Fibroblast Growth Requires CT10 Regulator of Kinase (Crk) and Crk-like (CrkL)

Abstract

CT10 regulator of kinase (Crk) and Crk-like (CrkL) are the cellular counterparts of the viral oncogene v-Crk. Elevated levels of Crk and CrkL have been observed in many human cancers; inhibition of Crk and CrkL expression reduced the tumor-forming potential of cancer cell lines. Despite a close relationship between the Crk family proteins and tumorigenesis, how Crk and CrkL contribute to cell growth is unclear. We ablated endogenous Crk and CrkL from cultured fibroblasts carrying floxed alleles of Crk and CrkL by transfection with synthetic Cre mRNA (synCre). Loss of Crk and CrkL induced by synCre transfection blocked cell proliferation and caused shrinkage of the cytoplasm and the nucleus, formation of adherens junctions, and reduced cell motility. Ablation of Crk or CrkL alone conferred a much more modest reduction in cell proliferation. Reintroduction of CrkI, CrkII, or CrkL individually rescued cell proliferation in the absence of the endogenous Crk and CrkL, suggesting that Crk and CrkL play overlapping functions in regulating fibroblast growth. Serum and basic FGF induced phosphorylation of Akt, MAP kinases, and S6 kinase and Fos expression in the absence of Crk and CrkL, suggesting that cells lacking Crk and CrkL are capable of initiating major signal transduction pathways in response to extracellular stimuli. Furthermore, cell cycle and cell death analyses demonstrated that fibroblasts lacking Crk and CrkL become arrested at the G₁-S transition and undergo a modest apoptosis. Taken together, our results suggest that Crk and CrkL play essential overlapping roles in fibroblast growth.

Introduction

The chicken tumor virus oncoprotein, v-Crk, induces a substantial increase in protein tyrosine phosphorylation through protein-protein interactions mediated by its SH2² and SH3 domains, leading to cell transformation (1). CrkII and Crk-like (CrkL) represent closely related cellular homologues of v-Crk and have similar structures and functions (2). CrkI is a splice variant of CrkII that lacks the C-terminal SH3 domain. Crk family proteins (CrkI, CrkII, and CrkL) interact with many proteins through their SH2 and SH3 domains. Overexpression of Crk family proteins induces transformation of cultured cells (3, 4). CrkI is a more potent transforming gene than CrkII and CrkL in part because, like v-Crk, it lacks a C-terminal regulatory phosphorylation site (3). Overexpression of Crk and CrkL has been reported in several human cancers, including oral squamous cell carcinoma (5), ovarian carcinoma (6), colon cancer (7), lung cancer (8, 9), breast cancer (10, 11), gastric cancer (12), and glioblastoma (13, 14). Reduced expression of either Crk or CrkL by RNA interference-mediated knockdown lowered the in vivo tumor formation of human ovarian (15), synovial sarcoma (16), glioblastoma (14), breast cancer (10), head and neck squamous cell carcinoma (17), and rhabdomyosarcoma (18) cell lines. Taken together, these reports imply that elevated levels of Crk family proteins promote cell transformation and enhance tumor cell growth (for review see Refs. 19 and 20).

To investigate functions of endogenous Crk and CrkL in biological processes at the cellular level, we developed mouse strains and cell lines harboring individual and combined floxed alleles of Crk and CrkL. Application of the Cre-loxP recombination-based conditional knock-out approach to cultured fibroblasts enabled us to demonstrate that endogenous Crk and CrkL are important for maintaining cell structural integrity and proper cell motility (21). We also discovered that Crk and CrkL are required for cell transformation induced by viral oncogenes (22). Here we utilized the conditional knock-out system to examine whether endogenous Crk and CrkL are required for cell growth. Our findings clearly demonstrate that Crk and CrkL play essential, overlapping roles in fibroblast proliferation by enabling cells to progress from the G₁ phase to the S phase. Our study also demonstrates that expression of any one protein among CrkI, CrkII, and CrkL at physiological levels is sufficient to secure cell proliferation.

Results

Crk and CrkL Are Essential for Fibroblast Proliferation

We used the Neon system (Life Technologies) to achieve rapid and efficient gene expression in growing fibroblasts by transduction of synthetic mRNA (synRNA). To reduce innate immune responses and increase ectopic protein expression, we synthesized mRNA using the modified ribonucleotides, pseudo-UTP, and methyl-CTP (23, 24). When fibroblasts immortalized by T antigen or the 3T3 protocol were transfected with synthetic green fluorescent protein mRNA (synGFP), both cell types in monolayer cultures showed evidence of green fluorescence at 1 day post transduction (DPT) (Fig. 1A). As cells proliferated, the intensity of fluorescence in individual cells gradually decreased (Fig. 1A). Similarly, the protein level of GFP in the cell population as measured by Western analysis displayed a similar trend (Fig. 1B). Time-lapse analysis of live fibroblasts transfected with synGFP showed strong GFP signals at 5 h post transduction, with the fluorescence reaching a maximum around 24 h post transduction, and thereafter, it gradually declined (Fig. 1, C and D). For quantification of the synRNA transfection efficiency, T antigen-immortalized fibroblasts were transfected with synGFP and fixed at different time points, and their nuclei were stained with DAPI. As shown in Fig. 1, E and F, 99 and 98% of fibroblasts were GFP-positive at 1 and 2 DPT, respectively. At 3 DPT, 69% of cells were positive for GFP. In contrast, only 5% of fibroblasts were GFP-positive at 2 and 3 DPT after GFP DNA transfection. The decrease in average GFP fluorescence intensity per cell over time is likely a consequence of the rapid proliferation of fibroblasts (Fig. 1G), thereby diluting the signal as the total intensities for all the fields of view (intensity × cell count) and the entire well (intensity x WST-1) stayed relatively constant over time (Fig. 1H). These observations suggest that electroporation of fibroblasts with synRNA is an effective method for rapid and efficient introduction of exogenous proteins into growing fibroblasts.

