Toxofilin, a Novel Actin-binding Protein from Toxoplasma gondii, Sequesters Actin Monomers and Caps Actin Filaments

Abstract

Toxoplasma gondii relies on its actin cytoskeleton to glide and enter its host cell. However, T. gondii tachyzoites are known to display a strikingly low amount of actin filaments, which suggests that sequestration of actin monomers could play a key role in parasite actin dynamics. We isolated a 27-kDa tachyzoite protein on the basis of its ability to bind muscle G-actin and demonstrated that it interacts with parasite G-actin. Cloning and sequence analysis of the gene coding for this protein, which we named Toxofilin, showed that it is a novel actin-binding protein. In in vitro assays, Toxofilin not only bound to G-actin and inhibited actin polymerization as an actin-sequestering protein but also slowed down F-actin disassembly through a filament end capping activity. In addition, when green fluorescent protein-tagged Toxofilin was overexpressed in mammalian nonmuscle cells, the dynamics of actin stress fibers was drastically impaired, whereas green fluorescent protein-Toxofilin copurified with G-actin. Finally, in motile parasites, during gliding or host cell entry, Toxofilin was localized in the entire cytoplasm, including the rear end of the parasite, whereas in intracellular tachyzoites, especially before they exit from the parasitophorous vacuole of their host cell, Toxofilin was found to be restricted to the apical end.

Full text

Molecular Biology of the Cell
Vol. 11, 355–368, January 2000
Toxofilin, a Novel Actin-binding Protein from
Toxoplasma gondii, Sequesters Actin Monomers and
Caps Actin Filaments
Olivier Poupel,* Haralabia Boleti,
†
Sophie Axisa,* Evelyne Couture-Tosi,
‡
and Isabelle Tardieux*
§
*Laboratoire de Biochimie et Biologie Mole´culaire des Insectes,
†
Unite´ de Biologie des Interactions
Cellulaires, Unite´ de Recherche Associe´e, Centre National de la Recherche Scientifique 1960, and
‡
Station Centrale de Microscopie E
´lectronique, Institut Pasteur, 75724 Paris Cedex 15, France
Submitted June 18, 1999; Revised October 13, 1999; Accepted October 14, 1999
Monitoring Editor: Tim Stearns
Toxoplasma gondii relies on its actin cytoskeleton to glide and enter its host cell. However, T.
gondii tachyzoites are known to display a strikingly low amount of actin filaments, which
suggests that sequestration of actin monomers could play a key role in parasite actin dynam-
ics. We isolated a 27-kDa tachyzoite protein on the basis of its ability to bind muscle G-actin
and demonstrated that it interacts with parasite G-actin. Cloning and sequence analysis of the
gene coding for this protein, which we named Toxofilin, showed that it is a novel actin-
binding protein. In in vitro assays, Toxofilin not only bound to G-actin and inhibited actin
polymerization as an actin-sequestering protein but also slowed down F-actin disassembly
through a filament end capping activity. In addition, when green fluorescent protein-tagged
Toxofilin was overexpressed in mammalian nonmuscle cells, the dynamics of actin stress
fibers was drastically impaired, whereas green fluorescent protein-Toxofilin copurified with
G-actin. Finally, in motile parasites, during gliding or host cell entry, Toxofilin was localized
in the entire cytoplasm, including the rear end of the parasite, whereas in intracellular
tachyzoites, especially before they exit from the parasitophorous vacuole of their host cell,
Toxofilin was found to be restricted to the apical end.
INTRODUCTION
Eukaryotic cells remodel their actin cytoskeleton continu-
ously in response to both intracellular and extracellular sig-
nals. This remodeling is crucial in mediating not only cell
motility but also many other fundamental cellular functions.
Extensive in vitro work has led to the current understanding
at the molecular level of the in vivo actin dynamics (Welch
et al., 1997; Ayscough, 1998).
Reorganization of filamentous actin varies according to
the cell type and the stimulus. It can be an extremely local-
ized, discrete, and transient phenomenon but can also result
in sustained reorganization involving extended parts of the
cell. However, despite the variation in its manifestation, the
reorganization of actin cytoskeleton relies primarily on the
intrinsic properties of the actin molecule (G-actin) to assem-
ble and disassemble into filaments (F-actin) (Carlier, 1990,
1991). The organization of the cytoskeleton is controlled by
a cohort of actin-associated proteins, which tightly coordi-
nate the intrinsic dynamic properties of actin and support its
spatial organization into various dynamic structures within
the cells (Stossel, 1993).
Coccidies are protozoan parasites that belong to the
phylum of Apicomplexa, members of which are human
pathogens of major medical importance. The causative
agent of malaria, Plasmodium falciparum, causes death of
⬎2 million children every year (Marsh and Snow, 1997),
and other Apicomplexa such as Toxoplasma gondii and to a
lesser extent Cryptosporidium parvum are devastating hu-
man pathogens when they parasitize immunocompro-
mised hosts (Kasper and Buzoni Gatel, 1998). Certain
developmental stages of these parasites including the
sporozoites of Plasmodium, Cryptosporidium, and Toxo-
plasma as well as the tachyzoites of Toxoplasma, move by a
gliding motion across either a mucous layer or an extra-
cellular matrix before encountering their host cells. They
In memory of O. Tardieux.
§
Corresponding author. E-mail address: tardieux@pasteur.fr.
Abbreviations used: ADF, actin-depolymerizing factor; DMEM,
Dulbecco’s modified Eagle’s medium; EST, expressed sequence
tag; GFP, green fluorescent protein; GST, glutathione S-transferase;
HFF, human foreskin fibroblast; 1,5-IAEDANS, N-(iodoacetyl)-N⬘-
1-sulfo-5-naphthylethylenamine; Ig, immunoglobulin; r-Toxofilin,
recombinant Toxofilin.
© 2000 by The American Society for Cell Biology 355

