increase in fluorescence (44). Incubation of fragmin60 with labeled actin monomers at different molar ratios induced an increase in actin fluorescence (Fig. 3C), reaching a plateau at a molar ratio of 1:2 fragmin60:actin. This suggests that fragmin60 is able to bind two actin monomers. Subsequent addition of EGTA caused a decrease in fluorescence, this time reaching a maximum at 1:1 fragmin60:actin (Fig. 3C), indicating that one actin monomer was released resulting in a 1:1 EGTA-resistant complex. These findings are similar to those reported for fragmin (34). Addition of EGTA before complex formation prevented actin binding (Fig. 3C), indicating that fragmin60 binds actin in a calcium-dependent manner. Likewise, PIP₂ inhibited complex formation between actin and fragmin60 (Fig. 3C). The PIP₂-binding region in gelsolin-like actin-binding proteins is conserved in fragmin60 (³³⁰RLLHLKGGK³³⁸) and differs by only one amino acid from the other fragmin isoforms (13, 14).

Characterization of the Fragmin60 C2 Domain

The C2 Domain Contains Three Calcium-binding Sites-- Ca²⁺-dependent C2 domains coordinately bind 2-3 Ca²⁺ ions at flexible loops on top of the -sandwich (45, 46). Phospholipids can increase the affinity for calcium binding (47-49), probably by providing additional coordination sites for calcium ions (50, 51). We investigated calcium binding by the fragmin60 C2 domain by isothermal titration calorimetry (ITC). Preliminary ITC experiments suggested that the C2 domain contains multiple calcium binding sites with different affinities. Therefore, varying amounts of calcium were injected in the sample cell to avoid immediate saturation of the expected high affinity site(s). Successive injections of calcium yielded a significant heat release (Fig. 4A, top panel), indicating that the calcium binding process is exothermal. The binding isotherm is characterized by two breakpoints, one at molar ratio 1 and another breakpoint at molar ratio 3 (Fig. 4A, bottom panel). In combination with the thermodynamic parameters (Table I), it can be deduced that, unlike most other C2 domains, the fragmin60 domain shows a very high affinity for calcium as even the "low" affinity sites show submicromolar affinities.

View larger version (19K): [in this window] [in a new window]   Fig. 4.   Characterization of the fragmin60 C2 domain. A, calcium binding measured by isothermal titration calorimetry. The top panel shows raw heat data obtained from 28 injections of calcium into a sample cell containing 25 µM GST-C2 (see "Experimental Procedures"). The last eight peaks correspond with heat of dilution and show that all of C2 is saturated with calcium. Bottom panel, binding isotherm created by plotting areas under the peaks against the molar ratio of calcium added to GST-C2. The line represents the optimal fit for a multiple ligand binding. B, phospholipid binding assays. Liposomes and indicated GST fusion proteins were incubated in the presence of 2 mM EGTA or 1 mM free Ca2+ for 15 min at room temperature. After centrifugation, the supernatants (S, non-binding fraction) and pellets (P, phospholipid-binding fraction) were separated as described under "Experimental Procedures." Equal proportions of the supernatants and pellets were subjected to 15% SDS-PAGE and then stained with Coomassie. C, influence of the C2 domain on actin binding by fragmin60. Results are shown for G-actin binding by full-length fragmin60 (black bars) and fragmin60C2 (white bars), as a function of the free calcium concentration. Values represent mean ± S.E. of three independent experiments.

The Fragmin60 C2 Domain Does Not Bind Phospholipids-- Approximately 50% of all C2 domains studied so far bind to phospholipids in a calcium-dependent way. To determine whether the fragmin60 C2 domain displays the same property, we examined its ability to bind small unilamellar vesicles consisting of PC and PS (see "Experimental Procedures"). Experiments were performed in parallel with several other previously studied GST-tagged C2 domains. As expected, Doc2 C2A (52), C2A domains from synaptogamins I and II (53), and NEDD4-C2 (54) were predominantly found in the pellet fraction following incubation with small unilamellar vesicles in the presence of 1 mM calcium, indicating that they bound liposomes in a Ca²⁺-dependent manner (Fig. 4B). By contrast, for Doc2 C2A and fragmin60 C2, no significant binding to PC/PS was detected under these assay conditions. As it is known from previous data that Doc2 does not bind phospholipids (55), it serves as a negative control in this assay. Furthermore, no differences were observed with vesicles containing other phospholipids, such as phosphatidylinositol or phosphatidylethanolamine (data not shown). We therefore conclude that the fragmin60 C2 domain does not interact with phospholipids.