FIGURE 1.

Efficient and rapid expression of GFP after synGFP transfection. A–C, Crk/CrkL double-floxed fibroblasts immortalized by T antigen or the 3T3 protocol were transfected without synRNA (None) or with synGFP (2 μg in 10 μl of cell suspension). A, both phase-contrast and GFP images of cells were taken at the indicated DPT. B, total cell lysates were obtained from cells transfected with synGFP for Western blot analyses. Protein levels of GFP at different DPT were compared. Vinculin levels were used as a control. C and D, T antigen-immortalized Crk/CrkL double-floxed fibroblasts were transfected with synGFP (2 μg in 10 μl cell suspension). At 5 h post transduction (HPT), GFP-positive cells were chosen, and time-lapse images of GFP and phase contrast were recorded every 4 h. The number of GFP-positive cells and their sum fluorescence intensities in the field of view were calculated using the Object Count function of the Nikon NIS element program. Four separate experiments were done, and a representative field of view was chosen for both image display (C) and GFP fluorescence quantification (D). E–H, T antigen-immortalized Crk/CrkL double-floxed fibroblasts were transfected without synRNA (None), with synGFP (2 μg in 10 μl cell suspension) or with GFP DNA (2 μg in 10 μl cell suspension). E, cells were fixed at the indicated DPT and stained with DAPI to visualize the nucleus. Representative images of GFP and DAPI are shown. Scale bar: 100 μm. F, the number of GFP-positive cells was counted (see “Experimental Procedures”). G, proliferation of fibroblasts after transfection was measured using WST-1. H, the average intensity of GFP fluorescence per cell for the synGFP-transfected cells (intensity) was calculated by combining all the sum fluorescence intensities in all the fields of view and dividing it by the total number of DAPI-positive objects. The total intensity was obtained by multiplying the average GFP fluorescence intensity per cell by the total number of DAPI-positive objects (intensity × cell count) or by the WST-1 measurement (intensity × WST-1).

To investigate the contribution of endogenous Crk and CrkL to cell growth, we immortalized fibroblasts from Crk/CrkL double-floxed mice using SV40 T antigen. We transfected the fibroblasts with synthetic Cre mRNA (synCre) to induce a rapid and efficient ablation of Crk and CrkL. As shown in Fig. 2A, staining of fibroblasts with crystal violet indicated substantial increases in cell number between 1 and 3 DPT for control and synGFP-transfected cells. In contrast, the cell number did not increase in cells transfected with synCre. Quantitative analyses of cell proliferation using WST-1 revealed similar exponential rates of cell proliferation in control and synGFP populations (slopes of 1.25 and 1.30, respectively) (Fig. 2, B–D). In contrast, synCre cells failed to proliferate (a slope of 0.05). Western analyses of cell lysates indicated that synCre induced efficient ablation of CrkI, CrkII, and CrkL. The steady state levels of Crk family proteins and other cellular proteins increased as control and synGFP cells recovered from trypsin-EDTA treatment and electroporation and continued to grow (Fig. 2E), probably due to increased cell-to-cell connections. It has been reported that expression and tyrosine phosphorylation of many cellular proteins, including focal adhesion proteins, increase at higher cell densities in Balb 3T3 fibroblasts (25). In contrast, in synCre cells, the levels of Crk family proteins declined after electroporation (Fig. 2, E–G). Tyrosine phosphorylation of p130Cas, a well known protein that associates with Crk family members, decreased as Crk family proteins disappeared from the cells (Fig. 2, E and H). Phosphorylation of Akt was also reduced in the absence of Crk and CrkL (Fig. 2, E and I). Although the level of β-actin showed a modest decrease, vinculin expression was not affected by the loss of Crk and CrkL (Fig. 2, J and K). The requirement of Crk and CrkL for cell proliferation and p130Cas phosphorylation was also observed in fibroblasts immortalized by the 3T3 protocol, which grew more slowly than T antigen-immortalized fibroblasts (data not shown). Our results indicate that endogenous Crk and CrkL are required for proliferation of immortalized fibroblasts.

FIGURE 2.

Lack of cell proliferation in the absence of Crk and CrkL. T antigen-immortalized Crk/CrkL double-floxed fibroblasts were transfected without synRNA (None) or with synGFP (GFP) or synCre (Cre) (2 μg in 10 μl of cell suspension). A, crystal violet images of cells were taken at the indicated DPT. Whereas control and synGFP-transfected fibroblasts had substantial proliferation, synCre-transfected cells failed to proliferate. Scale bar: 500 μm. B and C, cell proliferation was quantitatively measured using WST-1, and the A_450–690 values are presented in linear (B) and logarithmic (C) scales. D, exponential trend lines for the WST-1 assay graphs were drawn, and their slopes, the coefficients of x, are presented as the rates for exponential cell growth. E, total cell lysates were obtained at the indicated DPT for Western blot analyses. Protein levels were compared among cells transfected without synRNA (None) or with synGFP (GFP), or synCre (Cre). F–K, protein bands were quantified using the Odyssey system and calculated as percentages of the control (none at 3 DPT), and their mean ± S.D. values are shown in the graph.

