Journal of Cell Science partnership with Dryad

Journal of Cell Science makes data accessibility easy with Dryad

The P2Y2 nucleotide receptor requires interaction with αv integrins to access and activate G12
Zhongji Liao, Cheikh I. Seye, Gary A. Weisman, Laurie Erb

Summary

The P2Y2 nucleotide receptor (P2Y2R) interacts with αv integrins to activate Go and induce chemotaxis in human 1321N1 astrocytoma cells. In this study, it was determined that the P2Y2R also requires interaction with αv integrins to activate G12 and associated signaling pathways that control chemotaxis in 1321N1 cells. Mutation of the Arg-Gly-Asp (RGD) integrin-binding sequence in the first extracellular loop of the human P2Y2R to Arg-Gly-Glu (RGE), which prevents integrin interaction, did not inhibit Gq or ERK1/2 signaling by the P2Y2R agonist UTP but completely inhibited activation of G12 and G12-mediated events, including Rho activation, cofilin and myosin light chain-2 phosphorylation, stress fiber formation and chemotaxis towards UTP. The involvement of G12 in all these events was verified by using a dominant negative Gα12 construct. G12 activation by the P2Y2R also was inhibited by anti-αvβ5 integrin antibodies and αv integrin antisense oligonucleotides, suggesting that αv integrin activity and expression are required for the P2Y2R to activate G12. Co-immunoprecipitation experiments confirmed that Gα12 protein associates with the wild-type P2Y2R and with αv integrins but not with the RGE mutant P2Y2R or with α3 integrins. Collectively, these results suggest that αv integrin complexes provide the P2Y2R with access to G12, thereby allowing activation of this heterotrimeric G protein that controls actin cytoskeletal rearrangements required for chemotaxis.

Introduction

Chemotaxis is the movement of a cell towards a chemoattractant or away from a chemorepellant and is a fundamental feature of eukaryotic cells; it is important for many physiological and pathological processes, such as embryogenesis, neurogenesis, angiogenesis, wound healing and homing of leukocytes to a site of infection. The ability of a cell to undergo chemotaxis requires the cell to assume a polarized morphology that is controlled by cell surface receptors that activate the Rho family of small GTPases, including Cdc42, Rac and Rho (Burridge and Wennerberg, 2004). Upon activation of a chemoattractant receptor, Cdc42 and Rac localize at the leading edge of a cell and control directional cell movement and the formation of a lamellipodium, respectively (Burridge and Wennerberg, 2004). Rho localizes at the rear and sides of a cell and controls the formation of contractile actin-myosin stress fibers (Xu et al., 2003). Together, these GTPases promote cell migration towards a chemoattractant by mediating extension of the actin cytoskeleton at the front edge of the cell and retraction of the cytoskeleton at the rear.

A variety of compounds can act as chemoattractants by stimulating Rho family GTPases: these include growth factors that activate tyrosine kinase receptors, extracellular matrix proteins that activate integrins, and compounds that activate certain G-protein-coupled receptors (GPCRs) (Burridge and Wennerberg, 2004). Recent studies have shown that GPCRs activate Rac and Rac-dependent lamellipodia formation through Gi/o, whereas the activity of Rho and Rho-dependent stress fiber formation is controlled by G12/13 (Hart et al., 1998; Kozasa et al., 1998; Xu et al., 2003). Furthermore, studies have shown that the βγ subunits of Gi/o are responsible for activation of Rac guanine nucleotide exchange factors (RacGEFs), which, in turn, activate Rac, whereas the α subunits of G12/13 are responsible for activation of RhoGEFs (Neptune et al., 1999; Welch et al., 2002).

In this study, we examined the mechanism of Rho activation and Rho-dependent stress fiber formation mediated by the P2Y2 nucleotide receptor (P2Y2R), an integrin-associated GPCR activated by ATP or UTP that mediates stress fiber formation and chemotaxis in a variety of cell types (Bagchi et al., 2005; Kaczmarek et al., 2005; Satterwhite et al., 1999; Sauzeau et al., 2000; Wang et al., 2005) and plays an important role in monocyte (Seye et al., 2002) and neutrophil homing (Chen et al., 2006) to injured or infected tissues. Previously, we found that an RGD integrin-binding domain in the first extracellular loop of the P2Y2R enables the receptor to interact with αvβ3 and αvβ5 integrins and this interaction is prevented by mutation of the RGD sequence to RGE (Erb et al., 2001). The RGD domain also was found to be necessary for coupling of the P2Y2R to Go-mediated but not Gq-mediated Ca2+ signaling (Erb et al., 2001) and a recent study by our group indicated that the P2Y2R requires interaction with αv integrins to activate Go and to initiate Go-mediated events, including activation of Rac and the RacGEF Vav2, upregulation of vitronectin expression and increased cell migration (Bagchi et al., 2005). Here we report that αv integrin interaction is required for the P2Y2R to access and activate G12, leading to Rho activation and Rho-dependent stress fiber formation.

