Lipid Raft–Specific Knockdown of Src Family Kinase Activity Inhibits Cell Adhesion and Cell Cycle Progression of Breast Cancer Cells

Cholesterol-enriched nanodomains called lipid rafts are thought to act as a platform for protein signaling in cells, but the physiologic significance of lipid rafts in cells and tissues has remained unsolved (1, 2). The main point of the present work is to show the physiologic significance of kinase activity in lipid rafts. Src family kinases (SFK) are known to be ubiquitously distributed in the cell membranes and regulate many biological processes (3, 4). To locate with high resolution where the SFK activation occurs in the cell membranes, we developed a genetically encoded transmembrane fluorescent indicator for detecting the SFK activation at the cell membranes and named it TM-Srcus (Fig. 1A; refs. 5, 6). TM-Srcus can monitor the substrate phosphorylation by activated SFK as the decrease in the cyan fluorescent protein (CFP)/yellow fluorescent protein (YFP) emission ratio through fluorescence resonance energy transfer (FRET; Fig. 1A). The detailed principle of this indicator is described in Materials and Methods. The total internal reflection fluorescence imaging of the SFK activation on the plasma membrane by TM-Srcus showed that SFK activation takes place in lipid rafts.

Based on this finding, we herein extended to develop a lipid raft–targeted SFK inhibitory fusion protein (LRT-SIFP) to inhibit SFK activity in lipid rafts. The LRT-SIFP contains the peptide inhibitor of SFK and the targeting sequence for localizing the SIFP to lipid rafts. The significance about the subcellular locations of kinase activity has hardly been studied by conventional inhibition methods such as small interfering RNA (siRNA) and chemical or peptide inhibitors, which ubiquitously inhibit kinase activity in cells. The present LRT-SIFP elucidated the importance of the SFK activation in lipid rafts for function of breast cancer cells.

The previously developed peptide inhibitor of SFK does not affect cell functions of MCF-7 cells derived from human breast cancer, although it has high potency. In contrast to this conventional peptide inhibitor, the present LRT-SIFP inhibits cell adhesion and cell cycle progression of MCF-7 and MDA-MB231 cells. In addition, these inhibitory effects of LRT-SIFP on cell functions are specific for tumor cell lines. The lipid raft–specific knockdown of SFK activity would potentially be useful for selective cancer therapy to prevent tumorigenesis and metastasis of breast cancer cells.

The principle of TM-Srcus. TM-Srcus is a fusion protein that contains a SFK substrate domain, a phosphorylation recognition domain, a flexible linker sequence, and two green fluorescent protein (GFP) mutants. When phosphorylated at the tyrosine in the SFK substrate domain of TM-Srcus by activated SFK, the adjacent phosphorylation recognition domain specifically binds to this phosphorylated tyrosine. When this binding occurs, the distance between CFP and YFP in TM-Srcus is expected to become shorter, and an intramolecular FRET response results (Fig. 1A). The emission ratio is the CFP fluorescence intensity divided by the YFP fluorescence intensity. If FRET is induced, the CFP fluorescence intensity decreases and the YFP fluorescence intensity increases. The emission ratio is decreased as a result. The indicator thereby monitors the SFK activation in vivo as the decrease in the CFP/YFP emission ratio through FRET. To directly detect the SFK activation in the whole region of cell membranes, TM-Srcus has the transmembrane domain that possesses a high affinity for cell membranes. The indicator can thus monitor SFK activation throughout the cytoplasmic surfaces of the cell membranes (Fig. 1A).

Plasmid construction. To construct the genes of TM-Srcus and TM-SrcusY314A, cDNA fragments of a transmembrane domain from a transmembrane phosphoprotein (Cbp, amino acids 1–52; ref. 7), the Y314 domain from a phosphopeptide containing Tyr³¹⁴ of Cbp fused with Ln, the Y314A domain fused with Ln, the SH2 domain from a COOH-terminal Src kinase (Csk, amino acids 80–162; ref. 8), CFP, and YFP with nuclear export signal sequence (NES) from a HIV-derived protein (9) were generated by PCR and cloned into pBlueScript SK(+).

