Collagen phagocytosis is regulated by the guanine nucleotide exchange factor Vav2

Abstract

Collagen phagocytosis is a crucial α2β1-integrin-dependent process that mediates extracellular matrix remodeling by fibroblasts. We showed previously that after initial contact with collagen, activated Rac1 accelerates collagen phagocytosis but the Rac guanine nucleotide exchange factors (GEFs) that regulate Rac are not defined. We examined here the GEFs that regulate collagen phagocytosis in mouse fibroblasts. Collagen binding enhanced Rac1 activity (5–20 min) but not Cdc42 or RhoA activity. Analysis of collagen bead-associated proteins showed enrichment with Vav2, which correlated temporally with increased Rac1 activity. Knockdown of Vav2 prevented Rac activation, recruitment of Rac1 to collagen bead binding sites, and collagen bead binding, but knockdown of Sos-1 or β-Pix had no effect on Rac activation or collagen binding. Vav2 was associated with the nucleotide-free Rac1 mutant (G15ARac1) after collagen binding. Collagen bead binding promoted phosphorylation of Vav2, which temporally correlated with Rac1 activation and which required Src kinase activity. Blockage of Src activity prevented collagen bead-induced Rac activation and collagen bead binding. Collectively these data indicate that Vav2 regulates the Rac1 activity associated with the binding step of collagen phagocytosis.

in health, collagen synthesis and degradation are balanced to preserve tissue homeostasis of soft connective tissues. Fibroblasts are the predominant cells of soft connective tissues and mediate matrix remodeling, which is dependent in part on intracellular collagen degradation by phagocytosis (14). Collagen phagocytosis is an important pathway for collagen turnover, and disruption of the intracellular collagen degradation pathway leads to imbalances in matrix homeostasis resulting in tissue overgrowth and fibrosis.

Cell adhesion and migration involve tightly controlled regulation of actin cytoskeletal proteins. Many of the processes involved in regulation of the actin cytoskeleton are also shared by phagocytosis, a receptor-driven process. In response to a diverse group of extracellular signals, small GTPases are activated to co-ordinate actin assembly and integrin activation (13, 24), and this has also been observed in phagocytosis (10, 11, 26). As the initial events in phagocytosis may be considered as a specialized form of cell adhesion (11), studies of actin- and integrin-dependent collagen phagocytosis and its associated regulatory systems can be facilitated by models that use loading of fibroblasts on collagen-coated latex beads (3). These models recapitulate some of the processes observed in the spreading of cells on tissue culture plates and show that alteration of integrin-mediated cell adhesion to the extracellular matrix is a critical determinant of collagen phagocytosis (2). We have shown previously that within 5 min of cell binding to collagen beads through α2β1-integrins, there is greatly accelerated bead binding and Rac activation (4). Similarly, in leukocytes, CD31 stimulates an integrin-dependent amplification of adhesion (22). However, the mechanisms that mediate this amplification and result in increased integrin affinity and collagen binding are not defined.

One of the consequences of signaling arising from integrin-mediated adhesion and cell spreading is the activation of Rac and Cdc42. These small Rho family GTPases in turn regulate extension of lamellipodia and filopodia, processes that involve complex and dynamic rearrangements of the actin cytoskeleton. The recruitment and function of Rho GTPases, specifically at sites of cell attachment to the extracellular matrix, are crucial for phagocytosis and spreading (10, 17, 24). Localized Rac1 activation at particle attachment sites triggers particle internalization, a process that requires rearrangement of the actin cytoskeleton leading to actin assembly and membrane protrusion (8). Integrin-mediated cell adhesion may affect interactions of GTP-Rac with its effectors by controlling the membrane targeting of GTP-Rac (13). Currently, the regulation of integrin affinity after initial particle attachment, which leads to Rac1 activation and accelerated cell adhesion, are not well understood.