The C2 Domain Increases the [Ca²⁺] Threshold for Actin Binding by Fragmin60-- By virtue of their ability to bind calcium and/or phospholipids, C2 domains can exert control on the properties inherent in other subdomains of the protein (56). To investigate whether the fragmin60 C2 domain influences the actin binding properties of the fragmin moiety, we performed pyrene-actin binding experiments under non-polymerizing conditions using full-length fragmin60 or a deletion mutant lacking the C2 domain. Actin binding by the full-length protein was saturated at 20 µM free calcium (Fig. 4C). Surprisingly, only 6 µM free calcium was sufficient to saturate actin binding by a deletion mutant that lacked the C2 domain.

The C2 Domain of Fragmin60 Specifically Interacts with the Physarum Homologue of S. cerevisiae Skp1 in a Calcium-independent Way

The lack of calcium-dependent phospholipid binding activity prompted us to search for a putative protein partner. Indeed, some C2 domains are able to mediate protein-protein interactions, such as the interaction between C2B of synaptotagmin I and the clathrin adaptor complex AP-2 (57) or the phospholipase A₂ C2 domain and vimentin (58). To this end, we expressed four kinds of GST-fusion proteins in E. coli: GST-C2 with hinge (GST-C2h), GST-C2 without hinge (GST-C2), GST-Frg60, and a fragmin60 construct lacking the C2 domain (GST-Frg60C2); see "Experimental Procedures"). The recombinant proteins were incubated with Physarum lysates, and associated proteins were separated by SDS-PAGE and visualized by silver staining. GST-C2h, GST-C2, and GST-Fr60 fusion proteins, but not GST-Frg60C2, specifically pulled down a protein with a molecular mass of 21 kDa (data not shown). This observation indicates that p21 associates specifically with the C2 domain.

The p21 band was subjected to trypsin digestion, and internal peptide sequences of the 21-kDa protein were generated by Q-TOF mass spectrometry. Several peptides were 100% identical to, or showed very high similarity with, the D. discoideum cytosolic glycoprotein FP21 (59) and budding yeast S-phase kinase associated protein Skp1p (60, 61). Skp1 is an intrinsic component of the kinetochore complex in yeast but is also a constituent of the proteasome pathway and is involved in the recycling of the SNARE Snc1p in yeast (62). A full-length Skp1 cDNA clone was obtained by PCR from a Physarum cDNA library through reversed genetics. The deduced amino acid sequence of Physarum Skp1 (PpSkp1, GenBank^TM accession no. AY050559) showed 75% similarity with mouse Skp1, indicating a high degree of conservation.

To confirm the interaction between fragmin60 C2 and PpSkp1, pull-down experiments were performed on Physarum lysates with different fragmin60 GST fusion constructs (GST-C2h, GST-C2, GST-Frg60, and GST-Frg60C2) and analyzed by Western blotting using polyclonal anti-PpSkp1 antibodies. As shown in Fig. 5A, PpSkp1 was detected in pull-down experiments with GST-Frg60 (lane 2), GST-C2h (lane 4), and GST-C2 (lane 5), but not with Frg60C2 (lane 3), confirming that endogenous Skp1 is specifically retained by GST fusions that minimally contain the C2 domain. Binding was independent of calcium (Fig. 5A, compare lane 1 with lane 2) This interaction appeared to be specific for the fragmin60 C2 domain because Doc2 C2A, synaptogamin I C2A and II C2A, NEDD4-C2, or Doc2 C2A did not associate with PpSkp1 under these conditions (Fig. 5B).