To construct the genes of LRT-SIFP and its alanine and deletion mutants, cDNA fragments of the lipid raft–targeting sequence (COOH-terminal H-Ras, CMSCKCVLS) fused with Ln, the mutated lipid raft–targeting sequence (GGGS with COOH-terminal Rho-A, GCLVL) fused with Ln, the peptide inhibitor for SFK (MIYKYYF) fused with a flag tag, the alanine-mutated peptide inhibitor for SFK (MIYKYAF) fused with a flag tag, and YFP were generated by PCR and cloned into pBlueScript SK(+).

All cloning enzymes (Takara Biomedical) were used according to the manufacturer's instructions. The PCR fragments were sequenced with an ABI310 genetic analyzer (Applied Biosystems). The constructs were subcloned into pcDNA3.1(+) (Invitrogen) behind a Kozak sequence using HindIII and XhoI sites.

Imaging of cells. MCF-7 cells expressing TM-Srcus were starved with a steroid-free medium [phenol red–free Eagle's MEM with 2% dextran-coated charcoal (DCC)–treated serum] for 12 h, then washed twice with HBSS (Sigma). For a total internal reflection fluorescence imaging, cells were observed under a total internal reflection fluorescence microscope (TIRFM; IX70, Olympus) with a charge-coupled device camera CoolSnap ES (Roper Scientific) controlled by MetaFluor (Universal Imaging). The exposure time at 440 ± 10-nm excitation was 300 ms. The fluorescent images of CFP and YFP were obtained with filters at 480 ± 15 and 535 ± 12.5 nm with a 60× oil immersion objective PlanApo60 (Olympus). For a conventional fluorescence imaging, cells were pretreated with PP2 (Calbiochem), stimulated with 17β-estradiol (E₂; Sigma), and imaged at room temperature under a Carl Zeiss Axiovert 135 microscope (Carl Zeiss) with a cooled charge-coupled device camera MicroMax (Roper Scientific) controlled by MetaFluor. The exposure time at 440 ± 10-nm excitation was 200 ms. The fluorescent images were obtained with filters at 480 ± 15 and 535 ± 12.5 nm with a 40× oil immersion objective (Carl Zeiss).

Density gradient fractionation. MCF-7 cells in a 10-cm dish expressing TM-Srcus and LRT-SIFP were scraped in ice-cold HBSS and spun down at 2,000 rpm 4°C, lysed thoroughly by pipetting through a 200-μL yellow pipetting tip with 180 μL TNE [10 mmol/L Tris-HCl (pH 7.6), 500 mmol/L NaCl, 1 mmol/L EDTA], 1% Triton X-100, 10% sucrose, 2 mmol/L sodium orthovanadate, 1% protease inhibitor cocktail (Sigma) at 4°C, and incubated for 20 min on ice. Three hundred sixty microliters of cold 60% Optiprep (Axis-Shield PoC AS) were added to the extract and incubated for 10 min on ice. The mixture of the extract and 60% Optiprep was transferred into thick-walled polycarbonate tubes (Beckman Coulter). The sample was overlaid with a 540-μL step of each of 35%, 30%, 25%, 20%, 0% Optiprep in TNE, 1% Triton X-100, 10% sucrose, 2 mmol/L sodium orthovanadate. The gradients were spun for 4 h at 200,000 × g at 4°C and collected into six 540-μL fractions.

Immunoprecipitation and immunoblot analysis. For immunoprecipitation, the six gradient fractions were diluted with an equal volume of TNE, 1% Triton X-100, 10% sucrose, and 2 mmol/L sodium orthovanadate (10). The diluted samples were immunoprecipitated with anti-phosphotyrosine antibody (PY20, Santa Cruz Biotechnology). For immunoblotting, the samples were separated by electrophoresis on a 10% SDS-acrylamide gel and the electrophoresed proteins were transferred onto a nitrocellulose membrane. The membranes were blocked with 1% dry milk or 3% bovine serum albumin (BSA) in TBST and was incubated with primary antibodies: anti-phosphotyrosine (PY20), anti-Fyn (Fyn 3, Santa Cruz Biotechnology), anti–transferrin receptor (Zymed Laboratories), anti-GFP (Clontech), anti–caveolin-1 (BD Transduction Laboratories), anti-Yes (1B7, WAKO Biochemicals), and anti-Src (GD11, Upstate Biotechnology). The anti–active Src antibody (clone 28) was kindly provided by Dr. K. Owada (Kyoto Pharmaceutical University, Yamashinaku, Kyoto, Japan). Bands were visualized with horseradish peroxidase (HRP)–conjugated antirabbit or antimouse immunoglobulin G (Amersham Life Science).