Regulation of GTPases involves transition between GDP/GTP-bound activation states that are regulated by several systems, including guanine nucleotide exchange factors (GEFs) and activating proteins (GAPs). The Vav family of GEFs activates Rho GTPases (Rho, Rac, Cdc42) by catalyzing the exchange of GDP for GTP. In mammals the Vav family of proteins consists of three known members (Vav-1,-2, and -3) that differ in their tissue distribution. Vav1 is expressed predominantly in hematopoietic cells, whereas Vav2 and Vav3 are more broadly expressed. Vav phosphorylation is functionally linked to Vav activation, and in vitro studies with purified proteins demonstrate that Vav phosphorylation is the critical regulatory event triggering Vav activation and Rac activation (12). Of the known GEFs that control the activation state of Rac, Vav2 is considered important for FcγR but not CR3 (integrin)-mediated phagocytosis in macrophages and COS cells (23). Whereas in fibroblasts Vav2 is necessary for integrin, but not growth factor-dependent activation of Rac leading to lamellipodia formation (20), Vav2 is not apparently involved in activation of Rac in response to β1-integrin ligation (18).

We have examined here the role of Rac1 GEFs in the binding step of collagen phagocytosis. The data show that collagen bead binding promotes phosphorylation of Vav2, which temporally correlates with Rac1 activation and requires Src kinase activity.

MATERIALS AND METHODS

Reagents.

Latex (2 μm diameter) beads were purchased from Polysciences (Warrington, PA). Antibodies to β-actin (clone AC-15), fluorescein isothiocyanate (FITC)-conjugated goat anti-mouse (GAM) antibody, and tetra-methyl rhodamine isothiocyanate (TRITC)-phalloidin were from Sigma (St. Louis, MO). Blocking antibody for α2β1 (clone no. BMA2.1) was from Millipore. T7 antibody was from Novagen (Madison, WI), and Vav2 (clone MAC 410) antibody was from Babraham Bioscience Technology. FITC-goat anti-rat antibody was purchased from Cedarlane Laboratories (Hornby, ON). Antibody to phosphorylated Vav2 (Tyr172) was from Santa Cruz Biotechnology. Antibody to Rac (clone 23A8) was obtained from Upstate Biotechnology. Rhotekin-RBD beads and RhoA antibody were purchased from Cytoskeleton (Denver, CO). Immobilized Protein G was obtained from Pierce (Rockford, IL).

Cells.

NIH 3T3 cells were cultured in DMEM (GIBCO BRL) supplemented with 10% fetal calf serum and 10% antibiotics. Jurkat cells were obtained from ATCC.

Collagen bead binding.

Collagen-coated latex beads (2 μm) were applied to nontissue culture dishes and dried down for attachment as described previously (3) followed by being washed with PBS. The number of beads plated per dish was adjusted to produce final bead-to-cell ratios specific for each experiment. Cells were counted electronically, and the cell concentration was adjusted before cells were plated on dishes containing collagen-coated beads. The plates were maintained at room temperature for 10 min to allow the cells to settle and attach to the collagen beads, and subsequently, the cells were washed with fresh medium at 37°C. Detached cells were removed by repeated washes. Those cells that were attached, spread and rapidly internalized the collagen beads (5).

Isolation of bead attached proteins and immunoprecipitation.

In some experiments to examine the proteins that were found at the phagocytic site, collagen-coated latex beads (6 μm) were attached to 100-mm nontissue culture plastic dishes. Cell suspensions were allowed to attach to beads for 20 min. Unattached, floating cells were aspirated and replaced with media warmed to 37°C to synchronize phagocytosis. Cells were collected at discrete time points thereafter. Cells and collagen-coated latex beads were collected with a cell scraper in extraction buffer (1% Triton X-100, 150 mM NaCl, 10 mM Tris·HCl at pH 7.2, 1 mM Na₃VO₄, 20 μg/ml aprotonin, and 1 μg/ml Pefabloc). The samples were sonicated (2 s) and centrifuged for 5 min at 8,000 g to remove unbroken cells. After clarification, equal amounts of proteins were incubated with antibodies to T7 or Vav2 to form immunocomplexes that were captured on Sepharose-G beads (Pierce Biotechnology, Rockford, IL) for 1 h at 4°C. The samples were boiled and separated on SDS-PAGE gels. Immunoblotted samples were probed with appropriate antibodies and quantified by scanning densitometry.

Immunofluorescence, confocal microscopy, and SEM.