View larger version (20K): [in this window] [in a new window]   Fig. 5.   The fragmin60 C2 domain interacts specifically and directly with the Physarum homologue of yeast Skp1p. A, left panel, Western blot. Polyclonal anti-PpSkp1 antibodies recognize one protein in Physarum lysates. Right panel, pull-down experiments of Skp1 from Physarum lysates using different fragmin60 GST fusion constructs. Skp1 was detected by Western blotting. Lane 1, GST-fragmin60 in the presence of 1 mM calcium. Lanes 2-6, no calcium. GST-fragmin60 (lane 2), GST-fragmin60C2 (lane 3), GST-C2h (lane 4), GST-C2 (lane 5). Lane 6, total lysate. B, other C2 domains investigated do not bind PpSkp1. Upper panel, Western blot of proteins pulled down by different GST-C2 domains, stained with anti-Skp1 antibodies, and detected by ECL. Lane 1, fragmin60C2; lane 2, SytIC2A; lane 3, SytIIC2A; lane 4, SytVIIC2A; lane 5, DocC2A; lane 6, DocC2A; lane 7, NEDD4C2. Proteins were separated by 20% SDS-PAGE. Lower panel, corresponding Coomassie-stained gel showing equal loading of different GST-C2 domains. C, recombinant Physarum Skp1 directly interacts with the fragmin60 C2 domain. 20 µg of GST-C2 was incubated with 5 µg (lane 1), 10 µg (lane 2), or 20 µg (lane 3) of His6-tagged Skp1 and after several washes electrophoresed. The Western blot was probed with anti-His6 antibody. Lane 4 shows 8% of input Skp1 (20 µg), and lanes 5-7 represent the same experiment but with 20 µg of GST. D, the C2 domain of fragmin60 and PpSkp1 interact in vivo. HEK cells were transfected with pEGFP-C2 (lanes 1-6) or pEGFP-C2 + pcDNA6Myc/His- Skp1 (lanes 2, 5, and 6). Proteins were immunoprecipitated with anti-Myc antibody (lanes 4 and 5) or mouse IgG1 (lanes 3 and 6). Retained proteins were eluted from the beads, separated on 15% SDS-PAGE, transferred to nitrocellulose, and immunoblotted with anti-PpSkp1 antibodies (lanes 2, 5, and 6) or anti-C2 antibodies to detect transfected GFP-C2 (lanes 1-6). Anti-PpSkp1 antibodies did not cross-react with endogenous Skp1.

We next tested whether binding of fragmin60 C2 with Skp1 is direct. In pull-down assays, factor X_a-digested Skp1 was retrieved by beads that were pre-coated with GST-C2, but not with GST alone (Fig. 5C), indicating a direct interaction.

To confirm the interaction between the C2 domain and Skp1 in vivo, co-immunoprecipitation experiments were carried out following double transfection of GFP-tagged C2 domain and Myc-tagged Skp1 cDNA constructs in HEK293T cells. GFP-tagged C2 domain (Fig. 5D, lanes 1 and 2) as well as PpSkp1 (Fig. 5D, lane 2) were readily detected in crude lysates of cells. Immunoprecipitation of Skp1 using anti-Myc antibodies revealed the presence of a complex between Skp1 and the C2 domain (Fig. 5D, lane 5). An irrelevant antibody was unable to precipitate the complex (Fig. 5D, lane 6) or GFP-C2 (Fig. 5D, lane 3), nor did anti-Myc antibodies precipitate GFP-C2 (Fig. 5D, lane 4). Taken together, all these findings demonstrate that the C2 domain of fragmin60 specifically and directly binds to PpSkp1.

PpSkp1 is able to interact not only with fragmin60, but also with the fragmin60-actin complex. Additionally, phosphorylation of actin in the fragmin60-actin complex by the AFK (see below) was not affected by PpSkp1 (data not shown). Thus, the role of this interaction is not clear at present.

Actin in the Actin-Fragmin60 Complex Is Phosphorylated in Vitro and in Vivo on Thr-203 by the Actin-Fragmin Kinase

In Vitro Phosphorylation of Actin-Fragmin60-- Previous studies have shown that actin in complex with fragminP or fragminA (13, 14, 16) can be phosphorylated in vitro by the AFK, a protein kinase that is very likely restricted to Physarum (63). In vivo phosphorylation of actin occurs only in complex with fragminP (37). Phosphorylation of the actin moiety blocks the F-actin capping activity of the complex in vitro (34) but also after microinjection into living CV1 and PtK2 epithelial cells (64). In view of the similarity between fragmin60 and the amoebal/plasmodial fragmin isoforms, we investigated whether actin could be phosphorylated in the actin-fragmin60 complex.

