|
|||||||||
|
|
|
|
|
|
|
|
|
|
|
HEMOSTASIS, THROMBOSIS, AND VASCULAR BIOLOGY
From the Department of Pathology and Laboratory
Medicine, Brown University, Providence, RI.
Platelets undergo a series of actin-dependent morphologic
changes when activated by thrombin receptor activating peptide (TRAP) or when spreading on glass. Polymerization of actin results in the
sequential formation of filopodia, lamellipodia, and stress fibers, but
the molecular mechanisms regulating this polymerization are unknown.
The Arp2/3 complex nucleates actin polymerization in vitro and could
perform this function inside cells as well. To test whether Arp2/3
regulated platelet actin polymerization, we used recombinant Arp2
protein (rArp2) to generate Arp2-specific antibodies ( Activation of platelets produces a
reproducible sequence of morphologic events, whether in suspension or
during spreading on glass: rounding, filopodial projection, attachment,
spreading, and ultimately contraction.1-6 These
morphologic changes depend on the reorganization of the actin
cytoskeleton, including severing of existing filaments, which causes
the discoid platelet to round and depends on
gelsolin,3,7,8 and polymerization of actin monomers into
new filaments.3,9-11 These new actin filaments organize
into 4 distinct structures: filopodia, lamellipodia, stress fibers, and
a contractile ring.4 Each of these structures performs a
different function, and each contains a different complement of
actin-binding proteins.4-6
The Arp2/3 complex is likely to regulate the polymerization of
actin during shape change in the platelet. Arp2/3 is a 7-member protein
complex isolated by poly-proline chromatography from the soil amoeba,
Acanthamoeba castellani.12 Although the Arp2
subunit was isolated from human platelets, first by F-actin affinity
chromatography4 and later as a member of the Arp2/3
complex that mediates Listeria-induced actin
assembly,13 neither its function nor its location in
platelets has been previously reported. In vitro, Arp2/3 nucleates
actin filament assembly14,15 and is required, together
with 2 other proteins, to reconstitute the actin-based motility of
Listeria bacteria.16 Arp 2/3 is reported to
have at least 2 binding sites for actin: one that binds to the sides of
actin filaments and the other that binds to the pointed ends of actin
monomers nucleating barbed-end elongation.14,17,18 In
vitro, this could produce networks of filaments that branch at 70°
angles. In soil ameba and in cultured cells, Arp2/3 is found in the
lamellipodia12,15,17 where filaments branch at 70°
angles.19 Antibodies to the p34 subunit of Arp2/3 inhibit
this branching activity in vitro and in vivo but do not inhibit the
incorporation of actin monomer.17 Antibodies to the Arp2,
but not the Arp3, subunit inhibit actin-polymerizing activity in
extracts of Acanthamoeba.20 Based on
correlations between this in vitro behavior and its cellular
location, Arp2/3 is considered the best candidate to regulate
actin dynamics physiologically at the membrane-cytoplasm
interface.21-27
Because platelets are anucleate and are very small, they can
neither be transfected nor injected. Permeabilization has been used to
load platelets with pyrene-actin to study agonist-stimulated actin
polymerization biochemically.3,11 In platelets,
actin-polymerizing activity increases in response to the agonist,
thrombin receptor activating peptide (TRAP), and 60% of this increase
is preserved after permeabilization with detergent. We discovered that
permeabilized platelets undergo normal morphologic events when exposed
to glass: they adhere and reorganize their actin filaments into the
typical 4 structures. We have used these techniques together with a set of molecular tools, including new inhibitory antibodies generated against recombinant Arp2 (rArp2), to test the role of Arp2/3 in actin
dynamics and in the morphologic events of surface-activated shape change.
