| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
* Institut d’Investigacions Biomèdiques August Pi i Sunyer, Department of Physiological Sciences I, and
Department of Cellular Biology, Facultad de Medicina-University of Barcelona, Barcelona, Spain
2Correspondence: IDIBAPS-Department of Physiological Sciences I, Facultad de Medicina-University of Barcelona, Barcelona, Spain. E-mail: miguelmorales{at}ub.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: actin cytoskeleton • cell spreading • lamella • protein transduction
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. Lamellipodia spreading is the first step in the process of locomotion and is provided by a rapid, site-specific nucleation and polymerization of actin in a complex network that pushes the cell membrane outward. Rac1, a small GTPase of the Rho family, plays a central role in inducing actin-dependent lamellipodia formation (2
; for an extensive review see ref. 3
).
Rac1 activation stimulates lamellipodia extension by activating WAVE proteins (members of the WASP/WAVE family). The mechanism, still under heavy scrutiny, involves the formation of a multimeric protein complex with HSPC300, Abi1, Nap1, PIR121, Nck, and the actin nucleation factor Arp2/3, which, in turn, generates branched actin filaments and new free barbed ends. This complex promotes F-actin polymerization, which will eventually push the membrane (4)
.
Profilins (14–17 kDa) are a family of conserved ubiquitous proteins considered to be important regulators of F-actin dynamics. Among the four mammalian isoforms, Profilin I (Pfn I) is present in all of the tissues with the exception of the striated muscle. Pfn I functions are critical for development, and knockout mice die during the early embryonic stages (5)
. Profilin II (PfnII), the neuronal isoform, is widely expressed in the nervous system. Pfn I and II have slight species-dependent variations in the actin and polyprolines affinities (for a review, see ref. 6
). Profilins III and IV have a more restrictive expression, being expressed only in testis (7
, 8)
. Profilins were originally described as G-actin-binding proteins, and therefore, they were thought to be mainly responsible for actin monomer sequestration (9)
. In the presence of capping filaments, profilins can sequester actin monomers forming a 1:1 complex and inhibit actin polymerization. In living cells, when actin barbed ends are free to polymerize, profilin promotes actin filament dynamics by catalyzing nucleotide exchange from ADP- to ATP-actin (10)
, thereby contributing to actin polymerization at the leading edge of lamellipodia. Moreover, it is believed that through their polyproline (PLP) binding site, profilins can interact with a variety of proteins involved in the actin-driven mechanism, such as Arp3, Ena/VASP, diaphanous, N-WASP, and WAVE1 among others (6)
.
We studied the role of profilin in lamellipodia generation using primary cultures of bovine trabecular meshwork (BTM) cells as a model. The trabecular meshwork is an ocular structure located at the iridocorneal angle and is believed to play an important role in the regulation of aqueous humor outflow and intraocular pressure (IOP) (11
, 12)
. Previously, by fusing Pfn I to a transduction domain (PTD4) that allows the whole recombinant protein to cross the cell membrane, we showed that the PTD4-profilin I (PTD4-PfnI) construct was able to penetrate into trabecular meshwork cells and modulate the outflow of aqueous humor through that tissue (13)
. Furthermore, topical-corneal application of PTD4-PfnI was able to reduce intraocular pressure (14)
. The fact that BTM cells exhibit a high degree of motility when cultured in vitro combined with the role of profilin promoting actin polymerization prompted us to study the role of Pfn I in cytoskeletal dynamics and lamellipodia formation.
Here, we describe a novel function of Pfn I, promoting the formation and growth of lamellipodia in the absence of growth factors. The effects are concentration and time dependent. This ability is Pfn I specific, requiring its actin-binding domain and appearing to be mediated by an interaction through the PLP binding domain, since a synthetic PLP domain stops lamellipodia induction. We also demonstrate that PTD4-PfnI effects do not appear to increase levels of Rac1 activation.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. Both bacterial expression vectors were created previously to allow the generation of recombinant proteins fused to the PTD4 (protein transduction domain 4, sequence; YARAAARQARA), a more efficient version of the transactivating transduction (TAT) domain (15)
. Two recombinant proteins containing point mutations of profilin fused with the transduction domain were also created: H133S (mutation in the polyproline binding site) and H119E (mutation in the binding site for actin, unable to bind actin). The coding sequence of both mutations was obtained by digesting CMV-pEGFP-C1-H133S vector and CMV-pEGFP-C1-H119E with BsrGI/MfeI and inserted into both the pRSETA-PTD4 and the pRSETA-mod vectors, digested with EcoRI and Acc651. Human Pfn I, H133S, and H119E mutations cDNAs were generously provided by Drs. Hitomi Mimuro and Tadaomi Takenawa, University of Tokyo, Japan.
Oligonucleotides were purchased from Sigma-Genosys (Haverhill, UK). Restriction enzymes were from New England Biolabs (Ipswich, MA, USA). All constructions were transformed into competent Escherichia coli DH5
bacteria. Prior to protein expression, positive clones were isolated, and inframe cloning was confirmed by sequencing.
Transduction proteins were added to the cell culture medium at the final concentration described in the results. A PTD4-hemagglutinin peptide (YARAAARQARA-HA) was used for control experiments. Pfn II cDNA sequence, PTD4-GP5, a polyproline transduction peptide containing three polyproline-binding domains spared by glycine residues [sequence: YARAAARQARA-(GPPPPP)3], and PTD4-HA peptide were purchased from GenScript (Piscataway, NJ, USA).
Expression and purification of proteins
Protein expression and purification was carried out as described previously (13)
. All recombinant proteins were expressed in E. coli BL21-PLys bacteria. Briefly, bacterial pellets were isolated and lysated by a freezing and thawing protocol in liquid N2 followed by sonication on ice. The soluble fraction was loaded onto a 25-ml column packed with Ni-NTA resin (Qiagen, Hilden, Germany) in purification buffer (PBS plus DNase and Protein inhibitor cocktail from Sigma (Madrid, Spain), washed with 20 and 50 mM imidazole, and eluted with 250 mM in purification buffer. Buffer exchange and concentration of eluted proteins was performed by centrifugation in Amicon Ultra-15 10000 MWCO centrifugal filters (Millipore Iberica, Madrid, Spain). Proteins were frozen in liquid N2 and stored at –80°C in 10–15% glycerol-PBS. After being thawed, proteins were kept at 4°C up to 1 wk.