subsequently enter these cells by an active process, and
once in a suitable intracellular niche, they either multiply
and/or differentiate (for reviews, see Silverman and
Joiner, 1997; Dubremetz, 1998), two steps required for
parasite spreading before transmission to a new host.
The strategies selected by these parasites for either gliding
onto a substratum or for invading their host cells depend on
the dynamics of their actin cytoskeleton (King, 1988; Do-
browolski and Sibley, 1996; Preston and King, 1996). How-
ever, although it has been established that actin dynamics is
required for the progression of the parasite’s life cycle, stud-
ies on the molecular basis of parasite actin dynamics have
been hampered by the transient and discrete nature of actin
cytoskeleton remodeling. As a consequence, only limited
knowledge is available for Apicomplexa compared with
what is known for other eukaryotic cells.
Recent data from our and other laboratories demonstrated
that Toxoplasma has a strikingly low amount of assembled
actin compared with the usual F- to G-actin ratio observed in
other eukaryotic cells (Dobrowolski et al., 1997b; Poupel and
Tardieux, 1999). The estimated amount of extractable F-actin
in our study represented ⬍5% of the total actin, and this
yield increased slightly when parasites were treated before
extraction with the actin-stabilizing drug jasplakinolide. In-
deed, tachyzoite actin filaments have not been detected in
situ either by phalloidin staining or by classical electron
microscopy. However, Heuser and Sibley report that micro-
filaments can be observed by freeze—etch microscopy be-
neath the plasma membrane of gliding tachyzoites (in Do-
browolski and Sibley, 1997). Our recent data demonstrate
that, although remarkably small, the pool of F-actin in T.
gondii tachyzoites remains competent for assembly and dis-
assembly, and for coupling to a myosin-type motor activity
(Poupel and Tardieux, 1999).
Parasite molecules, capable of controlling actin monomer
sequestration and desequestration in association with mol-
ecules regulating actin filament turnover, are expected to be
among the major effectors responsible for the unusual low
F-actin content in Toxoplasma. Therefore, stimuli inducing
actin polymerization during parasite gliding and host cell
entry will have to locally decrease the critical actin concen-
tration required for filament assembly. In the eukaryotic
cells analyzed thus far, such changes in the critical actin
concentration are known to be elicited by fluctuations in the
activity of capping proteins and profilin, whereas they are
amplified by other G-actin-binding proteins (Schafer and
Cooper, 1995; Perelroizen et al., 1996).
We present here the identification and characterization of
aT. gondii novel actin-binding protein, which we named
Toxofilin. Toxofilin sequesters muscle G-actin and inhibits
its polymerization in vitro. Additionally, it associates with
muscle F-actin by capping the actin filament end. When
Toxofilin was ectopically overexpressed as green fluorescent
protein (GFP)-tagged protein in mammalian nonmuscle
cells, it clearly disrupted the actin cytoskeleton and caused
disassembly of actin stress fibers. In tachyzoites, Toxofilin
binds G-actin and copurifies with a parasite F-actin-contain-
ing fraction, suggesting that it may control parasite actin
dynamics as well. Such a role was further suggested by the
highly variable localization pattern of Toxofilin in the mov-
ing parasite, i.e., during gliding or host cell entry.
MATERIALS AND METHODS
Parasite Production and Recovery
The RH T. gondii strain was propagated in female Swiss mice as
described by Poupel and Tardieux (1999). The parasites were pel-
leted in PBS
⫺
containing 0.1% (vol/vol) protease inhibitor stocks 1
and 2 and were stored at ⫺70°C until use. The protease inhibitor
stocks were composed, for stock 1, of 4-[2-aminoethyl]benzenesul-
fonylfluoride (5 mg/ml), aprotinin (2 mg/ml), leupeptin (2 mg/
ml),and benzamidin (16 mg/ml) in H
2
O and, for stock 2, of pepsta-
tin A (5 mg/ml in DMSO).
Parasite Handling for Further Protein Extraction
and Affinity Chromatography
Frozen tachyzoites (10
9
) were thawed on ice and lysed by five liquid
nitrogen freezing and defreezing cycles in 2 ml of buffer A (20 mM
Tris-Cl, pH 8.0, 50 mM KCl, 0.1 mM EGTA, 0.1 mM EDTA) supple-
mented with 0.5% (vol/vol) protease inhibitor stocks. The extract
was centrifuged (15 min, 1000 ⫻g, 4°C) to remove nuclei and
unbroken cells. The supernatant was centrifuged (30 min, 100,000 ⫻
g, 4°C) in a TL100 table-top ultracentrifuge (Beckman Instruments,
Palo Alto, CA) using the TLA 100.3 rotor. The final supernatant was
applied to either a monomeric actin column prepared as described
by Miller and Alberts (1989) with actin purified from a rabbit muscle
acetone powder (Pardee and Spudich, 1982) or a DNase1 affinity
column as previously described (Fahrni, 1992). Both columns were
washed with 20 column volumes of buffer A supplemented with
0.05% Nonidet P40. G-actin-bound proteins were recovered by elu-
tion with 3 column volumes of buffer B (buffer A supplemented
with 3 mM MgCl
2
and 1 mM Na
2
ATP) followed by precipitation
with trichloroacetic acid containing 0.04% (wt/vol) sodium deoxy-
cholate. DNase1-bound proteins were directly eluted in SDS-PAGE
sample buffer. Samples were boiled and kept at ⫺70°C until analysis
by SDS-PAGE electrophoresis. To recover parasite F-actin, proteins
were extracted in 2 ml of actin-stabilizing buffer C (60 mM 1,4-
piperazinediethanesulfonic acid, 25 mM HEPES, 125 mM KCl, 2
mM MgCl
2
, 5 mM EGTA, 100
␮
MNa
2
ATP, 1
␮
M phalloidin, pH 7.2)
containing 0.1% (wt/vol) saponin (Sigma, St. Louis, MO) and 0.5%
(vol/vol) protease inhibitor stocks (Poupel and Tardieux, 1999).
Peptide Microsequencing, cDNA Library Screening,
and DNA Sequencing
The gel slice containing the 27-kDa actin-binding protein from the
parasite was subjected to tryptic digestion (30°C, 18 h, 0.3 mg of
trypsin in 0.1 M Tris-Cl, pH 8.6, 0.01% Tween 20). The tryptic
peptides were recovered by HPLC on a DEAE-C18 column. The
sequence of each peptide was found in different clones from the T.
gondii database of expressed sequence tags (ESTs; Washington Uni-
versity–Merk Toxoplasma EST Project, St. Louis, MO). Nondegen-
erate primers were synthesized for amplification of the target se-
quence from the clone identified as TgESTzy57g11. The
oligonucleotide with the sequence 5⬘-CG GAG NAG CCC TAG TTC
CTG-⬘, corresponding to the amino acid sequence QELGLLR, was
used as the downstream PCR primer, whereas 5⬘-TCA AGT GAC
CAA GGC GGT CGA-3⬘was chosen as the upstream primer. The
PCR conditions for amplification of the 294-bp DNA product were
as follows: a hot start of 2 min at 94°C followed by 35 cycles (1 min,
94°C; 30 s, 61°C; and 45 s, 72°C) and a final elongation step at 72°C
for 10 min. The 294-bp fragment recovered was
32
P labeled using
random priming (Megaprime kit; Amersham, Arlington Heights,
IL), purified on a Sephacryl S-400 HR column (Pharmacia, Uppsala,
Sweden), and used as a probe to screen a T. gondii tachyzoite cDNA
library (kindly provided by J.W. Ajioka, Cambridge University,
Cambridge, United Kingdom). After two rounds of screening, 10
independent overlapping clones were selected, and their cDNA was
prepared for nucleotide sequencing performed by Genset (Paris,
O. Poupel et al.
Molecular Biology of the Cell356