Constructs and antibodies
Affinity-purified antipeptide rabbit antibodies against p34 were a gift
from John Condeelis.17 Rabbit anti-kaptin was raised in
our laboratory,5 and mouse anti-actin is from the
monoclonal hybridoma JLA-4.4,5 Western blotting was
performed as previously described.4-6
Quantification of the amount of Arp2 in platelets was performed as
follows. Platelet-rich plasma (PRP) was used on the day of draw as
described.4-6 The platelet count was determined (usually approximately 5.5 × 1011 platelets/U). After washing,
the platelets from one unit of PRP were resuspended in gel sample
buffer at a volume calculated to give a final concentration of
2.5 × 1011 platelets/mL. Various amounts of platelets
were loaded into separate lanes on sodium dodecyl
sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) in parallel with
different amounts of purified rArp2 and were probed for Arp2 by Western
blotting. The amount of Arp2 in the platelet sample was determined by
comparing the density of resultant bands in the platelet samples with
the density of bands containing a known amount of rArp2 using the
GelDoc with Quantity One software (BioRad, Hercules, CA). The amount of
Arp2 molecules per platelet was determined by dividing the amount of Arp2 detected on the blot by the number of platelets loaded on the gel.
Molar concentration was determined by assuming that the average volume
of a platelet was 7 fL.30 It required
1.6 × 108 platelets (6 µL) per platelet sample
to detect a band on the blot, with an anti-Arp2 detection limit of 15 ng.
Precipitation of Arp2 with Pyrene actin polymerization kinetics Pyrene actin (10% labeled; Cytoskeleton, Denver, CO) was aliquoted, lyophilized from G-buffer (5 mM Tris [pH 8.1], 0.2 mM CaCl2, 0.2 mM adenosine triphosphate [ATP], 0.5 mM dithiothreitol [DTT]), and stored in liquid nitrogen. On the day of use, the actin was thawed, incubated at 10 µM in water for 4 hours at room temperature, and centrifuged for 1.5 hours at 100 000g. Immediately before each experiment, this stock of actin was diluted to 1.3 µM in 200 µL polymerization buffer (100 mM KCl, 2 mM MgCl2, 0.1 EGTA, 10 mM Tris [pH 7.0], 0.5 mM ATP, 0.5 mM DTT). The cuvette containing the diluted actin was placed in the fluorometer for a baseline reading before the platelet suspension was added.Washed platelets were resuspended in platelet buffer (145 mM NaCl, 5 mM KCl, 0.5 mM Na2HPO4, 2 mM MgCl2, 10 mM glucose, 10 mM HEPES [pH 7.4], and 0.3% globin-free bovine serum albumin more than 99% pure) and were maintained at 37°C. For sonication, 90 µL platelets were treated with 10 µL PHEM buffer (60 mM PIPES [pH 6.9], 25 mM HEPES [pH 6.9], 10 mM EGTA, 2 mM Mg Cl2, and protease inhibitors) with or without 1 µL of 1.2 mM TRAP (Sigma, St Louis, MO) as described.6 Sonicates were prepared individually just before each assay, antibodies were added immediately thereafter, and the suspension was injected into the cuvette containing pyrene-actin. For Triton permeabilization, a 1:10 dilution of 7.5% Triton X-100 with and without TRAP (1 µL of 1.2 mM) was added to the PHEM buffer, and no sonication was performed. Otherwise, the procedure was essentially the same as for sonication. For sonication or Triton permeabilization, fewer than 30 seconds elapsed between TRAP stimulation and measurement in the fluorometer. After platelets were added, actin polymerization was measured for 10 to 20 minutes at 1-second intervals with excitation at 365 and emission 407 nm in a luminescence spectrometer (model LS50B; Perkin-Elmer, Wellesley, MA). Data were collected using the Winlab program and were analyzed with Microsoft Excel. Each series of experiments was performed on the same day with the same preparations of actin and platelets, beginning with baseline measurements of actin alone, and of platelets with or without TRAP. Morphometric analysis of Immunofluorescence to determine endogenous Arp2/3 location PRP without permeabilization was dotted onto glass coverslips, and the coverslip was flooded with Tyrode buffer as described above. After 20 minutes, spread platelets were fixed in 4% paraformaldehyde in PHEM buffer containing a 1/50 dilution of FITC-phalloidin with or without 0.25% Triton for 20 minutes at room temperature. Fixed platelets were blocked with 1% bovine serum albumin (Sigma) in phosphate-buffered saline with 0.1% Triton X100. Coverslips were then stained with Arp2 followed by secondary
antibody. Images were obtained on a Nikon fluorescence microscope
equipped with an RT-Spot liquid crystal digital camera (Diagnostic
Instruments, Starling Heights, IL).
