Phosphatidylinositol Phosphate 5-Kinase Iγi2 in Association with Src Controls Anchorage-independent Growth of Tumor Cells

Background: PIPKIγ isoforms play roles in cell migration, polarization, and membrane trafficking and are overexpressed in triple-negative breast cancers, indicating protumorigenic functions.

Results: PIPKIγi2 associates with the C terminus of Src, and this interaction interdependently controls their functioning.

Conclusion: PIPKIγi2 and Src synergistically control anchorage-independent tumor cell growth.

Significance: This study shows unexpected mechanisms for a phosphatidylinositol 4,5-biphosphate-generating enzyme that synergizes with the proto-oncogene Src to regulate oncogenic signaling.

Abstract

A fundamental property of tumor cells is to defy anoikis, cell death caused by a lack of cell-matrix interaction, and grow in an anchorage-independent manner. How tumor cells organize signaling molecules at the plasma membrane to sustain oncogenic signals in the absence of cell-matrix interactions remains poorly understood. Here, we describe a role for phosphatidylinositol 4-phosphate 5-kinase (PIPK) Iγi2 in controlling anchorage-independent growth of tumor cells in coordination with the proto-oncogene Src. PIPKIγi2 regulated Src activation downstream of growth factor receptors and integrins. PIPKIγi2 directly interacted with the C-terminal tail of Src and regulated its subcellular localization in concert with talin, a cytoskeletal protein targeted to focal adhesions. Co-expression of PIPKIγi2 and Src synergistically induced the anchorage-independent growth of nonmalignant cells. This study uncovers a novel mechanism where a phosphoinositide-synthesizing enzyme, PIPKIγi2, functions with the proto-oncogene Src, to regulate oncogenic signaling.

Introduction

The ability of tumor cells to defy anoikis, cell death caused by lack of cell-matrix interaction, and grow in an anchorage-independent manner determines their capacity to survive in the vasculature and lymphatic circulation during tumor metastasis (1, 2). In adherent cells, focal adhesions are the contact points of cells to their underlying substratum and also serve as signaling hubs. Anchorage dependence of normal cells stems from the fact that they derive a large part of their proliferative and survival signals from their substratum via focal adhesions (3). Contradicting this, many focal adhesion proteins, including FAK, integrin-linked kinase, paxillin, Src, talin, and pCAS130, are actively involved in oncogenic processes that allow tumor cells to survive/grow in an anchorage-independent manner and promote tumorigenesis (4–7). However, the precise mechanism for how tumor cells assemble the signaling molecules at the plasma membrane following the disruption of cell-matrix interaction, thus bestowing anchorage independence for survival and growth, remains poorly understood (8–11).

Phosphatidylinositol 4,5-biphosphate represents the major phosphoinositide in the plasma membrane, where it functions as a pleiotropic lipid-signaling molecule regulating many cellular functions including cell survival, cell cycle progression, cell migration, vesicle trafficking, and actin cytoskeleton reorganization (12–19). The spatiotemporal generation of phosphatidylinositol 4,5-biphosphate regulates the targeting and/or functional activity of different molecules in specified subcellular compartments (12, 13, 20). Spatial signaling is achieved by an association of phosphatidylinositol 4,5-biphosphate-generating enzymes with molecules that are often phosphatidylinositol 4,5-biphosphate effectors (12, 20). For example, PIPKIγi2 is specifically targeted to focal adhesions via interaction with talin, where it generates phosphatidylinositol 4,5-biphosphate required for focal adhesion assembly and cell migration (21, 22). Phosphatidylinositol 4,5-biphosphate generation lies at a junction point in the phosphoinositide signaling, where it can function as a lipid messenger or be used as substrate for PI3K and PLC to generate the second messengers: phosphatidylinositol-3,4,5-triphosphate, inositol-1,4,5-triphosphate, and diacylglycerol (20).

In mammalian cells, phosphatidylinositol 4,5-biphosphate is largely generated by type I PIPK² enzymes, which are classified into PIPKIα, PIPKIβ, and PIPKIγ isoforms (12–14). PIPKIα is targeted to nucleus and controls nuclear events (14). PIPKIβ is localized at vesicular structures in the perinuclear region (12, 13). Mammalian cells express at least five splice variants of PIPKIγ (e.g., PIPKIγi1, PIPKIγi2, PIPKIγi4, and PIPKIγi5); among them, PIPKIγi2 is targeted to both focal adhesion sites and endosomal membranes (12, 24). Studies are emerging that implicate PIPKIγ in tumor progression, where increased expression of PIPKIγ in breast cancer tissues correlated with poor patient survival (25). Increased PIPKI activity is observed in hepatocellular carcinoma (26), and PIPKIγ regulates the transcriptional activity of β-catenin downstream of growth factor signaling (27). Combined, these studies indicate a role for PIPKIγ in tumor progression.

Src, non-receptor tyrosine kinase and proto-oncogene, regulates the PIPKIγi2 interaction with talin (28). Src activation is a hallmark of many cancers, and several mechanisms are implicated in its activation of tumorigenic processes (29, 30). The plasma membrane recruitment and activation of Src is primarily mediated by myristolylation of glycine residue in its N terminus (31), although highly conserved basic residues in its N-terminal part also play an important role in Src function and its plasma membrane recruitment via electrostatic interaction with anionic lipid molecules (32). Here, we show that PIPKIγi2 and Src, both focal adhesion molecules, form a signaling complex following the disruption of cell-matrix interaction and support the anchorage-independence of tumor cells.