Bacteria and proteins were handled following the Safety Guidance for Laboratory Personnel Working with TAT Protein Transduction Domains (13)
.
Cell lines and bovine trabecular meshwork cultures
BTM cells were cultured using a modification of the technique described by Stamer et al. (16)
. As described previously (11)
, eyes from 3- to 6-month-old cows were obtained from the local abattoir 0.5 to 2 h after killing and kept in PBS at 4°C. BTM strips were digested with 1–2 mg/ml collagenase and 0.5 mg/ml BSA (all from Sigma, Madrid, Spain) at 37°C for 2 h. After mechanical disruption, the supernatant was collected, centrifuged, resuspended, and seeded in culture flasks in Dulbecco’s modified Eagle’s medium (DMEM) containing: 10% fetal bovine serum, 100 mg/ml L-glutamine, 100 U/ml penicillin, 100 µg/ml streptomycin, and 2.5 µg/ml amphotericin-B (all from Sigma), cells reached confluence 7–10 days later. After confluence, cells were plated in serum containing media, at a density of 6 x 104 onto 12-mm-diameter glass acid-washed coverslips. Cell passages were performed using trypsin-EDTA.
Three days after seeding, prior to the beginning of the experiment, cells were serum-starved for 16 h. Recombinant proteins and drugs were added to DMEM with or without serum depending on the experimental protocol. Phalloidin oleate, jasplakinolide, ML7, and bebblistatine were purchased from Calbiochem (EMB Biosciences, San Diego, CA, USA).
Human fibroblasts (MRC-5) were kindly provided by Dr. A. Angulo [Institut d’Investigacions Biomèdiques August Pi i Sunyer (IDIBAPS), Barcelona] and normal rat kidney (NRK) cells were generously provided by Dr. F. Tebar and Dr. A. Lledó (IDIBAPS, Barcelona). All cells were grown in DMEM containing the supplements mentioned above.
Time-lapse experiments
Confocal phase-light images were acquired using a Leica true confocal scanner (TCS) laser scanning confocal microscope (Leica Microsystems, Heidelberg, Germany). Cells were mounted in a thermoregulated chamber for the duration of the experiments. Cells were illuminated at 450 nm using an x63 oil-immersion objective lens (NA 1.4) and the confocal pinhole set at 1 Airy unit. Pictures were taken every 10–15 s, commonly 10 min for control and 30–50 min after stimulation. We used kimographs generated with ImageJ software (National Institutes of Health, Bethesda, MD, USA) to quantify the dynamic and speed cell membrane spreading at a single point (17)
. Membrane spreading velocity was calculated following the cell edge contour and adjusting a linear regression to the initial slope. Ascending and descending contours of the membrane edge indicate membrane protrusions and withdrawal events, respectively; steeper angles of the slope correspond to high speed rates. Membrane ruffles are closely related to lamellipodia: when the membrane fails to spread, the lamellipodium retracts toward the cell body, and this retraction is usually accompanied by a "membrane wave" formed by actin-folding rising above the level of the lamellipodium, visualized as a dark shadow when using phase-light microscopy (17)
.
Immunocytochemistry
After treatment, cells were washed 2x with PBS, fixed with 4% paraformaldehyde in PBS for 10 min, and then washed 4x with PBS. Blocking and permeabilization was performed with the following solution: 0.1% TX-100, 2% BSA, 2% goat serum, incubating the cells for 10 min. Next, cells were incubated at room temperature with a primary antibody in blocking solution for 1 h, washed 4x with PBS, and incubated with the appropriate secondary antibody conjugated with fluorescence. Finally, cells were washed 5x with PBS and preserved with mowiol mounting solution for fluorescence microscopy or confocal imaging. The following antibodies were used: monoclonal against Pfn I was from Synaptic Systems (Goettingen, Germany). A mouse anti-T7-tag antibody from Novagen (San Diego, CA, USA) was used to detect the recombinant profilins. A polyclonal antibody against T7 tag was from Abcam (Cambridge, UK). F-actin was visualized by adding Oregon-green, Alexa 568 or Alexa-350 Phalloidin (Molecular Probes, Eugene, OR, USA).
Fluorescence images were obtained with an inverted microscope (IX70; Olympus, Tokyo, Japan) with a Polychrome IV (Till Photonics GmbH, Gräfelfing, Germany) as a source of illumination. Pictures were taken with an attached cooled CCD camera (Orca II-ER, Hammantsu, Japan). Photographs were taken from representative fields. Pictures were analyzed using Leica and Orca II software.
Membrane lamellipodia, stained with fluorescent conjugated Phalloidin, were easily identified as a thin protrusive sheet at the leading edge of the cell membrane, lacking actin stress fibers. Using immunocytochemistry techniques, we did not distinguish between membrane ruffles and lamellipodia, but only cells exhibiting clearly discernible flat lamellipodia were scored as positive. For each condition, an average of 250–300 cells from three independent experiments were analyzed and quantified. Images were collected in a random fashion. Values are expressed as means ± SE. Lamellipodia density was estimated as the number of lamellipodia divided by the total number of cells with lamellipodia.
Nuclear plasmid microinjection
Nuclear plasmid injection of the plasmids expressing Pfn I-GFP was performed with an Eppendorf 5246 microinjector. 12–16 h after cell-plating plasmids were injected at a final concentration of 100 nM. To keep a low level of profilin-GFP expression, cells were fixed and processed for inmunostraining 5 h after injection. A plasmid coding for the GFP protein (CMV-pEGFP-C1) at a concentration of 5 nM was employed as a negative control.
AKT and Rac activation
Cells lysates were collected in lysis buffer (50 mM Tris HCl, pH 7.4; 150 mM NaCl; 1% Nonidet P-40; 5 mM EDTA; 5 mM NaF; 5 mM Na3VO4; 1 mM PMSF; 5 g/ml leupeptine; and 5 g/ml aprotinin) and clarified by low-speed centrifugation. The supernatants were subjected to SDS-PAGE and Western blot analysis as described previously (18)
and probed with antibodies against AKT and Phospho-AKT (Cell Signaling Technology, Danvers, MA, USA). To quantify Rac activation levels, supernatants were incubated for 1 h with GST-Cdc42-GTP interactive binding domain (CRIB) (Upstate Biotechnology, Waltham, MA) corresponding to the p21-binding domain (residues 67–150 of human PAK-1; expressed in E. coli and bound to glutathione agarose). The samples were washed 4x with lysis buffer and centrifuged. Pellets were boiled in 2x Laemmlli buffer. Samples were separated on 10% SDS-PAGE gels, transferred to nitrocellulose paper, and probed with Rac1 antibody (clone 23A8; Upstate Biotechnology). Supernatants from samples were also run on SDS-PAGE gels and probed for total Rac1. The signal was detected using a chemoluminescent detection system (Pierce, Rockford, IL, USA). Rac1 inhibitor, NSC23766, was purchased from Calbiochem-EMD (San Diego, CA, USA).