France), using both vector and T. gondii Toxofilin-specific primers
(Genset).
Native Electrophoresis with IAEDANS-labeled
Muscle Actin
Polymerized rabbit muscle actin was prepared from clarified G-
actin by adding KCl to 50 mM, MgCl
2
to 2 mM, and Na
2
ATP to 1
mM. The Cys-373 of F-actin (50
␮
M) was labeled with N-(io-
doacetyl)-N⬘-1-sulfo-5-naphthylethylenamine (1,5-IAEDANS, 500
␮
M; Molecular Probes, Eugene, OR) as described by Dos Remedios
and Cooke (1984). A cytosolic fraction from 5 ⫻10
8
frozen
tachyzoites was prepared as described in the previous paragraph by
using 400
␮
l of buffer D (2 mM Tris-Cl, pH 8.0, 0.2 mM CaCl
2
, 0.1
mM dithiothreitol, 0.2 mM Na
2
ATP) supplemented with 0.5% (vol/
vol) protease inhibitor stocks. One hundred twenty-five microliters
of cytosol were incubated (15 min, 4°C) with 50
␮
g of labeled
G-actin (in a 25-
␮
l volume) under rotating agitation and analyzed
by native gel electrophoresis as described by Safer (1989). In some
experiments, the interaction of recombinant Toxofilin with G-actin
was also tested by native gel electrophoresis. For this, 30
␮
gof
recombinant toxofilin (in 125
␮
l buffer A), obtained as described
below, were incubated (15 min, 4°C) with 50
␮
g of labeled G-actin
(in a 25-
␮
l volume) before analysis on a native gel. Fluorescent
bands were visualized under a UV lamp (
␭
⫽321 nm).
Expression of Glutathione S-Transferase
(GST)–Toxofilin Fusion Protein and Recombinant
Toxofilin Purification
The fragment for expression of Toxofilin was prepared by PCR
amplification of a full-length Toxofilin-encoding cDNA using prim-
ers introducing a BamHI restriction site at position 5⬘and a SalI
restriction site at position 3⬘. For amplification of the upper strand
(5⬘-GGCCGGATCCATGGCGCAATACAAGTCACGC-3⬘) and of
the lower strand (5⬘-GGCCGTCGACTTACGACGAGGGCAT-
AGCGCC-3⬘), the amplified fragment was cloned into the expres-
sion vector pGEX6-P3 (Pharmacia) after digestion with BamHI and
SalI of both fragment and vector. For expression of the GST-Toxo-
filin, an Escherichia coli clone (BL21 strain) positive for the plasmid
was grown up to OD of 1 and induced with isopropylthio-
␤
-d-
galactoside (1 mM, 3 h, 37°C). At the end of the induction period,
the bacteria were pelleted and subsequently lysed in buffer D (50
mM Tris, pH 8.0, 150 mM NaCl, 1 mM EDTA, Triton X-100 [0.5%
vol/vol], N-lauryl sarcosyl [1.5% vol/vol; Sigma]) at 4°C. The su-
pernatant recovered after centrifugation (15,000 ⫻g, 15 min, 4°C)
was incubated with glutathione-Sepharose (Pharmacia) (4°C, over-
night), and the beads were washed with 40 bead volumes of buffer
C supplemented with 0.1% Triton X-100. The bound GST-polypep-
tide was cleaved with prescission protease to recover the recombi-
nant Toxofilin (r-Toxofilin) without GST (Pharmacia). r-Toxofilin
was lyophilized as 100-
␮
g aliquots and stored at ⫺70°C until use.
Antibodies
The mouse monoclonal antibody anti-P30 (Biogenex, San Ramon,
CA) recognizes a major tachyzoite surface. The mouse monoclonal
anti-MIC3 (Tg42F3.2F5) and the anti-GRA1 (Tg 17–43) recognize,
respectively, the MIC3 protein from tachyzoite micronemes and the
GRA1 protein from tachyzoite-dense granules. These antibodies
were kindly provided by M.F. Cesbron (Institut Pasteur, Lille,
France). Antibodies against r-Toxofilin were obtained by immuniz-
ing two rats and a rabbit with r-Toxofilin, and the positive sera were
affinity purified on an r-Toxofilin HiTrap-N-hydroxysuccinimide-
activated affinity column (Pharmacia).
Electrophoresis and Western Blotting
Proteins were separated by SDS-PAGE (12% gel; Laemmli, 1970)
and transferred to Hybond-C nitrocellulose (Amersham) using a
semidry blotting apparatus (Hoefer, San Francisco, CA; 45 min, 1.5
mA/cm
2
). The membrane was blocked and treated after incubation
with the appropriate antibodies as described by Tardieux et al.
(1998). Membranes were incubated overnight at 4°C either with a rat
anti-T. gondii Toxofilin serum diluted 1:5000 in Tris-buffered saline
and 0.1% (vol/vol) Tween 20 and/or a rabbit anti-T. gondii actin
antibody diluted to 1:4000. After several washes in Tris-buffered
saline and 0.1% (vol/vol) Tween 20, the blots were incubated (1 h,
23°C) with HRP-labeled anti-rat immunoglobulin (Ig) diluted to
1:5000 (Jackson ImmunoResearch, West Grove, PA) or HRP-labeled
anti-rabbit Ig diluted to 1:3000 (Amersham). In the actin cosedimen-
tation experiment, rabbit actin was detected using the mouse mono-
clonal antibody clone C4 (dilution 1:4000) followed by the Western
blot kit reagents (ECF, Amersham).
Pyrene-labeled Muscle Actin Polymerization Assay
Actin was prepared as described by Tardieux et al. (1998). Pyrene-
labeled actin was obtained as described by Kouyama and Mihashi
(1981). The final concentration of rabbit actin was 100
␮
g/ml. The
concentration of r-Toxofilin in the assay varried between 1 and 2
␮
M. To eliminate potential aggregates, r-Toxofilin was ultracentri-
fuged before its addition to actin (4°C, 30 min, 100,000 ⫻g). Poly-
merization was carried out at 23°C in buffer P (100 mM KCl, 2 mM
MgCl
2,
1 mM ATP, 5 mM Tris-HCl, pH 8.0), and monitored in a
Spex fluorimeter (F-2000; Hitachi, Tokyo, Japan) at
␭
ex
⫽350 nm
and
␭
em
⫽390 nm. The steady-state values of F-actin were obtained
after 16 h incubation at 4°C and the samples were subsequently
centrifuged (4°C, 2 h, 100,000 ⫻g, in a TL100.A rotor; Beckman) and
analyzed by SDS-PAGE and Western blot.
Protein Immunoprecipitation
Toxofilin was immunoprecipitated from the final 100,000 ⫻gsu-
pernatant of the parasite lysate supplemented with 0.5% (vol/vol)
Triton X-100 and 0.5% (wt/vol) BSA (fraction V; Life Technologies,
Gaithersburg, MD) using rabbit anti-Toxofilin followed by protein
A-Sepharose (Pharmacia). Washes were successively performed in
buffer A containing both 0.5% (wt/vol) BSA and 0.5% (vol/vol)
Triton X-100 and then in buffer A containing 0.5% (vol/vol) Triton
X-100 and buffer A. Immunocomplexes were eluted in SDS-PAGE
sample buffer. Samples were boiled and kept at ⫺70°C until ana-
lyzed by SDS-PAGE.
Human Cell Culture
Human foreskin fibroblasts (HFFs) were cultured between the 11th
and 25th passages in Dulbecco’s modified Eagle’s medium (DMEM)
with Glutamax medium (Life Technologies) supplemented with
10% (vol/vol) FBS (Dutscher, Brunath, Israel) and penicillin-strep-
tomycin (10
␮
g/ml, Life Technologies). HeLa cells were cultured
under the same conditions as a laboratory-established cell line.
Expression and Detection of GFP-Toxofilin Fusion
Protein in HeLa Cells
The fragment for expression of Toxofilin was obtained by PCR
amplification of a full-length Toxofilin-encoding cDNA using prim-
ers introducing 5⬘-SalIanda3⬘-BamHI restriction sites. The primers
used were 5⬘-GGCCGTCGACATGGCGCAATACAAGTCACGC-3⬘
for the upper strand and 5⬘-GGCCGGATCCTTACGACGAGGG-
CATAGCGCC-3⬘for the lower strand.
The amplified fragment was cloned into the pEGFP-C1 MCS
(Clontech, Cambridge, United Kingdom) expression vector after
digestion of both insert and vector with BamHI and SalI enzymes.
The GFP-Toxofilin coding plasmid was prepared from a positive
clone of E. coli (strain DH5
␣
; Qiagen [Hilden, Germany] method)
and used for transfections of 50% confluent HeLa cells. Transfec-
tions were performed with the CaCl
2
phosphate transfection mix-
Toxofilin Controls Actin Dynamics
Vol. 11, January 2000 357