EXPERIMENTAL PROCEDURES

Antibodies

Antibodies for Src (antibody 2108S), ERK1/2 (antibody 9194), and pERK1/2 (antibody 4370) were purchased from Cell Signaling; antibodies for Src (antibodies 44660G and 44662G) and pFAK (antibody 44-624G) were purchased from Invitrogen; antibodies for talin (antibody 05-385), FAK (antibody 05-537), Src (antibody 05-184), and phosphotyrosine (antibody 4G10) were purchased from Millipore; antibodies for paxillin (antibody 61005) and Csk (antibody 610079) were from BD Bioscience. Other antibodies used were: HA (antibody MMS-101R; Covance), talin (antibody HPA004748; Sigma), cortactin (antibody 05-180; Upstate), and anti-GFP (antibody GF28R; Thermo Scientific). Antibodies for PIPKIγ, PIPKIγi2, and PIPKIα were developed in the laboratory (28, 33, 34). Antibody specific for tyrosine-phosphorylated PIPKIγi2 was kind gift of Dr. Dianqing Wu (Yale University).

DNA Constructs, Mutagenesis, and siRNA

PIPKIγ isoforms or PIPKIγi2 mutants were subcloned into MluI and SalI sites in frame with HA tag in the N terminus of PWPT vector (Addgene) as described previously (22). Full-length chicken Src or Src mutants used in the study were cloned into BamHI and SalI sites of PWPT vector. All the mutations used in the study were created using a QuikChange site-directed mutagenesis kits (Stratagene) followed by DNA sequencing to confirm the integrity of the DNA sequence. For cloning the C terminus of Src, oligonucleotides used for annealing were: TCGAGGAGGACTACTTCACGTCCACCGAGCCCCAGTACCAGCCCGGGGAGAACCTCTAGG (sense) and TCGACCTAGAGGTTCTCCCCGGGCTGGTACTGGGGCTCGGTGGACGTGAAGTAGTCCTCC (antisense). After annealing of these oligonucleotides, 5′-prime and 3′-prime ends harbor XhoI and SalI sites, respectively, for cloning into pEGFP-C3 vector (Clontech) in frame with GFP in the N terminus. This construct was further subcloned into PWPT-GFP vector for retroviral infection.

siRNA

The following oligonucleotides were used: control siRNA, CCUUGGUGACUCGUAGUUU; siPIPKIγi2, GAGCGACACAUAAUUUCUA; siPIPKIγi2 (second), CGGCGAGAGCGACACAUAA siPIPKIγ, GCCACCTTCTTTCGAAGAA; siPIPKIα, GAAGUUGGAGCACUCUUGG; and siSrc, GGCUCCAGAUUGUCAACAA. siRNA for talin and FAK were purchased from Santa Cruz Biotechnology.

Cell Culture

MDA-MB-231, NIH3T3, HEK293, and HEK293FT cells were cultured in DMEM containing 10% FBS. T47D and HCC1954 cells were cultured in RMPI 1640 containing 10% FBS. SUM1315 cells were culture in Ham's F12 supplemented with 5% FCS. Suspended cells in the study refer to the cells resuspended in medium containing 0.1% BSA and 1.0% FBS after trypsinization/detachment and incubated at 37 °C in the incubator for 2–3 h except for time course study. For overnight culture in suspension condition, cells suspended in the medium were seeded into the culture plate coated with 0.3% agar to avoid cell attachment. For stimulation of cells in adherent condition, cells were serum-starved overnight before stimulating the cells with 10% FBS or EGF (50 ηg/ml) for the indicated time periods. For the stimulation of cells in suspension condition, cells were resuspended as described above and incubated for 2–3 h before stimulating with FBS (10% FBS) or EGF (50 ηg/ml) or extracellular matrix protein (combination of fibronectin/collagen type I, 25 μg/ml each) for the indicated time periods.

Transfection or Lentiviral Infection

For siRNA-mediated knockdown of genes, Lipofectamine RNAiMAX (Invitrogen) was used following the protocol provided by manufacturer, and cells were used 48–72 h post-transfection. For transient transfection into HEK293 cells, Lipofectamine 2000 (Invitrogen) was used. Cells were harvested 24 hours post-transfection. For the expression or co-expression of genes into MDA-MB-231 or NIH3T3 or MCF-7 cells, the lentiviral system was used as described previously (22). Cells were harvested 24–48 h post-infection (70–80% infection efficiency were achieved for the experiments).

Immunoprecipitation and Immunoblotting

Cells were lysed using lysis buffer (50 mm Tris-HCl, pH 7.4, 150 mm NaCl, 0.5% Triton X-100, 1 mm EDTA, 10 mm NaF, 5 mm Na₃VO₄, and protease inhibitors). Clear supernatants were incubated with indicated antibodies for 3–4 h to overnight at 4 °C followed by isolation of immune complexes using protein G-Sepharose 4B beads (Amersham Biosciences). The beads were washed three times with lysis buffer before eluting the immune complexes with 2× sample buffer and then subjected to immunoblotting using indicated antibodies.