Results

P2Y2R-mediated Rho activation and stress fiber formation require interaction with αvβ5 integrin

To determine whether integrin interaction with the P2Y2R is important for receptor signaling, we mutated the Arg95-Gly96-Asp97 (RGD) sequence in the P2Y2R to Arg95-Gly96-Glu97 (RGE), a sequence that does not have high affinity for integrins (Erb et al., 2001), and compared signaling events mediated by the wild-type and RGE mutant receptors that were expressed in human 1321N1 astrocytoma cells. These cells are devoid of endogenous G-protein-coupled P2Y receptors (Parr et al., 1994) and, thus, provide a suitable null background for this study. Also, we incorporated a hemagglutinin (HA) epitope tag at the C-termini of the wild-type and RGE mutant P2Y2Rs and used flow cytometry to verify that the cell surface expression levels of these receptor constructs were equivalent (Fig. 1A). We found that mutation of the RGD sequence to RGE did not prevent ERK1/2 phosphorylation in response to the P2Y2R agonist UTP (Fig. 1B) but completely prevented Rho activation (Fig. 1B) and stress fiber formation (Fig. 1C) induced by UTP at all concentrations tested (100 nM to 2 mM). A UTP dose-response curve indicated that the EC50 value for UTP-induced Rho activation by the wild-type P2Y2R was ∼1 μM (supplementary material Fig. S1), which is similar to the EC50 value of UTP for activation of other P2Y2R-mediated responses (Erb et al., 2001). These results suggest that integrin interaction with the P2Y2R, via the RGD integrin-binding domain, is necessary for the P2Y2R to activate Rho and cause stress fiber formation. Likewise, Rac activation and chemotaxis mediated by the P2Y2R in 1321N1 cell transfectants were found to require expression of the RGD integrin-binding domain in the P2Y2R (Bagchi et al., 2005). As expected, untransfected 1321N1 cells that do not express P2Y receptors do not form stress fibers, undergo chemotaxis or exhibit Rho or Rac activation in response to the P2Y2R agonists ATP or UTP (data not shown).

We have found that the RGE mutant P2Y2R can fully stimulate ERK1/2 phosphorylation (Fig. 1B) and Gq-mediated calcium signaling (Erb et al., 2001); however, these responses require agonist concentrations that are three orders of magnitude higher than for the wild-type receptor. This has raised questions as to whether the RGE mutation affects P2Y2R signaling by preventing integrin interaction or by affecting P2Y2R agonist binding affinity. For example, a recent study by Qi et al. confirmed that mutation of the RGD sequence to RGE in the P2Y2R decreases agonist potency for inositol phosphate formation by 1000-fold but mutation of the RGD sequence to AHN did not alter agonist potency, leading the authors to speculate that the RGD to RGE mutation was affecting agonist binding affinity (Qi et al., 2005). To verify that the loss of Rho signaling observed for the P2Y2R RGE mutant was due to decreased integrin binding and not to alteration in agonist binding affinity we constructed a P2Y2R mutant in which the entire RGD integrin-binding sequence was substituted with three alanines (AAA). We found that the AAA mutant P2Y2R had a similar UTP dose response for ERK1/2 phosphorylation as the wild-type receptor (Fig. 2A) but was unable to activate Rho in response to UTP (Fig. 2B). Interestingly, 1321N1 cells expressing the AAA mutant did display a higher basal level of ERK1/2 phosphorylation and Rho activity compared with cells expressing the wild-type P2Y2R (Fig. 2A,B). The reason for the increased basal activity exhibited by the AAA mutant is unknown; nonetheless, these studies support the hypothesis that integrin interaction via the RGD domain of the P2Y2R is important for controlling Rho activation by the P2Y2R.

Fig. 1.

P2Y2R-mediated Rho activation and stress fiber formation require P2Y2R interaction with αv integrins. (A) Cell surface expression of HA-tagged wild-type (WT) or RGE mutant (RGE) P2Y2Rs in 1321N1 cells was determined by flow cytometry. Cells stained with only secondary antibody were used as a negative control. (B) Cells expressing the WT or RGE mutant P2Y2R were treated with or without 1 mM UTP for 5 minutes prior to measuring ERK1/2 phosphorylation and Rho activity. Results shown are representative of three experiments. (C) Cells expressing the WT or RGE mutant P2Y2R were treated with or without 1 mM UTP or 20% FBS (positive control) for the indicated time and stress fibers were visualized by staining filamentous actin (shown in green) with Oregon-Green-labeled phalloidin. Texas-Red-labeled DNase I was used to label monomeric actin (shown in red). More than 100 cells were examined in three separate experiments, and representative images are shown. Bar, 50 μm.

Fig. 2.

ERK1/2 and Rho activation mediated by WT and AAA mutant P2Y2Rs. Human 1321N1 cells were transiently transfected with either WT (RGD) or AAA mutant P2Y2Rs in pcDNA3.1(–). Transfected cells were starved overnight and stimulated with the indicated concentration of UTP for 5 minutes prior to measuring ERK1/2 phosphorylation (A) or Rho activity (B). (A) UTP-induced with 1 mM UTP for 5 minutes before measuring ERK1/2 phosphorylation is expressed as fold increase over basal level in cells transfected with WT P2Y2R. Data points represent the mean ± s.e.m. of results from three experiments. (B) The Rho-GTP band density was normalized to total Rho and relative intensities are shown below each band as mean ± s.e.m. of results from four experiments.

