ARTICLE IN PRESS

European Journal of Cell Biology 84 (2005) 951–960

0171-9335/$ - se

doi:10.1016/j.ej

�Correspondfax: +334 7272

E-mail addr

www.elsevier.de/ejcb

A novel phosphatidylinositol 4,5-bisphosphate-binding domain targeting the

Phg2 kinase to the membrane in Dictyostelium cells

Cedric Blanca, Steve Charetteb, Nathalie Cherixb, Yaya Lefkira, Pierre Cossonb,Francois Letourneura,�

aIFR 128 BioSciences Lyon-Gerland, Institut de Biologie et Chimie des Proteines, UMR5086, CNRS/Universite Lyon I,

7 Passage du Vercors, F-69367 Lyon cedex 07, FrancebUniversite de Geneve, Centre Medical Universitaire, Departement de Physiologie Cellulaire et Metabolisme,

CH-1211 Geneve 4, Switzerland

Received 6 July 2005; received in revised form 13 September 2005; accepted 14 September 2005

Abstract

Phg2 is a ser/thr kinase involved in adhesion, motility, actin cytoskeleton dynamics, and phagocytosis inDictyostelium cells. In a search for Phg2 domains required for its localization to the plasma membrane, we identified anew domain interacting with phosphatidylinositol 4,5-bisphosphate (PI(4,5)P2) and phosphatidylinositol 4-phosphate(PI(4)P) membrane phosphoinositides. Deletion of this domain prevented membrane recruitment of Phg2 and properfunction of the protein in the phagocytic process. Moreover, the overexpression of this PI(4,5)P2-binding domainspecifically had a dominant-negative effect by inhibiting phagocytosis. Therefore, plasma membrane recruitment ofPhg2 is essential for its function. The PI(4,5)P2-binding domain fused to GFP (green fluorescent protein) (GFP-Nt-Phg2) was also used to monitor the dynamics of PI(4,5)P2 during macropinocytosis and phagocytosis. GFP-Nt-Phg2disappeared from macropinosomes immediately after their closure. During phagocytosis, PI(4,5)P2 disappeared evenbefore the sealing of phagosomes as it was already observed in mammalian cells. Together these results demonstratethat PI(4,5)P2 metabolism regulates the dynamics and the function of Phg2.r 2005 Elsevier GmbH. All rights reserved.

Keywords: Phagocytosis; Macropinocytosis; Phosphoinositides; Dictyostelium

Introduction

Phagocytosis is the process of ingestion of largeparticles (41 mm in diameter). This process is essentialfor the homeostasis of multicellular organisms because itis involved, for example, in the elimination of damagedcells and of pathogenic microorganisms. Upon recogni-

e front matter r 2005 Elsevier GmbH. All rights reserved.

cb.2005.09.014

ing author. Tel.: +334 7272 2681;

2604.

ess: f.letourneur@ibcp.fr (F. Letourneur).

tion of a particle by receptors on the surface of aphagocytic cell, activation of signaling pathways leadsto a major remodeling of the actin cytoskeleton, theformation of pseudopodial extensions and the seques-tration of the particle in a phagosome (reviewed inAderem and Underhill, 1999).Several approaches have led to a better understanding

of the formation and the subsequent maturation ofphagosomes. Isolation of phagosomes and total proteincontent analysis by mass spectrometry have revealed keyproteins in the phagocytic process (Garin et al., 2001;

ARTICLE IN PRESSC. Blanc et al. / European Journal of Cell Biology 84 (2005) 951–960952

Desjardins, 2003; Okada et al., 2005). However,confirmatory genetic evidences (e.g. gene knock-out)for the function of the identified proteins in thephagocytic process are very often missing. Moreover,systematic genetic approaches aimed at dissecting themolecular events in the phagocytic process have beenhindered by the absence of simple genetic tools to isolateand characterize mammalian phagocytic mutant cells.One substitutive model is the amoeba Dictyostelium

discoideum, which is a professional phagocyte that canfeed upon yeast and bacteria. The phagocytic mecha-nism operating in this unicellular eukaryotic organism isvery similar to that described in mammalian cells(Cardelli, 2001; Rupper and Cardelli, 2001; Duhonand Cardelli, 2002). Because Dictyostelium is haploidand easily amenable to genetic analysis, it has allowedthe detailed characterization of the role of many geneproducts in the phagocytic process (Niewohner et al.,1997; Rivero et al., 1999; Titus, 1999; Cornillon et al.,2000; Rupper et al., 2001; Seastone et al., 2001; Fey etal., 2002; Schreiner et al., 2002; Gebbie et al., 2004).In addition to proteins, phospholipids have been

shown to play a major role in phagocytosis, especially inthe regulation of the actin cytoskeleton during the earlystage of phagosome formation (reviewed in Botelho etal., 2004). In mammalian cells, the plasma membrane-associated phosphatidylinositol 4,5-bisphosphate(PI(4,5)P2) serves as a docking factor for skeletonproteins thus allowing actin polymerization underneaththe contact zone with the particle. PI(4,5)P2 can be alsohydrolyzed by phospholipase Cg into IP3 (inositol 1,4,5-trisphosphate) and diacylglycerol (DAG) but it can alsobe converted to PI(3,4,5)P3 (phosphatidylinositol 3,4,5-trisphosphate) by class I PI3 kinase (phosphatidylinosi-tol 3 kinase). This metabolism leads to PI(4,5)P2disappearance from phagosomal membranes and con-comitant actin depolymerization (Scott et al., 2005)suggesting that PI(4,5)P2 might be critical for actindetachment from the phagosomal membrane. Thedynamics of PI(4,5)P2 membrane recruitment andrelease in Dictyostelium cells has never been assessedfor phagocytosis and macropinocytosis.We recently identified a new kinase, Phg2, involved in