ture overnight (37°C, 5% CO
2
). Cells after two washes in DMEM
containing 10% FBS, were plated on glass coverslips and incubated
overnight (37°C, 6% CO
2
) for 20 h. Expression of the GFP-Toxofilin
was analyzed by fluorescence microscopy, and F-actin was stained
with rhodamine-phalloidin. Forty to 60% of the transfected cells
expressed GFP-tagged Toxofilin.
Immunofluorescence of Extracellular and
Intracellular Parasites
Free T. gondii tachyzoites freshly isolated from mice were allowed to
glide on glass slides in PBS
⫺
. The parasites were fixed in situ by
paraformaldehyde (2% in PBS
⫺
, 15 min, 23°C), washed in PBS
⫺
,
treated with NH
4
Cl (50 mM in PBS
⫺
, 5 min), and deposited on a
glass slide at a density of 10
7
/ml to air dry.
HFF cells were plated in 3.5-cm-diam Petri dishes at a density of
2⫻10
5
cells per dish on acid-washed 12-mm circular glass cover-
slips. After overnight culture in DMEM and Glutamax (Life Tech-
nologies) containing 10% (vol/vol) FBS and antibiotics, tachyzoites
were added to the cells (2 ⫻10
7
per dish) for 15 min (37°C, 5% CO
2
).
Some coverslips were then washed in PBS
⫺
and fixed (2% parafor-
maldehyde, 0.1 M lysine, 0.05 M NaPO
4
, pH 7.4) for 15 min, whereas
others were further incubated in fresh 10% medium for 20 h (37°C,
5% CO
2
) before fixation. Cells were labeled with the mouse mono-
clonal antibody anti-P30 (dilution 1:300 in PBS
⫺
supplemented with
2.5 mg/ml goat serum). After several washes, cells were incubated
with the Alexa-488 goat anti-mouse IgG (heavy and light chain,
Molecular Probes; 7
␮
g/ml, 1 h, 23°C). The excess of conjugate was
washed away with PBS
⫺
, and cells were permeabilized with Triton
X-100 (0.1% (vol/vol) in PBS
⫺
containing 2.5 mg/ml normal goat
serum). Free extracellular parasites were permeabilized for 1 min
(23°C), and HFFs were infected with parasites for 3 min (23°C). The
detergent was removed by several PBS
⫺
washes, and cells were
incubated with nonimmune or affinity-purified immune sera. Both
the rat and the rabbit anti-Toxofilin affinity-purified sera were di-
luted at 1:200 in PBS
⫺
containing 2.5 mg/ml goat serum. The
staining was subsequenctly revealed by incubation with the Alexa-
568 goat anti-rat or anti-rabbit IgG (heavy and light chain, Molec-
ular Probes; 7
␮
g/ml, 1 h, 23°C). Unbound fluoroconjugate was
removed with PBS
⫺
washes, and DAPI (Sigma; 5
␮
g/ml) was added
for 5 min before the last PBS
⫺
wash. When needed, F-actin of HFF
or HeLa cells was stained with rhodamine-phalloidin (0.66 mg/ml,
30 min, room temperature; Molecular Probes).
The samples were examined under an epifluorescence microscope
(Axiophot; Zeiss, Oberkochen, Germany) attached to a cooled
charge-coupled device camera (Photometrics, Tucson, AZ), using a
63⫻Plan-Apochromat lense. Images were aquired using the IPlab
software (Signal Analytics, Vienna, VA) and analyzed by NIH Im-
age (National Institutes of Health, Bethesda, MD) or Adobe (Moun-
tain View, CA) Photoshop software.
For the quantitation of F-actin fluorescence, cells from several
fields were analyzed using the NIH Image program. The area
around each cell was delineated, and the mean fluorescence inten-
sity was measured in pixels. The results were presented as the mean
of the fluorescence intensity for each group of cells ⫾SD.
Electron Microscopy
Actin filaments at steady state (2
␮
M) were incubated with purified
r-Toxofilin (1
␮
M) for 15 min (23°C). Affinity-purified anti-Toxofilin
antibodies were added (2
␮
g/ml, 15 min, 23°C), and the r-Toxofilin
was revealed by anti-rabbit antibodies conjugated to 5-nm gold
particles. The actin filaments were pelleted through a 30% sucrose
cushion in F-actin stabilizing buffer (50,000 ⫻g, 30 min, 4°C) and
were further processed for electron microscopy as described by
Tardieux et al. (1998). The samples were analyzed using a Phillips
(Eindhoven, The Netherlands) CM12 transmission electron micro-
scope.
RESULTS
T. gondii Tachyzoite Cytosolic Extracts Contain a
27-kDa Protein That Binds Mammalian G-actin
To isolate T. gondii tachyzoite proteins sequestering actin
monomers, we initially searched for tachyzoite cytosolic
proteins that could bind to muscle rabbit G-actin. For these
studies we used affinity chromatography and native gel
electrophoresis. When a 100,000 ⫻gcytosolic extract from T.
gondii tachyzoites (Figure 1A, lane b) was passed through an
affinity column of rabbit muscle G-actin, a protein migrating
in SDS-PAGE with a molecular mass of ⬃27 kDa was repro-
ducibly recovered in the eluate and visualized on blots by
Ponceau S staining (Figure 1A, lane a, triangle). Moreover,
when rabbit G-actin fluorescently labeled with IAEDANS
was incubated with a 100,000 ⫻gcytosolic extract from T.
gondii tachyzoites and the mixture was submitted to native
gel electrophoresis, several fluorescent bands were detected
(Figure 1B, lane a). Only one fluorescent band was observed
when the same amount of G-actin was analyzed alone (lane
c). When actin polymerization was promoted, a much fainter
band of fluorescently labeled G-actin was detected (lane d).
The additional fluorescent bands detected in Figure 1A,
lane a, represent complexes of labeled G-actin with parasite
polypeptides. To dissociate and separate the complexes
composing the two major shifted bands (Figure 1B, lane a, *1
and *2) SDS-PAGE was performed (Figure 1C). Both com-
plexes were shown to contain a comigrating 27-kDa protein
(marked with a triangle), which was visualized after transfer
to nitrocellulose membrane and Ponceau S staining (Figure
1C, lanes a and b). Two other minor bands of ⬃40 and 47
kDa were also detected in both complexes. No fluorescent
bands were observed in the T. gondii cytosolic fraction (Fig-
ure 1B, lane b), and the 40- and 47-kDa parasite proteins
were not detected from the gel slice that migrated at the
same position as the fluorescent actin-containing band
marked as *1 (Figure 1C, lane c). This strengthened our
conclusion that muscle G-actin interacts with at least three
parasite proteins. As expected, the fluorescent band marked
as *0, which migrated faster during native gel electrophore-
sis, contained only G-actin (Figure 1C, lane d). In addition,
the 27-kDa protein recovered from either of the two purifi-
cation procedures (affinity chromatography [Figure 1A, lane
a] and native gel electrophoresis [Figure 1B, lane a]) appears
to be the same. First, its digestion with trypsin yielded
identical peptide profiles. Second, antibodies made against
the recombinant Toxofilin recognized the same 27-kDa
T.gondii actin-binding protein isolated by either of the two
above-described procedures (our unpublished results).
Identification of Toxofilin cDNA and Protein
Sequence
Peptide microsequencing of two tryptic peptides from the
isolated 27-kDa T. gondii actin-binding protein gave rise to
the following partial sequences: QAALAGQILNEQR and
QQELGLLRPEER. Each peptide was encoded by several
different T. gondii ESTs (www.cbil.upenn.edu/ParaDBs/;
Ajioka et al., 1998). Using nondegenerate primers, we ob-
tained by PCR a 294-bp fragment, which was used as a
probe to screen a tachyzoite cDNA library. We isolated 10
overlapping clones from which we obtained a single open
O. Poupel et al.
Molecular Biology of the Cell358

reading frame coding for a 27,085-Da protein (GenBank
accession number AJ132777). This 246-amino-acid protein is
predicted to be basic (pI ⫽9.63) and to contain two coiled-
coil domains (amino acids 120–149 and 206–234). No con-
sensus nucleotide-binding or actin-binding sites were pre-
dicted by the sequence. Southern blot analysis was
performed on genomic T.gondii DNA after digestion with
either BamHI, EcoRI, or HindIII restriction enzymes. In all
cases, only one signal was obtained, strongly suggesting that
this protein is encoded by only one gene copy in the T. gondii
(our unpublished results). The DNA and protein sequence
of this newly identified protein was compared with all the
known protein sequences in the nonredundant GenBank,
but no significant similarity with other known proteins was
revealed. Therefore the 246-amino-acid T. gondii acting-
binding protein is novel, and we named it Toxofilin.
The Recombinant Toxofilin Binds Directly to G-
actin and Controls Actin Polymerization In Vitro
To investigate the activities of Toxofilin in vitro, we ex-
pressed it in E. Coli as a recombinant protein (r-Toxofilin). To
study whether r-Toxofilin binds directly to G-actin and
whether it affects actin dynamics, we performed 1) a native
gel electrophoresis, 2) a pyrene-labeled actin polymerization
assay, and 3) an actin cosedimentation assay.
When r-Toxofilin was added to IAEDANS-labeled mus-
cle G-actin and the mixture was separated on a native gel,
Figure 1. The T. gondii 27-kDa protein binds to G-actin. (A) G-actin chromatography. A 100,000 ⫻g T. gondii cytosolic extract was subjected
to affinity chromatography over a rabbit muscle G-actin column. Proteins bound to G-actin were eluted by 3 mM MgCl
2
and 1 mM ATP.
These proteins were further separated by SDS-PAGE on a 12% gel, transferred to a nitrocellulose membrane, and visualized with red Ponceau
S stain. Lane b, cytosolic fraction corresponding to 5 ⫻10
7
T. gondii tachyzoites; lane a, eluate. The polypeptide migrating with a molecular
mass of ⬃27 kDa is marked with a triangle. (B) Native gel electrophoresis. A 100,000 ⫻gparasite cytosolic extract (T.g) was incubated with
IAEDANS-labeled G-actin (G), and the protein complexes were separated by native gel electrophoresis. The bands were visualized under a
UV lamp (
␭
⫽312 nm). Lane a, three bands were detected and marked, respectively, *0, *1, and *2; lane b, parasite cytosolic extract without
fluorescent actin; lane c, fluorescent G-actin; lane d, fluorescent G-actin incubated under polymerization conditions to reach steady state
before native gel electrophoresis. (C) Ponceau S staining on blot. Slices of the native gel (see B) containing either complex *0, *1, or *2 or from
the adjacent lanes from positions corresponding to the migration of *0, *1, or *2 were boiled in SDS-PAGE sample buffer and analyzed on
a 12% gel. The separated proteins were subsequently transferred onto a nitrocellulose membrane and stained with Ponceau S. Lanes a and
b, material from bands *1 and *2, respectively. Apart from actin, three other similar bands were detectable in the complexes *1 and *2,
including a band of ⬃27 kDa (triangle); lane c, sample extracted from a slice cut from lane b on the native gel. The slice corresponded to
polypeptides comigrating with *1; lane d, material from the band *0 from the native gel; lane e, material containing G-actin from lane c of
the native gel.
Toxofilin Controls Actin Dynamics
Vol. 11, January 2000 359