Quantification of intracellular profilin concentration
Calculation of intracellular amounts of profilin was done by Western blot densitometry. Briefly, diluted samples from a 25-cm2 flask cell lysate were run in an SDS polyacrylamide gel. Diluted amounts of purified recombinant PTD4-PfnI, were loaded in the same wells. Both proteins, endogenous profilin, and recombinant PTD4-PfnI easily distinguished by its molecular weight, were probed with an monoclonal antibody against profilin. Optical density was measured with a digital S-FluoImager (Bio-Rad, Barcelona, Spain). Cell volume estimation was performed as follows: BTM cells were detached using an isotonic solution free of divalent cations plus 10 mM EGTA (isotonic solution in mM: 125 NaCl, 5.3 KCl, 1.3 Ca Cl2, 1.25 MgCl2, 20 HEPES, pH 7.4). Cells kept in suspension adopted a spherical shape. Mean cell diameter was directly calculated from phase-light microscopy pictures. Solutions were equilibrated by adding dextrane to match isotonic cell medium osmolarity using a Vapor pressure osmometer, VAPRO 5520 from Wescor (Logan, UT, USA).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. The recombinant protein, PTD4-PfnI, with an estimated molecular weight of 21 kDa, was purified under nondenaturalizing conditions (Supplemental Fig. S1) and validated by Western blot analysis with antibodies against the T7 tag and profilin (Fig. 1
A) (13)
. A control version of profilin, without the transduction domain, was also generated following the same protocol as the PTD4-PfnI (Fig. 1A
). Immunocytochemistry experiments showed a significant intracellular delivery as early as 10 min and a high degree of intracellular delivery was observed after 1 h of incubation
(Fig. 6)
(13)
.
|
|
|
|
|
|
Early research in Swiss 3T3 fibroblasts demonstrated that Rac1 activation by growth factors present in serum promotes polymerization of actin at the cell membrane, inducing the formation of lamellipodia and membrane ruffles (2)
. Since that initial study, this effect has been confirmed in a wide variety of cell types. Accordingly, we first investigated the ability of BTM cells to generate lamellipodia in response to FBS. After BTM cells were serum-starved for 16 h, a condition known to induce a quiescent state (Fig. 1Ba
), cells were incubated in the presence or absence of media containing 10% FBS. As described previously (19)
, serum reintroduction increases lamellipodia density. Lamellipodia, weakly stained with phalloidin, were formed along the cell periphery, where a ring of cortical actin protrudes over the stress fibers domain (Fig. 1Bb
). Interestedly, when PTD4-PfnI was tested in serum-starved BTM cells, the number of lamellipodia increased significantly despite the absence of growth factors. Several examples for different concentrations are shown in Fig. 1Ca-d
).
The time course after FBS addition shows a biphasic component, reaching a peak 5 min after the start of FBS application (lamellipodia density of 71%). This stimulatory effect was short lived, and lamellipodia density declined after 10 min. After 30 min, lamellipodia density fell to 46%, and after 120 min, 32% of the cells exhibited lamellipodia (Fig. 2
A, open circles). By contrast, in BTM cells incubated with DMEM alone, only a small population of cells (11.6±2.9%) exhibited lamellipodia. The time course of the PTD4-PfnI effect was studied at a concentration of 10 µM. The proportion of cells developing lamellipodia rose quickly to 20.6 ± 1.7% after 1 min of PTD4-PfnI addition, to 47 ± 2.0% after 2 min, reaching a steady state after 10 min (67±0.3%). In contrast to the results obtained with FBS, the PTD4-PfnI effect did not show a down-regulation process, and 70.6 ± 2.1% of the cells still exhibited lamellipodial structures after 120 min (Fig. 2A
, closed circles). Cells still presented lamellipodia 4 h (42%, n=2) and 16 h later (30%, n=2), although during these long periods of exposure to PTD4-PfnI the cytoskeleton showed clear symptoms of depolymerization and retraction, probably due to a G-actin sequestration effect (data not shown).
Remarkably, when 10% FBS was combined with 10 µM PTD4-PfnI, the time course of lamellipodia formation resembled the one with PTD4-PfnI alone. After 30 and 120 min, the amount of cells showing lamellipodia was 74.6 ± 3.0 and 67.9 ± 4.5%, respectively (Fig. 2A
, open squares). Interestingly, when no starved cells were treated with a concentration of 10 µM PTD4-PfnI, the number of cells exhibiting lamellipodia increased to a 64.7 ± 3.8% after 120 min of treatment, compared with a 42.1 ± 1.6% of cells showing lamellipodia in the continuous presence of FBS (Fig. 2C
).
These results suggest that an increase in the intracellular levels of profilin was sufficient to induce lamellipodia formation. To confirm this point, we microinjected the cells with an expression plasmid coding for pEGFP-C1-PfnI (20)
. After nuclear microinjection, the cells rapidly express the fluorescent protein (Fig. 2B
). Cells were processed for inmunostaining 5 h after microinjection to avoid a deleterious high level of profilin expression that could induce actin depolymerization. Lamellipodia quantification, showed that 62 ± 4.7% of the microinjected cells generated one or more than one lamellipodia (from a population of 145 cells from 6 different cultures). As a control, cells were microinjected with a pEGFP-C1-expressing plasmid; in these conditions, 36.8 ± 0.8% (n=75 from three different cultures) of the cells exhibited lamellipodia, a proportion similar to that obtained in the presence of FBS or after 120 min of FBS reintroduction (34.9±4.5%; Fig. 2C
).