Cell culture and transfection. MCF-7 cells were cultured in MEM (Sigma) supplemented with 10% fetal bovine serum (FBS), 1% penicillin/streptomycin, and 0.1 mmol/L nonessential amino acids at 37°C in 5% CO₂. MDA-MB231 cells were cultured in DMEM supplemented with 10% FBS and 2 mmol/L l-glutamine at 37°C in 5% CO₂. HEK293 cells were cultured in DMEM supplemented with 10% FBS at 37°C in 5% CO₂. The cells were transfected with LipofectoAMINE 2000 (Invitrogen). Cells were plated 24 to 36 h before transfection onto glass-bottom dishes for fluorescence imaging or onto plastic culture dishes for immunoblot analysis.

Cell staining with cholera toxin B subunit-Alexa647. MCF-7 cells expressing TM-Srcus and LRT-SIFP were washed with HBSS and incubated at 37°C for 1 h with 2 μg/mL cholera toxin B subunit (CTXB)-Alexa647 (Molecular Probes) in MEM supplemented with 0.01% BSA, 1% penicillin/streptomycin, 25 mmol/L HEPES, and 0.1 mmol/L nonessential amino acids (11). After staining, cells were washed twice with HBSS and observed at room temperature under a TIRFM or a confocal microscope. The exposure time at 633 ± 10-nm excitation was 200 ms. The fluorescent images of CTXB-Alexa647 under a TIRFM or a confocal microscope were obtained with a 60× oil immersion objective PlanApo 60 and a 100× oil immersion objective, respectively.

Cell adhesion assay. MCF-7 cells expressing LRT-SIFP were washed with PBS, trypsinized, and resuspended with PBS. Aliquots of the cell suspension (10⁶ per mL) were placed on a 24 × 24-mm micro-cover glass (Matsunami) and on the bottom of a six-well plate (Nunc). After 24-h incubation, 80% confluent cells both on the cover slide and on the bottom of the well were transfected with each of constructed cDNAs encoding the SIFPs and were incubated for 24 h at 37°C. For counting the number of originally plated fluorescence-positive cells, the transfected cells on the cover slide were washed with HBSS and fixed with 3% PFA for 10 min at 4°C. The fluorescence-positive cells on the cover slide were directly counted under a fluorescence microscope (×40). For counting the number of adhered fluorescence-positive cells, the transfected cells on the bottom of the well were trypsinized and resuspended in culture medium. The whole-cell suspension (10⁶ cells/mL) was placed on the cover slide coated with 33 μg/mL fibronectin in PBS overnight at 4°C and was incubated for 4 h at 37°C with medium containing 10% FBS (12). The adhered cells on the fibronectin-coated cover slide were washed with HBSS and fixed with 3% PFA for 10 min at 4°C. The adhered fluorescence-positive cells were directly counted under a fluorescence microscope (×40). The adhesion index is the number of the adhered fluorescence-positive cells divided by that of the originally plated fluorescence-positive cells. The adhesion index of MCF-7 cells with and without incubation with 5 μmol/L PP2 for 4 h at 37°C (Supplementary Fig. S1A) was measured under a microscope (×40) as described above.

Propidium iodide staining and fluorescence-activated cell sorting analysis. MCF-7 cells (1.5 × 10⁶) expressing LRT-SIFP were washed with PBS, trypsinized, and resuspended with ice-cold PBS. The resuspended cells were fixed and permeabilized with 70% ethanol for 30 min at 4°C. The permeabilized cells were washed twice with PBS and incubated with 100 μg/mL RNase A (Qiagen) in PBS for 30 min at room temperature, and then subjected to propidium iodide staining (50 μg/mL) for 30 min at 4°C. The cell cycle profile of MCF-7 cells with and without incubation with 5 μmol/L PP2 (Supplementary Fig. S2) was obtained as described above.

MCF-7 cells (1.5 × 10⁶) expressing SIFP or YFP were fixed with 1% PFA in PBS for 15 min at 4°C to avoid a drain of the expressing protein from the cells through permeabilization. The PFA-fixed cells were then permeabilized and stained with propidium iodide using 500 μL of low-salt solution (0.1% Triton X-100, 3% polyethylene glycol-8000, 50 μg/mL propidium iodide, 100 μg/mL RNase, 4 mmol/L sodium citrate, pH 7.1–7.2) for 30 min at 4°C, and subsequently added with an equal volume of high-salt solution (0.1% Triton X-100, 3% polyethylene glycol-8000, 50 μg/mL propidium iodide, 100 μg/mL RNase, 400 mmol/L NaCl, pH 7.1–7.2) for 30 min at 4°C (13). The propidium iodide–stained cells were analyzed by fluorescence-activated cell sorting (Beckman Coulter).