we observed a shift in the migration profile of actin (Fig. 2A,
lane b) compared with the migration of actin alone (lane a). In
the mixture with r-Toxofilin actin ran slower. Of note, at
equimolar concentrations, all the G-actin detected by fluores-
cence was complexed to r-Toxofilin. This result strongly argues
for a direct binding of r-Toxofilin to G-actin.
To further analyze the effect of r-Toxofilin on the kinetics of
muscle actin polymerization and depolymerization, we used a
pyrene-labeled actin polymerization assay. The effect of
r-Toxofilin on the extent of inhibition of actin polymerization
was concentration dependent, as demonstrated by the steady-
state values given in parentheses (Figure 3B). These steady-
state values were obtained 16 h after the assay was begun. The
insets in Figure 2B show that, at steady state, there was a
significant decrease of sedimented actin in the presence of
r-Toxofilin (upper panel), whereas r-Toxofilin copelleted with
F-actin (lower panel). For this determination we used an anti-
Toxofilin antibody raised against the r-Toxofilin. The copellet-
ing of r-Toxofilin could not be an artifact attributable to aggre-
gation, because we routinely ultracentrifuged r-Toxofilin-
containing solutions before using them in the assay.
The Recombinant Toxofilin Has F-actin Capping
Activity In Vitro
When r-Toxofilin was added to preassembled actin fila-
ments that were allowed to depolymerize spontaneously,
the rate of monomer dissociation was clearly slower com-
pared with the control (Figure 2C). Similarly, the effect
was concentration dependent, as illustrated by the steady-
state values in parentheses. The insets in Figure 2C illus-
trate that F-actin was protected from dissociation (upper
panel) by the r-Toxofilin found associated with it (lower
Figure 2. Recombinant Toxofilin controls actin dynamics in vitro: it binds directly to actin monomers and caps actin filaments. (A)
Native gel electrophoresis. IAEDANS-labeled G-actin was incubated with or without r-Toxofilin (lanes b and a, respectively) before
electrophoresis on a native gel. (B) Pyrene-labeled G-actin polymerization assay. G-actin at 3
␮
M and 1% pyrene labeled was incubated
with or without r-Toxofilin before induction of polymerization. The kinetics of actin polymerization was monitored in a Spex
fluorimeter (F-2000; Hitachi) at
␭
ex
⫽350 nm and
␭
em
⫽390 nm. ⴙ,kinetics of actin polymerization in the absence of r-Toxofilin; ⴱ, with
1
␮
M r-Toxofilin; E, with 2
␮
M r-Toxofilin. Fluorescence intensity is measured in arbitrary units. Steady-state values are given in
parentheses. Inset, upper panel, the amount of G and F-actin pelleted after ultracentrifugation (2 h, 100,000 ⫻g, 4°C; as described by
Tardieux et al., 1998); lower panel, amount of r-Toxofilin recovered in the supernatant (G) and pellet (F). (C) Pyrene-labeled F-actin
depolymerization assay. Ten micromolar steady-state F-actin was induced to spontaneously depolymerize upon dilution to 2
␮
M. The
kinetics of actin depolymerization was illustrated by the curve labeled ⴙ. Kinetics of actin depolymerization was also followed in the
presence of r-Toxofilin at 1.5
␮
M(ⴱ) and 2
␮
M(E). Inset, upper panel, amount of G and F-actin; lower panel, amount of r-Toxofilin
recovered in the supernatant (G) and pellet (F).
O. Poupel et al.
Molecular Biology of the Cell360

panel). These data suggested that Toxofilin could function
as a capping protein.
Additionally, electron micrographs displaying localization of
Toxofilin on actin filaments strongly suggested that it caps one
end of the filaments (Figure 3). Quantitative analysis was per-
formed on 50 randomly selected actin filaments from three sepa-
rate experiments. By statistical analysis we compared the fre-
quency by which gold particles appeared either at one end or
along the side of actin filaments. First, for filaments on which only
one gold particle was bound (46 of 50, 45 of 50, and 48 of 50,
respectively, for the three replicates), the frequencies for each gold
particle to be found in either position were not different among the
three replicates (
␹
2
⫽1.40; df ⫽2; p ⫽0.497). Second, when the
values from the three replicates were pooled together, the associ-
ation of gold particles with one end of a filament was found to be
highly significant (pooled frequencies compared with half of the
particles in each position:
␹
2
⫽8.02; df ⫽1;p⫽0.0046). And
finally, in the few filaments that had several gold particles bound
(4 of 50, 5 of 50, and 2 of 50, respectively, for the three replicates),
there was always one localized at one filament end.
Figure 3. Toxofilin localizes at
one end of actin filaments. Actin
filaments at steady state (2
␮
M)
were incubated with purified r-
Toxofilin (1
␮
M). Immunolocal-
ization of r-Toxofilin was re-
vealed by anti-rabbit antibodies
conjugated to 5-nm gold (see ar-
rows). Actin filaments were neg-
atively stained with 2% uranyl ac-
etate.
Toxofilin Controls Actin Dynamics
Vol. 11, January 2000 361

Ectopically Expressed GFP–Toxofilin Disrupts F-
actin Cytoskeleton in Mammalian Nonmuscle Cells
We have presented above biochemical evidence that r-Toxo-
filin can control muscle actin dynamics in vitro. To examine
whether Toxofilin could affect actin dynamics in vivo, we
overexpressed GFP-tagged Toxofilin in epithelial cells
(HeLa). When GFP was expressed alone (Figure 4A), the
F-actin stained with phalloidin was similar to that observed
in nontransfected cells, with well-organized stress fibers (see
Figure 4.
O. Poupel et al.
Molecular Biology of the Cell362

arrows). In contrast, cells expressing GFP-tagged Toxofilin
no longer displayed actin stress fibers (Figure 4, B–D). In-
stead, they had a punctate staining pattern resembling what
is usually seen in cytochalasin d-treated cells (in particular,
Figure 4B).
We quantified the amount of F-actin in a sample of HeLa
cells expressing GFP-Toxofilin fusion protein by measuring
the intensity of fluorescent phalloidin in each cell using the
NIH Image analysis program. By comparing the fluores-
cence associated with F-actin in cells that express or do not
express GFP-Toxofilin, we found a reduction of ⬃45% in the
fluorescence of the cells expressing GFP-Toxofilin. The same
comparison done between the cells expressing or not express-
ing GFP did not show any significant difference (Table 1).
In addition, it was clear that the extent of the effect of the
GFP-tagged Toxofilin on the actin morphology depended on
the levels of expression of the fusion protein (Figure 4D).
Only ⬃43% of the transfected cells showed clear defects on
the actin cytoskeleton, whereas the rest of the cells, usually
expressing lower levels of GFP-Toxofilin, had smaller dis-
appearance of the actin stress fibers. The expression of flu-
orescent Toxofilin in HeLa cells transfected with the plasmid
encoding GFP-Toxofilin was confirmed by immunostaining
with the anti-Toxofilin antibody (Figure 4E).
In parallel, we could recover the GFP-Toxofilin protein
from transfected HeLa cells by immunoprecipitation of actin
(our unpublished results), demonstrating that the pheno-
type observed in the transfected cells was at least partially
due to direct binding of GFP-Toxofilin to G-actin.
T. gondii Toxofilin Binds to Parasite G-actin and Is
Also Associated with a Parasite Fraction
Containing F-actin
To characterize Toxofilin in the parasite, we first examined
whether the protein was associated with parasite actin that
was isolated on the basis of its affinity to DNase 1. Using
affinity chromatography, we observed that the recovered
43-kDa band copurified with other products of which the
major band migrates as a 27-kDa polypeptide (Ponceau
S-stained blot; Figure 5A, lane a, *). We identified by West-
ern blot analysis that the 43-kDa band corresponds to actin
(lane b), and the 27-kDa copurifying product corresponds to
Toxofilin (lane c). An anti-T. gondii actin and the anti-Toxo-
filin andibodies were respectively used for this determina-
tion. In addition, direct immunoprecipitation of parasite
actin coprecipitated Toxofilin (our unpublished results),
whereas actin (marked with a sphere) coprecipitated with
immunoprecipitated Toxofilin (Figure 5B, lanes b and d).
Based on the information obtained from the studies with
mammalian actin, we subsequently investigated whether
Toxofilin was also present in the detergent-insoluble fraction
containing parasite F-actin. As assessed by Western blot
(Figure 5C), Toxofilin indeed cofractionated with a parasite
F-actin (lanes b and d). Interestingly, Toxofilin migrated as a
doublet in the detergent-insoluble fraction (lane d), suggest-
ing the presence of a modified form of Toxofilin.
Figure 4 (facing page). Expression of GFP-Toxofilin fusion protein affects actin dynamics in mammalian nonmuscle cells. HeLa cells
(50% confluent) were transfected by the CaCl
2
phosphate method with plamids encoding either GFP-Toxofilin or GFP (see MATERIALS
AND METHODS) and were incubated overnight before replating on glass coverslips. The cells were incubated for 20 h (37°C, 5% CO
2
)
before processing for immunofluorescence (see MATERIALS AND METHODS). For revealing the F-actin or Toxofilin, the cells were
permeabilized after fixation and stained with either rhodamine-phalloidin or with an anti-Toxofilin antibody, respectively. (A)
Phalloidin staining of F-actin in HeLa cells expressing GFP. Well-organized actin stress fibers are visible. (B–D) Phalloidin staining of
F-actin in cells expressing GFP-Toxofilin. Actin stress fibers are disorganized. (D) Different levels of expression of the GFP-Toxofilin.
(E) Detection of Toxofilin with an anti-Toxofilin antibody in HeLa cells transfected with the GFP-Toxofilin-encoding plasmid.
Table 1. F-actin quantitation
No. of
cells
analyzed Expression of
GFP-toxofilin Expression
of GFP
Mean
fluorescence
intensity
(pixels)
% Control
fluorescence
intensity
29 ⫹⫺1180 ⫾250 55.8
42 ⫺⫺2115 ⫾582
33 ⫺⫹1784 ⫾311 107.0
39 ⫺⫺1668 ⫾416
The fluorescence associated with HeLa cells from several fields was
analyzed using the NIH Image program. The area around each cell
was delineated, and the mean fluorescence intensity was measured
in pixels. The results are presented as the mean of the fluorescence
intensity for each group of cells ⫾SD. Control, cells not expressing
GFP or GFP-toxofilin.
Toxofilin Controls Actin Dynamics
Vol. 11, January 2000 363