After transfection (24 h) with different constructs, cells were plated on beads and were allowed to spread and bind to beads for 30 min. In some experiments cells were fixed with 3% formaldehyde in PBS, permeabilized with 0.2% Triton-X100, and stained with appropriate antibodies. The spatial distribution of staining around beads was determined by confocal microscopy (×40 oil immersion lens; Leica, Heidelberg, Germany). Transverse optical sections were obtained at 1-μm nominal thickness. For morphological assessments of cellular interactions with beads, cells were fixed in 4% formaldehyde, dehydrated through an ethanol series, and critical point dried in a Polaron CPD7501. Samples were mounted on aluminum stubs, and plasma was sprayed with a 5-nm thick coat of platinum in a Polaron SC 515 scanning electron microscopy coating system and examined with a Hitachi C-2500 scanning electron microscope.

Plasmids and siRNA transfections.

Green fluorescence protein (GFP) Vav2 and T7 tagged Vav2 were generated by one of us (P. Marignani). The Rac1 nucleotide-free mutant G15ARac1 was kindly provided by K. Burridge (University of North Carolina). Constitutively activated GFP Rac was obtained from A. Kapus (St. Michael's Hospital Research Institute, Toronto, Canada). For small interfering RNA (siRNA) knockdown, cells (10⁶) on tissue culture plates (100 mm) were transfected with 5–20 pmol of Smart Pool siRNA oligonucleotides specific to Vav2, Sos1, and β-PIX (Dharmacon) using Oligofectamine (Invitrogen) without antibiotics. To estimate transfection efficiency and to serve as a transfection control, pEGFPluc (Clontech) was used, and the number of fluorescent cells were counted.

Rac, Cdc42, and RhoA activation.

To detect Rac, Rho, and cdc42 activity, cells were serum starved and lysed. Cell lysates were collected in lysis buffer with protease inhibitors (Upstate) and clarified by low-speed centrifugation, and supernatants were incubated with a glutathione-S-transferase (GST) fusion-protein (Upstate) corresponding to the p21-binding domain (PBD), residues 67–150 of human PAK-1; expressed in Escherichia coli and bound to glutathione agarose. The samples were washed four times with wash buffer after an hour of incubation at 4°C. Pellets were boiled in 2× Laemlli buffer. Samples were separated on 10% SDS PAGE gels, transferred to nitrocellulose paper, and probed with Rac antibody (clone no. 23A8; Upstate) or cdc42 antibody (Upstate). Similarly, Rho activation assays were performed with Rhotekin beads (Cytoskeleton) and blotted with RhoA antibody. Supernatants from samples were also run on SDS-PAGE gels and immunoblotted for total Rac, cdc42, and Rho proteins. Samples stimulated with lysophosphatidic acid (LPA) were used as a positive control to detect active Rho.

Statistical analyses.

For continuous variables, means and SE were computed, and differences between groups were evaluated by Student's unpaired t-test or ANOVA for multiple comparisons with statistical significance set at P < 0.05. Post hoc comparisons were performed with Tukey's test. For all experiments, at least three independent experiments were evaluated, each performed in triplicate.

RESULTS

Role of Rac in collagen bead binding.

Rac activity regulates growing lamellae, mediates cell spreading (24), and greatly increases collagen phagocytosis (4). To study the role of Rac activity in the initial steps of collagen phagocytosis and to determine its regulation of the binding step, we assessed whether Rac1 localized to collagen beads after bead ligation in GFP-Rac1-transfected fibroblasts. Collagen beads, but not BSA-coated beads, showed enhanced bead-associated accumulation of Rac1 (Fig. 1A, a–d). The fluorescence intensity around the beads was compared by line scans (Fig. 1A, e and f), and background fluorescence was subtracted from the fluorescence intensity around beads. Data from 25 cells observed at each time interval and from three different experiments showed threefold higher fluorescence intensity around collagen beads compared with poly-l-lysine beads (Fig. 1A,g; P < 0.01). We established that collagen bead binding was specifically mediated by the α2β1-integrin with the use of a blocking monoclonal antibody (clone no. BMA2.1; at 0.2 mg/ml) followed by incubation with fluorescent collagen-coated beads. At bead cell ratios of 4:1, there was 85% reduction in the percentage of cells with bound beads in the antibody-treated samples. Western blot analysis of bead-associated proteins showed maximal association of active Rac1 with beads between 2 and 20 min, whereas Rho and Cdc42 did not detectably associate with beads (Fig. 1B). Similarly, in cells transfected with hemagglutinin-tagged Cdc42 and GFP RhoA, there was no significant localization to collagen-coated beads (Fig. 1C, a–d) as quantified by fluorescence intensity line scans (Fig. 1C, e,f).