Generation of rArp2 and anti-Arp2 antibodies To develop tools to test the role of Arp2/3 in platelet actin dynamics, we cloned the Arp2 subunit and raised antibodies to full-length rArp2. Immunization of 8 different rabbits produced 3 with high-titer antiserum that was sensitive and specific for Arp2. We characterized theArp2 antibodies against bacterially expressed
rArp2 (Figure 1A) and against platelet
extracts (Figure 1B-C). Affinity-purified Arp2 detected a single
band in induced bacterial homogenates (Figure 1A, lane 3) and
recognized purified rArp2 (Figure 1A, lane 6). The antibody could
detect as little as 15 ng purified rArp2. In platelet extracts, Arp2
detected a single band of the appropriate molecular weight (44 kd)
(Figure 1B).
Quantitative Western blot analysis comparing purified rArp2 with platelet extracts revealed that the concentration of Arp2 in platelets is 0.3 µM, or 1300 molecules per platelet. This amount is sufficient to account for all the new free actin filament ends (410-570 barbed ends/platelet) produced during activation,3,11,31 even if only half the Arp2/3 is activated. This is significantly less protein than VASP or gelsolin, which are estimated at 5 µM each,6,31,32 and this amount is consistent with the yield of Arp2/3 complex obtained from platelets in purification protocols.13 Arp2 was depleted from Triton-treated, sonicated platelet extracts by
incubation with Effect of Arp2 inhibition on platelet actin-polymerizing activity The effect ofArp2 on actin-polymerizing activity of platelet
extracts was measured in the pyrene assay. To perform this assay,
platelets must be permeabilized sufficiently for enough pyrene-labeled
actin to be loaded for detection in the fluorometer.8,11 Such permeabilization also allows entry of the Arp2 antibody, thus
permitting measurements of the effect of antibody on actin polymerization.
TRAP produced a robust increase in actin-polymerizing activity in
platelets permeabilized by sonication (6.3-fold, Figure 2A, C) or detergent (Figure 2B). This
TRAP-stimulated increase was exquisitely sensitive to
Three additional experiments confirmed the specificity of the Because Arp2/3 is expected to increase the initial rate of
polymerization, we calculated the rates during the first 120 seconds and compared the effect of different treatments (Table 1). Only In the absence of TRAP, Adherence and spreading of permeabilized platelets on glass To examine whether the formation of particular actin structure(s) such as filopodia and lamellipodia requires Arp2/3 activity, we developed a technique to load platelets withArp2 and
then to observe the morphology of actin after spreading on glass
(Figure 3). Platelets in plasma were
exposed briefly to detergent with or without antibody, mounted on
glass, and rapidly diluted with an approximately 500-fold excess
of Tyrode solution. After spreading, platelets were fixed and stained
for F-actin.
Permeabilization alone did not alter adherence to the glass. The same
average number of platelets adhered with or without permeabilization
(19.9 ± 7.6 and 19.5 ± 6.5 platelets per high-power field,
respectively). The presence of any antibody increased adherence, regardless of antibody type (27.7 ± 5.2 platelets per high-power field
after preimmune treatment and 28.2 ± 8.0 after Permeabilized platelets exhibited all the typical actin structures as imaged by FITC-phalloidin staining (Figure 3, top row, middle panel), though some minor variations were observed. Filopodia seemed longer, and the diffuse staining in the cytoplasm was decreased. Permeabilized platelets treated with preimmune antibodies also formed filopodia, lamellipodia, and stress fibers indistinguishable from platelets permeabilized without antibody (Figure 3, top row, right panel). Morphologic effect of Arp2 inhibition on platelet shape change Inactivation of Arp2/3 had a striking effect on morphology and on actin structures (Figure 3, lower row). Four dose-dependent effects ofArp2 on morphology were observed and quantified: (1) decreased
proportion of spread platelets; (2) decreased proportion with
filopodia; (3) increased proportion in the rounded stage; and (4)
presence of a novel morphology, blebbing, not seen in the absence of
Arp2. These effects were quantitative and Arp2 dose dependent
(Figure 4). Antibodies from 2 different
rabbits were each used in parallel with the preimmune antibodies from the same rabbit.