The 10 min time point was used to study the dose-response relationship (0.03 to 30 µM) of PTD4-PfnI effects (Fig. 2D
), exhibiting an EC50 of 4.50 µM with a Hill coefficient of 0.4. The highest concentration tested, 30 µM, appeared to induce some degree of cellular retraction probably due to a sequestering effect of actin monomers by profilin (an example is shown in Fig. 1Cd
).
Recently, it has been reported that the interaction of positive charges from arginine-rich peptides, such as TAT, with membrane proteoglycans induce lamellipodia formation (21)
. A PTD4 peptide fused to a hemagglutinin (PTD4-HA) domain was assayed to test whether PTD4 membrane binding by itself was responsible for lamellipodia spreading. 10 µM PTD4-HA did not induce lamellipodia formation in serum-starved BTM cells in any of the time intervals studied. After raising the concentration to 30 µM (Fig. 1D
) only a small increase in lamellipodia density was observed (17.2±6.2 and 18.08±2.0 after 10 and 30 min, respectively; Fig. 2A
, black inverted triangles). The effect of the same concentration of PTD4-HA was studied in the continuous presence of FBS, quantification of lamellipodia density showed that the proportion of cells exhibiting lamellipodia (47.6±7.6%) was not different to the control conditions (Fig. 2C
).
To rule out that a contaminant present in the purification media or macropinocytosis by itself would be the responsible of the lamellipodia induction, a series of controls were performed. First, recombinant profilin, expressed and purified under identical conditions as PTD4-PfnI and containing the same putative contaminations (Fig. 1A, E
), was added to serum-starved BTM cells. After 10 min of incubation, the percentage of cells showing lamellipodia was similar to that found in starving conditions. Similar results were obtained after 30 min incubation (11.2±2.4 and 12.2±1.2%, 10 and 30 min, respectively; Fig. 2E
).
Second, PTD4-galactosidase, a 120-kDa protein (13)
, at 10 µM was also tested. No significant increment in the lamellipodia density was found at any of the time intervals studied (12.3±3.4 and 16.1±3.9%, 10 and 30 min, respectively; Fig. 2E
). Finally, glycerol at the maximum concentration left after resuspending the protein stock (0.3%) was also tested without any observable increase in the basal number of lamellipodia (11.4±0.56%, at 10 min; Fig. 2E
).
Western blot quantification was employed to calculate the amount of internalized PTD4-PfnI. The estimated level of PTD4-PfnI delivered was calculated at one interval time (30 min) and only one concentration (10 µM). First, the intracellular concentration of endogenous profilin in BTM cellular volume was calculated by detaching the cells and measuring their diameter. The average cell volume was 3.4 pl (range 2–13.6 pl; Supplemental Fig. S2), and the mean profilin concentration was 125 ± 12.5 µM (range 31–150 µM). Transduction peptides of the TAT family enter cells via macropinocytosis mediated by a strong interaction with membrane proteoglycans (22)
. Thus, after cells were incubated with PTD4-PfnI, trypsin treatment was used to remove the protein attached to the cell membrane surface (22)
. Because PTD4-PfnI and profilin can be easily differentiated by molecular weight (Fig. 3
; Supplemental Fig. S2), we have compared the amount of both proteins in a cell lysate with or without trypsin treatment (Fig. 3)
. The result indicates that after 30 min of incubation with 10 µM PTD4-PfnI, the percentage of recombinant protein was 
16–20% (n=2) of the total amount of profilin (Fig. 3)
. Without trypsin treatment, the proportion increased to 33–45% (n=2), probably due to the amount of protein that remains attached to the extracellular membrane. The intracellular concentration of profilin in BTM was calculated to be 125 µM; thus, the intracellular concentration of PTD4-PfnI after 30 min of incubation would reach a maximum close to 25 µM.
To examine whether profilin’s effects on lamellipodia extension are specific to BTM cells or generally to other cell types, we next investigated the effects of PTD4-PfnI transduction in cell lines from human and rat sources. The results, summarized in Table 1
, show that in MRC-5 (lung human fibroblast cell line) and NRK cell types, PTD4-PfnI induced the formation of lamellipodia, although the sensitivity in each cell type was different. In MRC5 fibroblasts, PTD4-PfnI (3 µM) lamellipodia induction was significantly higher than in FBS, with a maximum formation of lamellipodia after 10 minutes (61%).
|
PTD4-PfnI at 3 and 10 µM produced a profound effect on NRK cells, which showed evident depolymerization of their cytoskeleton and cellular detachment. At a lower concentration (1 µM) PTD4-PfnI increased lamellipodia density (30% at 10 min and 48% at 60 min); similar results were obtained with FBS induction (30 and 55% at 10 and 60 min).
These results indicate that PTD4-PfnI is able to cross the cell membrane to induce membrane lamellipodia formation in less than 2 min. This effect is specific of the construction carrying PTD4-PfnI and independent of the presence of growth factors in the culture media.
PTD4-PfnI induces multiple lamellipodium
In cultures treated with PTD4-PfnI, cells often exhibited multiple lamellipodia formation. We therefore quantified the number of lamellipodia per cell, comparing PTD4-PfnI with FBS induction (an example of a cell with multiple lamellipodia is shown in Fig. 1Cc
). The mean number of lamellipodia per cell was higher in PTD4-PfnI than in FBS treated cells (3.8 vs. 3.0 lamellipodia/cell, respectively, P<0.05). The lamellipodia distribution shifted to high mean values after PTD4-PfnI treatment, and the population of cells exhibiting multiple lamellipodia, >4, was clearly increased (Fig. 2F
).
To induce lamellipodia formation, profilin must be able to bind actin and polyprolines
To gain insight into profilin’s mode of action, we generated membrane-permeable versions of Pfn II isoform and two mutated profilins: H119E, that impairs interaction with actin, and H133S with a decreased affinity for polyprolines but with normal affinity for actin (20
, 23)
. Purification of both recombinant proteins was confirmed using antibodies against profilin or the T7 tag sequence (Fig. 4
A). Addition of PTD4-H119E on serum-starved cells induced hardly any lamellipodia formation (18 and 19% after 10 and 30 min; closed circles in Fig. 4B
). The observation that PTD4-H119E incubation for 120 min reduced lamellipodia density below the basal level led us to think that after a long period of incubation, this mutated protein could be affecting actin polymerization by competing with endogenous profilin. We therefore tested the ability of H199E to displace endogenous profilin and inhibit FBS lamellipodia induction. Serum-starved cells were incubated with 10 µM PTD4-H119E for 120 min before the addition of FBS, and lamellipodia density was quantified after 10 min. In these conditions, lamellipodia formation was completely inhibited, even below the basal level in serum-starved conditions (Fig. 4C
).