GST Pulldown Assay

Different regions of chicken Src were PCR-amplified and cloned into pGEX-6P-1 (Novagen). Proteins were expressed into BL21 and purified using glutathione-Sepharose 4B beads (Amersham Biosciences). For in vitro binding study, purified GST fusion proteins immobilized on the Sepharose beads were incubated with His-tagged PIPKIγi2 purified from bacteria or with cell lysates prepared from HEK293 cells transfected with HA-PIPKIγi2 at 4 °C for 1 h followed by elution of bound proteins with 2× sample buffer for immunoblotting.

Cell Proliferation and Anchorage-independent Growth

For cell proliferation assay, MDA-MB-231 cells were seeded into 12-well culture plate (1,000 cells/well) in DMEM containing 10% FBS. Cells were manually counted every second day for up to 8 days.

For anchorage-independent growth, cells were suspended in medium containing 0.3% agar and seeded into 24-well culture plates. To avoid cell attachment, culture plates were precoated with 0.5% agar before cell seeding. Cultures were fed with fresh medium in every 3–5 days and cultured for 10–28 days depending upon the cells type used. Similarly, cell numbers used for seeding were also adjusted depending upon the efficiency of cells to form colonies. In some cases, Src inhibitor (PP1, 0.5 μm) was added into the medium. Colonies developed were fixed with 3.7% paraformaldehyde and stained with 0.1% crystal violet to facilitate the visualization and counting.

Immunofluorescence Microscopy (IF)

For examining the co-localization of PIPKIγi2 and Src at focal adhesions, cells were seeded into collagen type I- or fibronectin-coated coverslip and incubated for 30 min before fixing the cells with 3.7% paraformaldehyde. Cells were permeabilized with 0.1% Triton X-100 before blocking with 3% BSA in PBS. The same procedures were used for IF study of the colonies developed in the soft agar. Cells were incubated with primary antibody overnight at 4 °C followed by incubation with Alexa 555- and/or Alexa 488-conjugated secondary antibody (Molecular Probes) for 1 h at room temperature. Slides were mounted using Vectashield and visualized with a Nikon TE2000-U microscope using 63× objective lenses. The images were acquired using MetaMorph and processed using adobe Photoshop.

For examining the phosphatidylinositol 4,5-biphosphate distribution in the PIPKIγ or PIPKIγi2 knockdown cells, MDA-MB-231 cells were transfected with siRNA as described above. After 24–36 h, cells were retransfected with plasmids for the expression of GFP-PLCδ-PH or GFP-PLCδ-PH mutant. Cells were processed for IF study following the overnight culture.

Statistical Analysis

The data are presented as means ± S.D. from at least three-independent experiments. Unpaired t test was conducted to determine the p value, and the statistical significance between two groups (p value equal to or less than 0.05 were considered significant).

RESULTS

PIPKIγ/PIPKIγi2 Regulate the Anchorage-independent Growth of Tumor Cells

PIPKIγi2 is a phosphatidylinositol 4,5-biphosphate-generating enzyme targeted to cell-matrix interaction sites via an interaction with talin (33, 35). Src phosphorylation of tyrosine residues at the C terminus of PIPKIγi2 (red Tyr residues in Fig. 1A) regulates its interaction with talin (28). Increased expression of PIPKIγ in breast cancer tissues inversely correlates with patient survival, indicating its potential role in tumor progression (25). To define an oncogenic role for PIPKIγ, pan-PIPKIγ or PIPKIγi2 was knocked down, and the effect on anchorage-independent growth of breast cancer cell lines in soft agar was examined. The knockdown of endogenous PIPKIγ or PIPKIγi2 significantly impaired the ability of MDA-MB-231, SUM1315, and T47D cells to grow in an anchorage-independent manner (Fig. 1, B–D). However, cell proliferation was not affected by PIPKIγi2 knockdown but was affected by pan-PIPKIγ knockdown in MDA-MB-231 cells in two-dimensional culture (Fig. 1E). Consistently, the loss of PIPKIγi2 was not sufficient to affect the localization of GFP-PLCδ-PH, a biosensor of phosphatidylinositol 4,5-biphosphate, whereas pan-PIPKIγ knockdown resulted in impaired localization of GFP-PLCδ-PH in the plasma membrane (Fig. 1F). These results indicate the collective function of different PIPKIγ variants in phosphatidylinositol 4,5-biphosphate synthesis in the plasma membrane. Furthermore, the knockdown of endogenous PIPKIγi2 in MDA-MB-231 cells expressing siRNA-resistant PIPKIγi2 did not affect the anchorage-independent growth (Fig. 1G). Further, the kinase-dead mutant of PIPKIγi2 poorly rescued anchorage-independent growth, signifying the importance of kinase activity. The expression level of ectopically expressed siRNA-resistant PIPKIγi2 was severalfold higher than that of endogenous PIPKIγi2 and resulted in induced anchorage-independent growth.

FIGURE 1.