Anti-Arp2 loading increased the proportion of rounded platelets in a
dose-dependent manner (2.85%, 22%, 27%, 32%, and 63% for 0, 0.1, 0.2, 0.3, and 1 µg/mL, respectively), with commensurate decreases in
the proportion that had produced filopodia or spread (Figure 4). In the
few The proportion of platelets with blebs, which are never found in
control preparations, increased in direct relationship to the amount of
Fab fragments of The efficiency of permeabilization was determined by staining platelets
with secondary antibody after fixation (Figure
5). Primary antibody was detected in 79%
of platelets permeabilized in the presence of preimmune antibodies and
76% of those treated with
Location of Arp2/3 in normal platelets during spreading on glass We used double-label immunofluorescence to compare the normal location of Arp2/3 with that of filamentous actin (Figure 6). Platelets were first allowed to spread and then were fixed with paraformaldehyde in buffer with or without detergent. Fixation without detergent preserves a large pool of Arp2/3 that is lost when fixation includes Triton, whereas this pool is extracted in fixatives that include detergent.33 Thus, the timing of exposure to detergent allowed us to distinguish between extractable and actin-bound Arp2/3 in spread platelets.
Arp2/3 was detected at the edge of the lamellipodium of spread
platelets fixed in the presence of detergent, as has been reported for
other cells fixed this way (Figure 6).12,15,17,19,33 Superimposition of images of the spread platelet double labeled for
Arp2 (Figure 6B, red) and F-actin (Figure 6B, green) shows that Arp2/3
was coincident in most places with the actin frill of the
lamellipodium. To determine whether the actin frill was at the edge of
the platelet, phase microscopy was needed. However, when detergent was
included in the fixative, effective imaging by phase microscopy was
difficult. Therefore, we fixed platelets without detergent, double
labeled them for F-actin (Figure 6C, green, FITC-phalloidin) and for
Platelets captured during early stages of activation revealed a
relationship between Arp2/3 location and filopodia. In the early
stages, Arp2/3 was concentrated in the cortex, and there was brighter
staining over the short, F-actin-rich protrusions (Figure
7A-B, far left). At later stages, the
platelet contracts and growing filopodia stained for Arp2/3 at their
tips and their bases (Figure 7A-B, second from left). As activation
progressed, filopodia elongated and Arp2/3 was often found at the tip
and at the base (Figure 7A-B, third and fourth from the left). In some
cases, an area of Arp2/3 staining was also present on the shaft of the
filopodium. Such filopodia are too small to be imaged by phase
microscopy, containing as few as 3 actin
filaments.1,3,9,30 Higher magnification of these filopodia
demonstrated the relationship between Arp2/3 and actin filaments
(Figure 7C). The small size of platelets at the filopodial stage of
activation is apparent when they were compared to a fully spread
platelet at the same magnification (Figure 7D).