A 45-kDa protein co-purified with fragmin60 in equimolar ratio when the latter was expressed in insect cells (Fig. 6A). This protein was identified by mass spectrometry-based peptide fingerprinting as actin (data not shown). When added to the actin-fragmin60 complex, recombinant actin-fragmin kinase promoted a time-dependent and strong phosphorylation of actin (Fig. 6B). Thus fragmin60 is the third member of the fragmin gene family that promotes actin phosphorylation by the AFK.

View larger version (20K): [in this window] [in a new window]   Fig. 6.   Fragmin60 promotes actin phosphorylation in the actin-fragmin60 complex by the AFK in vitro and is associated with phospho-actin in sclerotia, like fragminP. A, recombinant fragmin60 binds to actin when isolated from insect cells. SDS-PAGE analysis (10%) of recombinant fragmin60 purified by nickel-nitrilotriacetic acid affinity chromatography. The gel was stained with Coomassie. B, fragmin60 and fragminP increase actin phosphorylation by AFK in vitro through actin binding. One µg of actin-fragmin60, purified from insect cells, or fragminP were incubated with ATP/Mg2+ and recombinant AFK, and phosphorylation was allowed to proceed for 10 min. Actin phosphorylation was detected by Western blotting using anti-phospho-actin antibodies. The amino acid sequence surrounding the phosphorylation site in insect actin is identical to the sequence in Physarum actin. C, left panel, detection of phospho-actin in a lysate of plasmodia (lane 1) or sclerotia (lane 2) using anti-actin phosphopeptide antiserum. Middle and right panels, Western blots showing expression of fragmin60 and fragminP in plasmodia and sclerotia. 15 ng of GST-tagged recombinant fragmin60 (lane 1), plasmodia extract (lane 2), sclerotia extract (lane 3), 35 ng of recombinant fragminP (lane 4), plasmodia extract (lane 5), and sclerotia extract (lane 6) were probed with anti-C2 antibodies (lanes 1-3) or anti-fragminP antiserum (lanes 4-6). Pl, plasmodia; Scl, sclerotia; Con, control. D, immunoprecipitation of actin-fragmin60/actin-fragminP from plasmodia and sclerotia (top panel, Coomassie stain), and Western blotting with anti-phospho-actin antibodies (bottom panel). Lanes 1 and 3, immunoprecipitation of actin-fragmin60 and actin-fragminP, respectively, from plasmodia. Lanes 2 and 4, immunoprecipitation of the respective complexes from sclerotia. IgGH, immunoglobulin heavy chain. Phospho-actin is indicated. Anti-C2 antibodies were chemically cross-linked onto G-Sepharose.

Fragmin60 Is Associated with Phosphorylated Actin in Physarum Sclerotia and in Plasmodia-- In sclerotia of Physarum, a dormant stage of the organism formed from plasmodia under dry stress, no less than 50% of all actin is phosphorylated (31). The majority of phosphorylated actin is present as a monomer. Actin phosphorylation in sclerotia prevents polymerization into filaments until more favorable environmental conditions enable the organism to become motile again. Because neither G-actin nor F-actin is phosphorylated by the AFK in vitro or in vivo (16, 37), it is believed that actin phosphorylation always occurs through intermediary complex formation with a fragmin-like protein.

In crude extracts, no phospho-actin could be detected in plasmodia but a strong signal was observed in sclerotia (Fig. 6C, left panel). This may suggest that actin is not phosphorylated in extracts of plasmodia or that phospho-actin levels are below the detection limit and masked by a large excess of unphosphorylated actin. We therefore investigated whether fragmin60 controls actin phosphorylation in vivo in plasmodia and sclerotia by immunoprecipitation of the complex and Western blotting with anti-phospho-actin antibodies. For comparison, phosphorylated actin in the actin-fragminP complex was also studied. Western blots using specific anti-fragminP and anti-C2 antibodies show that fragminP and fragmin60 are both expressed in sclerotia (Fig. 6C, middle and right panels). We observed that actin complexed with fragmin60 was phosphorylated to a similar extent compared with actin complexed with fragminP (Fig. 6D, lanes 1 and 3). In sclerotia we found that actin-fragmin60 as well as actin-fragminP complexes were strongly phosphorylated at actin Thr-203 (Fig. 6D, lanes 2 and 4). From these findings we conclude that both fragmin60 and fragminP participate in controlling actin phosphorylation in plasmodia as well as in sclerotia.