Cells treated with 10 µM PTD4-H133S still induced some lamellipodia formation, but in lower percentages than those treated with PTD4-PfnI. Time course analysis revealed a lamellipodia density from 35% at 10 min to 24% at 120 min (open circles in Fig. 4B
). Since profilin is thought to be recruited via binding to polyproline-rich motifs (6)
, the reduced ability of H133S to induce lamellipodia suggests that this domain is involved in the targeting of profilin, although the partial induction of lamellipodia density encountered with H133S is quite puzzling. To study further the role of the PLP-binding domain on profilin effects, we used a PTD4 fused to a peptide sequence comprising three repeats of the polyproline core from VASP; The [PTD4-(GP5)3] (24)
. Overexpression of this sequence has previously been used to inhibit Listeria monocytogenes motility and the activity-dependent translocation of Pfn II to the head of hippocampal spines (25)
. Mixing both, 100 µM PTD4-(GP5)3 and 10 µM PTD4-PfnI, reduced the percentage of cells with lamellipodia to 17.5 ± 1.5% (n=9) after 10 min of the application (Fig. 4C
). To exclude any interference in the transduction process due to the mixing of two charged proteins, cells were preincubated with PTD4-(GP5)3 for 15 min before the addition of PTD4-PfnI. Using this protocol, lamellipodia density was similar to that in the previous experiment when both proteins were added together (16.3%; n=2) . When the H133S was tested in combination with the PTD4-(GP5)3, the lamellipodia density was reduced to a similar extent (19.9±3.1%; Fig. 4D
).
If the PLP-binding domain is crucial for PfnI function, we would expect an inhibitory effect on FBS-induced lamellipodia due to an inhibition of the endogenous profilin. To test this hypothesis, serum-starved BTM cells were preincubated for 15 min with 100 µM of PTD4-(GP5)3, the subsequent addition of FBS increased lamellipodia density only to 18.7 ± 1.5% (Fig. 4C
).
Human Pfn I and Pfn II have similar biochemical properties, except that the affinity of human Pfn I for actin is higher than that of Pfn II. To determine whether the two isoforms have similar functions in the promotion of lamellipodia extension, we generated a transduction version of Pfn II (PTD4-PfnII; Fig. 4A
). Again, the expression and purification were confirmed by Western-blot analysis using an antibody against the T7 tag present in the recombinant protein. The addition of 10 µM PTD4-PfnII to BTM serum-starved cells induced only a small increase in lamellipodia density, (25±1.2% after 10 min). The addition of Pfn II, without the transduction domain, resulted in a lamellipodia density similar to that of starving conditions (14.3±1.4%; Fig. 4D
). In summary, these results suggest that lamellipodia induction is specific to the Pfn I isoform, and both the PLP and the actin-binding domains are crucial in this process.
Effect of different inhibitors of actin dynamic on lamellipodia formation
To broaden our understanding of the PTD4-PfnI-induced lamellipodia formation, several different well-known inhibitors of the cytoskeleton dynamics were assayed. It is known that lamellipodia extension requires rapid actin polymerization and continuous F-actin turnover and that phalloidin oleate (PO), a cell membrane-permeable phalloidin, inhibits F-actin conversion to G-actin and stabilizes the actin filament. In BTM cells treated with 10 µM PO, lamellipodia induction by FBS was strongly inhibited, with only 6.6 ± 0.5% of the cells exhibiting lamellipodia. Interestingly, PTD4-PfnI effects were less sensitive to PO inhibition, inducing lamellipodia formation in 27.8 ± 1.4% of the cells. Jasplakinolide, a compound that promotes net actin polymerization, blocks lamellipodia protrusion in chick fibroblasts when used at a concentration of 1 µM (26)
. We tested the effects of Jasplakinolide in BTM cells on lamellipodia formation at a concentration of 0.2 µM. After 30 min incubation, cells showed an abnormal actin distribution with no clear actin stress fibers and numerous clusters of aggregated actin in the cell body cytoplasm. In these conditions, lamellipodia induction was stopped completely both in the presence of PTD4-PfnI or FBS (data not shown).
5-Iodonaphthalene-1-sulfonyl homopiperazine (ML7), a myosin light-chain kinase inhibitor, suppresses EGF-induced lamellipodia and membrane ruffling in Cos1 cells (27)
. To determine whether actin-myosin-based contractility is required for the PTD4-PfnI lamellipodia induction, we tested the effects of ML7 at a concentration of 50 µM. In these conditions, FBS- or PTD4-PfnI-induced lamellipodia formation was completely inhibited without any visible perturbation in cell shape or in the actin cytoskeleton. Blebbistatin (5 µM), a membrane-permeable inhibitor of myosin II ATPase activity (28)
also blocked lamellipodia induction by both FBS and PTD4-PfnI, although, in contrast to the effect of ML7, blebbistatin visibly damaged the BTM cell cytoskeleton (data not shown), making any interpretation of the results extremely complicated.
Membrane dynamics after PTD4-PfnI treatment
To improve our understanding of the lamellipodia extension by PTD4-PfnI, time-lapse microscopy was used to analyze real-time membrane dynamics. In serum-starved conditions, BTM cells are not motile. In these conditions, the peripheral regions of the cellular edge were highly dynamic, exhibiting a continuous cycle of protrusion and withdrawal and membrane ruffling (we called these areas "ruffle areas") (17
, 29
, 30)
. Membrane ruffles are closely related to lamellipodia: when the membrane fails to spread, the lamellipodium retracts toward the cell body and this retraction is usually accompanied by a "membrane wave" formed by actin-folding rising above the level of the lamellipodium, visualized as a dark shadow when using phase-light microscopy (17)
.