Morphological Alteration of Fibroblasts in the Absence of Crk and CrkL

Staining of fibroblasts with Crk and CrkL antibodies suggests that Crk is expressed in the nucleus and cytoplasm with a higher expression in the nucleus, whereas CrkL is expressed in both the cytoplasm and the nucleus at similar levels (Fig. 3B, data not shown). When cells were transfected with synCre, the level of Crk substantially declined. On the other hand, the decrease in the CrkL level was not obvious in part because a polyclonal antibody for CrkL was used for the staining, although Western blot analysis clearly showed a substantial change in CrkL levels between the control and synCre cells (Fig. 2, E and G, versus Fig. 3B). Staining polymerized actin with phalloidin indicated that actin stress fiber formation was defective in the absence of Crk and CrkL. Cytoplasmic staining with a heat shock protein 90 antibody showed well developed cellular processes in control fibroblasts that were lost in the absence of Crk and CrkL, resulting in a more rounded appearance (Fig. 3, A and B). Without Crk and CrkL, cells formed tight clusters (Fig. 3A) that could be dissociated by treatment with trypsin-EDTA (data not shown), suggesting that they are not multinucleated cells. Because overexpression of Crk caused breakdown of adherens junctions and cell spreading as well as decreased levels of two key components of the adherens junction, E-cadherin and p120-catenin, in Madin-Darby canine kidney epithelial and non-small cell lung cancer cells (26, 27), we tested whether adherens junction formation changed in fibroblasts lacking both Crk and CrkL. As shown in Fig. 3B, staining of control fibroblasts with a p120-catenin antibody did not exhibit any junctional structure between neighboring cells. In contrast, staining of synCre-transfected fibroblasts with the same antibody showed well developed adherens junctions between the neighboring cells (yellow arrows in Fig. 3B). However, the E-cadherin antibodies we tested did not show such a clear formation of adherens junctions. On the other hand, a pan-cadherin antibody displayed cell-to-cell junctions in both control and synCre-transfected cells, and the junctions looked more concentrated in synCre cells (white arrows in Fig. 3B). It is unclear whether they were adherens junctions because the shapes (white arrows) appeared to be different from the adherens junctions stained with p120-catenin (yellow arrows). Our results are consistent with Crk-induced breakdown of adherens junctions reported by Lamorte et al. (26) and suggest that adherens junctions were formed between neighboring cells in the absence of Crk and CrkL, contributing to formation of cell clusters. Furthermore, both nuclear and cytoplasmic areas of the cells significantly decreased in the absence of Crk and CrkL (Fig. 3, C and D). In addition, the cytoplasmic circularity increased in the absence of Crk and CrkL (Fig. 3E), suggesting that cells shrank and became rounded. Next, we examined whether the altered cell structure and formation of adherens junctions affect cell motility. The scratch wound-healing assay revealed directional migration of wild-type fibroblasts to the wound site for 24 h with 44 ± 4% of cells located in the region of interest (ROI) (Fig. 4, A–C). In contrast, only 4 ± 4% of synCre-transfected fibroblasts migrated to the wound site. These results confirm our previous findings (21) and demonstrate that Crk and CrkL play essential roles in cell structure, formation of cell-to-cell junctions, and cell motility.

FIGURE 3.

Morphological alterations of cells in the absence of Crk and CrkL. T antigen-immortalized Crk/CrkL double-floxed fibroblasts were transfected without synRNA (None) or with synCre (0.2 μg in 10 μl cell suspension). At 3 DPT, cells were fixed and stained with antibodies and DAPI to visualize distribution of proteins and identify the cytoplasm and the nucleus. A, cells were immunostained with a HSP90 antibody to visualize the cytoplasm. Scale bar: 50 μm. B, cells were immunostained with Crk, CrkL, p120-catenin, pan-cadherin, and E-cadherin antibodies to visualize distribution of the proteins. Cells were also co-stained with a HSP90 antibody and phalloidin to visualize the cytoplasm and actin stress fibers, respectively. Yellow and white arrows indicate formation of cell-to-cell junctions. C, areas of DAPI-stained objects were calculated (see “Experimental Procedures”), and their mean ± S.D. values are shown. D, areas of HSP90-stained objects were calculated (see “Experimental Procedures”), and their mean ± S.D. values are shown. E, the circularity (roundness) of the cytoplasm (Cyto) and nucleus (Nuc) was measured (see “Experimental Procedures”), and their mean ± S.D. values are shown. ***, p < 0.001; **, p < 0.01, compared with control.

FIGURE 4.

Reduced cell motility in the absence of Crk and CrkL. T antigen-immortalized Crk/CrkL double-floxed fibroblasts were transfected without synRNA (None) or with synCre (0.2 μg in 10 μl cell suspension). At 2 DPT, a wound was created by scratching through the cell monolayer with a micropipette tip, and cells were allowed to migrate into the gap at the wound site for 24 h, fixed, and stained with a HSP90 antibody and DAPI to visualize the cytoplasm and nucleus. A, after fluorescence images were taken, a ROI of 2230 × 1000 μm as the original wound gap was drawn (red rectangles) using the Nikon NIS element program. Six separate experiments were done for each group, and representative images are shown. Scale bar: 500 μm. B, DAPI-stained objects were counted using the Nikon NIS element program for the whole image and for the ROI. Data are shown as the mean ± S.D. (bars) values. C, the percentage of cells that migrated to the ROI was calculated for each image. Data are shown as the mean ± S.D. (bars) values. ***, p < 0.001, compared with control.