adhesion, motility, actin cytoskeleton dynamics andphagocytosis (Gebbie et al., 2004). Sequence analysisrevealed that Phg2 is not an integral membrane protein.However, Phg2 mainly localized to the plasma mem-brane (Gebbie et al., 2004), suggesting that Phg2interacts with components of the plasmalemma. Inter-estingly, in spite of the continuous remodeling of theplasma membrane during macropinocytosis, Phg2 wasnot seen associated with membranes of intracellularcompartments. Therefore, there must exist a specificmechanism for the tight control of Phg2 localization.In this study, as a first step toward a better under-

standing of Phg2 function, we determined the domain of

Phg2 required for its localization to the plasmamembrane. Within the Phg2 N-terminal (Nt) part, weidentified a new domain interacting with PI(4,5)P2 andPI(4)P (phosphatidylinositol 4-phosphate) membranephosphoinositides. We established that plasma mem-brane recruitment of Phg2 is necessary for the properfunction of the protein in the phagocytic process. Inaddition, we described for the first time the regulation ofthe PI(4,5)P2 level in Dictyostelium membranes duringmacropinocytosis and phagocytosis.

Materials and methods

Cell culture, immunofluorescence and live-cell

microscopy

D. discoideum strain DH1-10 (Cornillon et al., 2000)was grown at 22 1C in HL5 medium and subculturedtwice a week. Cells were not allowed to reach a densityof more than 2� 106 cells/ml. For immunofluorescenceanalysis, cells were applied on a glass coverslip for 2 h,then fixed with 4% paraformaldehyde for 30min. Cellswere observed by laser scanning confocal microscopy(Zeiss LSM 510). For live-cell microscopy, DH1-10 cellsexpressing adenylyl cyclase-pleckstrin homology-greenfluorescent protein (CRAC-PH-GFP) or GFP-Nt-Phg2were pre-cultured in HL5. Three hours before observa-tion under a microscope equipped with a dual-discmicro-lens scanning technology for confocal imaging(QLC100 Confocal Scanner from Visitron Systems,Puchheim, Germany), cells were resuspended to aconcentration of about 106 cells/ml in low fluorescence(LF) medium and 0.5ml of the cell suspension wasplated on circular glass coverslips. The LF mediumcontains per liter 5 g casein peptone (bioMerieux, Lyon,France), 0.35 g yeast extract (Becton, Dickinson andCompany, Basel, Switzerland), 11 g glucose, 2.2 gKH2PO4, 1 g K2HPO4, 100ml tap water. Duringobservation under the microscope, GFP-Nt-Phg2 cellswere incubated in LF medium containing 2.5mg/mlTRITC-dextran (Sigma Aldrich, Saint QuentinFallavier, France) while CRAC-PH-GFP cells wereincubated in LF medium alone. Images were takenevery second and every 2.5 s for cells expressing GFP-Nt-Phg2 or CRAC-PH-GFP, respectively, usingMetamorph 6.0 application software (UniversalImaging Corporation, Downingtown, PA).

Plasmids and cell transfection

All Phg2 mutants were produced by PCR using pairsof oligonucleotides containing BamHI and XhoI sites in50 and 30, respectively. Positions of the truncations areindicated in Figs. 1 and 2. PCR fragments were digested

ARTICLE IN PRESS

Fig. 1. The Nt domain of Phg2 is necessary and sufficient for

its plasma membrane recruitment. (A) Schematic representa-

tion of Phg2 mutants used to determine the membrane-binding

domain of Phg2. GFP was fused to the Nt part of the indicated

protein domains. Each construct was given a name (e.g.

pFL712) indicated on the left of the figure. Positions of the

mutations are indicated on the black line representing protein

domains. (B) Confocal microscopy study of wild-type cells

(DH1-10) expressing the GFP constructs. The fusion of the

first 245 Nt residues of Phg2 to GFP (pFL684) was sufficient

to address the chimera to the plasma membrane. Reciprocally,

the deletion of the first 198Nt residues (pFL726) prevented

membrane recruitment of Phg2. Bar, 10 mm.

Fig. 2. A minimal domain of 113 residues was sufficient for

membrane recruitment of GFP. (A) Schematic representation

of Phg2 mutants used to narrow down the minimal region of

Phg2 required for membrane recruitment. The indicated

constructs (on the left) were stably transfected in wild-type

cells, and their plasma membrane localization assessed by

confocal microscopy analysis (see Fig. 1 for pFL754) and

indicated on the left (+, plasma membrane association; �, no

localization on membranes). (B) Protein sequence of the

minimal domain. This region contains several positively

charged residues (10 Arg, 13 Lys; underlined characters) and

aromatic residues (4 Tyr, 4 Trp, 4 Phe; bold characters).