To further assess the role of P2Y2R-integrin interaction in UTP-induced Rho activation and stress fiber formation mediated by the wild-type P2Y2R, we used function-blocking antibodies directed against αvβ5, an integrin that interacts with the P2Y2R, or α3, an integrin that does not interact with the P2Y2R (Erb et al., 2001). Flow cytometry experiments indicate that 1321N1 cells express immunodetectable cell surface αv, α3 and β5 integrin subunits, but not β3 (Bagchi et al., 2005). Pretreatment of P2Y2R-expressing cells with anti-αvβ5 integrin antibodies inhibited UTP-induced Rho activation (Fig. 3A) and stress fiber formation (Fig. 3B), whereas anti-α3 integrin antibodies had no effect (Fig. 3A,B), suggesting that αv integrin activity is important for P2Y2R-mediated Rho activation and stress fiber formation. By contrast, anti-αvβ5 integrin antibodies did not inhibit Rho activation induced by fetal bovine serum (FBS) in 1321N1 cells (supplementary material Fig. S2), suggesting that αv integrins are not involved in Rho activation by growth factors present in serum.

Integrin-P2Y2R interaction regulates Rho-dependent signaling

Activation of Rho leads to stress fiber formation by causing the phosphorylation of myosin light chain-2 (MLC-2) and cofilin, an actin-depolymerizing protein that is inhibited when phosphorylated on Ser3 (Kimura et al., 1996; Moriyama et al., 1996). To verify that integrin-P2Y2R interaction is important for Rho signaling, we analyzed Rho-dependent signaling events in 1321N1 cells expressing wild-type P2Y2R or RGE mutant. UTP caused a dose-dependent increase in phosphorylation of MLC-2 (Fig. 4A) and cofilin (Fig. 4B) in cells expressing the wild-type P2Y2R but not in cells expressing the RGE mutant. Moreover, cofilin phosphorylation mediated by the wild-type P2Y2R was inhibited by pretreatment of the cells with anti-αvβ5 antibodies but not anti-α3 integrin antibodies (Fig. 4C), further demonstrating that the P2Y2R interacts selectively with αv integrins to activate Rho-mediated signaling events. Since cofilin phosphorylation can be regulated by proteins other than Rho, including Rac and testicular protein kinase 1 (Burridge and Wennerberg, 2004; Toshima et al., 2001), we used the ROCK-Rho pathway inhibitor Y-27632 to determine whether cofilin phosphorylation mediated by the P2Y2R occurs through the Rho signaling pathway. UTP-induced cofilin phosphorylation was completely inhibited when the cells were pretreated with Y-27632 (Fig. 4D). Furthermore, pretreatment of the cells with the Gi/o inhibitor pertussis toxin (PTX) did not inhibit UTP-induced cofilin phosphorylation (Fig. 4D) but did inhibit UTP-induced Rac activation (Bagchi et al., 2005), suggesting that Go-mediated activation of Rac is not involved in cofilin phosphorylation by the P2Y2R.

Fig. 3.

P2Y2R-mediated Rho activation and stress fiber formation require αv integrin activity. Human 1321N1 cells expressing the WT P2Y2R were incubated overnight at 37°C in serum-free medium with 10 μg/ml anti-αvβ5 or anti-α3 antibody. Cells were then treated with or without 100 μM UTP for 5 minutes before analysis of Rho activity (A) or 30 minutes before analysis of stress fiber formation (B). (A) Representative blots from three experiments are shown. (B) More than 100 cells were examined in three separate experiments and representative images are shown. Bar, 50 μm.

Fig. 4.

P2Y2R-αv integrin interaction regulates Rho-mediated signaling. (A,B) Human 1321N1 cells expressing the WT or RGE mutant P2Y2R were incubated with the indicated concentration of UTP for 5 minutes. Cell lysates were prepared and analyzed by immunoblotting with (A) anti-phospho-myosin light chain 2 (p-MLC-2) or (B) anti-phospho-cofilin antibodies. (C) Cells expressing the WT P2Y2R were incubated overnight with or without 10 μg/ml anti-αvβ5 or anti-α3 antibody, then stimulated with 100 μM UTP for 5 minutes. Cell lysates were prepared and analyzed by immunoblotting with anti-phospho-cofilin antibodies. (D) Cells expressing the WT P2Y2R were pretreated at 37°C in serum-free medium with the Rho-dependent kinase inhibitor Y-27632 (10 μM) for 1 hour or 200 ng/ml Bordetella pertussis toxin (PTX) overnight, then stimulated with 100 μM UTP for 5 minutes. Cell lysates were analyzed by immunoblotting with anti-phospho-cofilin antibodies. Protein loading in each lane was evaluated by stripping the membrane of antibodies and re-probing with anti-actin antibodies. Blots representative of 3-5 experiments are shown.