Dose-dependent Effects of synCre

The highly efficient transduction of cells with synCre raised a potential concern of nonspecific toxicity of CRE. Previously, other groups associated high levels of CRE expression with inhibition of cell growth and cytopathic effects (28, 29). Therefore, we conducted a dose-response analysis of synCre in wild-type and Crk/CrkL double-floxed fibroblasts. Transfection of wild-type fibroblasts with synCre up to 0.2 μg did not affect exponential growth of cells (Fig. 5, A and C). However, use of 0.66 μg and more of synCre reduced proliferation of wild-type cells in a dose-dependent manner. In contrast, as little as 0.066 μg of synCre inhibited proliferation of Crk/CrkL double-floxed cells, and at 0.2 μg there was essentially no growth (Fig. 5, B and C). Crystal violet staining of Crk/CrkL double-floxed cells showed that both inhibition of cell proliferation and morphological alteration occurred at a level of 0.2 μg synCre, indicating that both phenotypes are related (Fig. 5D). Examination of CrkI, CrkII, and CrkL indicated that all were effectively ablated using as little as 0.2 μg of synCre in double-floxed cells (Fig. 5, E–G). Furthermore, the changes in the growth rate and in the Crk family protein levels were closely correlated in Crk/CrkL double-floxed cells (Fig. 5, C versus F and G), suggesting that Crk family proteins play an essential role in cell proliferation. Phosphorylated forms of p130Cas and Akt were decreased to similar extents by 0.2 μg of synCre (Fig. 3, H and I). On the other hand, the decrease in β-actin levels induced by 2 μg synCre in wild-type cells seems to reflect a nonspecific effect of synCre. These results suggest that 0.2 μg is the optimal amount of synCre to investigate the effects of Crk/CrkL ablation without confounding nonspecific effects of high levels of CRE. Therefore, we chose to use 0.2 μg of synCre for further analyses.

FIGURE 5.

synCre concentration-dependent inhibition of cell proliferation. Wild-type (WT, A) and Crk/CrkL double-floxed (DF, B) fibroblasts immortalized by T antigen were transfected with increasing concentrations of synCre (μg in 10 μl cell suspension), and their cell proliferation was measured using WST-1. C, an exponential trend line for the WST-1 assay graph was drawn, and its slope, the coefficient of x, is presented as the rate for exponential cell growth. D, cells that were transfected with indicated concentrations of synCre were fixed at 3 DPT, and crystal violet images of cells were taken. Scale bar: 200 μm. E, total cell lysates were obtained from cells transfected with indicated concentrations of synCre at 3 DPT for Western blot analyses. Protein levels were compared among different synCre concentrations. F–K, protein bands were quantified using the Odyssey system, and their mean ± S.D. values are shown.

Independent Roles of Crk or CrkL in Cell Growth

Because transfection of Crk/CrkL double-floxed cells with synCre ablated both Crk and CrkL, it was unclear which of these proteins was actually required for cell proliferation. To ablate Crk or CrkL alone, we derived immortalized fibroblasts from Crk or CrkL floxed mice and transfected them with synCre. As shown in Fig. 6, A and B, loss of either Crk or CrkL only slightly reduced cell proliferation (10.8% and 4.9% inhibition for single-floxed cells compared with 92.4% for double-floxed cells). Furthermore, loss of either Crk or CrkL did not result in any detectable morphological alterations (Fig. 6C). Western analyses verified the loss of Crk or CrkL alone upon synCre transfection in single-floxed cells (Fig. 6, D and E). It is noteworthy that phosphorylated p130Cas decreased only when cells lost Crk (or Crk plus CrkL). The result suggests that Crk predominantly mediates phosphorylation of p130Cas. Phosphorylation of Akt decreased in the absence of both Crk and CrkL, but it did not decrease significantly when cells lacked either Crk or CrkL, suggesting that both Crk and CrkL contribute to Akt phosphorylation, with either sufficient for the phosphorylation to occur. The results from single-floxed fibroblasts demonstrate that Crk and CrkL play overlapping functions in cell proliferation and that ablation of one does not cause obvious defects.

FIGURE 6.

Cell proliferation in the absence of either Crk or CrkL. A, wild-type (WT), Crk/CrkL double-floxed (DF), Crk floxed (CF), and CrkL floxed (LF) fibroblasts immortalized by T antigen were transfected with synCre (0.2 μg in 10 μl of cell suspension), and their cell proliferation was measured using WST-1. B, an exponential trend line for the WST-1 assay graph was drawn, and its slope, the coefficient of x, is presented as the rate for exponential cell growth. C, crystal violet images of cells transfected without synRNA (None) or with synCre (Cre) were taken at 3 DPT. Scale bar: 200 μm. D, total cell lysates were obtained from cells at 3 DPT for Western blot analyses. Protein levels were compared between control and synCre-transfected cells. E, protein bands were quantified using the Odyssey system, calculated as percentages of the control (None at 1 DPT), and their mean ± S.D. values are shown.