C. Blanc et al. / European Journal of Cell Biology 84 (2005) 951–960 953

by BamHI and XhoI, cloned into BamHI/XhoI sites ofpDXA-GFP2 (Levi et al., 2000) and sequenced (Ge-nome express, Grenoble, France). Plasmids were linear-ized by ScaI and transfected in Dictyostelium byelectroporation as described (Cornillon et al., 2000).Plasmid WF38 was a kind gift from P. Devreotes (JohnsHopkins Medical Institutions, Baltimore, MD). Thisplasmid (Parent et al., 1998) contains the sequencecoding for the PH domain of the cytosolic regulator ofCRAC fused to GFP under the control of the actinpromoter, for constitutive expression in Dictyostelium

cells. GST-Nt-Ph2 (GST, glutathione S-transferase) wasprepared by cloning a PCR fragment corresponding tothe first 245 Nt residues of Phg2 into pGEX4T1(Amersham, Saclay, France). The GST-PH domain ofPLC-d, used as a control for PI(4,5)P2 binding, was akind gift of M.A. Lemmon (University of PennsylvaniaSchool of Medicine, Philadelphia, PA). GST fusion

proteins were prepared as described previously (Cossonand Letourneur, 1994) and eluted in phosphate-bufferedsaline (PBS) containing 10mM reduced glutathione.

Lipid dot-blot

Lipid dot-blot assays were done on a PIP strip(Echelon Biosciences Inc., Salt Lake City, UT) accord-ing to the manufacturer’s instructions. Purified GSTfusion proteins were incubated at 2 mg/ml with the PIPstrip at room temperature for 3 h. After several washes,protein binding on the membrane was analyzed usinganti-GST antibodies (Sigma Aldrich).

Macropinocytosis and phagocytosis uptake assays

To assess internalization, 105 cells were transferredinto 1ml fresh HL5 medium containing 1 ml FluoresbriteYG carboxylate microspheres (1 mm diameter; Poly-sciences Inc., Warrington, PA) or 0.5mg/ml FITC-dextran (Molecular Probes, Eugene, OR). The cells wereincubated with shaking (200 rpm) for 1 h, then washedtwice with ice-cold HL5 and analyzed using a fluores-cence spectrofluorometer (FACSCalibur, BecktonDickinson, San Jose, CA) as previously described(Cornillon et al., 2002).

ARTICLE IN PRESSC. Blanc et al. / European Journal of Cell Biology 84 (2005) 951–960954

Results

The Nt domain of Phg2 is necessary and sufficient

for its plasma membrane recruitment

The protein kinase Phg2 is a peripheral membraneprotein, which contains several domains required for itsfunction (e.g. the Ras-binding domain, RBD). Todetermine which domain of Phg2 is responsible for itsspecific localization to the plasma membrane, differentregions of Phg2 covering the whole protein were fused toGFP (Fig. 1A). These mutants were transfected in DH1cells and their localization assessed by confocal micros-copy analysis of fixed cells (Fig. 1B). The fusion of the first245 Nt residues of Phg2 to GFP was sufficient to addressthe chimera to the plasma membrane. In contrast,chimeras containing either the kinase domain (pFL695),the RBD (pFL683) or the Ct domain (pFL716) of Phg2were not targeted to the cell surface. Interestingly, thedeletion of the first 198 Nt residues (pFL726) preventedmembrane recruitment of Phg2. Together these resultsestablish that the Nt domain of Phg2 is necessary andsufficient for its docking to the plasma membrane.To narrow down the minimal region of Phg2

implicated in membrane recruitment, additional trunca-tions were made (Fig. 2). Microscopy studies of cellsexpressing the resulting GFP proteins established that aminimal domain of 113 residues was sufficient formembrane recruitment of GFP (Figs. 1 and 2). Theamino-acid sequence analysis revealed that this region isrich in positively charged residues (10 Arg, 13 Lys) andpresents several aromatic residues (4 Tyr, 4 Trp, 4 Phe).Blast searches did not reveal any significant similarities withalready described domains involved in membrane targeting.

Plasma membrane recruitment of Phg2 is essential

for its function

To determine whether the specific localization of Phg2to the plasma membrane was required for its function,

Fig. 3. Plasma membrane localization of Phg2 is essential for

phagocytosis. The indicated transfected cells were incubated

for 1 h with FITC-labeled latex beads (A, B) or with FITC-

dextran (C). The internalized fluorescence was measured in a

fluorescence-activated cell sorter. Data are the means plus

standard deviation of at least four independent experiments.

The expression level of chimeric proteins was tested by

Western blotting with anti-GFP and was comparable in all

strains (not shown). Phg2� cells exhibit severe defects in their

ability to phagocytose latex beads. Whereas GFP-Phg2

complemented the phagocytic defect of phg2� cells, the

deletion of the PI(4,5)P2-binding domain prevented phg2�

functional complementation. Overexpression of the Phg2

PI(4,5)P2-binding domain (pFL684) but not of the kinase

domain (pFL695) in wild-type cells inhibits phagocytosis but

not macropinocytosis.

full-length (pFL712) or Phg2 deleted for the first 198 Ntresidues (pFL726) was expressed in Phg2 minus cells(phg2�). Phagocytosis of latex beads was assayed foreach type of transfected cells in HL5 culture medium.Whereas the expression of GFP-Phg2 complemented thephagocytic defect of phg2� to a level comparable toparental wild-type cells, the deletion of the Nt domainprevented this complementation (Fig. 3A and B).

ARTICLE IN PRESS

Fig. 4. Phg2 binds membrane phosphoinositides in vitro.