Activation of Gα12 by the P2Y2R requires interaction with αv integrin

Rho activation and Rho-dependent stress fiber formation mediated by GPCRs are controlled by heterotrimeric G proteins in the G12/13 family (Buhl et al., 1995; Xu et al., 2003). Generally, GPCRs that stimulate stress fiber formation also couple to Gq/11 but regulate stress fiber assembly through activation of either G12 or G13 (Gohla et al., 1999). Here, we directly investigated whether the RGD integrin-binding domain of the P2Y2R is required for activation of specific G proteins (i.e. G12 and Gq). Results indicated a 2.5-fold increase in [35S]GTPγS binding to Gα12 immunoprecipitated from UTP-treated membrane extracts of 1321N1 cells expressing the wild-type P2Y2R compared with untreated controls, but extracts from cells expressing the RGE mutant receptor did not exhibit an increase in [35S]GTPγS binding to Gα12 in response to UTP (Fig. 5A). By contrast, UTP induced a two- to threefold increase in [35S]GTPγS binding to Gαq upon activation of either the wild-type or RGE mutant P2Y2R (Fig. 5B). Activation of G12/13 and Gq/11 proteins by the P2Y2R was also verified by analyzing serine or threonine phosphorylation of Gα12/13 (Kozasa and Gilman, 1996) and tyrosine phosphorylation of Gαq/11 (Umemori et al., 1997), as previously described. We found that UTP caused phosphorylation of both Gα12 and Gαq in 1321N1 cells expressing the wild-type P2Y2R (Fig. 5C), whereas no phosphorylation of Gα13 was detected in these cells (data not shown). UTP caused phosphorylation of Gαq but not Gα12 in cells expressing the RGE mutant P2Y2R (Fig. 5C), suggesting that αv integrin interaction with the P2Y2R is required for UTP-induced activation of G12 but not Gq.

To further assess whether αv integrins are involved in P2Y2R-mediated activation of Gα12, we tested the effects of inhibition of αv activity or expression using anti-αv integrin antibodies or αv antisense oligonucleotides, respectively. We found that Gα12 phosphorylation by UTP was inhibited by pretreatment with anti-αv but not with anti-α3 integrin antibodies in 1321N1 cells expressing the wild-type P2Y2R (Fig. 6A). Likewise, transfection of αv antisense oligonucleotides in 1321N1 cells expressing the wild-type P2Y2R, which significantly suppressed αv expression (Fig. 6B), completely inhibited Gα12 activation by UTP, as assessed by GTPγS binding (Fig. 6B). Transfection of these cells with αv sense oligonucleotides did not inhibit Gα12 activation by UTP (Fig. 6B). Together, these results suggest that αv integrin expression and activity are required for P2Y2R-Gα12 coupling.

P2Y2R accesses Gα12 in a complex with αv integrins

To determine whether the P2Y2R interacts with Gα12 in a complex with αv integrins, co-immunoprecipitation experiments were performed. Results indicated that endogenous αv integrin co-immunoprecipitated with the HA-tagged wild-type P2Y2R to a much greater extent than with the HA-tagged RGE mutant P2Y2R expressed in 1321N1 cells (Fig. 7A). This association between αv integrin and the wild-type P2Y2R occurred with or without activation of the P2Y2R, although 5 min stimulation with UTP did cause a slight but reproducible reduction in association between these proteins (Fig. 7A). We also found that endogenous Gα12 co-immunoprecipitated with the wild-type P2Y2R but not with the RGE mutant, whereas endogenous Gαq co-immunoprecipitated with both P2Y2R constructs (Fig. 7A). Co-immunoprecipitation of both Gαq and Gα12 with the wild-type P2Y2R was inhibited after UTP treatment (Fig. 7A), consistent with the concept that activation of GPCRs causes the release of receptor-coupled G protein subunits, thus triggering various downstream responses. Although we were unable to detect any association between Gα12 and αv integrins when endogenous levels of these proteins were used, we did find that wild-type Gα12 overexpressed in 1321N1 cells co-immunoprecipitated with endogenous αv integrins but not with endogenous α3 integrins (Fig. 7B), suggesting that G12 selectively associates with complexes containing αv integrins. Interestingly, UTP treatment did not cause dissociation of Gα12 and αv integrins (Fig. 7B) and fluorescence microscopy images indicated that UTP treatment caused a slight redistribution of Gα12 and Gαo onto membrane protrusions or lamellipodia in 1321N1 cells (Fig. 8).

Fig. 5.

P2Y2R-αv integrin interaction is required for G12 coupling. (A,B) Membrane preparations from 1321N1 cells expressing the WT or RGE mutant P2Y2R or pLXSN vector-transfected cells (negative control) were used in [35S]GTPγS binding assays in the presence or absence of 1 mM UTP. After termination of the assay, samples were immunoprecipitated with antiserum against (A) Gα12 or (B) Gαq/11 and radioactivity in the immunoprecipitates was calculated. Data are the means ± s.e.m. of results from three separate experiments and are shown as fold increase over untreated cells. (C) Human 1321N1 cells expressing the WT or RGE mutant P2Y2R were incubated with 1 mM UTP for 2 minutes. G12 activation was detected by immunoprecipitation (IP) of Gα12 with anti-Gα12 antibody and immunoblotting (IB) of Gα12 with anti-phosphoserine/threonine antibody. Gq/11 activation was detected by IP with anti-phosphotyrosine antibody and IB with anti-Gαq/11 antibody. Blots representative of three experiments are shown.