Rescue of Cell Proliferation by Reintroduction of Crk Family Proteins

To assess the contribution of individual Crk family proteins as well as SH2 and SH3 domains to cell proliferation, we co-transfected Crk/CrkL double-floxed cells with synCre plus various Crk and CrkL constructs (Fig. 7A) to introduce exogenous genes in the absence of the endogenous Crk and CrkL. Transfection of Crk/CrkL double-floxed cells with various Crk/CrkL constructs without synCre did not affect cell morphology or proliferation (top panels of Fig. 7, B and C and Fig. 7, D and F). Both the morphological alterations and failure of cell proliferation that were caused by loss of Crk and CrkL upon synCre introduction were effectively reversed when cells were co-transfected with wild-type CrkI, CrkII, or CrkL (bottom panels of Fig. 7, B and C, and Fig. 7, E and F). Western analyses confirmed the loss of endogenous CrkI, CrkII, and CrkL in the presence of synCre plus any one of the Crk family proteins (Fig. 7, G and H). Exogenous CrkII and CrkI, whose protein levels were 145 and 61%, respectively, of the endogenous CrkII level, were able to fully rescue cell proliferation. The full rescue by both CrkII and CrkI suggests that the C-terminal SH3 domain that is absent in CrkI does not play a significant role in cell proliferation. Exogenous CrkL, whose protein level was 108% of the endogenous CrkL, also rescued cell proliferation. Taken together, the results indicate that cells can proliferate normally if any of the Crk proteins is present at a level approximately half that of endogenous CrkII or CrkL.

FIGURE 7.

Recovery of cell proliferation by reintroduction of Crk and CrkL. T antigen-immortalized Crk/CrkL double-floxed (DF) fibroblasts were co-transfected with synCre (0.2 μg in 10 μl cell suspension) plus various Crk and CrkL constructs (2 μg of synRNA in 10 μl cell suspension), and their cell proliferation was measured using WST-1. A, schematic diagrams of various Crk and CrkL constructs. B, crystal violet images of cells transfected with various Crk and CrkL constructs with or without synCre were taken at 3 DPT. Scale bar: 200 μm. C, cells were fixed at 3 DPT and stained with phalloidin and DAPI to visualize actin stress fibers and the nucleus, respectively. D, proliferation of cells transfected with various Crk and CrkL constructs without synCre was measured using WST-1. E, proliferation of cells transfected with various Crk and CrkL constructs plus synCre was measured using WST-1. F, an exponential trend line for the WST-1 assay graph was drawn, and its slope, the coefficient of x, is presented as the rate for exponential cell growth. G, total cell lysates were obtained from cells at 3 DPT for Western blot analyses. Protein levels were compared among control, synCre-transfected cells and cells transfected with synCre plus various Crk and CrkL constructs. H, protein bands were quantified using the Odyssey system, and their mean ± S.D. values are shown in the graph. The protein levels of endogenous CrkII and CrkL in control cells served as references and were assigned 100% each, and relative protein levels of endogenous and exogenous proteins in cells expressing CRE and various Crk and CrkL constructs are normalized to their references. The steady state protein level of endogenous CrkI was ∼5% that of endogenous CrkII in normal cells.

Both CrkII and CrkL, but not CrkI, have conserved tyrosine residues (Tyr-221 and Tyr-207, respectively) in the linker region between the two SH3 domains. Phosphorylation of these tyrosine residues by kinases such as Abl and Arg followed by an intramolecular interaction with the SH2 domain and reduced accessibility of both the SH2 and N-terminal SH3 domains to their binding partners is known to contribute to decreased transforming activities of CrkII and CrkL (30, 31). Therefore, we tested whether phosphorylation of the regulatory tyrosine residues affects cell proliferation. Phosphotyrosine mutants between the N- and C-terminal SH3 domains of CrkII and CrkL, CrkIIY221F and CrkLY207F, rescued cell proliferation substantially but to a lesser extent than the wild-type proteins (Fig. 7, E and F). The substantial rescue by CrkIIY221F suggests that phosphorylation of Tyr-221 may not be required for cell proliferation. Considering that the level of CrkIIY221F was 131% that of the endogenous CrkII, the phosphotyrosine mutant seems to be less effective than the wild-type despite lack of an intramolecular interaction between the tyrosine residue and the SH2 domain. The substantial rescue by CrkLY207F, although its expression level was 32% that of the endogenous CrkL, suggests that phosphorylation of CrkL at Tyr-207 is not essential for cell proliferation. Next we tested whether cells require both SH2 and SH3 domains to proliferate. The presence of both the SH2 and N-terminal SH3 domains of Crk is essential for growth as both CrkIIR38K and CrkIIW169K failed to rescue cell proliferation (Fig. 7, E and F). The modestly higher growth rate of cells expressing CrkIIW275K, a mutant in the C-terminal SH3 domain, may reflect higher residual levels of endogenous Crk family proteins and/or a lower expression level of the mutant. Taken together, our results suggest that CrkI, CrkII, and CrkL play essential overlapping roles in cell proliferation through their SH2 and N-terminal SH3 domains.