Fig. 1.

A: cells incubated with collagen-coated beads (a and b) and BSA-coated beads (c and d) were fixed and immunostained for Rac1. Note localization of endogenous Rac1 to collagen-coated beads but not to BSA beads. Fluorescence intensity was measured across line scans (h,i and j,k) and shown in line graphs e,f marked by arrows at the point of intersection with beads. Histograms (g) show means and SE of fluorescence intensity above background from three experiments of 25 cells each. Bar graph shows threefold difference in fluorescence intensity around beads between collagen and BSA-coated beads (P < 0.01). B: Western blot analysis of proteins associated with collagen beads isolated at different time points show enhanced Rac1 association at 2–10 min, whereas active Rho and cdc42 were not associated. Lysophosphatidic acid-stimulated sample was used as a positive control to detect and validate Rho activation assay. C: endogenous Cdc42 (a,b) and RhoA (c,d) were immunostained in fixed cells, but there is no localization to collagen beads. Data are also presented as line graphs (e,f) of fluorescence intensity measured across line scans g,h and i,j at the bead site. D: fibroblasts transfected with constitutively active green fluorescence protein (GFP)-tagged Rac or GFP-empty vector were isolated by flow cytometry, incubated with collagen-coated beads for 30 min, and prepared for SEM. Cells expressing constitutively active Rac (GFP-RacQ61) show extended lamellae that bind to a large number of collagen beads (a,b) compared with cells transfected with empty vector that show relatively fewer bound beads (c,d). E: collagen bead binding (means ± SE) over time in cells transfected with small interfering RNA (siRNA) for Rac1. These cells show reduced collagen binding compared with cells transfected with control siRNA (P < 0.05 at 30 and 60 min).

We used scanning electron microscopy to study the effect of constitutively active GFP-Rac (Q61) on cell interactions with collagen beads. Transfected cells were isolated by flow cytometry, and the GFP-Rac-expressing cells were incubated with collagen beads. Cells with active Rac1 spread and bound to collagen beads (Fig. 1D, a,b), whereas cells transfected with empty vector showed reduced lamellar extensions and fewer bound beads (Fig. 1D, c,d). Furthermore, in cells transfected with siRNA for Rac1 compared with cells transfected with nontargeted siRNA, there was a twofold reduction in the mean numbers of collagen beads bound over time (Fig. 1E; P < 0.05).

Role of Vav2 in collagen binding.

Although many of the GEFs share homologous common domains that promote their interactions with small GTPase-binding proteins, the distribution and biological activity of GEFs varies depending on cell type. Therefore, we examined protein expression levels of different GEFs in total cell lysates and proteins isolated from collagen beads after 5 min of incubation with NIH 3T3 fibroblasts (Fig. 2 A,a). We found Vav2, Sos1, and β-Pix in total cell lysates and abundant Vav2 and Sos1 in samples from collagen bead-associated proteins. β-Pix was barely detectable in collagen bead-associated proteins. Consistent with its restricted distribution to hematopoietic cells, there was no detectable Vav1. We did not detect Vav3, and C3G was only just detectable. Since these findings could be dependent on antibody reactivity, we assessed protein expression for all of these GEFs in Swiss 3T3 cells, human embryonic kidney (HEK) cells, Cos cells, and Jurkat cells. These cells were chosen on the basis of their known expression of the expected proteins. Immunoblots of whole cell lysates prepared from these cells indicated that our failure to detect Vav1 and Vav3 was not because of detection problems because these proteins were found in Jurkat and HEK cells, and C3G was readily detected in the Swiss 3T3 cells (Fig. 2, A,b).

Fig. 2.

A,a: proteins isolated from collagen beads after 5 min of incubations with NIH 3T3 cells, and total cell lysates were probed with different antibodies to identify expression levels of various guanine nucleotide exchange factors (GEFs). A,b: different cell lines were examined to validate the antibodies used. Note that 3T3 cells here are Swiss 3T3 cells and exhibit more C3G. B: cells tranfected with siRNA and control siRNA show 60–80% reduction in expression levels of Vav2, β-PIX, and Sos-1 by densitometry. C: Western blot analysis of collagen bead-associated proteins shows time-dependent association of endogenous Vav2 in cells tranfected with control siRNA. Knockdown of Vav2 expression with siRNA did not affect Sos-1 and β-PIX association with collagen-coated beads at 30 min. D: cells transfected with GFP-Vav2 and incubated with collagen beads show accumulation of Vav2 around beads at 5 min. E: collagen bead binding in cells transfected with Vav2, Sos-1, β-PIX, and control siRNA. There was 50% reduction in collagen bead binding at 30 and 60 min in cells with knockdown of Vav2 expression (P < 0.01), but there was no change of collagen bead binding after knockdown of Sos-1 or β-PIX.