The addition of PTD4-PfnI (10 µM) produced a lamellipodia extension from either ruffle area or preformed lamellipodia. Figure 5
A shows a growing lamellipodium from a preexisting one (top) and a lamellipodium formation that expands from a ruffle area (bottom; see Supplemental Videos 1 and 2). Kymographic analysis of the membrane revealed that shortly after PTD4-PfnI addition (black arrow), the membrane initiates a process of extension at a constant speed. Membrane extension spreading decreased gradually until finally stopping, usually after 40–50 min (the maximum time recorded was 60 min) with a mean velocity obtained from the linear regression of the initial slopes of 0.53 ± 0.10 µm/min (n=11; Fig. 5C
). No statistical differences were found between the membrane spreading originated from a ruffle area or from a lamellipodium. In both situations, periodic extensions and retractions of the leading edge of the cell membrane were observed, reflected on the kymograph as ascending and descending profiles (sharp peaks in Fig. 5A
, right panels). These membrane extensions profiles were similar to the anisotropic spreading previously described (29
, 31)
.
Serum reintroduction showed a more complex behavior than PTD4-PfnI. Two types of membrane growth were observed, characterized by different membrane profiles at the leading edge. In some cases after FBS addition, the cell withdrew, probably because of the retraction of the stress fibers mediated by Rho activation. Afterward, the cell started a process of lamellipodia extension followed by cell movement (Fig. 5Ba
, top). The spreading of a new lamellipodium was characterized by an exponential membrane extension and a lack of centripetal waves, followed by a steady state in which the extension movement stopped, similar to the isotropic spreading described by Dubin-Thaler et al. (31)
. An example of this type of membrane extension is shown in Fig. 5B
top (see also Supplemental Video 3). This type of membrane movement was infrequent; less than 20% of the cells examined presented this sort of membrane spreading. The mean growth velocity of this type of spreading was 2.07 ± 1.2 µm/min (n=6).
The second type of membrane growth, which occurred more frequently (in approximately 80% of the cells), was the anisotropic profile characterized by the periodic extension and retraction of the cellular edge (Fig. 5Ba
, bottom). This growing type was more frequent than the isotropic type, and it originated in a ruffle area (Supplemental Video 4). In this case, the membrane slowly grows, with a mean velocity of 0.29 ± 0.04 µm/min (n=8), significantly different from the speed rate induced by PTD4-PfnI (Fig. 5C
).
A notable difference between the effects of profilin and FBS was the induction of cell motility. Addition of FBS started the whole process of cell spreading and motility. In contrast, the addition of PTD4-PfnI was not sufficient to generate cell movement by itself, despite the persistence of lamellipodia.
The kinetics of the periodic extensions and retractions were also analyzed (Fig. 5D
). Protrusion extension frequency was estimated as the time between two consecutive peaks on the kymograph. Periodicity between lamellipodia extensions was not statistically different in control compared with PTD4-PfnI or FBS stimulation (Fig. 5D
). Lamellipodia protrusion velocity, the ascending part of the peaks, was 2.19 ± 0.25 µm/min in control conditions and 2.17 ± 0.16 and 2.81 ± 0.5 µm/min, after FBS and PTD4-PfnI addition, respectively, although the mean value obtained was slightly higher in the presence of PTD4-PfnI, the mean was not statistically different when it was compared with the control or FBS group (Fig. 5D
). Retraction membrane velocity, the descending part of the peak, was also not statistically different among the different situations (Fig. 5D
). Membrane ruffles, rearward actin waves, were often visible, although at low frequency, as black shadows that move centripetally to the main cell body (17)
. In control conditions, the velocity was of a 1.47 µm/min, compared with 1.42 ± 0.13 in the presence of PTD4-Pfn I, in contrast in FBS conditions, ruffles velocity was significantly slower, with a mean velocity of 0.93 ± 0.12; similar changes after serum stimulation were previously reported in Giannone et al. (29)
.
Membrane dynamics was also studied after the addition of 10 µM PTD4-H119E and after mixing PTD4-PfnI (10 µM) with 100 µM of the PTD4-(GP5)3 peptide (n=2). Neither of these conditions induced the extension of new lamellipodia. Interestingly, the curling movement was not stopped by the PTD4-(GP5)3 peptide transduction or by addition of H199E (data not shown).
Intracellular distribution of PTD4-PfnI
PTD4-PfnI intracellular distribution was studied by immunocytochemistry using antibodies against the T7 tag sequence present in the protein. In spreading cells (Fig. 6
A), PTD4-PfnI (3 µM for 10 min) showed staining in punctae concentrated mainly in perinuclear localizations likely at the Golgi membranes (32)
and inside the lamellipodia domain (Fig. 6A
), always located between the leading edge of the lamellipodia and the stress fiber domain (see detail in Fig. 6Ae
), localization consistent with the process of macropynocytic internalization. Endogenous Pfn I, detected using a monoclonal antibody, showed a homogeneous distribution throughout the cell and an evident colocalization with the cortical cytoskeleton at the leading edge (Fig. 6Ba-c
) (33
, 34)
. Higher magnification revealed that profilin is distributed in fine puncta or vesicular networks clearly visible in the lamellipodia (Fig. 6Ba-d
). To increase the amount of PTD4-PfnI delivery, longer incubations were employed (0.3 µM for 120 min). In this conditions, double inmunostainnig for PTD4-PfnI and endogenous profilin revealed that a large amount of vesicular-like PTD4-PfnI was present into the cell, showing a high degree of colocalization with the endogenous profilin (Fig. 6Ca-e
), suggesting that PTD4-PfnI escape the macropynocytic vesicle and diffuse into the cytoplasm. Especially with long incubations, PTD4-PfnI was also found at the cell nucleus (Fig. 6Cb
) corroborating previous reports of a role for profilin in the exportin6-dependent transport pathway (35)
.
Profilin induces lamellipodia extension independently of Rac activation
PI3K activation in response to growth factors leads to a polarized accumulation of phosphatidylinositol 3,4,5-triphosphate (PIP3), which, in turn, stimulates small GTPases of the Rho family, such as Rac and Cdc42, inducing filopodia and lamella formation (36
37
38)
. It has been reported that PfnI may activate PI3K in vitro by direct binding to the p85 regulatory subunit (39
, 40)
. To rule out a possible activation of PI3K by PTD4-PfnI in the lamellipodia extension process, we carried out a series of experiments quantifying AKT phosphorylation levels, a kinase located downstream of PI3K (41)
. As shown in Fig. 7
A, the phosphorylation level of AKT at 10, 30, and 60 min increased after FBS stimulation. In contrast, PTD4-PfnI did not show any significant change in AKT phosphorylation levels at the same time points.