Redistribution of Toxofilin in Tachyzoites during
Gliding and Host Cell Entry
Using immunofluorescence microscopy, we observed that
Toxofilin was mostly found apically in the cytoplasm of
tachyzoites and excluded from the nucleus (Figure 6). How-
ever, in extracellular gliding parasites the staining varied
strikingly among individuals. It was observed either as
densely organized in the apical cytoplasm (Figure 6A, 1,
large arrowhead) or distributed as patches in the apical part
of the cytoplasm (Figure 6A, 1 and 3, small arrows). How-
ever, it was often found in an apical, arrowlike distibution
(Figure 6A, 2, small arrowhead) with occasional additional
staining behind the nucleus at the distal part of the parasite
(Figure 6A, 2, large arrow). Interestingly,images catching
active host cell entry always showed a positive staining for
Toxofilin at the distal end of the parasite membrane (Figure
6A, 4 and 5, small arrows). Entering parasites were detected
on the basis of the incomplete staining of the parasite (Figure
6A, 4 and 5, small arrowheads). The partial staining of
parasite surface was achieved because of the restricted ac-
cessibility of an antibody recognizing a major Toxoplasma
surface molecule (P30), to only the part of the parasite that
still remains extracellular, i.e., its posterior end. To further
characterize the localization of Toxofilin, we costained with
antibodies recognizing secretory dense granules (Figure 6B,
1) or micronemes (Figure 6B, 2). Toxofilin did not colocalize
with either of these organelles.
In contrast to motile parasites, Toxofilin presented a
strong and uniform staining in all the parasites once they
have entered the host cell (our unpublished results). At the
end of the parasite intracellular life before exit from the host
cell, the newly formed tachyzoites still display a strong
uniform pattern for Toxofilin at their apical side (Figure 7
with inset 2⫻).
DISCUSSION
T. gondii tachyzoites, as other Apicomplexan parasites, rely
on the assembly and disassembly of their actin cytoskeleton
to invade their host cells. Actin assembly and disassembly
(Dobrowolski and Sibley, 1996) and actin–myosin coupling
activity (Dobrowolski et al., 1997a) underlie the process by
which surface molecules are capped during parasite motility
and invasion. Both of these mechanisms provide the energy
necessary for the capping process and therefore generate the
force driving parasite movement on a substratum or into its
host cell. The molecules controlling these mechanisms re-
main to be identified. However, it is clear that the changes in
the abundance, location, and organization of actin filaments
are expected to be under the control of actin monomer
sequestration and desequestration and of filament turnover.
Because tachyzoites display an unusually low amount of
F-actin, we hypothesized that sequestration of monomeric
actin through the activity of actin-associated proteins may
Figure 5. T. gondii Toxofilin binds to parasite G-actin and is associated with a parasite fraction containing F-actin. (A) DNase 1 affinity
chromatography. DNase 1-bound proteins were eluted in SDS-PAGE sample buffer and separated on a 12% acrylamide gel. After transfer
to a nitrocellulose membrane, eluted proteins were visualized with Ponceau S staining (lane a). Lane b, the eluate was probed for actin with
an anti-actin antibody specific for T. gondii actin; lane c, the eluate was probed for Toxofilin with an anti-Toxofilin antibody. Toxofilin is
marked with a star. (B) Toxofilin immunoprecipitation. A 100,000 ⫻gparasite cytosolic extract was prepared from [
35
S]methionine-cysteine-
labeled tachyzoites (lane a, load) and immunoprecipitated with anti-Toxofilin antibody (lane b, eluate). Toxofilin is marked with a star. Both
the cytosolic extract and the immunoprecipitate contained actin (E) as assessed by Western blot using an anti-T. gondii actin-specific antibody
(lanes c and d, respectively). (C) Protein fractionation. Parasite proteins were extracted in F-actin-stabilizing conditions to recover fractions
containing G-actin (lane a) or F-actin (lane b; E. Both fractions contained Toxofilin (ⴱ) as assessed by Western blot analysis using an
anti-Toxofilin antibody (lanes c and d, respectively).
O. Poupel et al.
Molecular Biology of the Cell364

play a key role in parasite actin dynamics. Such actin-se-
questering proteins have been well described in systems
ranging from yeast to mammals, and they are mostly mem-
bers of three large families, the profilins, the thymosins, and
the actin-depolymerizing factor/cofilins (Maciver, 1998). In
T. gondii, a 13.5-kDa depolymerizing factor that shares a
high degree of similarity with the ubiquitous actin-depoly-
merizing factor (ADF)/cofilin family has been described
(Allen et al., 1997). Proteins of this family are able to increase
the rate of depolymerization of ADP-bound F-actin from the
Figure 6. Toxofilin localization during tachyzoite
motility and host cell entry is dynamic. (A, 1–3)
Tachyzoites were allowed to glide on glass slides
before being fixed and stained for the major sur-
face protein P30 with mouse monoclonal anti-P30
followed by a secondary anti-mouse antibody con-
jugated to Alexa-488. Cells were then permeabil-
ized and stained for Toxofilin using the anti-Toxo-
filin antibody followed by an anti-rabbit antibody
conjugated to Alexa-568. Nuclei were stained by
DAPI (see MATERIALS AND METHODS). In
green, the surface of the parasite; in red, Toxofilin
distribution; in blue, nuclei. (A, 4 and 5)
Tachyzoites were incubated for 10 min with HFF
cells plated on glass coverslips. Coverslips were
washed in PBS and processed for immunofluores-
cence (see MATERIALS AND METHODS) either
immediately or after further incubation for 10 or
20 h (37°C, 5% CO
2
). Extracellular tachyzoites were
stained with the anti-P30 followed by the anti-
mouse antibody conjugated to Alexa-488. The cells
were subsequenty permeabilized and stained for
Toxofilin followed by an anti-rabbit antibody con-
jugated to Alexa 568. Nuclei were stained by DAPI.
Extracellular parasites were detected by the green
fluorescent staining on their surface, whereas
tachyzoites in the process of entering the host cell
present an incomplete staining of their surface (ar-
rows). (B) Gliding tachyzoites were processed as
for A. After fixation they were permeabilized and
stained for either their micronemes, using the
mouse monoclonal antibody anti-MIC1, or for their
dense granules, using the mouse monoclonal anti-
body anti-GRA3. In both cases, the secondary anti-
mouse antibody was conjugated to Alexa-488, and
the staining is shown in green. After removal of the
unbound conjugated antibody, the cells were
stained for Toxofilin and for DNA as described
above.
Toxofilin Controls Actin Dynamics
Vol. 11, January 2000 365