Frg60 Expression Is Developmentally Regulated

We have previously shown that two other members of the Physarum fragmin gene family are encoded by separate genes and that their expression is developmentally regulated; fragminA is amoebae-specific, whereas fragminP is plasmodium-specific (13-14). We analyzed the expression of fragmin60 during apogamic development using a ³²P-labeled cDNA encoding C2 domain as probe. Amoebae that develop apogamically into plasmodia develop from a single amoeba without fusion with another compatible amoeba and therefore remain haploid throughout development (65). Northern blots against RNA isolated from apogamically developing cells showed that fragmin60 (1.8-kb mRNA) is not expressed in amoebae (Fig. 7, top panel). This was confirmed by Western blots on amoebal extracts, where no reaction with a 60-kDa protein was observed (data not shown). Fragmin60 mRNA was detected when as few as 4% of the cells were developing, and the overall pattern of expression is similar to the fragminP (1.2 kb) expression profile (13). The strongest hybridization signals were observed in micro- and macroplasmodia.

View larger version (62K): [in this window] [in a new window]   Fig. 7.   Expression analysis of frg60 during development. Northern blots, using a C2 domain cDNA probe, are shown against total RNA (10 µg/lane) isolated from the apogamic strain Colonia Leicester (CL). A, amoebae; Mi, microplasmodia; Ma, macroplasmodia. The numbers on top indicate the percentage of haploid developing cells in each sample. Frg60 is a 1.8-kb mRNA expressed in committed cells and in micro- and macro plasmodia, but not in amoebae (top panel). Similar results were obtained with the entire fragmin60 cDNA as probe. Middle panel, the same blot was probed with a Physarum cDNA encoding actin. Bottom left panel, expression of the actin-fragmin kinase gene during development. mRNA size markers are shown on the left.

Because the fragmin60 and fragminP isoforms control phosphorylation of actin by the actin-fragmin kinase, and in light of the observation that AFK recognizes only this type of substrate, we studied the expression of the AFK during apogamic development for correlation with the expression patterns of fragmin60 and fragminP. In Fig. 7 (bottom panel), we show that AFK expression is undetectable in amoebae, but expression is initiated early in development because the 2.3-kb mRNA signal is detected in a population of cells that contains 1% committed cells (cells that are irreversibly committed to develop further into plasmodia). These results suggest that the expression profile of the protein kinase and accessory proteins involved in actin phosphorylation is correlated.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

We report the cloning and functional characterization of fragmin60, a protein with a novel variation on the classical domain structure encountered in gelsolin and related proteins from higher or lower organisms. Fragmin60 associates with actin and forms a 1:2 complex in a calcium-dependent manner under non-polymerizing conditions. In addition, is it also functionally related to gelsolin because it severs actin filaments. However, the presence of an N-terminal C2 domain in fragmin60 highlights a major structural difference because such domains are not encountered in F-actin-severing proteins. Fragmin60 C2 does not bind phospholipids but interacts with Skp1. The lack of phospholipid binding is not a singular observation because the middle C2 domain of mammalian unc-13 homologues does not bind phospholipids either, although it harbors all calcium-coordinating residues (66).

The C2 domain from fragmin60 binds three calcium ions with relatively high affinity in comparison to other C2 domains. For instance, K_d values of 0.8 and 24 µM have been reported for plant PLD C2, and 590 µM for plant PLD C2 using isothermal titration calorimetry (67). The fragmin60 C2 domain harbors 5 aspartate residues (Asp-22, Asp-28, Asp-91, Asp-93, Asp-98) also present in other C2 domains, which are involved in coordinate binding of calcium ions. Binding of calcium by the synaptotagmin I C2A domain in the absence of phospholipids was studied here as an internal control during ITC experiments, yielding a K_d value of ~100 µM for the binding site with the highest calcium affinity, which is close to what has been reported earlier (68, 69). It seems likely that the calcium-binding site with nanomolar affinity would be constantly occupied under physiological conditions because calcium concentrations in the cytosol of resting cells approximately amount to 50-100 nM (70). At present there is no evidence in the literature suggesting that regulation of calcium-sensitive proteins in Physarum is controlled by calcium fluctuations significantly different from those observed in mammalian cells. The high affinity binding site in this C2 domain does not, however, exclude a regulatory role, as it may prime binding of the two other calcium ions.