12 activity is required for P2Y2R-mediated stress fiber formation and cell migration

To verify that the G12 protein is responsible for P2Y2R-mediated Rho activation and downstream signaling events, a dominant negative mutant of Gα12 (Gα12DN, Q231L/D299N) was used (Yang et al., 2005). Overexpression of Gα12DN in 1321N1 cells expressing the wild-type P2Y2R completely inhibited UTP-induced Rho activation, cofilin phosphorylation, stress fiber formation and cell migration (Fig. 9A,C,D), but did not inhibit UTP-induced ERK1/2 phosphorylation (Fig. 9B), indicating that G12 is specifically required for P2Y2R-mediated Rho activation and Rho-dependent signaling events leading to stress fiber formation and ultimately, cell migration.

Fig. 6.

P2Y2R-mediated G12 activation requires αv integrin expression and activity. (A) Cells expressing the WT P2Y2R were incubated overnight at 37°C in serum-free medium with 10 μg/ml anti-αvβ5 or anti-α3 antibody, then stimulated with 100 μM UTP for 2 minutes. Gα12 phosphorylation was detected as described in Fig. 5. Relative intensities are shown below each band as mean ± s.e.m. of results from four experiments. (B) Cells expressing the WT P2Y2R were transfected with 5 μg of αv antisense (AS) or sense (S) oligonucleotides, as described previously (Bagchi et al., 2005). Lipofectamine transfected cells served as a negative control. Expression of αv integrin in these cells was determined by immunoblot analysis and a representative blot is shown. UTP-induced binding of [35S]GTPγS to Gα12 was determined, as described in Fig. 5. Data are the mean ± s.e.m. of results from three separate experiments and are shown as fold increase over untreated cells.

Fig. 7.

The P2Y2R accesses G12 in a complex with αv integrins. (A) RGD-dependent association of the P2Y2R with αv integrins and Gα12. Cells expressing the HA-tagged WT or RGE mutant P2Y2R and endogenous Gq/11, G12 and αv integrins were treated with or without 1 mM UTP for 5 minutes and lysates were subjected to IP with mouse anti-HA antibody conjugated to agarose beads and IB with anti-Gα12, anti-Gαq/11, anti-αv integrin, or anti-HA antibody. Cells transfected with the pLXSN vector served as a negative control. (B) Gα12 associates with αv integrins. Human 1321N1 cells expressing cDNA for the WT P2Y2R and Gα12 were treated with or without 100 μM UTP for 5 minutes and lysates were subjected to IP with IgG, anti-α3 integrin, or anti-αv integrin antibody followed by IB with anti-Gα12, anti-α3 integrin or anti-αv integrin antibody. Blots representative of three experiments are shown.

Discussion

The G-protein-coupled P2Y2R is known to activate several heterotrimeric G proteins, including Go and Gq (Boarder et al., 1995; Erb et al., 2001). In this study, we show for the first time that the P2Y2R is also able to activate G12 and to initiate chemotactic signaling events downstream of G12, including Rho activation, cofilin phosphorylation, stress fiber formation and directional cell migration (Figs 1, 5 and 9). Furthermore, we demonstrate here and in previous studies that an RGD integrin-binding domain in the first extracellular loop of the P2Y2R is necessary for the P2Y2R to activate Go and G12, but not Gq (Fig. 5) (Bagchi et al., 2005; Erb et al., 2001), suggesting that integrin complexes provide the P2Y2R with access to select pools of heterotrimeric G proteins. Since the P2Y2R is known to interact with αv integrins (Erb et al., 2001), we performed a series of co-immunoprecipitation experiments to verify that this interaction is required for the P2Y2R to access G12 protein. These experiments showed that (1) Gα12 co-immunoprecipitated with αv integrins but not with α3 integrins; (2) the P2Y2R co-immunoprecipitated with Gαq, Gα12 and αv integrins; and (3) mutation of the P2Y2R integrin-binding domain (i.e., substitution of RGD with RGE that does not bind integrins) did not affect the ability of P2Y2R to co-immunoprecipitate with Gαq but inhibited P2Y2R co-immunoprecipitation with Gα12 and αv integrins (Fig. 7). Although many studies suggest that amino acids located in intracellular loop 2 and the N- and C-terminal portions of intracellular loop 3 are the key elements responsible for GPCR selectivity of G protein recognition (Wess, 1997), the results presented here suggest that αv integrin complexes are also important for establishing interaction between select G proteins and a GPCR. And, although it is well known that GPCRs require integrin activity to stimulate chemotaxis (Miettinen et al., 1998; Till et al., 2002), this is the first indication that a GPCR requires interaction with an integrin to provide access to specific heterotrimeric G proteins that regulate chemotaxis.

Similar to GPCR-mediated chemotaxis, there is some evidence that growth factor receptors and integrins use heterotrimeric G proteins to stimulate chemotaxis. Pertussis toxin (PTX), which specifically inactivates Gi/o proteins by covalent modification of the α subunits, has been found to inhibit VEGF-induced monocyte migration (Barleon et al., 1996) as well as growth-factor-induced Rac1 and Cdc42 activation by a chimeric EGF/VEGF receptor (Zeng et al., 2002). In addition, the latter study determined that overexpression of the Gβγ-sequestering minigene, hβARK1, inhibited growth-factor-induced Rac1 and Cdc42 activation, suggesting that βγ subunits of PTX-sensitive Gi/o proteins are involved in growth-factor-induced activation of Rac and Cdc42 (Zeng et al., 2002). Vitronectin-induced chemotaxis of human melanoma cells mediated by the αvβ3 integrin is also inhibited by PTX (Aznavoorian et al., 1996), suggesting a role for Gi/o proteins in this process. Although the mechanism of heterotrimeric G protein activation by growth factor receptors and integrins is unclear, it is possible that these chemoattractant receptors stimulate the release of ATP, thus triggering activation of G-protein-coupled nucleotide and nucleoside receptors involved in chemotaxis. In support of this idea, studies have demonstrated that ATP is released from epithelial cells (McNamara et al., 2006; McNamara et al., 2001) and migrating neutrophils (Chen et al., 2006) upon exposure to chemotactic bacterial proteins.