Failure of Cell Proliferation Despite c-Jun N-terminal Kinase (JNK) Reintroduction

JNK family members have often been implicated in Crk signaling (32). For example, JNKs and Crk proteins participate in endothelin-1 signaling in cardiomyocytes (33), vascular endothelial growth factor receptor-3-induced signaling in primary human umbilical vein endothelial cells (34), and the Drosophila homolog of the PDGF/VEGF receptor (PVR)-induced thorax closure during Drosophila metamorphosis (35). Cell transformation induced by v-Crk requires JNK activity (36, 37) and Met-induced fibroblast transformation is correlated with Crk-mediated JNK activation (38, 39). Therefore, we examined whether JNK phosphorylation was altered in the absence of Crk and CrkL. In wild-type fibroblasts, JNK phosphorylation was not altered in the presence of up to 0.66 μg synCre (Fig. 8, A and B). In contrast, JNK phosphorylation decreased dramatically in a dose-dependent manner in the presence of synCre, similar to the drop in CrkII and CrkL in double-floxed fibroblasts (Fig. 8, A and B, versus Fig. 5, C, F, and G). Vinculin levels, however, were not affected by synCre (Fig. 8A). Next we tested to see whether the expression of constitutively active JNKs could rescue cell proliferation in the absence of Crk and CrkL. Although transfection of fibroblasts with synthetic mRNA for MKK-JNK1 and MKK-JNK2 (40) induced low levels of protein expression and phosphorylation, transfection of the corresponding DNAs led to higher levels of expression (Fig. 8F). Co-transfection of Crk/CrkL double-floxed fibroblasts with constitutively active JNK1 and JNK2 synRNA or DNA together with synCre led to various levels of expression and phosphorylation of JNK1 and JNK2 (Fig. 8F). Co-transfecting cells with CrkII, but not constitutively active JNK1, rescued proliferation of synCre-transfected cells (Fig. 8, C and D). Surprisingly, we found that expression of activated JNK2 interfered with CRE recombinase activity, resulting in incomplete elimination of Crk proteins in transfected cells. To avoid interference between JNKs and CRE, we first transfected Crk/CrkL double-floxed fibroblasts with synCre then 12 h later we transfected the cells with constitutively active JNKs. Western analyses showed that endogenous Crk family proteins were effectively removed by the initial synCre transfection. The subsequent transfection of cells with constitutively active JNKs enabled expression and phosphorylation of both JNK1 and JNK2 (Fig. 8G); however, expression of constitutively active JNKs was not sufficient to rescue cell proliferation in the absence of Crk and CrkL (Fig. 8H). Furthermore, a second transfection with wild-type, dominant negative, and constitutively active forms of JNK1, JNK2, or both did not rescue cell proliferation in the absence of Crk and CrkL (data not shown). These results suggest that although JNK phosphorylation dramatically declines in the absence of Crk and CrkL, activation of JNKs alone is not sufficient to drive cell proliferation in the absence of Crk and CrkL.

FIGURE 8.

Failed rescue of cell proliferation by JNK activation. A and B, T antigen-immortalized wild-type (WT) and Crk/CrkL double-floxed (DF) fibroblasts were transfected with different concentrations of synCre (μg in 10 μl cell suspension), and protein levels of phosphorylated JNKs were compared. Protein bands for phosphorylated JNKs were quantified using the Odyssey system, and their mean ± S.D. values at different synCre concentrations are shown. C–H, proliferation of T antigen-immortalized Crk/CrkL double-floxed (DF) fibroblasts co-transfected with synCre (0.2 μg) plus CrkII (2 μg synRNA) (D) or different concentrations (μg) of synRNA or DNA for constitutively active JNKs, MKK-JNK1 (D) and MKK-JNK2 (E) was measured using WST-1. F, total cell lysates were obtained from cells co-transfected with synCre plus constitutively active JNKs for Western blot analyses. Protein levels of JNKs and phosphorylated JNKs at 1 DPT, and those of Crk and CrkL at 3 DPT are shown. G, total cell lysates were obtained from cells serially transfected with synCre first and 16 h later with constitutively active JNKs for Western blot analyses. Protein levels of JNKs and phosphorylated JNKs at 1 DPT and those of Crk and CrkL at 3 DPT are shown. H, proliferation of cells that were serially transfected first with synCre (0.2 μg) and 16 h later with different concentrations (μg) of constitutively active JNKs was measured using WST-1.

Capability of Activating Signal Transduction Pathways in the Absence of Crk and CrkL

To gain insights into the mechanism underlying the blockage of cell proliferation in the absence of Crk and CrkL, we examined whether cells lacking Crk and CrkL were capable of responding to extracellular stimuli. Several growth factors have been reported to stimulate proliferation of fibroblasts (41–43). In addition, Crk and CrkL are known to be required for PDGF-stimulated remodeling of the actin cytoskeleton and cell migration (44). At first we cultured fibroblasts in low serum (0.5% FBS) for 4 days. Whereas control fibroblasts grew for 3 days and stopped growing probably due to depletion of nutrients, synCre-transfected cells failed to grow in low serum (Fig. 9A). The addition of EGF, PDGF, and basic FGF stimulated proliferation of control fibroblasts, which became apparent 4 days after growth factor treatment (Fig. 9, B–D). We also tested IGF-I, but it did not stimulate cell proliferation under our experimental conditions (data not shown). Although both PDGF and basic FGF were able to stimulate proliferation of synCre-transfected cells, the effect of EGF on cell proliferation diminished 4 days after treatment. The results suggest that serum and EGF, but not PDGF or basic FGF, require Crk and CrkL to induce cell proliferation. Because cells were capable of proliferating in response to some growth factors, we examined whether major signal transduction pathways function in the absence of Crk and CrkL. When control fibroblasts were stimulated with high serum (20% FBS) after a day with low serum (0.5% FBS), there was a rapid onset of phosphorylation of Akt, MAP kinases (ERK, JNK, and p38), and S6 kinase in 5 min as well as induction of Fos expression in an hour (Fig. 9E). Cells transfected with synCre exhibited similar biochemical changes. Basic FGF, which evoked the highest response in cell proliferation among the tested growth factors, induced phosphorylation of MAP kinases and Fos induction in both control and synCre-transfected cells. These results suggest that cells lacking Crk and CrkL are capable of initiating major signal transduction pathways.