(A) protein-lipid overlays. Glutathione S-transferase (GST)

fusion proteins comprising the Phg2 Nt domain (left panel) or

the PLC-d PH domain (right panel) were used to probe two

identical nitrocellulose membranes containing various immo-

bilized phosphoinositides (PIP strip, Echelon Biosciences). For

each strip, lipids were deposed in two rows and only labeled

outside. The GST-Nt-Phg2 domain binds synthetic PI(4)P and

PI(4,5)P2 whereas no interaction was detected with control

phospholipids, e.g. phosphatidylethanolamine (PE), phospha-

tidylcholine (PC), phosphatidylserine (PS), and phosphatidic

acid (PA). (B) Confocal microscopy study of cells expressing

pFL684 treated overnight (HL5+neo) or not (HL5) with

10mM neomycin sulfate, a drug that competes with intracel-

lular ligands of PI(4,5)P2. Upon neomycin treatment, GFP-Nt-

Phg2 membrane staining is lost, and the chimera mainly

localizes in a dotted pattern throughout the cytosol instead.

Bar, 10 mm.

C. Blanc et al. / European Journal of Cell Biology 84 (2005) 951–960 955

Therefore, plasma membrane localization of Phg2 isrequired for its function. Interestingly, overexpression ofthis Nt domain alone fused to GFP in wild-typecells inhibited phagocytosis, whereas overexpression ofthe Phg2 kinase domain also fused to GFP had noeffect (Fig. 3B). The exact mechanism of this dominant-negative effect remains uncertain, although onelikely possibility is that GFP-Nt-Phg2 could competewith endogenous Phg2 for membrane localization.Noteworthy, the overexpression of GFP-Nt-Phg2did not inhibit macropinocytosis (Fig. 3C), thusreinforcing the hypothesis that the dominant-negative effect is due to the recruitment of thischimeric protein on membranes, preventing the properfunctioning of endogenous Phg2 because of defectivelocalization.

Phg2 binds membrane phosphoinositides in vitro

The fact that the Phg2 Nt domain was sufficient tolocalize GFP to the plasma membrane suggested thatthis domain was able to interact with a docking factoron the internal face of the plasma membrane. Using anaffinity chromatography column, we were not able todetect any Phg2 Nt domain-binding protein in Dictyo-

stelium lysates. Furthermore, a yeast two-hybrid screenusing the Phg2 Nt domain as a bait failed to identify anybinding proteins. We therefore examined the possibilitythat the Phg2 Nt domain bound to membrane phos-phoinositides.To directly assess the binding property of the Phg2 Nt

domain to phosphoinositides, we carried out protein–lipid overlays. A GST fusion protein with Phg2 Ntdomain (GST-Nt-Phg2) was used to probe a nitrocellu-lose membrane containing various immobilized phos-phoinosides (PIP strip, Echelon Biosciences). As shownin Fig. 4A, GST-Nt-Phg2 specifically bound syntheticPI(4)P and PI(4,5)P2, whereas no interaction wasdetected with control phospholipids (PA, PS, PE, PI,PI(3)P, PI(5)P, PI(3,4)P2). The absence of interactionwith animal-derived PI(4,5)P2 could be due either to thepurification process of PI(4,5)P2 from animals leading toan overestimate of the real amount of PI(4,5)P2immobilized on membranes or to the nature of fattyacid chains found in animal PI(4,5)P2. Interestingly, inthis dot-blot assay a similar pattern was observed withthe PH domain of PLC-d well-known to interact withPI(4,5)P2 (Fig. 4A).

Subcellular distribution of the Phg2 Nt domain

correlates with PI(4,5)P2 binding

The lipid dot-blot assay suggested that the plasmamembrane phospholipid PI(4,5)P2 could serve as adocking site for Phg2 in vivo. Consequently, reducing

the availability of PI(4,5)P2 on membranes shouldinterfere with Phg2 localization. Neomycin is a high-affinity ligand for polyphosphorylated phosphoinosi-tides (Wang et al., 1984), and in particular it has beenshown to compete with intracellular ligands of PI(4,5)P2(Bompard et al., 2003). To test the requirement ofPI(4,5)P2 for Phg2 membrane recruitment, cells wereincubated overnight with 10mM neomycin sulfate andthe localization of GFP-Nt-Phg2 was analyzed byconfocal microscopy. As shown in Fig. 4B, GFP-Nt-Phg2 membrane staining was lost in neomycin-treatedcells and the chimera was distributed in a dotted patternthroughout the cytosol instead. Note that, althoughneomycin treatment could have possible side effects,treated cells were still viable and washing out neomycinsulfate restored plasma membrane localization of GFP-Nt-Phg2 (data not shown).

ARTICLE IN PRESSC. Blanc et al. / European Journal of Cell Biology 84 (2005) 951–960956

Dynamics of Phg2 Nt membrane localization during

macropinocytosis

Observation of cells expressing GFP-Nt-Phg2 afterfixation revealed that in addition to the plasmamembrane, the fusion protein also localized to intracel-lular compartments. This minor localization was onlyseen in a small proportion of the cells and with anintensity lower than the one at the plasma membrane.We suspected that this intracellular localization of GFP-Nt-Phg2 could correspond to newly formed macropino-somes. To confirm this hypothesis, live confocal micros-copy was performed on cells expressing GFP-Nt-Phg2and incubated with TRITC-dextran in order to followfluid-phase endocytosis and macropinosome formation.As shown in Fig. 5A, the fusion protein decoratednascent macropinosomes but disappeared gradually andrapidly around this compartment.To compare the kinetics of disappearance of GFP-Nt-

Phg2 from macropinosomes with that of other macro-pinosome markers, the behavior of the PH domain ofthe cytosolic regulator of CRAC (Parent et al., 1998)fused to GFP (CRAC-PH-GFP) was also analyzed. ThePH domain of CRAC binds PI(3,4)P2 and PI(3,4,5)P3(Huang et al., 2003) and is recruited as patch to theplasma membrane which is active for pseudopod,macropinosome or phagosome formation (Fig. 5B)(Parent et al., 1998; Tuxworth et al., 2001). Aspreviously reported (Dormann et al., 2004), CRAC-PH-GFP remained about 1min around new macropino-somes while GFP-Nt-Phg2 disappeared in about 20–30 s(Fig. 5B and C). Interestingly, CRAC-PH-GFP recruit-ment still increased after the closure of the macropino-some. CRAC-PH-GFP staining reached its maximum7–10 s after closure and then started to disappeargradually (Fig. 5C). On the other hand, GFP-Nt-Phg2started to disappear immediately after the closure(Fig. 5C).