The experiments reported here provide direct functional evidence
that Arp2/3 is required for both filopodial and lamellipodial formation
in the platelet. Loading of platelets with Arp2 and the formation of filopodia The presence of Arp2/3 at the tips of the filopodia is surprising because Arp2/3 is reported to bind the pointed ends and the sides of actin filaments, whereas the other end, the barbed end, is at the tips of filopodia, as reported in cultured cells.34 Arp2/3 may be passively carried to the tip, along with membrane, as the filopodium grows. Arp2/3 may then nucleate new filaments for further growth of the filopodium. Alternatively, all platelet extensions that appear to be filopodia may not be identical to those produced by cells in culture. Indeed, in platelets, such extensions were originally termed pseudopodia, and some were found to contain microtubules.1,31 The actin filaments in some platelet filopodia may have reversed polarity, with the pointed end at the tip. Re-examination by electron microscopy of the actin organization and polarity in the platelet pseudopodia is needed to resolve these possibilities.Whether Arp2/3 plays a role in filopodial formation in other cells is
controversial. Evidence for Arp2/3 involvement is that cdc42, reported
to activate Arp2/3,20,24 initiates filopodial formation in
cultured cells.35,36 On the other hand, the Arp2/3 complex
nucleates branched filament networks in vitro14,18,37 and
is found at branch points by immunogold labeling in
cells.19 Filopodia are composed of parallel filaments and
do not have such branches, making such network formation by Arp2/3 seem
unlikely. However, 2 recently published experiments show that Arp2/3
can nucleate filament formation in the absence of branching. First, anti-p34 blocks branching and slows, but does not inhibit,
Arp2/3-induced actin polymerization;17 second,
polymerization by Arp2/3 in the presence of tropomyosin has a similar
effect37 At least at the early stage of filopodial projection in platelets, Arp2/3 seems to be required, but it may not be sufficient, and other proteins such as VASP may be involved at later stages. VASP is found at the tips of filopodia in neuronal growth cones,38,39 and recruitment of VASP by the proline-rich domain of N-WASp is required in addition to Arp2/3 for efficient actin polymerization at the cell surface of a hematopoietic cell line, RBL-2H3.40 Initially described in stress fibers in platelets spread on glass,41 VASP has recently been shown along the length of platelet filopodia and at the edge of the lamellipodia during spreading.6 Because VASP inhibits gelsolin severing, we have recently proposed that the release of VASP from sides of filaments could potentiate gelsolin activity.6 Many other actin-binding proteins that organize actin filament networks into three-dimensional structures are present in platelets and must act in concert on the newly formed filaments in a highly regulated choreography. Inhibition by Arp2 inhibition is
caused by its binding to Arp2 protein. Each of these 6 experiments was
performed both in the pyrene assay and in the spreading assay. First, 3 different rabbits produced inhibitory antibodies. Second, preimmune
antibodies, drawn from each rabbit before immunization, had no
detectable effect. Third, Arp2 was highly specific for Arp2 by
immunoprecipitation and Western blot. Anti-Arp2 did not precipitate
other proteins, including actin or p34, another Arp2/3 family member.
Fourth, inhibition was not detected with affinity-purified antibodies
against another actin-binding protein, kaptin, or against p34, another
Arp2/3 subunit. Fifth, Arp2 Fab fragments, which lack the ability to
cross-link antigen, had the same inhibitory effect as intact Arp2.
Sixth, pretreatment of the antibody with recombinant Arp2 protein
eliminated inhibitory activity.