|
Rac1 activation plays a central role in the process of lamellipodia (2)
. Next, we investigated whether PTD4-PfnI treatment induces Rac1 activation. Quantification of active Rac-GTP levels was performed by using an affinity-precipitation assay with a GST-tagged CRIB (42
, 43)
. Rac1 activation was quantified from subconfluent cellular cultures starved for 16 h and subsequently stimulated with PTD4-PfnI or FBS. As shown in Fig. 7B
, Rac-GTP was pulled down at different time points: 0, 10, 30 and 120 min. No increase in the amount of Rac-GTP was observed following addition of PTD4-PfnI (n=5); in fact, the amount of Rac-GTP remained low after long incubation times, and in some experiments, even falling below the starving level. In contrast, FBS induced a clear activation showing that Rac activity was nearly maximal at 10 min (200%) decreasing toward basal levels within 120 min (Fig. 7C
). Despite the lack of Rac activation after PTD4-PfnI treatment, Rac should be functionally active for lamellipodia formation since NSC23766, a Rac1 inhibitor at 150 µM (44)
, impairs lamellipodia formation after both PTD4-PfnI or FBS induction (Fig. 7D
).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The protein transduction technique we have used extensively in this study presents certain advantages over classical transfection techniques. For instance, transfection is very inefficient when used in primary cultured cells such as BTM cells. Using a protein transduction method, which is independent of cell type (45)
, we circumvent the problems of transfection or gene overexpression, additionally, the procedure is straightforward, allowing a control of the levels of expression and the time of induction.
The ability of PTD4-PfnI stimulating lamellipodia spreading was studied in serum-starved conditions and in the presence of growth factors. To confirm that the effects described are due to profilin and not related to unexpected effects of the membrane-permeable profilin construct, a series of microinjection experiments using a plasmid-expressing GFP-Pfn I were done. Short time expression of profilin-GFP induced lamellipodia formation in a similar manner than PTD4-PfnI transduction.
In the light of the reported induction of lamellipodia by arginine peptides (21)
, care was taken to establish that effects observed in lamellipodia extension were specific of PTD4-PfnI and not of the transduction itself. In our cells, PTD4-HA only at concentration of 30 µM was able to induce lamellipodia but in a reduced proportions (only 18% compared with a 70% for PTD4-PfnI). PTD4-βGal, a larger protein fused to the PTD4 was also inefficient, stimulating lamellipodia induction. In addition, profilin mutation studies provided further confirmation of their specificity. The mutation in the actin-binding domain (H119E; 23
, 46
) was unable to induce lamellipodia by itself and inhibited their formation after FBS stimulation. Additionally both H133S (20)
and PTD4-PfnII showed a reduced ability to induce lamellipodia. Taken together, the data indicate that the stimulatory effects on lamellipodia formation were in fact induced by PTD4-PfnI.
Internalization by a lipid raft-dependent macropinocytosis is the major pathways for TAT-fusion proteins and arginine-rich peptides uptake (21
, 22
, 47)
. As a prerequisite for internalization, TAT must bind strongly to cell-surface heparan sulfates proteoglycans. Indeed, a common feature of transduction domains is the presence of a high density of basic amino acid residues, arginines, or lysines. The PTD4 domain was designed changing three positive charged arginines residues by neutral alanines, reducing the overall positive charge of the domain and forcing an alpha helix structure for PTD (15)
. Although studies of internalization process of PTD4 have not yet performed, the reduced number of positive charges will not favor the interaction with membrane heparan sulfates (47)
. We can speculate that PTD4 used a macropynocytic process to enter the cell similarly to the TAT domain. Further work will be needed to clarify the transduction mechanism of the PTD4.
Incubation of cells with PTD4-PfnI for as little as 1–2 min was enough to induce lamellipodia spreading. Similar times of protein uptake were found using a TAT-Cre-mediated recombination domain as a transduction reporter; incubation for only 5 min was sufficient to cross the extracellular and nuclear membranes and induce recombination (22)
. Our results confirm that protein transduction is a rapid and efficient process.
Interestingly, variations in the cytoplasmatic concentration of profilin appear critical for its function. PC12 cell differentiation and neuritogenesis is stimulated by moderate Pfn I overexpression (2–3x above basal level), even in the absence of NGF (48)
. In contrast, higher levels of profilin expression (30–86x above basal levels) reduce PC12 differentiation, probably because at these high concentrations, Pfn I acts mainly as a sequestering agent, preventing actin polymerization (48)
. In other cell types, Pfn I overexpression alters actin cytoskeleton distribution and thick stress fibers tend to disappear (23
, 49
, 50)
. In our experiments, PTD4-PfnI induced fiber retraction and alterations in the actin cytoskeleton only after long incubation times (16 h) or at high concentrations (30 µM). Intracellular concentration of Pfn I has been estimated in platelets (55 µM; ref. 51
), aortic endothelial cells (10. 6 µM; ref. 50
), and embryonic chick brain (5–6 µM; ref. 52
). In BTM cells, Pfn I concentration varies between 30 and 150 µM, a range of concentrations in agreement with previous data. In contrast, the amount of transduced protein is difficult to estimate. When transduction techniques are employed, most of the protein trapped by macropinocytosis ends in the endosomal compartment where it will be degraded. However, macropynocytic vesicles are inherently leaky, and some of their contents is released to the cytoplasm (53)
. Our estimation of the intracellular delivery shows that after removal of the extracellular protein, about 20% of the total profilin was exogenous. Assuming that this amount of PTD4-PfnI is distributed homogenously through the entire cytoplasm, PTD4-PfnI will reach an intracellular concentration between 6 and 25 µM, 5x below the estimated concentration of endogenous profilin. In fact, the real concentration of delivered protein will be below these levels since part of the protein remains trapped in macropinocytic vesicles.
Profilin binding domains are involved in the mechanism of lamellipodia induction. Longer time incubations with the mutation in the actin-binding domain (PTD4-H119E) prevents lamellipodia spreading; probably, the intracellular accumulation of this mutated form will displace endogenous profilin inhibiting actin polymerization. In other cell types, overexpression of this mutated form of profilin also inhibited Rac1-dependent ruffle formation in 3T3 cells (46)
and blocked intracellular spreading of Shigella (20)
.