Dynamics of Phg2 Nt membrane localization during

phagocytosis

Recent studies on mammalian phagocytes haverevealed a transient biphasic distribution of PI(4,5)P2at the level of forming phagosomes (Botelho et al.,2000). To test whether this was the case in Dictyostelium,cells expressing GFP-Nt-Phg2 were incubated with yeastand fixed after a short incubation period to detect earlyevents. In contrast to a PI(4,5)P2 increase during theearly stage of phagocytosis in mammalian cells, we didnot detect any local increase in GFP-Nt-Phg2 fluores-cence on phagocytic cups (Fig. 6). However, as reportedfor PI(4,5)P2-binding proteins on mammalian phago-somes (Botelho et al., 2000), GFP-Nt-Phg2 staining waslost before the closure of phagosomes, and this

disappearance initiated at the base of nascent phago-somes. Therefore, these data suggest that in Dictyo-

stelium cells, PI(4,5)P2 is not produced or locallyaccumulated during the early stage of phagocytosis,but it is rapidly metabolized even before formation ofphagosomes is completed.

Discussion

A new PI(4,5)P2-binding domain

In this study we identified a new PI(4,5)P2-bindingdomain in the ser/thr kinase Phg2. Since the mainlocalization of PI(4,5)P2 is the plasma membrane, thisdomain allows the specific recruitment of Phg2 to thismembrane. In addition, PI(4,5)P2 is known to be rapidlymetabolized in membranes during internalization pro-cesses. Therefore, this lipid metabolism provides aconvenient system to control the dynamics of Phg2localization on membranes.There are many protein domains that interact with

PI(4,5)P2 and they all play important roles in controllingendocytosis, exocytosis and cytoskeleton rearrange-ment. These domains include PH domains, ENTHdomains (Epsin N-terminal homology), ANTH domains(AP180 N-terminal homology), AP-2 Nt domain, tubbyand FERM domains (4.1-ezrin-radixin-moesin) (re-viewed in Lemmon, 2003). In addition, several proteinsinvolved in recruiting the cytoskeleton to the plasmamembrane have been shown to interact with PI(4,5)P2through short stretches of basic residues (Sechi andWehland, 2000; Niggli, 2001;Yin and Janmey, 2003).For instance a polybasic domain in Wiskott-Aldrichsyndrome protein (WASP) binds PI(4,5)P2 (Papayanno-poulos et al., 2005). Basic and aromatic residues ofmyristoylated alanine-rich C kinase substrate(MARCKS) were shown to bind PI(4,5)P2 with highaffinity (Wang et al., 2001, 2002).The 113-residue domain of Phg2 interacting with

PI(4,5)P2 has no homology with any domains known tointeract with PI(4,5)P2. However, it contains severalstretches of basic residues as well as 12 aromaticresidues. Truncation of the 28 first Nt residues contain-ing two diaromatic stretches (Y91W92 and Y99W100,pFL755) of the Phg2 PI(4,5)P2-binding domain preventsplasma membrane localization of the correspondingGFP chimera. Furthermore, the deletion of the last 38Ct residues of this Phg2 domain, removing two basicstretches (R157RRK160 and K171RKK174, pFL785) alsoinhibits plasma membrane recruitment of GFP. Sinceany deletion of basic or aromatic regions in the 113-residue domain of Phg2 is deleterious for membraneinteraction, we propose that all these regions act inconcert for PI(4,5)P2 recognition, although we cannot

ARTICLE IN PRESS

Fig. 5. GFP-Nt-Phg2 and CRAC-PH-GFP have different kinetics of disappearance from newly produced macropinosomes. Time

series showing the formation of a macropinosome in cells expressing GFP-Nt-Phg2 in presence of (A) TRITC-dextran or (B) in cells

expressing CRAC-GFP. In (A) TRITC-dextran is used to identify the macropinosome after disappearance of GFP-Nt-Phg2 around

it. In (B) the green background caused by diffused CRAC-GFP is enough to distinguish the macropinosome even after the

disappearance of the strong CRAC-GFP signal around it. (C) Graphic representation of the loss of GFP-Nt-Phg2 and CRAC-PH-

GFP staining around macropinosomes. The time 0 corresponds to the macropinosome closure. For GFP-Nt-Phg2, the fluorescence

ratio of macropinosome staining was calculated as the (fluorescence on macropinosome/fluorescence on plasma mem-

brane)� 100%. For CRAC-PH-GFP, the fluorescence ratio of macropinosome staining was calculated as the [(fluorescence on

macropinosome minus background fluorescence)/highest corrected value]� 100%. The CRAC-PH-GFP curve is the mean of the

analysis of five different macropinosomes while the GFP-Nt-Phg2 curve is the mean of 16 different macropinosomes. The standard

error of the mean is shown for each curve.