Activation of Arp2/3 in platelets Another question will be how Arp2/3 is activated in platelets. Our results show that the thrombin and the glass-activated pathways converge on Arp2/3. In humans, N-WASp but not the WASp isoform activates Arp2/3.42-44 Platelets have little N-WASp and abundant WASp.45,46 Platelet extracts do not support the N-WASp-dependent actin polymerization induced by Shigella bacteria.47 Furthermore, platelets from patients with Wiskott-Aldrich syndrome have no detectable defect in actin assembly on activation though they are abnormally small,48 indicating that some protein other than WASp must activate Arp2/3 in platelets. Other members of the WASp/Scar family appear to be expressed in platelets (Oda, personal communication). If the WIP isoform turns out to be present in platelets, this could be the activator of Arp2/3 for filopodial production.49 Because WASp appears to be the downstream mediator of cdc42, it will be important to determine whether cdc42 is also involved in platelet filopodial formation.Binding to the sides of actin filaments can also activate Arp2/3.18,37,50 Actin filaments of the platelet membrane skeleton could thus serve as activation sites for Arp2/3 in platelets following agonist stimulation. The membrane skeleton of the nonstimulated platelet consists of submembranous microfilaments that line the inner surface of the platelet plasma membrane in an ordered array parallel to the membrane.51,52 In quick-freeze, deep-etch replicas of platelets captured in the early stages (1-2 seconds) after thrombin activation, this array becomes more prominent.52 Biochemical analysis of the resting platelet membrane skeleton demonstrates that actin, spectrin, myosin, and actin-binding protein are present.53-56 This membrane skeleton undergoes dramatic remodeling after agonist stimulation, including severing of the actin filaments.7,8,31,55,56 If severing is a consequence of the release of filaments by VASP, as we have proposed, then these filament sides could act as activation sites for Arp2/3. Evidence for other nucleators of actin polymerization Arp2/3 was not detected in all actin structures, and some actin-polymerizing activity remained in extracts treated withArp2. Thus, other mechanisms may exist to initiate polymerization. Indeed, actin polymerization seems too important to be mediated by a single mechanism. Evidence supporting an alternative mechanism for actin polymerization includes a report that only 40% of cold-induced, barbed-end formation in platelets is inhibited by the C-terminal of
N-WASp.57 Other candidates for nucleators include VASP
(discussed above), gelsolin, moesin, and kaptin. Gelsolin severing of
filaments is hypothesized to produce fragments that, once uncapped, act as nucleation sites for rapid elongation.8,58,59 Moesin, the only member of the ERM family found in platelets, is also present
at filopodial tips and induces filopodiallike membrane protrusions when
overexpressed in cultured cells.60 Kaptin, an
ATP-sensitive F-actin-binding protein, is also found at sites of new
filament formation in many cells and at the leading edges of the spread
platelet.5
Alternatively, Arp2/3 may indeed be the only nucleator. After all,
Contextual model of Arp2/3 role in platelet actin reorganization Based on these results, we propose the following model of Arp2-mediated platelet actin polymerization.61 A soluble pool of Arp2/3 is recruited to the membrane upon agonist stimulation in the first stage, rounding, of activation. Activation of this Arp2/3 would produce an explosive burst of polymerization of new filaments at the cortex. Polymerization is aided by an increase in actin monomers resulting from gelsolin severing and capping of existing filaments,6-8,11,58,59 possibly facilitated by cofilin-mediated depolymerization.16,19,27 Dissociation of side-binding proteins such as VASP would potentiate severing and depolymerization6 and expose filament sides to serve as activation sites for Arp2/3. Subsequent barbed-end capping by capping protein62 or 2E4/kaptin5 would limit the length and location of new filament elongation. Filopodia and lamellipodia would then form as new filaments become networked by other actin-binding proteins, such as VASP,6,41-actinin,63 or tropomyosin.64
Conclusions This study reports several significant advances. First, these results identify Arp2/3 as a major regulator of platelet actin dynamics, responsible for the formation of filopodia and lamellipodia. This represents a significant advance in our understanding of the molecular events leading to platelet shape change. Second, our new permeabilization method, which preserves the platelet's ability to respond to agonists after loading with molecules as large as immunoglobulins, at last makes it possible to manipulate the molecular composition of the platelet cytoplasm. This powerful new technology will allow us to investigate biochemical relationships between signaling pathways and morphologic changes occurring in platelets but common to all cells. Results from platelets are therefore likely to provide fundamental information about the principles and paradigms governing actin dynamics inside all cells since membrane-associated actin polymerization is also required for the formation of a wide number of physiologically significant structures in virtually all eukaryotic cells.
We thank Jem Prakash for technical assistance and Leslie Hunter for advice on bioengineering. We are grateful to Eric Fyrberg for the single-stranded cDNA of Arp2 and to John Condeelis for the anti-p34 antibody and helpful discussions. We thank Dee Bainton for her interest in our work and useful discussions.
|
Top
Abstract
Introduction
au/
not 