pointed filament ends as well as to induce the spontaneous
breaking of actin–actin contacts and thus play a major role in
controlling dynamics (Carlier, 1998).
On that basis, we were interested in identifying parasite
proteins with a sequestering function. Using complementary
biochemical approaches, we initially isolated a parasite 27-
kDa protein on the basis of its ability to bind rabbit muscle
G-actin and demonstrated that this protein formed a com-
plex with parasite G-actin. It was subsequently identified as
a novel actin-binding protein. This protein does not display
any significant similarity to any other known actin-seques-
tering proteins such as profilins or thymosins or with any
other known actin-binding protein. Additionally, it does not
contain any known consensus sequence for actin binding
site. We named this protein Toxofilin.
Actin-binding proteins often act in concert with each other
or with other proteins, some of which are regulatory mole-
cules (Pantaloni and Carlier, 1993; Mccollum et al., 1996;
Narumiya et al., 1997). Our data, in particular those obtained
after native electrophoresis, suggested that Toxofilin partic-
ipated in a protein complex containing G-actin. By analyzing
how r-Toxofilin interacted with actin, both with a native gel
assay and an actin polymerization assay, we demonstrated
that r-Toxofilin was capable by itself of sequestering G-actin
in vitro. The inhibition of actin assembly in presence of
r-Toxofilin was also observed using the cosedimentation
assay. In addition, r-Toxofilin is likely to cap actin filaments,
an activity that was further supported by immunoelectron
microscopy. r-Toxofilin was clearly localized at one end of
most of the actin filaments analyzed. Nevertheless, it does
not act as a heterodimer like the classical actin-capping
proteins described in every other eukaryotic cell studied
thus far (Schafer and Cooper, 1995), nor does it seem to
increase the initial rate of actin polymerization as do other
known capping proteins that promote actin nucleation (Coo-
per and Pollard, 1985).
Actin dynamics was also clearly affected when Toxofilin
was expressed in mammalian nonmuscle cells. GFP-Toxofi-
lin ectopic expression in epithelial cells (HeLa) resulted in
loss of actin stress fibers, a result that was probably due to
inhibition of the renewal of actin stress fibers. The extent of
the effect on actin cytoskeleton depended on the level of
GFP-Toxofilin expression. We claim that this is a direct effect
on the actin cytoskeleton, because immunoprecipitation of
actin from GFP-Toxofilin-expressing cells coprecipitated
Toxofilin as well. This piece of data further supports the
monomer-sequestering activity of Toxofilin suggested by
the in vitro kinetics of actin polymerization. Despite its
dramatic effect on the actin cytoskeleton, ectopic expression
of GFP-Toxofilin did not affect the microtubule network (our
unpublished results).
To localize endogeneous Toxofilin in tachyzoites we
used the anti-Toxofilin antibodies. As expected from pre-
vious reports on actin localization (Dobrowolski et al.,
1997b) and from our biochemical data, Toxofilin primarily
stained the cytoplasm in the anterior part of the T. gondii
tachyzoite. However, it is noteworthy that a different
distribution of Toxofilin was observed in gliding
tachyzoites with respect to the intracellular and immobile
parasites. Indeed, when tachyzoites were freshly isolated
from mice and allowed to glide, Toxofilin staining was
heterogeneous from one parasite to the other, sometimes
clearly in patches, whereas in other cases it was restricted
under the apical membranes with occasionally a patch at
the rear side. We examined whether the patchy distribu-
tion could be due to an association with organelles, espe-
cially with the numerous dense granules, but we did not
observe any colocalization.
Actin dynamics in T. gondii tachyzoite is not only required
during gliding but also during host cell entry. T. gondii
tachyzoites invade their host cells within few seconds once
they have reoriented their apical side to establish contact
Figure 7. In intracellular tachyzoites,
Toxofilin distributes uniformly in the
apical cytoplasm. Tachyzoites were in-
cubated for 10 min (37°C, 5% CO2)
with HFF cells plated on glass cover-
slips. The nonbound parasites were re-
moved by changing the medium, and
the infected cells were further incu-
bated for 20 h as described in MATE-
RIALS AND METHODS. At the end of
the incubation period the cells were
fixed, and extracellular tachyzoites
were stained with the anti-P30 anti-
body followed by the anti-mouse anti-
body conjugated to Alexa-488, seen in
red. The cells were subsequently per-
meabilized and stained for Toxofilin
followed by an anti-rabbit conjugated
to Alexa-568. Nuclei were stained by
DAPI, shown here in green (artificial
color). The actin cytoskeleton of the
host cell was stained with rhodamine-
phalloidin, shown in blue (artificial
color). The indicated area was enlarged
2⫻and is presented as an inset.
O. Poupel et al.
Molecular Biology of the Cell366

with the host cell membrane (Silverstein and Joiner, 1997).
Concomitantly, they secrete part of the content of their api-
cal organelles, mainly the rhoptries (Carruthers and Sibley,
1997). This is a regulated secretion event that participates in
the formation of the parasitophorous vacuole membrane
together with the invagination of the host cell plasma mem-
brane (Suss-Toby et al., 1996; Lingelbach and Joiner, 1998).
We detected tachyzoites when they were still “half in–half
out” of their host cell. In these parasites, we always observed
a redistribution of Toxofilin, including a strong patch at the
distal part of the parasite, which had not yet been internal-
ized. It is known that during gliding Apicomplexan para-
sites cap some of their surface molecules, which move away
from the apical side (the advancing part of the cell) toward
the distal end. It is accepted that parasite actin filaments
associated with cross-linked surface molecules move rear-
ward during the capping process. These filaments are
formed by assembly in the front of the cell. This implies that
actin monomers have to be locally desequestered under the
membranes to be competent for association through the
nucleation of new actin filaments or through elongation of
actively growing preexisting filaments. In the latter case,
preexisting filaments should uncap to lower the critical con-
centration and generate free ends for monomer assembly.
Toxofilin release may be involved during this uncapping
step as well, because it could regulate the pool of available
monomers for actin assembly. The newly elongated actin
filaments formed under the membrane would then rapidly
dissociate at the rear side, leaving monomers to be seques-
tered by Toxofilin.
Once into their host cell and during their entire intracel-
lular growth phase, tachyzoites are more or less immobile.
In these parasites, the staining of Toxofilin was significantly
stronger and homogeneously localized throughout the api-
cal cytoplasmic part for every parasite. In contrast to the
uniform pattern observed in immobile parasites, the broad
pattern of Toxofilin distribution observed in gliding
tachyzoites may reflect a highly dynamic behavior of Toxo-
filin, probably coordinated with the dynamic state of actin.
It remains to be clarified how Toxofilin could be involved
in the actin polymerization process controlling tachyzoite
gliding, invasion into host cells, and further intracellular
development. Genetic manipulation of the level of protein
expression in live T. gondii tachyzoites will help answer
these questions (Boothroyd et al., 1995; Donald and Roos,
1995). Finally, taking into account that the gliding motility of
Apicomplexan parasites displays some unique features, it is
expected that understanding how Toxofilin acts in the par-
asite in vivo should permit the identification of novel mol-
ecules critically involved in long-term survival of Apicom-
plexan parasites. This could create new possibilities for
targeted therapeutic approaches against Apicomplexan-
caused diseases.
In conclusion, Toxofilin, besides its great importance for
understanding Toxoplasma biology, will become an inter-
esting tool for the dissection of actin dynamics in other
systems, as is already the case for proteins from other mi-
croorganisms, in particular Listeria monocytogenes (Lasa and
Cossart, 1996; Carlier and Pantaloni, 1997, Moreau and Way,
1999).
ACKNOWLEDGMENTS
We deeply thank G. Milon, N. Guillen, and G. Langsley (Institut
Pasteur, Paris, France) for encouragement throughout this project
and P. Cossart for careful reading of the manuscript. Special thanks
to G. Langsley, who has participitated in the design of several
experiments. We are grateful to M.F. Cesbron and J.F. Dubremetz
(Institut Pasteur, Lille, France) for providing us with the anti-T.
gondii MIC3 and anti-T. gondii GRA1 antibodies and D. Sibley
(Washington University, St. Louis, MO) for the anti-T.gondii actin
antibody. We are grateful to G. Milon, O. Mercereau-Puijalon, and
P. Falanga (Institut Pasteur, Paris, France) as well as to M. Alizon
(Institut Cochin de Ge´ne´tique Mole´culaire, Cochin, Paris, France)
for supplying reagents and to A. Dautry for providing access to
microscope and camera facilities. This project was financed by the
Pasteur Institute (Dr. Genevieve Milon, Unite´ d’Immunophysiologic
et Parasitisme Intracellulaire), and we address a special thanks to J.
Castex (Pasteur Institute, Paris, France).
REFERENCES
Ajioka, J.W., et al. (1998). Gene discovery by EST sequencing in
Toxoplasma gondii reveals sequences restricted to the Apicomplexa.
PCR Methods Appl. 8, 18–28.
Allen, M.L., Dobrowolski, J.M., Muller, H., Sibley, L.D., and Man-
sour, T.E. (1997). Cloning and characterization of actin depolymer-
izing factor from Toxoplasma gondii. Mol. Biochem Parasitol. 88,
43–52.
Ayscough, K.R. (1998). In vivo functions of actin-binding proteins.
Curr. Opin. Cell Biol. 10, 102–111.
Boothroyd, J.C., Black, M., Kim, K., Pfefferkorn, E.R., Seeber, F.,
Sibley, L.D., and Soldati, D. (1995). Forward and reverse genetics in
the study of the obligate, intracellular parasite Toxoplasma gondii. In:
Methods in Molecular Genetics, vol. 6, ed. K.W. Adolph, San Diego:
Academic Press, 3–29.
Carlier, M.F. (1990). Actin polymerization and ATP hydrolysis.
Adv. Biophys. 26, 51–73.
Carlier, M.F. (1991). Actin: protein structure and filament dynamics.
J. Biol. Chem. 266,1–4.
Carlier, M.F. (1998). Control of actin dynamics. Curr. Opin. Biol. 10,
45–51.
Carlier, M.F., and Pantaloni, D. (1997). Control of actin dynamics in
cell motility.J. Mol. Biol. 269, 459–467.
Carruthers, V.B., and Sibley, L.D. (1997). Sequential protein secre-
tion from three distinct organelles of Toxoplasma gondii accompanies
invasion of human fibroblasts. Eur. J. Cell Biol.73, 114–123.
Cooper, J.A., and Pollard, T.D. (1985). Effect of capping protein on
the kinetics of actin polymerization. Biochemistry 24, 793–799.
Dobrowolski, J.M., Carruthers, V.B., and Sibley, L.D. (1997a). Par-
ticipation of myosin in gliding motility and host cell invasion by
Toxoplasma gondii. Mol. Microbiol. 26, 163–173.
Dobrowolski, J.M., Niesman, I.R., and Sibley. L.D. (1997b). Actin in
the parasite Toxoplasma gondii is encoded by a single copy gene,
ACT1 and exists primarily in a globular form. Cell Motil. Cytoskel-
eton 37, 253–262.
Dobrowolski, J.M., and Sibley, L.D. (1996). Toxoplasma invasion of
mammalian cells is powered by the actin cytoskeleton of the para-
site. Cell 84, 933–939.
Dobrowolski, J.M., and Sibley, L.D. (1997). The role of the cytoskel-
eton in host cell invasion by Toxoplasma gondii. Behring Inst. Mitt. 99,
90–96.
Donald, R.G.K., and Roos, D.S. (1995). Insertional mutagenesis and
marker rescue in a protozoan parasite: cloning of the uracil phos-
Toxofilin Controls Actin Dynamics
Vol. 11, January 2000 367