A particular feature of the fragmin60 C2 domain is its occurrence in a protein belonging to a family known to bind calcium and phospholipids (PIP₂), regulating their interaction with actin. As evidenced by the ITC and actin binding experiments, the C2 domain and the actin binding moiety of fragmin60 display different calcium binding affinities. In this context it may seem paradoxical that the high affinity calcium binding (C2) domain of fragmin60 delays calcium-dependent actin binding by the fragmin subdomain (Fig. 4C). This observation could imply that the C2 domain, at low micromolar calcium concentrations, "locks" the fragmin part in a conformation that prevents actin binding. The region separating the C2 domain from the fragmin part may in this respect function as a linker allowing the C2 domain to fold back onto the actin binding part. Surface regions of the C2 domain outside the calcium-binding loops may be involved in interdomain associations. In this scenario, calcium-binding site(s) of the gelsolin core could be responsible for releasing the inhibitory C2 domain. We were, however, unable to show a direct association between both parts of the protein by pull-down assays in calcium-free or low calcium conditions. Possibly, both subdomains interact only weakly; alternatively, regulatory intramolecular interactions may require the entire intact protein. Taking into account that fragmin60C2 requires calcium for actin binding, one can assume that activation of fragmin60 proceeds through several intramolecular rearrangements.

Different calcium binding affinities by C2 and fragmin subdomains could also imply that both subdomains function independently, and that calcium binding by the C2 domain serves an entirely different purpose, i.e. mediating interaction with another component. We are currently investigating calcium-dependent association between the fragmin60 C2 domain with proteins present in the insoluble fraction of Physarum lysates. Here we identified Skp1 as a partner for the fragmin60 C2 domain in the cytosol of Physarum. This interaction was independent of calcium, and specific for fragmin60 C2 because C2 domains from other proteins investigated here were unable to bind native Skp1 from Physarum, or interact with recombinant Skp1 in vitro. These findings further extend recent evidence showing that C2 domains and Skp1 are more versatile in their interactions than previously thought. Indeed, Lopez-Lluch et al. (71) recently reported that the C2 domain of PKC- binds filamentous actin and that this interaction is important in regulating actin redistribution in neutrophils. Skp1, on the other hand, was originally identified as an interaction partner of F-box proteins, which in their turn bind proteins destined for proteolysis through the ubiquitin degradation pathway (60-61). Interestingly, Seol et al. (72) have isolated a complex containing Skp1 (RAVE, a regulator of V-ATPase) that contains neither Cdc53 (cullin1) nor an F-box protein, indicating that Skp1 can form multiple complexes with functions other than protein degradation.

The interaction between Skp1 and fragmin60 C2 points to a possible regulatory role for Skp1 in the control of actin dynamics. This hypothesis is supported by the observation of Sassi et al. (73), who reported that Skp1 frequently co-localizes with F-actin in aggregation-competent Dictyostelium amoebae. Uetz et al. (74), on the other hand, showed that Skp1 interacts with the actin-binding protein coronin in yeast two-hybrid assays. Coronin is known as a promoter of rapid F-actin assembly (75). The finding that association between Skp1 and the C2 domain is of relatively low affinity could indicate that its potential function may involve a local enrichment of the former in areas of microfilament rearrangement.