PIPKIγ/PIPKIγi2 is required for anchorage-independent growth of tumor cells. A, schematic diagram of PIPKIγi2 showing the domain organization and the specific amino acid sequence at the C terminus. Red letters indicate Src phosphorylation sites crucial for talin binding. B–D, 48 hours after siRNA transfection, cells were cultured in soft agar for 2–4 weeks. Immunoblotting shows the knockdown of PIPKIγ and PIPKIγi2. Representative images of colonies developed by T47D cells are shown. E, for cell proliferation assay, MDA-MB-231 cells after siRNA transfection were seeded into 12-well culture plates (1,000 cells/well). Cells were counted manually every second day. F, 24–36 h after siRNA transfection, cells were retransfected with plasmids for expression of GFP-PLCδ-PH or GFP-PLCδ-PH mutant. Cells were processed for IF study after overnight culture (scale bar, 10 μm). G, siRNA was used to knock down endogenous PIPKIγi2 from MDA-MB-231 cells or MDA-MB-231 cells expressing siRNA-resistant wild type PIPKIγi2 (PIPKIγi2WT) or its kinase dead mutant (PIPKIγi2KD). Cells were cultured in soft agar 48 hours after transfection. Immunoblotting shows the knockdown of endogenous PIPKIγi2 in the cells. The expression of siRNA-resistant PIPKIγi2WT or PIPKIγi2KD was confirmed by immunoblotting using antibodies specific for PIPKIγi2 or HA tag. All the values are means ± S.D. from three-independent experiments. The error bars represent S.D. (p values are indicated).

To investigate the role of each PIPKIγ splice variant, PIPKIγ variants were ectopically expressed into MDA-MB-231 cells, which express a low level of PIPKIγ/PIPKIγi2 compared with other breast cancer cell lines examined (not shown). As shown in Fig. 2 (A–C), the expression of each variant significantly promoted anchorage-independent growth, although PIPKIγi2 expression had substantially greater effect, which correlated with its expression level (Fig. 2D). The knockdown of ectopically expressed PIPKIγi2 completely abrogated the growth promoted by PIPKIγi2 overexpression (Fig. 2E). Moreover, PIPKIγi1 or PIPKIγi2 expression into MDA-MB-231 cells did not show an obvious effect on cell proliferation in two-dimensional culture (Fig. 2F), indicating specificity for anchorage independent growth regulation. PIPKIγi2 overexpression also promoted anchorage-independent growth of MCF-7 cells (Fig. 2G).

FIGURE 2.

Overexpression of PIPKIγ/PIPKIγi2 promotes anchorage-independent growth. A, immunoblot showing the expression of HA-tagged PIPKIγ variants into MDA-MB-231 cells. B and C, oncogenic growth of MDA-MB-231 cells expressing different PIPKIγ variants were examined by counting the colonies developed after 2–3 weeks of culture in soft agar as described above. D, MDA-MB-231 cells expressing low and high levels of PIPKIγi2 were selected, and their oncogenic growth on soft agar was monitored as described above. E, siRNA was used to knock down ectopically expressed PIPKIγi2, and its effect on anchorage-independent growth was examined as described above. F, for cell proliferation assay, MDA-MB-231 overexpressing PIPKIγi1 or PIPKIγi2 was seeded into 12-well culture plates (1,000 cells/well). Cells were counted manually every second day. G, MCF-7 cells overexpressing PIPKIγi2 were cultured in soft agar for 21 days before counting the developed colonies (at least 10 randomly selected areas were used for counting the colonies). All the values are means ± S.D. from three-independent experiments. The error bars represent S.D. (p values are indicated).

PIPKIγi2 Regulates Src Activation Downstream of Growth Factor Receptors and Integrins

To delineate the signaling molecules involved in PIPKIγ regulation of anchorage-independent growth, we examined the impact of PIPKIγ or PIPKIγi2 knockdown on Src, a key signaling molecule with roles in cell survival, oncogenic, and/or anchorage-independent growth (4, 6, 29) that phosphorylates PIPKIγi2 (28). Suspension culture of different tumor cells displayed significantly impaired Src activation (tyrosine phosphorylated Src in its activation site) upon PIPKIγ or PIPKIγi2 knockdown (Fig. 3, A and B, and data not shown). Corroborating this, the overexpression of PIPKIγi2 increased the activation level of Src, without a noticeable change in FAK activation (assessed by autophosphorylation of FAK at Tyr³⁹⁷) (Fig. 3C). The knockdown of ectopically expressed PIPKIγi2 abrogated activation level of Src (Fig. 3D), and this coincided with significantly reduced anchorage-independent growth (Fig. 2E). However, the impact of PIPKIγ/PIPKIγi2 knockdown or overexpression on Src activation was less obvious in the adherent condition (not shown).

FIGURE 3.