C. Blanc et al. / European Journal of Cell Biology 84 (2005) 951–960 957

ARTICLE IN PRESS

Fig. 6. GFP-Nt-Phg2 membrane localization during phagocy-

tosis. Cells expressing GFP-Nt-Phg2 were incubated with yeast

cells labeled with TRITC and fixed after 10min of incubation.

Cells were then observed by confocal microscopy. Left and

right panels correspond to two representative examples. GFP-

Nt-Phg2 does not present any local enrichment on nascent

phagocytic cups. However, GFP-Nt-Phg2 staining is lost

before the closure of phagosomes and this disappearance

proceeded at the base of nascent phagosomes. Bar, 10 mm.

C. Blanc et al. / European Journal of Cell Biology 84 (2005) 951–960958

formally exclude that these deletions disrupt the foldingof this membrane-interacting domain. Further studieswill be required to determine whether each stretch bindsindividually with low-affinity PI(4,5)P2 or if this Phg2domain adopts a particular folding which brings allbasic and aromatic amino acid stretches to form a high-affinity lipid-binding site on Phg2. In addition toPI(4,5)P2, this same Phg2 domain binds PI(4)P in vitro.In mammalian cells, PI(4)P is mainly localized on Golgimembranes although a minor fraction is also describedon plasma membrane (Balla et al., 2005). Since Phg2 isnot associated with internal membranes (e.g. Golgiapparatus), it is not clear whether this in vitrointeraction has any physiological relevance.

Dynamics of PI(4,5)P2 during macropinocytosis and

phagocytosis

The dynamics of phophoinositides has been recentlydeciphered by the use of specific sensors of theirmetabolism during different cellular processes. Hence,studies on mammalian phagocytes have shown atransient biphasic distribution of PI(4,5)P2 at the levelof forming phagosomes (Botelho et al., 2000). Uponparticle recognition there is a rapid accumulation ofPI(4,5)P2 concomitant with the recruitment of type Iphosphatidylinositolphosphate kinase and actin poly-merization. This step is rapidly followed by thedisappearance of PI(4,5)P2 which directs actin disas-sembly. However in several cell types, the presumedlocal enrichment of PI(4,5)P2 on membranes appearsactually due to the poor resolution of membrane foldsby confocal microscopy (van Rheenen and Jalink, 2002).Therefore, PI(4,5)P2 levels on forming phagosomes hasto be carefully re-evaluated for macrophages.

In this study, we took advantage of the Phg2PI(4,5)P2-binding domain to monitor the dynamics ofthis phospholipid in two related mechanisms, macro-pinocytosis and phagocytosis in the model phagocyte,D. discoideum. In contrast to mammalian macrophages,PI(4,5)P2 does not concentrate at the site of interactionbetween plasma membrane and particles in Dictyoste-

lium. Although a local production of PI(4,5)P2 might betoo rapid for detection in our phagocytic assays, thisdiscrepancy may reflect the fact that the recruitment ofactin in Dictyostelium during phagocytosis might becontrolled by a PI(4,5)P2-independent mechanism.Interestingly, the actin cytoskeleton plays also an

essential role in macropinocytosis, and phosphoinosi-tides are important in this process as well (Cardelli,2001; Simonsen et al., 2001). As for phagocytosis, noaccumulation of PI(4,5)P2 was detected on sites ofpseudopod formation suggesting again that the localmobilization or synthesis of this phospholipid is also nota critical step for macropinocytosis in Dictyostelium. Itis known that PI(3,4,5)P3 is produced at plasmamembrane sites active for pseudopod, macropinosomeand phagosome formation through activation of PI3kinase, which converts PI(4,5)P2 in PI(3,4,5)P3 (Van-haesebroeck et al., 2001). The kinetics of CRAC-PH-GFP (a probe for PI(3,4,5)P3) recruitment suggests thatthe peak level of production is only reached after theclosure of the macropinosome. This result is inagreement with recently published work (Dormannet al., 2004). The concomitant release of GFP-Nt-Phg2and recruitment of CRAC-PH-GFP in the first 10 s afterthe closure of macropinosomes strongly suggests thatPI(4,5)P2 present at the plasma membrane and recog-nized by GFP-Nt-Phg2 is removed from macropino-somes by its conversion in PI(3,4,5)P3. In addition,PI(4,5)P2 can be metabolized to yield DAG and IP3 byphospholipase C, and therefore this reaction could alsoaccount for the loss of GFP-Nt-Phg2 staining onmacropinosome membranes.

Membrane localization of Phg2 is required for its

function

Phg2 is constitutively present at the level of theplasma membrane, and our results indicate that thislocalization is essential for its function in phagocytosissince deletion of the PI(4,5)P2-binding domain in Phg2prevents complementation of the phagocytic defect of aPhg2 minus strain. Kinases implicated in cell adhesionsuch as focal adhesion kinase (FAK) or integrin-linkedkinase (ILK) are recruited on transmembrane receptorsonly upon ligand recognition, leading to the subsequentactivation of these kinases (Schwartz, 2001; Wu andDedhar, 2001). Since Phg2 is already associated to theplasma membrane, there must exist a mechanism to

ARTICLE IN PRESSC. Blanc et al. / European Journal of Cell Biology 84 (2005) 951–960 959

regulate the activity of Phg2 and keep it inactive in theabsence of particles. This hypothesis is currently underinvestigation.In conclusion, phagocytosis and macropinocytosis

processes are finely tuned by the sequential apparition ofnew phosphoinositides on phagosome and macropino-some membranes, which assists the recruitment ofproteins essential for the formation of these organelles.We demonstrate here that PI(4,5)P2 metabolism pro-vides a rapid and efficient mechanism to regulate thelocalization of the Phg2 kinase on phagosomal mem-branes.