TM-Srcus can monitor SFK activation induced by a physiologic dose of 10 nmol/L estrogen in living MCF-7 cells. We constructed cDNA that encodes TM-Srcus (Fig. 1A,, right) and introduced the cDNA in MCF-7 cells, which are derived from human female breast cancer cells. TM-Srcus was observed in the plasma membrane and endomembranes (i.e., ER and Golgi membranes) owing to its transmembrane domain (Fig. 1A and B). We stimulated the cells expressing TM-Srcus with a physiologic dose of 10 nmol/L E₂, which is known to activate SFK through estrogen receptor outside the nucleus (14). The CFP/YFP emission ratio of TM-Srcus decreased on 10 nmol/L E₂ stimulation, and this decrease in the emission ratio reached a plateau within 20 min (Fig. 1B). After pretreatment of the cells with a specific SFK inhibitor, 5 nmol/L PP2 for 1 h, no significant change in the CFP/YFP emission ratio was observed even on 10 nmol/L E₂ stimulation (Fig. 1B and C). In addition, 10 nmol/L E₂ did not induce a change in the CFP/YFP emission ratio in the cells cotransfected with TM-Srcus and selective siRNA against Src (Fig. 1B and C). In addition, pretreatment of the cells with antiestrogenic agent ICI 182,780 for 1 h inhibited the E₂-induced change in the CFP/YFP emission ratio of TM-Srcus (Fig. 1B and C). TM-SrcusY314A in which the tyrosine was replaced with alanine that has no phosphoacceptor site (Fig. 1A,, right) did not show any change in the CFP/YFP emission ratio on 10 nmol/L E₂ stimulation (Fig. 1C). These results indicate that TM-Srcus can monitor substrate phosphorylation by activated SFK in MCF-7 cells stimulated with a physiologic dose of 10 nmol/L E₂.

Lipid raft–specific SFK activation in MCF-7 cells revealed by TM-Srcus. We next observed MCF-7 cells expressing TM-Srcus under a TIRFM. TIRFM provides a fluorescent image within a depth of ∼100 nm from the contact regions, enabling a selective visualization of the basal plasma membrane of cells (15). Pseudocolor images that represent the CFP/YFP emission ratio of TM-Srcus showed a blue shift in specific regions of the plasma membrane on 10 nmol/L E₂ stimulation (Fig. 1D). These spatially limited regions of the plasma membrane may be clarified by a course of experiments described in the following. A single lipid raft cannot be observed because of its nanometer size by conventional fluorescence microscopy with wavelength limitation of 200 nm, but a lipid raft cluster consisting of multiple rafts was what we observed. We stained MCF-7 cells with a lipid raft marker, CTXB-Alexa647 (1), and observed the stained cells under a TIRFM. The E₂-induced SFK activation occurred locally in the plasma membranes as was observed in the cells without staining by CTXB-Alexa647 (Figs. 1D and 2,A, left and middle). The blue-shifted regions that represent the SFK activation seem to be located where CTXB is concentrated (Fig. 2,A, middle and right). A scatterplot of the emission ratio of TM-Srcus versus the CTXB fluorescence intensity showed that a decrease in the emission ratio of TM-Srcus is correlated with an increase in the CTXB fluorescence intensity (R = −0.82; Fig. 2B). These results indicate that the activation of SFK on the stimulation of MCF-7 cells with E₂ is provoked in lipid raft clusters in the plasma membrane.