Recently, the general importance of the P2Y2R and the adenosine A3 receptor in mediating chemotaxis towards bacterial proteins was demonstrated in neutrophils (Chen et al., 2006). The authors showed that ATP is released at the leading edge of human neutrophils migrating towards the bacterial chemoattractant N-formyl-Met-Leu-Phe (FMLP) and is rapidly broken down to adenosine by ecto-ATPases on the cell surface. In vivo assessment of neutrophil infiltration into the peritoneal cavity of P2Y2R–/– and A3 receptor–/– mice injected with a murine chemotactic protein (Trp-Lys-Tyr-Met-Val-Met-NH2) or with Staphylococcus aureus bacteria indicated that both the P2Y2R and the adenosine A3 receptor are required for neutrophil recruitment. Furthermore, neutrophils lacking the adenosine A3 receptor migrated toward Trp-Lys-Tyr-Met-Val-Met-NH2, but with diminished speed, whereas neutrophils lacking the P2Y2R showed a loss in sensing of the chemoattractant gradient. In agreement with the work of Chen et al. on neutrophils, we found that the P2Y2R remains evenly distributed in the plasma membrane of 1321N1 cells after receptor activation (supplementary material Movie 1), which supports the conclusion that the P2Y2R controls chemotaxis by sensing and amplifying signals induced by chemoattractants.

Fig. 8.

Co-immunostaining of Gα12 and Gαo in cells exposed to a UTP gradient. Human 1321N1 cells expressing wild-type P2Y2Rs were plated on chambered coverslips and incubated overnight in 300 μl serum-free medium. UTP (1 μl of a 100 mM solution) was added at the bottom left and after 30 minutes, cells were washed, fixed, permeabilized and stained with rabbit anti-Gα12 and mouse anti-Gαo antibodies. Texas-Red-conjugated anti-rabbit IgG and Alexa Fluor 488-conjugated anti-mouse IgG were then used to localize Gα12 and Gαo, respectively. Images representative of three experiments are shown. Arrows indicate membrane protrusions indicative of lamellipodia formation. Bars, 50 μm (left column); 20 μm (enlarged images in right column).

Results presented in this study demonstrate that Gα12 selectively associates with complexes containing αv integrins (Fig. 7). Although the mechanism of this interaction is unclear, it is possible that the cadherin family of cell-surface adhesion proteins may be involved because Gα12 has been found to interact with the cytoplasmic tails of several cadherins (Meigs et al., 2001) and E-cadherin is known to associate with αv integrins (von Schlippe et al., 2000). Another mechanism of αv-Gα12 association may involve Tec tyrosine kinases. Members of the Tec family have been found to interact directly with Gα12 and with focal adhesion kinase (Chen et al., 2001; Jiang et al., 1998), thus linking G12 to integrin complexes. Furthermore, Gα12 can interact directly with leukemia-associated RhoGEF (LARG) and, upon phosphorylation of LARG by Tec, Gα12 effectively stimulates the RhoGEF activity of LARG (Suzuki et al., 2003). Although we were unable to detect LARG expression in 1321N1 cells (data not shown), LARG belongs to a subfamily of RhoGEFs (including Lsc/p115 RhoGEF and PDZ-RhoGEF), which, unlike other RhoGEFs, contains a regulator of G protein signaling (RGS) domain that facilitates binding to Gα12/13 (Francis et al., 2006; Fukuhara et al., 2000; Fukuhara et al., 1999; Reuther et al., 2001) and, in some instances, Gαq (Booden et al., 2002; Vogt et al., 2003). Members of this RhoGEF subfamily share the ability to specifically activate RhoA but not other Rho family GTPases, such as Rac1 and Cdc42 (Banerjee and Wedegaertner, 2004). Therefore, the link between this subfamily of RhoGEFs and the P2Y2R warrants further investigation to better define how P2Y2Rs access G12 in αv-containing complexes.

In summary, the present study indicates that the P2Y2R requires interaction with αv integrins to access G12, but not Gq, and to stimulate chemotactic signaling events mediated by G12, including Rho activation, cofilin and MLC-2 phosphorylation and stress fiber formation. Since our previous work indicated that the P2Y2R also requires interaction with αv integrins to activate Go and Go-mediated cell migration (Bagchi et al., 2005), these studies establish that αv integrin complexes are required for the P2Y2R to access select heterotrimeric G proteins involved in chemotaxis.