From the relative abundance of Vav2, Sos1, and β-pix in the collagen bead-associated proteins in the NIH 3T3 cells, we used siRNA knockdown to determine their importance in collagen bead phagocytosis. There was a 70–80% reduction in the relative abundance of these proteins after siRNA (Fig. 2B). Western blot analyses of bead-associated proteins at different time points after incubation with collagen beads showed that knockdown of Vav2 inhibited its association with collagen beads (as expected), whereas siRNA knockdown of Vav2 did not affect the accumulation of Sos-1 and β-Pix around collagen beads. In cells transfected with control siRNA, endogenous Vav2 associated with collagen beads transiently (5–20 min after collagen binding) and then the relative abundance of Vav2 diminished subsequently (Fig. 2C). Fibroblasts transfected with GFP-Vav2 and incubated with collagen beads showed increased fluorescence at the bead sites after 5 min of collagen bead incubation (Fig. 2D).

We reduced the expression of Vav2, Sos1, and β-Pix by siRNA and studied collagen bead binding. Knockdown of Vav2 resulted in a 50% reduction of collagen bead binding compared with controls (P < 0.01). Cells transfected with Sos1, β-PIX, or nontargeted siRNA showed no significant difference compared with controls (Fig. 2E; P > 0.2).

Vav2 targets Rac1.

Since Western blot analysis showed a temporal correlation between the relative abundance of Rac1 and Vav2 in bead-associated proteins after the initiation of bead binding, we examined their localization in more detail. Cells cotransfected with GFP Rac1 and wild-type Vav2-T7 were incubated with collagen beads for 5 min. We observed marked colocalization of the two proteins (Fig. 3A, a–d). We then studied the role of Vav2 on the recruitment of active Rac in cells cotransfected with siRNA and YFP-PBD (to detect active Rac). In cells treated with siRNA for Vav2, there was no evidence for recruitment of active Rac1 to collagen bead binding sites (Fig. 3B, a–c). Although YFP-PBD could potentially bind to cdc42 (6), since cdc42 was not associated with collagen beads (Fig. 1B), we assume that our results reflect predominantly, activated Rac1. Optical sectioning by confocal microscopy showed the presence of YFP-PBD and Vav2 around beads (Fig. 3B, d–f).

Fig. 3.

A: fibroblasts cotransfected with GFP-Rac1 (a) and Vav2-T7 (b) show accumulation of Rac1 and Vav2 at collagen bead sites. B: fibroblasts with siRNA knockdown of Vav2 were transfected with yellow fluorescent p21-GTPase-activated protein kinase (YFP-PBD) to show active Rac. After Vav2 knockdown, there was no detectable accumulation of YFP-PBD around beads (a–c), whereas cotransfections with Vav2-T and YFP-PBD show colocalization and accumulation at collagen bead sites (d–f). Histograms show means ± SE %fluorescence intensity above background of areas immediately around beads (g,h). C: serum-starved cells were stimulated with collagen-coated beads for indicated times. Cell lysates were clarified by centrifugation and incubated with PAK-PBD agarose beads for 1 h. Beads were washed and proteins immunoblotted for Rac1. Cells transfected with Vav2 siRNA show absence of Rac1 activation in response to collagen binding. D: Rac nucleotide-free mutant G15A pulldown shows association with Vav2 in response but not to collagen bead binding and not to β-PIX or Sos-1. There was also no association with BSA-coated beads.

Vav2 regulates Rac activation during collagen bead binding.

We examined a possible functional link between Vav2 and Rac1. After knockdown of Vav2 expression, there was impaired Rac activation in response to collagen bead binding compared with cells transfected with an irrelevant siRNA (to GFP), which showed transient increases of Rac activity at 5 min after collagen bead binding. Similarly, knockdown of Sos1 and β-PIX expression levels did not effect activation of Rac1 upon collagen binding in 3T3 fibroblasts (Fig. 3C).