We next carried out a density gradient fractionation that separates the total lysates of the cells expressing TM-Srcus into six gradient fractions (see Materials and Methods; refs. 16, 17). This fractionation method enables the isolation of detergent-resistant membranes (DRM) that are not equated to lipid rafts but include part of a raft-associated protein (18). The gradient fractions were blotted with HRP-conjugated CTXB and anti–caveolin-1 antibody, which are known to bind to positive markers for DRMs, GM1 and caveolin, respectively (19). These gradient fractions were also blotted with antibody against transferrin receptor, a negative marker for DRMs (20). The distribution of the positive and negative markers for DRMs in these fractions was similar to what was previously observed in another epithelial cell line (ref. 20; Fig. 2C and D). We identified fractions 2, 3, and 4 as DRM fractions, and fraction 5 as a non-DRM fraction (Fig. 2C and D). From the immunoblotting of the six fractions with anti-GFP antibody, TM-Srcus was found to distribute in all fractions (Fig. 2C). However, the immunoblotting of these six fractions with anti-phosphotyrosine antibody showed that phosphorylation of TM-Srcus by activated SFK on 10 nmol/L E₂ stimulation is induced only in fraction 4 (Fig. 2C). Furthermore, we carried out immunoblotting of these fractions with anti–active Src and anti-Src antibodies. The anti–active Src antibody is known to selectively recognize the COOH-terminal regulatory domain of the active Src (21). On 10 nmol/L E₂ stimulation, activated Src was localized in the DRM fraction (fraction 4) in which TM-Srcus is phosphorylated (Fig. 2C and D), although Src was distributed in both DRM and non-DRM fractions (Fig. 2D). These results support the above conclusion that E₂ provokes activation of SFK exclusively in lipid rafts in MCF-7 cells. In addition, this lipid raft–specific phosphorylation of TM-Srcus was elicited with other sex steroids, such as progesterone (P4) and androgen 5α-dihydrotestosterone (Fig. 2C), which are known to activate SFK through their receptors (14). It is thus concluded that the sex steroids, E₂, P4, and dihydrotestosterone, provoke activation of SFK only in lipid rafts in MCF-7 cells.

We also examined whether the distribution of SFKs such as Src, Yes, and Fyn in the gradient fractions is changed by E₂ stimulation. We carried out the immunoblotting of the gradient fractions with anti-Src, anti-Yes, and anti-Fyn antibodies. The distribution of Yes and Fyn in the gradient fractions was not changed on stimulation of MCF-7 cells with 10 nmol/L E₂, but Src was found to translocate into DRM fraction on E₂ stimulation (Fig. 2D). These results imply not only that Src in DRM fraction is activated by E₂ but also that other Src located in non-DRM fraction is recruited to DRMs on E₂ stimulation for its activation there.

The detailed structure of the LRT-SIFP. SFK activation has been known to be necessary for cell functions such as cell adhesion to integrin ligands and growth factor–induced mitogenesis (22, 23), which are closely related to growth and metastasis of cancer cells (24) and to homeostasis of normal cells. To determine whether the lipid raft–specific SFK activation has significant effects on these cell functions, we developed a LRT-SIFP that inhibits the SFK activity in lipid rafts. The LRT-SIFP consists of a pseudosubstrate-based peptide inhibitor of SFK (MIYKYYF; ref. 25), YFP, a flag tag, and a lipid raft–targeting sequence (Fig. 3A). The lipid raft–targeting sequence is the COOH-terminal nine amino acids of H-Ras (CMSCKCVLS) to which GFP was fused for visualizing lipid rafts in cells in previous studies (26–28). This targeting sequence is known to be posttranslationally double-palmitoylated at the cysteine residues to localize in lipid rafts in the cells (26, 27).

Characterization of LRT-SIFP. MCF-7 cells expressing LRT-SIFP were stained with the lipid raft marker CTXB-Alexa647 and were observed under a confocal microscope. Lipid raft clusters were observed in endosome-like vesicles and in the plasma membrane (Fig. 3B). LRT-SIFP was colocalized with the lipid raft clusters in the plasma membrane (Fig. 3B). We next analyzed the inhibitory effect of LRT-SIFP on SFK activity in lipid rafts by measuring the FRET response of TM-Srcus in the cells coexpressed with both TM-Srcus and LRT-SIFP. The cDNAs of TM-Srcus and LRT-SIFP lacking YFP were connected by the internal ribosome entry site (IRES) sequence and were transfected into MCF-7 cells (Fig. 3A). The E₂-induced FRET response of TM-Srcus that represents the SFK activation in lipid rafts was completely inhibited by LRT-SIFP (Fig. 3C). This result indicates the inhibitory effect of LRT-SIFP on SFK activity in lipid rafts. This complete inhibition of the E₂-induced SFK activation in lipid rafts by LRT-SIFP and the colocalization of LRT-SIFP with lipid raft clusters in the plasma membrane confirm that the FRET response of TM-Srcus observed with TIRFM (Fig. 2A) is due to the E₂-induced SFK activation in lipid rafts in the plasma membrane, not in the endosome-like vesicles.