Acknowledgements

We thank Dr. P. Rousselle (IBCP, France) for helpfuldiscussions and encouragements. This work was sup-ported by grants from the Association pour la Re-cherche sur le Cancer (ARC), and the ‘‘Emergence’’program of the Region Rhone-Alpes. P. Cosson’slaboratory is funded by the Fonds National Suisse pourla Recherche Scientifique. S.J. Charette received afellowship from the Human Frontier Science Program.

References

Aderem, A., Underhill, D.M., 1999. Mechanisms of phagocy-

tosis in macrophages. Annu. Rev. Immunol. 17,

593–623.

Balla, A., Tuymetova, G., Tsiomenko, A., Varnai, P., Balla,

T., 2005. A plasma membrane pool of phosphatidylinositol

4-phosphate is generated by phosphatidylinositol 4-kinase

type-III alpha: studies with the PH domains of the

oxysterol binding protein and FAPP1. Mol. Biol. Cell 16,

1282–1295.

Bompard, G., Martin, M., Roy, C., Vignon, F., Freiss, G.,

2003. Membrane targeting of protein tyrosine phosphatase

PTPL1 through its FERM domain via binding to

phosphatidylinositol 4,5-biphosphate. J. Cell Sci. 116,

2519–2530.

Botelho, R.J., Teruel, M., Dierckman, R., Anderson, R.,

Wells, A., York, J.D., Meyer, T., Grinstein, S., 2000.

Localized biphasic changes in phosphatidylinositol-4,5-

bisphosphate at sites of phagocytosis. J. Cell Biol. 151,

1353–1368.

Botelho, R.J., Scott, C.C., Grinstein, S., 2004. Phosphoinosi-

tide involvement in phagocytosis and phagosome matura-

tion. Curr. Top. Microbiol. Immunol. 282, 1–30.

Cardelli, J., 2001. Phagocytosis and macropinocytosis in

Dictyostelium: phosphoinositide-based processes, biochemi-

cally distinct. Traffic 2, 311–320.

Cornillon, S., Pech, E., Benghezal, M., Ravanel, K., Gaynor,

E., Letourneur, F., Bruckert, F., Cosson, P., 2000. Phg1p is

a nine-transmembrane protein superfamily member in-

volved in Dictyostelium adhesion and phagocytosis. J. Biol.

Chem. 275, 34287–34292.

Cornillon, S., Dubois, A., Bruckert, F., Lefkir, Y., Marchetti,

A., Benghezal, M., De Lozanne, A., Letourneur, F.,

Cosson, P., 2002. Two members of the beige/CHS

(BEACH) family are involved at different stages in the

organization of the endocytic pathway in Dictyostelium. J.

Cell Sci. 115, 737–744.

Cosson, P., Letourneur, F., 1994. Coatomer interaction with

di-lysine endoplasmic reticulum retention motifs. Science

263, 1629–1631.

Desjardins, M., 2003. ER-mediated phagocytosis: a new

membrane for new functions. Nat. Rev. Immunol. 3,

280–291.

Dormann, D., Weijer, G., Fowler, S., Weijer, C.J., 2004. In

vivo analysis of 3-phosphoinositide dynamics during

Dictyostelium phagocytosis and chemotaxis. J. Cell Sci.

117, 6497–6509.

Duhon, D., Cardelli, J., 2002. The regulation of phagosome

maturation in Dictyostelium. J. Muscle Res. Cell Motil. 23,

803–808.

Fey, P., Stephens, S., Titus, M.A., Chisholm, R.L., 2002.

SadA, a novel adhesion receptor in Dictyostelium. J. Cell

Biol. 159, 1109–1119.

Garin, J., Diez, R., Kieffer, S., Dermine, J.F., Duclos, S.,

Gagnon, E., Sadoul, R., Rondeau, C., Desjardins, M.,

2001. The phagosome proteome: insight into phagosome

functions. J. Cell Biol. 152, 165–180.

Gebbie, L., Benghezal, M., Cornillon, S., Froquet, R., Cherix,

N., Malbouyres, M., Lefkir, Y., Grangeasse, C., Fache, S.,

Dalous, J., Bruckert, F., Letourneur, F., Cosson, P., 2004.

Phg2, a kinase involved in adhesion and focal site modeling

in Dictyostelium. Mol. Biol. Cell 15, 3915–3925.

Huang, Y.E., Iijima, M., Parent, C.A., Funamoto, S., Firtel,

R.A., Devreotes, P., 2003. Receptor-mediated regulation of

PI3Ks confines PI(3,4,5)P3 to the leading edge of chemo-

taxing cells. Mol. Biol. Cell 14, 1913–1922.

Lemmon, M.A., 2003. Phosphoinositide recognition domains.

Traffic 4, 201–213.

Levi, S., Polyakov, M., Egelhoff, T.T., 2000. Green fluorescent

protein and epitope tag fusion vectors for Dictyostelium

discoideum. Plasmid 44, 231–238.

Niewohner, J., Weber, I., Maniak, M., Muller-Taubenberger,

A., Gerisch, G., 1997. Talin-null cells of Dictyostelium are

strongly defective in adhesion to particle and substrate

surfaces and slightly impaired in cytokinesis. J. Cell Biol.

138, 349–361.

Niggli, V., 2001. Structural properties of lipid-binding sites in

cytoskeletal proteins. Trends Biochem. Sci. 26,

604–611.