The Rac1 nucleotide-free mutant G15ARac1 can bind active GEFs with high affinity (15, 16). We examined binding of Vav2, β-PIX, and Sos1 to G15ARac1 beads in lysates prepared from cells that had been treated with collagen beads, BSA-coated beads, or no beads. We found that collagen bead binding enhanced the association of Vav2 with G15ARac1 after 5 min of bead incubation, whereas BSA-coated beads elicited no response (Fig. 3D). We found no evidence of β-PIX or Sos1 binding to G15ARac1 before or after collagen bead incubation with cells.

Src-mediated Vav2 tyrosine phosphorylation in response to collagen binding.

Tyrosine phosphorylation mediates both activation and downmodulation of the biological activity of Vav proteins (19), and integrin-mediated binding can induce tyrosine phosphorylation of various GEFs, including Vav2 (21). Accordingly, we examined tyrosine-phosphorylated proteins in response to collagen binding and observed a gradual increase in tyrosine phosphorylation of proteins with molecular mass of 72–130 kDa in total cell lysates (Fig. 4 A,a). We determined whether endogenous Vav2 was one of these phosphorylated proteins. After incubation with collagen beads, Vav2 was immunoprecipitated, and the immunoprecipitates were immunoblotted with an antibody that recognizes phosphorylated Y172 (19). Phosphorylation of Tyr172 in Vav2 relieves an intramolecular inhibition, thus resulting in enhanced Vav2 activity by exposing the Dbl homology domain (1). We observed increased phosphorylation of Y172 particularly at 2 and 5 min after incubation with collagen beads (Fig. 4 A,b). A similar experimental design was used for cells transfected with wild-type Vav2 tagged with T7. Vav2 was immunoprecipitated from lysates using agarose beads conjugated with T7 antibody and analyzed for tyrosine phosphorylation by immunoblotting with PY20. Tyrosine-phosphorylated Vav2 was maximal at 5 min after exposure to collagen beads. There was no tyrosine phosphorylation of Vav2 in response to cells binding BSA-coated beads at 5 min (Fig. 4 A,c).

Fig. 4.

A,a: total cell lysates were immunoblotted for tyrosine phosphorylated proteins (with PY20) at discrete times after collagen bead incubation. A,b: endogenous Vav2 was immunoprecipitated, and tyrosine phosphorylation at residue 172 was determined at different times in response to collagen bead binding with a phospho-specific Src antibody. A,c: fibroblasts transfected with wild Vav2-T7 were stimulated with collagen or BSA-coated beads. T7 was immunoprecipitated and tyrosine phosphorylation of immunoprecipitates was determined by Western blot analysis with PY20. B: treatment with Src inhibitor PP2 (25 μM) showed inhibition of Y416 tyrosine phosphorylation of Src in response to collagen bead binding but not with its analog PP3 or untreated cells. All cells were preincubated with collagen-coated beads. C: treatment with PP2 inhibited collagen bead binding by 50% (P < 0.05). D,a: immunoprecipitation of Vav2 in PP2-treated cells caused inhibition of phosphorylation of 172Y of Vav2 in cells incubated with collagen beads for 5 min. D,b: fibroblasts transfected with wild Vav2-T7 were treated with the Src inhibitor PP2 or its analog PP3 and stimulated with collagen-coated beads for 5 min. T7 was immunoprecipitated and tyrosine phosphorylation of immunoprecipitates showed inhibition of Src activity in tyrosine phosphorylation of Vav2. E: inhibition of Src activity with PP2 inhibited Rac activation in response to collagen bead binding.

Previous work has shown that Src phosphorylates exchange factors that activate Rho family proteins (20). To investigate the role of Src in Vav2 tyrosine phosphorylation, we determined whether Src activity could be blocked with PP2. Treatment with PP2 (25 μM) inhibited Src phosphorylation at Tyr416 in response to collagen bead binding (Fig. 4B). Treatment with PP2 for 30 min also caused a greater than twofold reduction in the mean proportion of cells that bound collagen beads compared with the inactive analog PP3 or to samples treated with vehicle alone (Fig. 4C). We repeated the experiments described in Fig. 4C and found that PP2 blocked collagen bead-induced phosphorylation of Vav2 on Tyr172, whereas PP3 treatment was not inhibitory (Fig. 4D, a and b). These data indicated that Vav2 phosphorylation depends in part on Src kinase activity.