In addition, to examine the effect of palmitoylation of LRT-SIFP on SFK activity in lipid rafts, we replaced the targeting sequence of LRT-SIFP with the consensus sequence for geranylgeranylation (CLVL). The GFP fused with this consensus sequence does not cluster in lipid rafts (26). We named this mutated construct mLRT-SIFP (Fig. 3A). The IRES-connected cDNAs of TM-Srcus and mLRT-SIFP devoid of YFP (Fig. 3A) were transfected into MCF-7 cells. The cells coexpressing TM-Srcus and mLRT-SIFP were observed under a conventional epifluorescence microscope. mLRT-SIFP did not inhibit the E₂-induced FRET response of TM-Srcus, which is due to the lipid raft–specific SFK activation (Fig. 3C). This result indicates that the palmitoylation of peptide inhibitor of SFK is necessary for the effective inhibition of the SFK activity in lipid rafts.

The inhibitory effect of LRT-SIFP on SFK activity in lipid rafts was also assessed by the density gradient fractionation method. LRT-SIFP contains the pseudosubstrate-based peptide inhibitor that is preferentially phosphorylated by activated SFK to inhibit the enzymatic activity of SFK (25). We examined whether LRT-SIFP was phosphorylated by SFK in MCF-7 cells. The six gradient fractions from MCF-7 cells expressing LRT-SIFP were immunoprecipitated with anti-phosphotyrosine antibody and were immunoblotted with anti-GFP antibody. LRT-SIFP was phosphorylated only in fraction 4, which is the DRM fraction, and this phosphorylation was blocked by treatment of cells with the specific SFK inhibitor PP2 for 2 h (Fig. 3D). In addition, we carried out immunoblotting of gradient fractions with anti–active Src antibody in the cells expressing LRT-SIFP. Compared with untransfected cells, the expression of LRT-SIFP significantly reduced the amount of active Src in fraction 4 (Fig. 3D). In contrast to active Src, the amount of Src in fraction 4 is not changed by the expression of LRT-SIFP (Fig. 3D). These results support that LRT-SIFP has a specific inhibitory effect on SFK activity in lipid rafts.

Physiologic effects of LRT-SIFP on tumor cell functions. Inhibition of SFK activity is known to suppress cell adhesion of many cell lines (22). We also confirmed that PP2, which ubiquitously inhibits SFK activity in cells, prevents cell adhesion of MCF-7 cells (Supplementary Fig. S1A). We then examined cell adhesion of MCF-7 cells expressing LRT-SIFP. The extent of adhesion of MCF-7 cells transfected with the SIFPs was measured and expressed in terms of the adhesion index, which is defined as the number of adhered fluorescence-positive cells divided by that of originally plated fluorescence-positive cells (see Materials and Methods). LRT-SIFP remarkably reduced the number of cells adhered to fibronectin, which is an integrin ligand (Fig. 4A). LRT-SIFP Del (=LRT-YFP), in which the peptide inhibitor of SFK is deleted (Fig. 3A), did not impair cell adhesion to fibronectin (Fig. 4A). LRT-SIFP Y6A contains the alanine-mutated peptide inhibitor MIYKYAF, whose inhibitory effect is reduced to about one-sixtieth (1/60; Fig. 3A; ref. 25). This LRT-SIFP Y6A did not suppress cell adhesion of MCF-7 cells (Fig. 4A). The reduction of cell adhesion by LRT-SIFP was not observed in the cells coexpressing wild-type (wt) Src and LRT-SIFP (Fig. 4A). In contrast to LRT-SIFP, SIFP, which is the YFP-fused peptide inhibitor of SFK lacking any targeting sequence (Fig. 3A), did not induce the reduction of MCF-7 cell adhesion (Fig. 4A). In addition, mLRT-SIFP, which has the mutated lipid raft–targeting sequence, did not inhibit MCF-7 cell adhesion (Figs. 3A and 4A). A control experiment was done to show that the effects of LRT-YFP and LRT-SIFP Y6A on cell adhesion can be modified by FBS stimulation. Cell adhesion of MCF-7 cells expressing LRT-SIFP and LRT-SIFP Y6A was inhibited in the absence of FBS stimulation (Supplementary Fig. S1B).