phoribosyltransferase locus from Toxoplasma gondii. Proc. Natl.
Acad. Sci. USA 92, 5749–5753.
Dos Remedios, C.G., and Cooke, R. (1984). Fluorescence energy
transfer between probes on actin and probes on myosin. Biochim.
Biophys. Acta 788, 193–205.
Dubremetz, J.F. (1998). Host cell invasion by Toxoplasma gondii.
Trends Microbiol. 6, 27–30.
Fahrni, J.F. (1992). Actin in the ciliated protozoan Climacostomum
virens: purification by DNase I affinity chromatography, electro-
phoretic characterization, and immunological analysis. Cell Motil.
Cytoskeleton 22, 62–71.
Kasper, L.H., and Buzoni Gatel, D. (1998). Some opportunistic par-
asitic infections in AIDS: candidiasis, pneumocystosis, cryptospori-
diosis, toxoplasmosis. Parasitol. Today 14, 150–156.
King, C.A. (1988). Cell motility of sporozoan protozoa. Parasitol.
Today 4, 315–319.
Kouyama, T., and Mihashi, K. (1981). Fluorimetry study of N-(1-
pyrenyl)iodoacetamide-labeled F-actin. Local structural change of
actin protomer both on polymerization and on binding of heavy
meromyosin. Eur. J. Biochem.114, 33–38.
Laemmli, U.K. (1970). Cleavage of structural proteins during the
assembly of the head of bacteriophage T4. Nature 227, 680–685.
Lasa, I., and Cossart, P. (1996). Actin-based bacterial motility: to-
ward a definition of the minimal requirements. Trends Cell Biol. 6,
109–114.
Lingelbach, K., and Joiner, K.A. (1998). The parasitophorous vacuole
membrane surrounding Plasmodium and Toxoplasma: an unusual
compartment in infected cells. J. Cell Sci. 111, 1467–1475.
Maciver, S.K. (1998). How ADF/cofilin depolymerizes actin fila-
ments. Curr. Opin. Cell Biol. 10, 140–144.
Marsh, K., and Snow, R.W. (1997). Host-parasite interaction and
morbidity in malaria endemic areas. Philos. Trans. R. Soc. Lond. B.
Biol. Sci. 352, 1385–1394.
Mccollum, D., Feoktistova, A., Morphew, M., Balasubramanian, M.,
and Gould, K.L. (1996). The Schizosaccharomyces pombe actin-related
protein, Arp3, is a component of the cortical actin cytoskeleton and
interacts with profilin. EMBO J. 15, 6438–6446.
Miller, K.G., and Alberts, B.M. (1989). F-actin affinity chromatogra-
phy: technique for isolating previously unidentified actin-binding
proteins. Proc. Natl. Acad. Sci. USA 86, 4808–4812.
Moreau, V., and Way, W. (1999). In vitro approaches to study actin
and microtubule dependent cell processes. Curr. Opin. Cell Biol. 11,
152–158.
Narumiya, S., Ishizaki, T., and Watanabe, N. (1997). Rho effectors
and reorganization of actin cytoskeleton. FEBS Lett. 410, 68–72.
Pantaloni, D., and Carlier, M.F. (1993). How profilin promotes actin
filament assembly in the presence of thymosin
␤
4. Cell 75, 1007–
1014.
Pardee, J.D., and Spudich, J.A. (1982). Purification of muscle actin.
Methods Enzymol. 85, 164–181.
Perelroizen, I., Didry, D., Christensen, H., Chua, N.H., and Carlier,
M.F. (1996). Role of nucleotide exchange and hydrolysis in the
function of profilin in actin assembly. J. Biol. Chem. 271, 12302–
12309.
Poupel, O., and Tardieux, I. (1999). Toxoplasma gondii motility and
host cell invasiveness are drastically impaired by jasplakinolide, a
cyclic peptide stabilizing F-actin. Microbes Infect. 1, 653–662.
Preston, T.M., and King, CA. (1996). Strategies for cell-substratum
dependent motility among Protozoao. Acta Protozool. 35, 3–12.
Safer, D. (1989). An electrophoretic procedure for detecting proteins
that bind actin monomers. Anal. Biochem. 178, 32–37.
Schafer, D.A., and Cooper, J.A. (1995). Control of actin assembly at
filament ends. In: Annual Review of Cell and Developmental Biol-
ogy, vol. 11, ed. J.A. Spudich, Palo Alto, CA: Annual Reviews,
497–518.
Silverman, J.A., and Joiner, K.A. (1997). Toxoplasma/host cell inter-
actions. In: Host Responses to Intracellular Pathogens, ed. S.H.E.
Kaufmann, Austin, TX: R.G. Landes, 313–338.
Stossel, T.P. (1993). On the crawling of animal cells. Science 260,
1086–1094.
Suss-Toby, E., Zimmerberg, J., and Ward, G.E. (1996). Toxoplasma
invasion: the parasitophorous vacuole is formed from host cell
plasma membrane and pinches of via a fission pore. Proc. Natl.
Acad. Sci. USA. 93, 8413–8418.
Tardieux, I., Baines, I., Mossakowska, M., and Ward, G.E. (1988).
Actin binding proteins of invasive malaria parasites: regulation of
actin polymerization by a complex of 32/34-Kda proteins. Mol.
Biochem. Parasitol. 93, 295–308.
Welch, M.D., Mallavarapu, A., Rosenblatt, J., and Mitchison, T.J.
(1997). Actin dynamics in vivo. Curr. Opin. Cell Biol. 9, 54–61.
O. Poupel et al.
Molecular Biology of the Cell368