PIPKIγi2 regulates Src activation downstream of growth factor receptors and integrins. A and B, MDA-MB-231 and T47D cells after siRNA transfection for PIPKIγ or PIPKIγi2 knockdown were incubated in suspension condition, and the activation status of Src, FAK, and ERK1/2 was examined by immunoblotting using their specific antibodies. ConsiRNA, control siRNA. C, MDA-MB-231 cells expressing different PIPKIγ variants were incubated in suspension condition before harvesting the cells to examine the activation status of signaling molecules. D, siRNA was used to knock down the expression of ectopically expressed PIPKIγi2 before examining the activation level of Src and other molecules in suspension condition. E, mock or PIPKIγi2-overexpressing MDA-MB-231 cells were incubated in suspension condition (Susp.) before harvesting the cells at different time points to examine the activation level of Src. F, MDA-MB-231 cells after siRNA transfection for PIPKIγ or PIPKIγi2 knockdown were serum-starved overnight before stimulating with EGF (50 ng/ml) or 10% FBS. Cells were harvested at different time points to examine the activation level of Src by immunoblotting. G, cells after siRNA transfection were incubated in suspension condition before stimulating with fibronectin/collagen type I (FN/Col.I). H, mock or PIPKIγi2-overexpressing cells incubated in suspension condition were stimulated with 10% FBS or fibronectin/collagen type I to examine the activated Src.

In suspension culture, MDA-MB-231 cells overexpressing PIPKIγi2 showed prolonged activation of endogenous Src (Fig. 3E). Src is activated downstream of integrins and growth factor receptors (29, 36). Consistently, EGF- or FBS-stimulated Src activation in MDA-MB-231 cells was significantly impaired upon PIPKIγ or PIPKIγi2 knockdown (Fig. 3F). Also, in suspension condition, PIPKIγ or PIPKIγi2 knockdown affected Src activation in response to stimulation with FBS or extracellular matrix proteins (Fig. 3G). In corroboration with knockdown studies, the increased expression of PIPKIγi2 promoted FBS- or extracellular matrix protein-stimulated activation of Src, indicating PIPKIγi2 regulation of Src activation downstream of integrins and growth factor receptors (Fig. 3H).

Src Is Required for PIPKIγi2-induced Anchorage-independent Growth and Vice Versa

In support of the above results, the knockdown of endogenous Src or the use of a Src inhibitor blocked the PIPKIγi2-induced anchorage-independent growth (Fig. 4, A–D), indicating the role for Src in PIPKIγi2-regulation of oncogenic growth. These results are consistent with a key role for Src in mediating cell survival and growth in both cellular and in vivo systems (6, 36).

FIGURE 4.

PIPKIγi2 and Src reciprocally regulate their oncogenic function. A–C, siRNA were used to knock down endogenous Src from MDA-MB-231 cells overexpressing PIPKIγi2. Cells were cultured in soft agar for 2–3 weeks before counting the colonies. ConsiRNA, control siRNA. D, PIPKIγi2-overexpressing cells were treated with a pharmacological inhibitor of Src (PP1 or control PP3, 0.5 μm) during suspension culture, and the colonies developed were counted after 2–3 weeks culture in soft agar. All the values are means ± S.D. from three-independent experiments. The error bars represent S.D. (p values are indicated). E, siRNA was used to knock down PIPKIγi2 or PIPKIα from MDA-MB-231 cells transfected with Src. Cells were fixed for IF study to examine the actin cytoskeleton (left panels) and Src localization (right panels). Scale bar, 20 μm. F, cortactin and FAK were immunoprecipitated from mock and Src-infected cells after siRNA transfection for PIPKIγi2 knockdown. Cells were harvested 48 hours post-transfection to immunoprecipitate the endogenous cortactin and FAK followed by immunoblotting using phosphotyrosine antibody. G, siRNA was used for PIPKIγi2 knockdown in MDA-MB-231 cells infected with lentivirus for Src overexpression. Cells were cultured in soft agar for 2 weeks before counting the colonies.

After demonstrating the Src function in PIPKIγi2-induced anchorage-independent growth, we inquired whether PIPKIγ/PIPKIγi2 is required for Src function. A feature of Src-transformed cells is a disorganized cytoskeleton with multiple cell protrusions (37). Consistently, the knockdown of PIPKIγi2, but not PIPKIα, abrogated the disorganized actin cytoskeleton phenotype induced by Src expression in MDA-MB-231 cells (Fig. 4E, left panels). Ectopically expressed Src localized to cell protrusions, and this targeting was also significantly reduced by PIPKIγ or PIPKIγi2 knockdown (Fig. 4E, right panels). Similarly, tyrosine phosphorylation of FAK and cortactin, Src substrates, induced by Src expression was reduced upon PIPKIγi2 knockdown (Fig. 4F). Decreased anchorage-independent growth of Src-expressing cells upon PIPKIγ or PIPKIγi2 loss was also accompanied by reduced Src activation (Fig. 4G), indicating that PIPKIγi2 is required for Src activation and function.

PIPKIγi2 and Src Synergistically Induce Anchorage-independent Growth

After demonstrating the Src requirement for PIPKIγi2-induced anchorage-independent growth and vice versa, we examined the ability of PIPKIγi2 (and other PIPKIγ variants) to induce oncogenic growth of the nontransformed NIH3T3 or MCF10A cells. Independent expression of PIPKIγi2 or Src poorly induced the anchorage-independent growth of NIH3T3 cells (Fig. 5A). Strikingly, co-expression of PIPKIγi2 and Src dramatically increased the anchorage-independent growth. The synergistic effect of PIPKIγi2 and Src was further demonstrated in MDA-MB-231 cells (Fig. 5B). Among PIPKIγ variants, PIPKIγi2 showed the most potent effect, emphasizing the functional specificity of PIPKIγi2 and Src in anchorage-independent growth regulation (not shown). However, a kinase dead mutant of PIPKIγi2 poorly induced anchorage-independent growth in synergy with Src, indicating the role of kinase activity of PIPKIγi2 enzyme (Fig. 5C). Furthermore, an analysis of the expression of PIPKIγ/PIPKIγi2 demonstrated the link between PIPKIγ/PIPKIγi2 expression and Src activation in many breast cancer cell lines examined (data not shown). In all tumor cell lines examined, the loss of PIPKIγ/PIPKIγi2 and/or Src inhibited oncogenic growth on soft agar (Fig. 5, D and E).