As previously demonstrated (20
, 46
, 54)
, the PLP-binding domain participates in the targeting and function of profilin since the H133S mutation shows a limited ability to induce lamellipodia. This mutated form of profilin was still able to induce some lamellipodia formation, indicating that either the mutation does not totally impair PLP binding or the domain is not critical for PTD4-PfnI lamellipodia induction. Many proteins containing proline-rich motifs (PRMs) are involved in regulating actin polymerization. The type 1 PRM domain, with a sequence XPPPPP (where X can be a G, L, I S or A), binds profilin to VASP, WASP/WAVE or Ezrim (for a review, see ref. 55
). This sequence has been previously used to inhibit actin-dependent motility of Listeria monocytogenes by blocking the binding of Pfn I to VASP (24)
. Overexpression of the same construction was used to inhibit activity-dependent Pfn II translocation to the head of the hippocampal spines (25)
. Our results showing that PTD4-(GP5)3 peptide inhibits the lamellipodia induced by both FBS and PTD4-PfnI confirm the efficiency of the (GP5)3 sequence in interfering with the profilin function and the importance of a functional PLP domain, although the ubiquitous presence of PRM domains in other proteins (SH3 in diaphanous (Dia); WW in formin or EVH1 in vinculin) cannot rule out other inhibitory effects of (GP5)3 peptide due to the high concentration used (55)
.
Endogenous profilin showed a homogenous distribution throughout the entire cell and was clearly found in lamellipodia/ruffles in spreading cells with a pattern in puncta. The puncta appeared in long rows or reticular networks that extends toward the leading edge of the lamellipodia. A similar distribution has been reported previously: profilin appears enriched in areas of membrane ruffling (46
, 56)
. In more detail, in PtK2 and BHK-21 cells, profilin was enriched at peripheral areas of ruffling activity distributed in fine vesicular structures (punctae) or forming a reticular network identified as the network of profilin:actin (33
, 34
, 57)
. The recombinant protein, soon after the addition, was found in vesicle-like structures at the lamellipodia membrane, probably associated with the process of macropynocytosis.
After longer incubation times, a large accumulation of the recombinant protein was found at the lamellipodia leading edge and associated with the reticular network of endogenous profilin, confirming that part of the transduced protein was released to the cytoplasm.
An important difference between FBS and PTD4-PfnI was that the latter only induced lamellipodia formation from ruffle areas or preexisting lamellipodia. Even after long starving periods (16 h), BTM cells still showed areas of ruffling activity at the cellular rim, where Rac1-GTP may be localized (37
, 58
, 59)
. In these areas, Rac1-GTP may participate together with the WAVE1/2-Arp2/3 complex regulating a continuous cytoskeletal polymerization and membrane extension. In serum-starved cells, reintroduction of FBS containing growth factors promotes a broad Rac1 activation, which in turn induces the formation of the WAVE1-Arp2/3 complex, catalyzing the formation and growth of de novo lamellipodia (60)
. The lack of de novo lamellipodia formation in PTD4-PfnI-stimulated cells suggests a role for profilin downstream from Rac1 activation. The absence of Rac1 and PI3K activation found after PTD4-PfnI addition supports this idea. Another result that supports the role of profilin downstream of Rac1 activation was provided by the pharmacological inhibition of Rac1, which inhibits PTD4-PfnI lamellipodia induction. This suggests that a certain level of active Rac1 is required to initiate membrane spreading triggered by PTD4-PfnI. In this regard, membrane spreading from ruffle areas suggests that the molecular machinery required (WAVE and Arp2/3 complex) must be present and active to account for the PTD4-PfnI effects.
Analysis of lamellipodia kinetics induced by FBS or PTD4-PfnI showed different membrane extension velocity and growth patterns. Isotropic and anisotropic membrane extension profiles have previously been described in mouse fibroblasts during cell spreading after plating (29)
. The isotropic profile, more frequent in serum-deprived cells, is characterized by a high rate of area increase and a low rate of centripetal actin waves. In contrast, anisotropic growing is characterized by reduced lamellipodia area spreading, stochastic and transient periods of membrane extension (31)
. Both profiles were observed in starved BTM cells after serum reintroduction, but only the anisotropic type was encountered after addition of PTD4-PfnI. Anisotropic membrane spreading was faster in PTD4-PfnI stimulated cell compared to FBS (0.53 vs. 0.29 µm/min), without any change in the other kinetic parameters. Both speeds are lower than others reported in the literature: 7 µm/min in fish keratinocytes (17)
, 3–6 µm/min in rodent fibroblasts (30
, 61)
, probably due to differences in cell type or culture substrate.
Membrane forward movement is accomplished by the incorporation of new actin monomers to the barbed end of actin filaments, which are always close to the inner side of the cell membrane. The growing process requires a rapid and coordinated depolymerization-polymerization of preexisting actin filaments. We hypothesize that the amount of PTD4-PfnI transduced would cause a local increase on profilin concentration just beneath the cell membrane close to the actin barbed ends. This recombinant profilin would bind ATP-actin accelerating actin turnover and increasing membrane spreading velocity. The fast membrane spreading and the PTD4-profilin localization support this model. In addition, it is possible that the presence of PTD4 domain favors the tethering and accumulation of profilin with the cellular membrane, in close vicinity of actin barbed ends enhancing actin polymerization and membrane spreading; in fact, PTD4-PfnI shows a great accumulation at the leading edge of the lamellipodia. In this sense, it has been demonstrated that profilin binding to PI(4,5)P2 is important for the protein targeting to the membrane (48
, 62
63
64)
Further work using profilin mutated in the lipid binding domain profilin will be in order to understand the mechanisms favoring lamellipodia spreading by profilin.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Received for publication November 21, 2006. Accepted for publication November 29, 2007.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  A. Chatterjee, L. Wang, D. L. Armstrong, and S. Rossie Activated Rac1 GTPase Translocates Protein Phosphatase 5 to the Cell Membrane and Stimulates Phosphatase Activity in Vitro J. Biol. Chem., February 5, 2010; 285(6): 3872 - 3882. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. Syriani, G. Cuesto, E. Abad, T. Pelaez, A. Gual, J. Pintor, M. Morales, and X. Gasull Effects of Platelet-Derived Growth Factor on Aqueous Humor Dynamics Invest. Ophthalmol. Vis. Sci., August 1, 2009; 50(8): 3833 - 3839. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
4 lamellipodia and accordingly reduced the population of cells with 
3 lamellipodia. Notice the presence of cells with 10, 12, and more lamellipodia in the presence of PTD4-PfnI.