Materials and Methods

Materials

Anti-human αv (Q20), αvβ5 (P1F76) and α3 (Ralph 3.2) monoclonal Abs, mouse IgG, polyclonal rabbit anti-human Gα12, anti-human Gαq/11, and anti-human ERK1/2 antibodies were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Polyclonal rabbit anti-human cofilin, anti-phospho-cofilin, anti-phospho-MLC-2, and anti-phospho-ERK1/2 antibodies were purchased from Cell Signaling (Beverly, MA). The polyclonal rabbit anti-human actin antibodies were purchased from Cytoskeleton (Denver, CO). The mouse antiphosphoserine/threonine antibody and the anti-αv integrin antibody for immunoblot analysis were purchased from BD Bioscience (San Jose, CA). The mouse anti-phosphotyrosine antibody was purchased from Upstate Biotechnology (Lake Placid, NY). Anti-HA conjugated agarose beads and anti-HA antibody were purchased from Covance (Berkeley, CA). Oregon-Green-conjugated phalloidin, Rhodamine-conjugated phalloidin and Texas-Red-conjugated DNase I were purchased from Molecular Probes (Eugene, OR). The Rho-dependent kinase inhibitor Y27632 was purchased from Calbiochem (Indianapolis, IN). All other reagents including nucleotides were obtained from Sigma-Aldrich (St Louis, MO), unless otherwise specified.

Cell culture and transfection

Human 1321N1 astrocytoma cells were cultured in Dulbecco's modified Eagle's medium (Invitrogen) containing 5% (v/v) fetal bovine serum, 100 U/ml penicillin and 100 μg/ml streptomycin and maintained at 37°C in a humidified atmosphere of 5% CO2 and 95% air. Cells were stably transfected with cDNA encoding either the wild-type or RGE mutant P2Y2R, as previously described (Erb et al., 2001). Both receptor constructs contained sequence encoding a hemagglutinin (HA) tag at the N-terminus of the P2Y2R, as previously described (Erb et al., 2001). To make the AAA-P2Y2R mutant, the RGD sequence was also substituted with Ala-Ala-Ala using the QuikChange XL site-directed mutagenesis kit (Stratagene, CA), and the HA-tagged P2Y2R cDNAs were excised from pLXSN vectors by digesting with EcoRI and BamHI, and ligated into the pcDNA3.1(–) mammalian expression plasmid. The Gα12 wild type and dominant-negative construct (Q231L/D299N) were obtained from Guthrie cDNA Resource Center (Sayre, PA). Human 1321N1 cells expressing the wild-type P2Y2R were cultured to 80% confluence and transiently transfected with the Gα12 constructs in the pcDNA3.1+ vector using the Lipofectamine 2000 reagent (Invitrogen, CA). The transfection efficiency for the Gα12 constructs was determined to be ∼60% using an indirect immunofluorescence assay. The day before experimental use of the cells, the growth medium was replaced with serum-free medium. The αv integrin was suppressed with αv antisense oligonucleotides, as described previously (Bagchi et al., 2005).

Actin stress fiber formation

Cells were plated on glass coverslips and treated as indicated at 37°C in serum-free DMEM. Then, cells were washed in PBS, fixed for 10 minutes in 3.7% (v/v) formaldehyde, treated with 0.5% (v/v) Triton X-100, and rinsed in PBS. For staining of F-actin, cells were incubated with 5 μg/ml 488 Oregon-Green-conjugated phalloidin for 45 minutes at room temperature and then washed with PBS. Texas-Red-labeled DNase I (5 μg/ml) was used to localize monomeric G-actin. In Fig. 9C, cells were incubated with rabbit anti-Gα12 antibody, washed and then stained with Rhodamine-conjugated phalloidin and Oregon-Green-labeled goat anti-rabbit IgG for 45 minutes. Coverslips were mounted on glass slides in ProLong antifade reagent (Molecular Probes) and examined using fluorescence microscopy (Nikon, Eclipse TE300) at room temperature. The objective was a Nikon Plan Fluor 40× lens. Images were acquired and processed with Northern Eclipse 6.0 software via a QImaging camera (Qimaging, British Columbia, Canada). Single cells were selected by cropping the image.

Fig. 9.

Effect of dominant-negative Gα12 (Gα12 DN) on signaling events mediated by the P2Y2R. (A-D) Human 1321N1 cells expressing the WT P2Y2R were cultured to 80% confluence and transiently transfected with the Gα12DN mutant in the pcDNA3.1+ vector. MOCK, vector transfected negative control. Transfected cells were incubated with 100 μM UTP for 5 minutes (A,B) or 30 minutes (C) and assayed for (A) Rho activity and cofilin phosphorylation, (B) ERK1/2 phosphorylation and (C) stress fiber formation, as described in the Materials and Methods. Total Gα12 was detected with anti-Gα12 antibody, as shown in A. Antibody-labeled Gα12 in (C) is shown in red and phalloidin-labeled filamentous actin is shown in green. (D) Cells (5×104) transfected with the indicated constructs were seeded into the upper chamber of Transwells. Lower chambers contained serum-free medium with or without 100 μM UTP. Cell migration was evaluated 16 hours after UTP stimulation and is expressed as the fold increase in the number of cells that moved across the Transwell membranes in response to UTP as compared to untreated controls. (A,B) Blots representative of three-five experiments are shown. (C) Images representative of three separate experiments are shown. Bar, 40 μm. (D) Data shown are the mean ± s.e.m. of results from five experiments.