We assessed whether blockade of Src kinase activity would interfere with activation of Rac1. Fibroblasts were pretreated with PP2 or PP3 and then incubated with collagen beads. Samples of bead-associated proteins isolated after 5 min showed inhibition of Rac activation in PP2-treated samples but not in PP3-treated samples (Fig. 4E). These data indicated that collagen bead binding induces tyrosine phosphorylation of Vav2 on Tyr172, which is mediated by Src. Therefore Src kinase activity and Vav2 are likely required for Rac activation in response to collagen binding.

DISCUSSION

Collagen phagocytosis by connective tissue fibroblasts mediates extracellular matrix remodeling, but unlike the phagocytosis of microorganisms by professional phagocytes such as macrophages, which is mediated by the Fc gamma receptor, collagen phagocytosis is dependent in part on α2β1-integrins (5). We previously observed that the initial interaction of collagen with α2β1-integrins activates Rac1 and greatly enhances collagen binding over time (2, 4), but how specific exchange factor signaling molecules orchestrate the initial interaction of collagen binding leading to acceleration of Rac1 activity is not defined.

In this study we have shown that the guanidine exchange factor Vav2 mediates Rac activation at collagen bead binding sites. We used the nucleotide-free Rac mutant G15ARac1 (15) to determine whether Vav2 interacts with active Rac1 after collagen binding. These experiments clearly indicated that Vav2 interacts with active Rac1 and is required for enhanced Rac1 activity around bound collagen beads. Furthermore, functional experiments showed that knockdown of Vav2 inhibited collagen binding over time and also inhibited enhanced Rac1 activation around the beads. These data are in contrast to an earlier report in which Vav2 activated Rac1 downstream of growth factor receptors but not after binding to β1-integrins (18). Thus while early steps in phagocytosis are appropriately modeled by adhesive cellular processes (11), adhesion is likely to be dependent on the type of cells that are examined. This may be particularly true for the binding step of phagocytosis in fibroblasts where small and spatially discrete activation sites are required for particle binding.

Role of Vav2 and Rac.

Our results with siRNA knockdown of Rac1 showed that enhancement of collagen bead binding required Rac1, whereas collagen binding did not affect Cdc42 or RhoA activation. Whereas these data do not rule out possible roles for Cdc42 or Rho in collagen internalization, the internalization of collagen is not detectable at early time points after collagen ligation (2). We also found that following bead binding, Rac1 activity increased. This is in agreement with previous studies showing that integrin-mediated cell adhesion promotes activation of Rac (9, 24). In contrast, FcR-mediated phagocytosis in macrophages is initiated by recruitment of activated Rac at particle attachment sites (7). Collectively our data show that Vav2 plays a crucial role in the initial binding step of collagen phagocytosis. Although other GEFs (i.e., Sos1 and β-PIX) were detected in the samples isolated from collagen bead binding sites, which raised the possibility of their involvement in Rac1 activation, knockdown of these exchange factors by siRNA demonstrated no effect on activation of Rac1 and enhanced collagen binding.

Role of Src kinase activity.

We found that tyrosine phosphorylation of Vav2 occurs specifically in response to collagen binding and that active Vav2 is required for collagen (but not BSA)-induced activation of Rac1. Notably, tyrosine phosphorylation of Vav2 is sufficient to catalyze GDP-GTP exchange on Rac1 (12) and Src is an effector of Vav2 phosphorylation (25). We found that inhibition of Src kinase activity with PP2 prevented Vav2 phosphorylation and Rac1 activity and greatly reduced collagen bead binding. Furthermore, our experiments with Vav2 knockdown demonstrated greatly reduced localization of Rac1 to collagen beads, suggesting that Vav2 is required for Rac1 targeting to collagen beads. These results are consistent with previous findings showing that in fibroblasts, Vav2 is necessary for fibronectin-dependent activation of Rac, leading to lamellipodia extension and cell spreading (20). Furthermore, Chinese hamster ovary and Jurkat cells transfected with Vav2 show β1-integrin mediated cell adhesion to fibronectin triggered by phosphorylation of Vav2 (27). Collectively, our data indicate a central role for Vav2 in collagen phagocytosis and Rac activation in this process.

GRANTS

This study was supported by CIHR Grant 416228 to C. A. McCulloch.