FIGURE 5.

PIPKIγi2 and Src show synergy in inducing oncogenic growth. A and B, PIPKIγi2 and Src were individually expressed or co-expressed into NIH3T3 (A) or MDA-MB-231 cells (B) using lentiviral infection. Their oncogenic growth was monitored by counting colonies developed in soft agar. C, MDA-MB-231 cells overexpressing PIPKIγi2WT or PIPKIγi2KD were infected with lentivirus for Src expression. Oncogenic growth of the infected cells were examined as described above. D and E, the effect of PIPKIγi2 and Src knockdown on oncogenic growth of T47D and HCC1954 cells. All the values are means ± S.D. from three-independent experiments. The error bars represent S.D. (p values are indicated). ConsiRNA, control siRNA.

PIPKIγi2, Src, and Talin Interdependently Regulate Their Subcellular Localization, and All Are Integrated into the Signaling Complex

Targeting of Src to focal adhesions is a key for its oncogenic activity (37–39). Immunofluorescence studies revealed extensive co-localization of endogenous active Src with PIPKIγi2, predominantly at focal adhesions (Fig. 6A, upper panels). Phosphatidylinositol 4,5-biphosphate-generating enzymes regulate intracellular vesicle trafficking and targeting of signaling molecules to the plasma membrane at sites of adhesion (40, 41). PIPKIγi2 knockdown modestly affected active Src localization at focal adhesions (Fig. 6, A, lower panels, and B). However, the loss of talin resulted in a more profound defect on Src localization at focal adhesions (Fig. 6, A and B) and is consistent with the role of talin in focal adhesion assembly.

FIGURE 6.

PIPKIγi2, Src, and talin mutually regulate their targeting to focal adhesions and form a complex in suspension culture. A, upper panels, IF study showing the co-localization of HA-PIPKIγi2 or HA-PIPKIγi1 (green) with endogenous active Src (red) at focal adhesions of MDA-MB-231 cells. Lower panels, MDA-MB-231 cells after siRNA transfection were fixed for IF study to examine the co-localization of active Src (red) with paxillin (green). ConsiRNA, control siRNA. Scale bar, 20 μm. B, the active Src localized at focal adhesions in the cell periphery was counted from at least 30 cells in each experiment. Paxillin was used as a focal adhesion marker. Immunoblot shows the knockdown of indicated proteins in the transfected cells. C, MDA-MB-231 or T47D after siRNA transfection for knockdown of PIPKIγ or PIPKIγi2 were culture in soft agar for 1 week. Colonies developed were processed for IF study to examine the effect on Src (red) and talin (green) localization at plasma membrane. Scale bar, 20 μm. D, MDA-MB-231 cells expressing PIPKIγi1 or PIPKIγi2 were cultured in soft agar for 1 week. The colony developed were processed for IF study to examine the co-localization of PIPKIγi2 or PIPKIγi1 (green) with active Src (red) at plasma membrane. Scale bar, 20 μm. E, HEK293 cells transfected with PIPKIγi1 or PIPKIγi2 or PIPKIγi2 mutant alone or co-transfected with Src were cultured in suspension condition overnight followed by immunoprecipitation (IP) to examine the co-immunoprecipitation of Src and talin with PIPKIγ variants or mutant. F, siRNA was used to knock down talin or FAK from MDA-MB-231 cells overexpressing PIPKIγi2. Their oncogenic growth was monitored by culturing the cells in soft agar as described above. All the values are means ± S.D. from three-independent experiments. The error bars represent S.D. (p values are indicated).

In three-dimensional suspension culture, PIPKIγi2 knockdown impaired both Src and talin localization at plasma membrane (Fig. 6C). Src extensively co-localized with PIPKIγi2 at cell-cell contact sites at the plasma membrane (Fig. 6D), whereas other PIPKIγ variants deficient in focal adhesion targeting (PIPKIγi1) were poorly localized with Src at plasma membrane. A PIPKIγi2 mutant deficient in Src phosphorylation and talin binding, thus defective in focal adhesion localization were also poorly co-localized with Src at plasma membrane in suspension culture (not shown). In suspension condition, PIPKIγi2 forms a stable complex with talin that is promoted by Src expression and PIPKIγi2 phosphorylation (Fig. 6E). All of these results indicate that PIPKIγi2 in coordination with talin regulates Src localization at focal adhesions and the plasma membrane in three-dimensional culture. Furthermore, impaired anchorage-independent growth in PIPKIγi2-overexpressing cells after talin knockdown (Fig. 6F) supports the coordinated roles of the focal adhesion molecules, PIPKIγi2, Src, and talin, in oncogenic signaling.