Okada, M., Huston, C.D., Mann, B.J., Petri Jr., W.A., Kita,

K., Nozaki, T., 2005. Proteomic analysis of phagocytosis in

the enteric protozoan parasite Entamoeba histolytica.

Eukaryot. Cell 4, 827–831.

Papayannopoulos, V., Co, C., Prehoda, K.E., Snapper, S.,

Taunton, J., Lim, W.A., 2005. A polybasic motif allows N-

WASP to act as a sensor of PIP(2) density. Mol. Cell 17,

181–191.

Parent, C.A., Blacklock, B.J., Froehlich, W.M., Murphy,

D.B., Devreotes, P.N., 1998. G protein signaling events are

activated at the leading edge of chemotactic cells. Cell 95,

81–91.

ARTICLE IN PRESSC. Blanc et al. / European Journal of Cell Biology 84 (2005) 951–960960

Rivero, F., Albrecht, R., Dislich, H., Bracco, E., Graciotti, L.,

Bozzaro, S., Noegel, A.A., 1999. RacF1, a novel member of

the Rho protein family in Dictyostelium discoideum,

associates transiently with cell contact areas, macropino-

somes, and phagosomes. Mol. Biol. Cell 10, 1205–1219.

Rupper, A., Cardelli, J., 2001. Regulation of phagocytosis and

endo-phagosomal trafficking pathways in Dictyostelium

discoideum. Biochim. Biophys. Acta 1525, 205–216.

Rupper, A., Grove, B., Cardelli, J., 2001. Rab7 regulates

phagosome maturation in Dictyostelium. J. Cell Sci. 114,

2449–2460.

Schreiner, T., Mohrs, M.R., Blau-Wasser, R., von Krempel-

huber, A., Steinert, M., Schleicher, M., Noegel, A.A., 2002.

Loss of the F-actin binding and vesicle-associated protein

comitin leads to a phagocytosis defect. Eukaryot. Cell 1,

906–914.

Schwartz, M.A., 2001. Integrin signaling revisited. Trends Cell

Biol. 11, 466–470.

Scott, C.C., Hobson, W., Botelho, R.J., Coady-Osberg, N.,

Chavrier, P., Knecht, D.A., Heath, C., Stahl, P., Grinstein,

S., 2005. Phosphatidylinositol-4,5-bisphosphate hydrolysis

directs actin remodeling during phagocytosis. J. Cell Biol.

169, 139–149.

Seastone, D.J., Harris, E., Temesvari, L.A., Bear, J.E., Saxe,

C.L., Cardelli, J., 2001. The WASp-like protein scar

regulates macropinocytosis, phagocytosis and endosomal

membrane flow in Dictyostelium. J. Cell Sci. 114,

2673–2683.

Sechi, A.S., Wehland, J., 2000. The actin cytoskeleton and

plasma membrane connection: PtdIns(4,5)P(2) influences

cytoskeletal protein activity at the plasma membrane. J.

Cell Sci. 113, 3685–3695.

Simonsen, A., Wurmser, A.E., Emr, S.D., Stenmark, H., 2001.

The role of phosphoinositides in membrane transport.

Curr. Opin. Cell Biol. 13, 485–492.

Titus, M.A., 1999. A class VII unconventional myosin is

required for phagocytosis. Curr. Biol. 9, 1297–1303.

Tuxworth, R.I., Weber, I., Wessels, D., Addicks, G.C., Soll,

D.R., Gerisch, G., Titus, M.A., 2001. A role for myosin VII

in dynamic cell adhesion. Curr. Biol. 11, 318–329.

Vanhaesebroeck, B., Leevers, S.J., Ahmadi, K., Timms, J.,

Katso, R., Driscoll, P.C., Woscholski, R., Parker, P.J.,

Waterfield, M.D., 2001. Synthesis and function of 3-

phosphorylated inositol lipids. Annu. Rev. Biochem. 70,

535–602.

van Rheenen, J., Jalink, K., 2002. Agonist-induced PIP(2)

hydrolysis inhibits cortical actin dynamics: regulation at a

global but not at a micrometer scale. Mol. Biol. Cell 13,

3257–3267.

Wang, B.M., Weiner, N.D., Takada, A., Schacht, J., 1984.

Characterization of aminoglycoside-lipid interactions and

development of a refined model for ototoxicity testing.

Biochem. Pharmacol. 33, 3257–3262.

Wang, J., Arbuzova, A., Hangyas-Mihalyne, G., McLaughlin,

S., 2001. The effector domain of myristoylated alanine-rich

C kinase substrate binds strongly to phosphatidylinositol

4,5-bisphosphate. J. Biol. Chem. 276, 5012–5019.

Wang, J., Gambhir, A., Hangyas-Mihalyne, G., Murray, D.,

Golebiewska, U., McLaughlin, S., 2002. Lateral sequestra-

tion of phosphatidylinositol 4,5-bisphosphate by the basic

effector domain of myristoylated alanine-rich C kinase

substrate is due to nonspecific electrostatic interactions. J.

Biol. Chem. 277, 34401–34412.

Wu, C., Dedhar, S., 2001. Integrin-linked kinase (ILK) and its

interactors: a new paradigm for the coupling of extra-

cellular matrix to actin cytoskeleton and signaling com-

plexes. J. Cell Biol. 155, 505–510.

Yin, H.L., Janmey, P.A., 2003. Phosphoinositide regulation

of the actin cytoskeleton. Annu. Rev. Physiol. 65,

761–789.