Rho activity assay

A Rho activation assay kit (Upstate Biotechnology) was used to assess Rho activity according to the manufacturer's instructions. Briefly, cells were cultured in 100-mm tissue culture dishes in culture medium and starved with serum-free medium for 24 hours before being stimulated with UTP for 5 minutes at 37°C. Cells then were washed three times with ice-cold PBS, suspended in Lysis Buffer containing 125 mM HEPES pH 7.5, 750 mM NaCl, 5% (v/v) Igepal CA-630, 50 mM MgCl2, 5 mM EDTA and 10% (v/v) glycerol, and the lysates were transferred to 1.5 ml tubes. Forty microliters of Rhotekin Rho binding domain (RBD)-agarose that only recognizes GTP-bound Rho were added to 500 μl lysate for 45 minutes at 4°C. The beads were pelleted by centrifugation (30 seconds at 14,000 g and 4°C) and washed three times with Lysis Buffer. Finally, the beads were resuspended in 40 ml of 2× Laemmli sample buffer [120 mM Tris-HCl pH 6.8, 2% (w/v) SDS, 10% (w/v) sucrose, 1 mM EDTA, 50 mM dithiothreitol, and 0.003% (w/v) Bromophenol Blue] and western blot analysis (see below) was performed with 1:1000 anti-Rho antibody (Upstate Biotechnology).

[35S]GTPγS binding assay

Membranes (80 μg protein) from 1321N1 astrocytoma cell transfectants expressing the wild type or RGE mutant P2Y2R or the pLXSN vector were isolated, as previously described (Tian et al., 1994) and incubated in assay buffer [50 mM Tris-HCl pH 7.4, 100 mM NaCl, 5 mM MgCl2, 1 mM EDTA, 1 mM DTT, 1 μM guanosine 5′-diphosphate, 1× protease inhibitor cocktail (Roche), and 50 nCi of [35S]GTPγS; 1250 Ci/mmol; Perkin Elmer, CA] containing the indicated concentration of UTP. Samples were incubated for 20 minutes at 30°C followed by addition of 0.5 ml ice-cold buffer containing 50 mM Tris-HCl (pH 7.4), 100 mM NaCl, and 5 mM MgCl2. The samples were centrifuged at 100,000 g for 15 minutes at 4°C and the resulting pellets were resuspended in 500 μl of solubilization buffer [100 mM Tris-HCl pH 7.4, 200 mM NaCl, 1 mM EDTA, 1.25% (v/v) NP40, 0.2% (w/v) SDS and 1× protease inhibitor cocktail]. Extracts were incubated overnight with 1:1000 anti-Gα12 or anti-Gαq antibody at 4°C. The extract was then incubated with 50 μl of a 50% protein-G-agarose suspension, and the immune complexes were collected by centrifugation and washed three times in wash buffer (50 mM Tris-HCl pH 7.4, 100 mM NaCl, and 5 mM MgCl2). [35S]GTPγS binding in the immunoprecipitates was quantified by liquid scintillation counting.

Immunoblot analysis and immunoprecipitation

Immunoblotting (IB) was performed as previously described (Liu et al., 2004). After the IB procedure, the membranes were stripped and reprobed with anti-actin or anti-ERK1/2 antibody to assess protein loading. Lysates from 1321N1 cell transfectants expressing the HA-tagged wild-type or RGE mutant P2Y2R, or the pLXSN vector were used for immunoprecipitation (IP) with anti-HA-conjugated agarose beads, as previously described (Liu et al., 2004). IP was also performed with cell lysates and anti-αvβ5 or anti-α3 antibody or normal goat IgG (negative control). The immune complexes were precipitated by protein-G-conjugated beads and analyzed by IB with anti-G12 antibody. Phosphorylation of G proteins was detected by IP of G12 or tyrosine-phosphorylated proteins using anti-G12 or anti-phosphotyrosine antibody, respectively. The immunoprecipitated samples were analyzed by IB with anti-phosphoserine/threonine or anti-Gq antibody, respectively.

Cell migration assay

Cell migration assays were performed with 8-μm pore size Transwells (Costar) as described (Bagchi et al., 2005). In brief, the cells were cultured at 37°C for 24 hours in DMEM supplemented with 5% fetal bovine serum, suspended by trypsinization, washed and resuspended in 100 μl of serum-free DMEM (5×104 cells) and placed in the upper chamber of the Transwells. The lower chamber was filled with 600 μl serum-free medium with or without 100 μM UTP. The cells were allowed to migrate for 16 hours at 37°C. Cells migrating to the lower side of the membrane were fixed with cold methanol and stained with Accustain. Cells were counted in 10 microscopic fields at 20× magnification.

Acknowledgments

We thank D. D. Zhang, S. Bagchi, J. Hamilton, J. R. Newton and J. Camden for technical assistance and R. Garrad for helpful comments on the manuscript. Human 1321N1 cells stably transfected with eGFP-hP2Y2R were a kind gift from Fernando A. González, Department of Chemistry, University of Puerto Rico. Z.L. was supported by a pre-doctoral fellowship from the American Heart Association-Heartland Affiliate. This work was supported by NIH grants AG18357 and DE07389 and the F21C program of the University of Missouri-Columbia.

Footnotes

  • Accepted March 7, 2007.

References

View Abstract