Our data suggest that fragmin60 as well as fragminP play a role in enhancing actin phosphorylation in sclerotia by the AFK, resulting in polymerization-incompetent actin. Combined, our data provide a model of how the actin phosphorylation level can vary dramatically between different phases of the life cycle of the organism (absent in amoebae, 2-5% in plasmodia, and 50-60% in sclerotia). The afk gene is not transcribed in amoebae, indicating that absence of actin phosphorylation in amoebae is the result of developmental regulation at the level of the protein kinase. Previous data (13), biochemical experiments and results presented here show that the expression profiles of afk, frgP, and frg60 are similar. This may be expected when the physiological role of the actin-fragmin kinase and actin-binding proteins are correlated with respect to actin phosphorylation. Phospho-actin in sclerotia, however, occurs mostly in free monomeric form (31). Because neither G- nor F-actin is phosphorylated by AFK (16), actin in sclerotia has to associate temporarily with fragminP or fragmin60 in order to be phosphorylated. We have shown here that both these intermediates (actin-fragmin and actin-fragmin60) exist in sclerotia. A subsequent dissociation step, mediated by an as yet unknown component although PIP₂ or lysophosphatidic acid (7) are likely candidates, would release monomeric phospho-actin that is no longer able to assemble into F-actin.

In conclusion, we have discovered a new gelsolin-related member with a distinct domain architecture. The C2 domain can engage the fragmin homologue in a network of new interactions important for spatio-temporal regulation of the actin skeleton.

	ACKNOWLEDGEMENTS

We acknowledge the contribution of colleagues who provided various C2 domain constructs. We thank Yvette De Ville for maintaining Physarum cultures.

	FOOTNOTES

^* This work was supported in part by Fund for Scientific Research-Flanders (Belgium) Grants G.0050.02, G.0060.96, and G.0044.97; by the concerted Research Actions Program of Ghent University; and by a grant from the Interuniversity Attraction Poles (to J. V.) (Federal Services for Science, Technology, and Cultural Affairs).The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

The nucleotide sequence(s) reported in this paper has been submitted to the GenBank^TM/EBI Data Bank with accession number(s) AF303112 and AY050559.

^§ Postdoctoral fellow of the Fund for Scientific Research-Flanders (Belgium).

Supported by Wellcome Trust Grant 042524.

^** Present address: Research and Business Development Office, University of Leicester, Leicester LE1 7RF, United Kingdom.

Present address: Medical Genetics, Cambridge University, WTCMMD Wellcome Trust, Addenbrookes Hospital, Cambridge CB2 2XY, United Kingdom.

^§§ To whom correspondence should be addressed: Dept. of Biochemistry, Faculty of Medicine and Health Sciences, Ghent University, Rommelaere Inst., Albert Baertsoenkaai 3, B-9000 Gent, Belgium. Tel.: 32-9-33-13340; Fax: 32-9-33-13597; E-mail: jan.gettemans@rug.ac.be.

Published, JBC Papers in Press, August 7, 2002, DOI 10.1074/jbc.M207052200

	ABBREVIATIONS

The abbreviations used are: PIP₂, phosphatidylinositol 4,5-bisphosphate; AFK, actin-fragmin kinase (EC 2.7.1.37); DTT, dithiothreitol; ECL, enhanced chemiluminescence; GST, glutathione S-transferase; PC, phosphatidylcholine; PMSF, phenylmethylsulfonyl fluoride; PpSkp1, Skp1 homologue from Physarum polycephalum; PS, phosphatidylserine; Q-TOF, quadrupole time-of-flight; MS, mass spectrometry; ITC, isothermal titration calorimetry; TBS, Tris-buffered saline.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

CiteULike

Complore

Connotea

Del.icio.us

Digg

Reddit

Technorati

This article has been cited by other articles:

				N. Kumar and S. Khurana Identification of a Functional Switch for Actin Severing by Cytoskeletal Proteins J. Biol. Chem., June 11, 2004; 279(24): 24915 - 24918. [Abstract] [Full Text] [PDF]

				K. Van Impe, V. De Corte, L. Eichinger, E. Bruyneel, M. Mareel, J. Vandekerckhove, and J. Gettemans The Nucleo-cytoplasmic Actin-binding Protein CapG Lacks a Nuclear Export Sequence Present in Structurally Related Proteins J. Biol. Chem., May 9, 2003; 278(20): 17945 - 17952. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) All Versions of this Article: 277/42/39840    most recent M207052200v1 Submit a Letter to Editor Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Sklyarova, T. Articles by Gettemans, J. Search for Related Content PubMed PubMed Citation Articles by Sklyarova, T. Articles by Gettemans, J. Social Bookmarking                      What's this?