An Interaction between PIPKIγi2 and Src Is Required for Anchorage-independent Growth

PIPKIγi2 is directly phosphorylated by Src (28). PIPKIs often associate with proteins they regulate (12, 20, 22) as such an interaction between PIPKIγ/PIPKIγi2 and Src was explored. The interaction between PIPKIγ variants and Src was observed in their endogenous levels and after co-expression and co-immunoprecipitation (Fig. 7A).

FIGURE 7.

PIPKIγi2 interaction with Src is required for their oncogenic growth control. A, upper panels, endogenous Src was immunoprecipitated (IP) from MDA-MB-231 or MDA-MB-435s cells, and co-immunoprecipitation of PIPKIγ and PIPKIγi2 was examined by immunoblotting. Lower panels, endogenous Src was immunoprecipitated from MDA-MB-231 cells expressing PIPKIγ variants followed by immunoblotting to examine the co-immunoprecipitation of PIPKIγ using anti-HA antibody. B, upper panels, schematic diagram of Src and the domains used for construction of GST fusion proteins. Lower panels, in vitro binding study performed using GST fusion proteins of Src and His-tagged PIPKIγi2 or cell lysates prepared from HA-PIPKIγi2 expressing cells. Bound proteins were examined by immunoblotting. C, HEK293 cells were co-transfected with HA-PIPKIγi2 and Src or its deletion mutants. Co-immunoprecipitation of PIPKIγi2 with Src and its mutant forms was examined by immunoblotting using anti-HA antibody. N-term., N-terminal; C-term., C-terminal. D, MDA-MB-231 cells expressing either PIPKIγi2 or Src or mutant Src alone or co-expressing them were culture in soft agar for 10–12 days before counting the colonies. E, MDA-MB-231 cells co-expressing PIPKIγi2 and Src were infected with lentivirus for the expression of GFP or GFP-C-tail of Src followed by immunoprecipitation of PIPKIγi2 to examine the co-immunoprecipitation of Src. F, these cells were cultured in soft agar for 10–12 days before counting the colonies. All the values are means ± S.D. from three-independent experiments (p values are indicated). NS, nonsignificant p value.

To define whether PIPKIγ directly interacts with Src and to determine the Src region required for PIPKIγi2 interaction, the GST pulldown assays were performed using recombinant GST fusion proteins of Src (kinase domain of Src was not soluble) and His-tagged PIPKIγi2 or cell lysates prepared from HEK293 cells transfected with HA-PIPKIγi2 (Fig. 7B). The full-length Src bound with His-tagged PIPKIγi2 or HA-PIPKIγi2, indicating that either the kinase domain or the C terminus mediates the Src interaction with PIPKIγi2. Co-expression and co-immunoprecipitation studies demonstrated that the Src C-terminal deletion mutant was defective in PIPKIγi2 binding (Fig. 7C), indicating that the C terminus is necessary for this interaction. The constitutively active Src (Y527F mutation) showed a modest increase in PIPKIγi2 interaction (Fig. 7C). Csk kinase that phosphorylates Src in Tyr⁵²⁷, promoting an autoinhibitory intramolecular interaction and Src inactivation, binds to C terminus of Src (29). However, PIPKIγ expression did not affect the Src association with Csk nor Tyr⁵²⁷ phosphorylation of Src (not shown).

Deletion of the Src C terminus leads to constitutive activation of Src (42). Remarkably, the co-expression of the C-terminal deletion mutant of Src with PIPKIγi2 failed to induce anchorage-independent growth in synergy with Src (Fig. 7D), indicating the importance of the PIPKIγi2 interaction with Src. Further, the expression of the GFP fusion protein with C-terminal tail of Src (GFP-C-tail) inhibited the localization of active Src at focal adhesions (not shown) and PIPKIγi2 interaction with Src (Fig. 7E). It also inhibited the anchorage-independent growth induced by co-expression of PIPKIγi2 and Src (Fig. 7F).

The highly conserved basic residues in the N terminus of Src play an important role in Src function and its recruitment to plasma membrane via electrostatic interaction with anionic phospholipids, including phosphatidylinositol 4,5-biphosphate and others (32). However, the precise role of phosphatidylinositol 4,5-biphosphate in Src function is not defined. As shown in Fig. 8A, mutations of all of these basic residues to neutral amino acids impaired Src association with PIPKIγi2 (Fig. 8B) and the ability of Src to induce oncogenic growth in synergy with PIPKIγi2 (Fig. 8C). Furthermore, kinase dead PIPKIγi2 showed impairment in inducing anchorage-independent growth in synergy with Src (Fig. 5C). Taken together, these results indicate a coordinated role of PIPKIγi2 and Src and phosphatidylinositol 4,5-biphosphate generation in oncogenic signaling and anchorage-independent growth (illustrated in schematic diagram in Fig. 8D).

FIGURE 8.