FIGURE 9.

Cell proliferation and activation of signaling molecules in response to serum and growth factors. A–D, T antigen-immortalized Crk/CrkL double-floxed (DF) fibroblasts were transfected without synRNA (None) or with synCre (0.2 μg in 10 μl cell suspension), plated onto 48-well plates (3000 cells/well), and cultured in normal serum (10% FBS) overnight. The medium was then replaced with DMEM plus low serum (0.5% FBS) with or without growth factors (100 ng/ml EGF, 50 ng/ml basic FGF, or 50 ng/ml PDGF), and cell proliferation for 4 days was measured using WST-1. Data are shown as the mean ± S.D. (bars) values. *, p < 0.05, **, p < 0.01, compared with low serum without growth factor treatment. E, T antigen-immortalized Crk/CrkL double-floxed (DF) fibroblasts were transfected with synCre (0.2 μg in 10 μl cell suspension) and cultured in normal serum (10% FBS) for 2 days. The medium was then replaced with DMEM plus low serum (0.5% FBS), and cells were cultured for additional 24 h. Then cells were stimulated with DMEM plus high serum (20% FBS) or 50 ng/ml basic FGF for 5 min or 1 h at 37 °C, and total cell lysates were obtained for Western blot analyses. F–H, protein bands were quantified using the Odyssey system, calculated as percentages of the control (None stimulated with 20% FBS for 1 h), and their mean ± S.D. values are shown. Whereas levels of CrkII, CrkI, CrkL, p130Cas, and phosphorylated p130Cas substantially decreased by synCre transfection, levels of vinculin, α-tubulin, JNK, phosphorylated cofilin, and phosphorylated TOR did not change by synCre transfection.

Cell Cycle Arrest and Apoptosis

We explored whether the failure in cell proliferation observed in the absence of Crk and CrkL was caused by changes in cell cycle. Cells were fixed with ethanol, stained with propidium iodide, and subjected to flow cytometry analysis. Cells transfected with synCre showed significant increases in the G₁ phase and significant decreases in the S and G₂/M phases compared with control cells at 2 and 3 DPT (Fig. 10, A and B). At 3 DPT, cells at G₁ increased from 35.3 ± 1.6% to 45.0 ± 3.7%, whereas cells at S decreasedfrom 22.1 ± 0.6% to 9.8 ± 0.1%. Cells at G₂/M also decreased from 34.0 ± 1.8% to 21.8 ± 1.2% at 3 DPT. These results suggest that cells lacking Crk and CrkL become arrested at the G₁-S transition. In addition, a modest but significant increase of sub-G₁ cells after synCre transfection (from 1.8 ± 0.4% to 6.9 ± 1.5% at 3 DPT) may indicate the cell death as a consequence of Crk and CrkL removal. Western blot analyses did not detect any of the two apoptosis markers, cleaved caspase 3 and cleaved PARV, in synCre-transfected cells at 3 DPT (data not shown). Then we stained live cells with annexin V and propidium iodide to examine whether cells undergo apoptosis and necrosis in the absence of Crk and CrkL. As shown in Fig. 11, A and C, annexin V-positive cells significantly increased after synCre transfection at 3 DPT (from 4.7 ± 0.4% to 9.8 ± 1.3%). On the other hand, propidium iodide-positive cells slightly, but not significantly, increased after synCre transfection at 3 DPT (from 4.3 ± 0.4% to 6.2 ± 1.1%) (Fig. 11, B and D). These results suggest that fibroblasts lacking Crk and CrkL undergo a modest apoptosis.

FIGURE 10.

Cell cycle analysis. T antigen-immortalized Crk/CrkL double-floxed (DF) fibroblasts were transfected without synRNA (None) or with synCre (0.2 μg in 10 μl cell suspension), cultured, and fixed with ethanol at the indicated DPT for flow cytometry analyses using propidium iodide. Experiments were done in triplicate for each group. A, histograms of representative propidium iodide staining are shown. B, percentages of events for cell cycle phases are calculated, and their mean ± S.D. values are shown. *, p < 0.05; **, p < 0.01, compared with control.

FIGURE 11.

Cell death analysis. T antigen-immortalized Crk/CrkL double-floxed (DF) fibroblasts were transfected without synRNA (None) or with synCre (0.2 μg in 10 μl cell suspension), cultured, and harvested at the indicated DPT for flow cytometry analyses using annexin V and propidium iodide to detect apoptosis and necrosis, respectively. Experiments were done in triplicate for each group. A and B, three separate experiments for annexin V staining were done, and representative histograms for annexin V (A) and propidium iodide staining (B) are shown. C and D, percentages of annexin V (AV)-positive (C) and propidium iodide (PI)-positive cells (D) are calculated, and their mean ± S.D. values are shown. *, p < 0.05, compared with control.