We also examined whether LRT-SIFP suppresses cell adhesion of other cell lines. LRT-SIFP significantly inhibited cell adhesion in other breast cancer cell lines such as MDA-MB231 and MCF-7 (Fig. 4B). However, this inhibition of cell adhesion by LRT-SIFP was not observed in the human normal cell line HEK293 (Fig. 4C). These results indicate that the inhibitory effect of LRT-SIFP on cell adhesion is specific for tumor cell lines.

Another effect of inhibition of SFK activity is suppression of the tyrosine phosphorylation of cyclin D2, which is important for cell cycle progression at G₁ phase (29). From cell cycle profiles obtained by flow cytometry with propidium iodide staining, we also confirmed that the 24-h incubation of MCF-7 cells with 5 μmol/L PP2 inhibits cell cycle progression (Supplementary Fig. S2). We next examined whether LRT-SIFP affects the cell cycle of MCF-7 cells. The cell cycle profiles of the YFP-positive cells expressing LRT-SIFPs were obtained by flow cytometry with propidium iodide staining (Fig. 5A). The cell cycle profiles of LRT-YFP, YFP, and mLRT-SIFP were used as a control for cell cycle progression of MCF-7 cells expressing LRT-SIFP, SIFP, and mLRT-SIFP, respectively (Fig. 5A,, C, and D). LRT-SIFP reduced the proportion of cells in G₂-M phase to those in G₁ phase, but LRT-SIFP Y6A did not reduce this proportion (Fig. 5A). As shown in Fig. 5B, LRT-SIFP significantly decreased the percentage of MCF-7 cells in G₂-M phase as compared with LRT-YFP control (P < 0.001). The reduction by LRT-SIFP of the proportion of cells in G₂-M phase to those in G₁ phase was not observed in cells coexpressing wt Src and LRT-SIFP (Fig. 5A). In contrast to LRT-SIFP, the cell cycle profile of YFP-positive cells expressing SIFP, which is the YFP-fused peptide inhibitor of SFK alone, is similar to that of cells expressing YFP control (Fig. 5C). In addition, no difference in cell cycle profile was observed between mLRT-YFP control and mLRT-SIFP (Fig. 5D).

We also examined the effect of LRT-SIFP on cell cycle progression of other cell lines. LRT-SIFP arrested cell cycle at G₁ phase in MDA-MB231 cells as well as in MCF-7 cells (Figs. 5A and 6A). As shown in Fig. 6B, LRT-SIFP significantly decreased the percentage of MDA-MB231 cells in G₂-M phase as compared with LRT-YFP control (P < 0.001). However, LRT-SIFP did not induce cell cycle arrest in normal HEK293 cells (Fig. 6C). These results indicate that the cell cycle of breast cancer cell lines is arrested by lipid raft–specific knockdown of SFK activity in a tumor-specific manner. LRT-SIFP was able to inhibit the SFK activity in lipid rafts, thereby preventing cell cycle progression and cell adhesion of breast cancer cells.

From the pharmacologic point of view, peptide inhibitors of SFK have been developed by several groups (25, 30). These inhibitors are cell permeable and showed their high inhibitory effect on Src activity in in vitro experiments. However, none of these peptide inhibitors showed significant effects on cell functions in in vivo experiments (25). In addition, we showed that one of previously developed peptide inhibitors of SFK, MIYKYYF (=SIFP), did not inhibit cell adhesion and cell cycle progression of MCF-7 cells derived from human breast cancer. In contrast to these conventional peptide inhibitors, the present LRT-SIFP inhibits cell adhesion and cell cycle progression of breast cancer cell lines MCF-7 and MDA-MB231. These results conclude that the lipid raft–targeting of SIFP accelerates its inhibitory effect on SFK activity in lipid rafts, thereby causing the effective inhibition of tumor cell functions. Furthermore, LRT-SIFP did not show these inhibitory effects in normal HEK293 cells, suggesting that the present lipid raft–specific knockdown of SFK activity would potentially be useful for selective cancer therapy to prevent tumorigenesis and metastasis of breast cancer cells.

Grant support: Japan Science and Technology Agency and Japan Society for the Promotion of Science.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

We thank Y. Imai, G. Wakamatsu, and M. Fujishiro for their help with the experiments and Dr. K. Owada for providing the anti–active Src antibody (clone 28).