Highly conserved basic amino acids residues in the N terminus of Src are required for oncogenic signaling in synergy with PIPKIγi2. A, schematic diagram depicting the highly conserved basic amino acid residues in N terminus of Src. Mutant forms of the chicken Src used for the study are indicated. N-term., N-terminal; C-term., C-terminal. B, HEK293 cells were co-transfected with PIPKIγi2 and Src or its mutant forms. Cells were harvested 24 hours post-transfection to immunoprecipitate PIPKIγi2. Co-immunoprecipitation of Src was examined by immunoblotting. C, Src or its mutant forms were expressed either individually or with PIPKIγi2 into NIH3T3 cells using lentivirus. 24–48 h post-infection, cells were harvested and cultured in soft agar for 10–12 days followed by counting of the colonies formed. All the values are means ± S.D. from three-independent experiments. The error bars represent S.D. (p values are indicated). NS, nonsignificant p value. D, schematic diagram depicting the collaborative function of PIPKIγi2, Src, and talin in oncogenic growth. PIPKIγi2 simultaneously binds with Src and talin independent of cell adhesion. PIPKIγi2 interaction with Src at its C terminus may help to alleviate intramolecular constraints, promoting Src activation. In turn, this enhances PIPKIγi2 interaction with talin, which promotes PIPKIγi2 and Src to localize at the proximity of integrin and growth factor receptors promoting oncogenic growth signaling.

DISCUSSION

The ability to grow in an anchorage-independent manner is one of the fundamental properties of tumor cells and is a key for metastasis, although the underlying mechanisms are poorly understood. Here, we show that the focal adhesion-targeted, phosphatidylinositol 4,5-biphosphate-synthesizing enzyme PIPKIγi2 coordinates with the pro-oncogenic molecule, Src, and the cytoskeletal adaptor molecule, talin, to regulate oncogenic growth of tumor cells. This is consistent with results showing that PIPKIγ expression correlates with poor breast cancer patient survival (25) and supports a potential role for PIPKIγ and PIPKIγi2 in tumor progression.

The activation of Src is a hallmark of many tumors, and several mechanisms are reported for Src activation (29, 30). Inactive Src largely remains in the perinuclear region, whereas active Src is targeted to the plasma membrane/cell adhesion sites in an actin-dependent manner (30). Myristoylation is required for Src recruitment to the plasma membrane and its activation (31). In addition, highly conserved basic residues at the N terminus of Src play an important role in Src function and its recruitment to the plasma membrane via electrostatic interaction with anionic phospholipids (32). Phosphoinositides, including phosphatidylinositol 4,5-biphosphate, constitute the pivotal lipid molecules in the plasma membrane that play key roles in the recruitment of signaling molecules possessing phosphoinositide-binding motifs and/or domains (43). The inability of Src mutants (substitution of basic residues to neutral) to interact with and function in synergy with PIPKIγi2 to induce anchorage-independent growth strongly suggests the phosphoinositide regulation of Src function and functional integration of Src into phosphoinositide signaling pathways. Binding data indicate that PIPKIγi2 interacts with the C terminus of Src. This interaction may relieve the intramolecular interaction between the C-tail and Src homology 2 domain, facilitating Src activation. However, PIPKIγi2 interaction with Src C terminus did not abrogate the Src association with Csk nor Csk phosphorylation of Src. In the absence of PIPKIγ/PIPKIγi2 expression or in cells expressing low levels of PIPKIγ, Src may remain in an inactive state. PIPKIγ/PIPKIγi2 might be playing a role in Src activation as well as its targeting to plasma membrane/focal adhesion sites in coordination with talin. This is consistent with the results that show a requirement for PIPKIγi2 and talin in localization of Src at focal adhesions in adherent conditions and at cell-cell contact sites in three-dimensional culture.

Talin, Src, and PIPKIγi2, are all targeted to focal adhesions in adherent cells. In the absence of cell-matrix interactions in suspension culture, all of them are assembled into a complex, presumably at the vicinity of integrins and growth factor receptors that sustain oncogenic signaling necessary for anchorage-independent growth. In three-dimensional culture, their mutual interaction may promote their organization at cell-cell contact sites in the plasma membrane, although their major fractions remain in cytosol (22, 44, 45). Furthermore, Src is required for PIPKIγi2 association with talin in different tumor cells (not shown), and Src expression significantly promoted PIPKIγi2 association with talin. Conversely, talin interaction with actin is regulated by phosphatidylinositol 4,5-biphosphate (46), suggesting their integrative and collaborative function. As talin is emerging as a potential regulator of oncogenesis, Src regulation of PIPKIγi2 interaction with talin is fully consistent with a collaborative role in anchorage-independent growth (5, 47, 48).

In the plasma membrane, PIPKIγi2 and activated Src may induce oncogenic signaling that contributes to anchorage-independent growth. Talin, a cytoskeletal protein and phosphatidylinositol 4,5-biphosphate effector protein, is an important component of this signaling nexus because it may selectively promote the PIPKIγi2 targeting to the plasma membrane in suspension culture because of its ability to interact with actin cytoskeleton. With these results, we have uncovered the mechanism of how focal adhesion molecules PIPKIγi2, Src, and talin converge into a signaling complex to support oncogenic growth of tumor cells. This could be an oncogenic axis required for in vivo tumor growth and metastasis, including that of triple negative breast cancers, where PIPKIγ and Src are predominantly overexpressed (23, 49). Furthermore, the elucidation of oncogenic signaling molecules downstream of PIPKIγi2 and Src and their functional relevance in vivo are future directions for study.