Discussion

Because their identification as cellular counterparts of the viral oncogene v-Crk, Crk and CrkL have been extensively studied for their roles in cell growth and tumorigenesis. Overexpression of Crk and CrkL in human cancers and their requirement for in vivo tumor formation by several cancer cell lines suggest that they contribute to cell transformation as well as aberrant growth and motility of tumor cells. However, it was unclear whether Crk and CrkL proteins, expressed at normal endogenous levels, contribute to cell growth and tumorigenesis. Previously, we reported that endogenous Crk and CrkL proteins are important for cell motility and structural integrity and that cells undergo severe structural alterations in the absence of Crk and CrkL (21). Furthermore, we demonstrated that endogenous Crk and CrkL are required for cell transformation induced by viral oncogenes such as v-fos and v-ras (22). Here, we demonstrate that endogenous Crk and CrkL proteins play essential roles in fibroblast growth. Loss of both Crk and CrkL caused a complete failure of cell proliferation, although the absence of either Crk or CrkL only slightly affected that process. Reintroduction of exogenous CrkII, CrkI, or CrkL individually, in the absence of endogenous Crk and CrkL, rescued cell proliferation, supporting the hypothesis that expression of any member of the Crk family proteins is sufficient for cell proliferation. Cell cycle analyses of fibroblasts lacking Crk and CrkL indicate that cells become arrested at G₁. The link between the cell cytoskeleton and cell cycle is not a new idea. Huang and Ingber (45) reported that prevention of human capillary endothelial cells from spreading and subsequent cell shape changes blocked cell cycle progression from G₁ to S, with a failure of the following biochemical changes: an increase in cyclin D1 protein levels, down-regulation of the cell cycle inhibitor p27Kip1, and phosphorylation of the retinoblastoma protein. In addition, pharmacological disruption of the actin cytoskeleton blocked the G₁-S transition in human capillary endothelial cells and mammalian fibroblasts (46, 47). It is unclear whether the decrease in actin protein levels we observed in the absence of Crk and CrkL contributes to the cell cycle arrest, partly because the change in actin levels was not always consistent and correlated with loss of Crk and CrkL unlike protein levels and phosphorylation of p130Cas. Those reports about the importance of the cell cytoskeleton in the cell cycle together with our previous report on the essential roles of Crk and CrkL in maintaining the cell cytoskeleton (21) suggest that Crk and CrkL play a critical role in cell cycle progression by controlling the cell cytoskeleton. Further studies would be needed to define the detailed mechanism. All in all, our previous and present results demonstrate that Crk and CrkL play essential roles in structural integrity, cell motility, cell transformation, and cell cycle in cultured fibroblasts. Our findings might help to develop a novel approach for therapeutic intervention in cancer. A clear understanding of the detailed mechanism responsible for cell proliferation mediated by endogenous or elevated levels of Crk and CrkL together with development of an effective and therapeutically feasible deletion of Crk and CrkL from cells will be essential for this goal. In addition, we need to investigate whether normal and transformed cells are different in the dependence on Crk and CrkL for proliferation.

Cell growth seems to require a threshold level of Crk family proteins; cells proliferate above the threshold, and increased expression does not stimulate cell proliferation further. Each Crk family protein alone is sufficient to drive cell proliferation. Previously, we were able to show that a human podocyte cell line does not distinguish between CrkII and CrkL for nephrine-induced lamellipodia formation, as we observed recovery of lamellipodia formation upon CrkII overexpression in CrkL knockdown cells and upon CrkL overexpression in CrkII knockdown cells (48). In that case, the threshold for Crk family proteins was higher, and removal of one protein caused defects, which could be rescued by overexpressing the other member. This study reaffirms that our approach is very effective in defining distinct and overlapping functions of Crk and CrkL. Many overlapping functions of Crk and CrkL have been uncovered by a close examination of individual and combined knock-out of Crk and CrkL (48–51).

The biochemical changes in the absence of Crk, CrkL, or both (Fig. 6) shed light on analyzing which signaling molecules or signal transduction pathways are involved in Crk/CrkL-mediated cell proliferation. Phosphorylation of p130Cas was rapidly decreased in the absence of both Crk and CrkL, a finding that correlated well with the levels of Crk and CrkL as well as the degree of cell proliferation. However, the observation that p130Cas phosphorylation, but not cell proliferation, was affected by the absence of Crk alone makes it difficult to conclude that p130Cas phosphorylation is a good indicator of Crk/CrkL-dependent cell proliferation. On the other hand, phosphorylation of Akt decreased in the absence of both Crk and CrkL, and the decrease was not significant when cells lacked either Crk or CrkL. However, Akt phosphorylation was not blocked completely in the absence of both Crk and CrkL, and it modestly increased at 3 DPT. Therefore, it is unclear whether Akt activity is directly connected to Crk family protein levels. MAP kinases have been implicated in Crk/CrkL-mediated signaling pathways (33–36, 38, 52). The levels of phosphorylated JNKs were reduced when cells lost Crk and CrkL (Fig. 8). Co-transfection of cells with synCre plus constitutively active JNKs suggested that JNK activation cannot rescue cell proliferation. These steady state levels of protein phosphorylation are different from the results from a short term stimulation of cells with serum and basic FGF. Increased phosphorylation of Akt, MAP kinases (ERK, JNK and p38), and S6 kinase and Fos expression were induced in response to serum and basic FGF regardless of Crk and CrkL ablation (Fig. 9, E–H). It is less clear what causes the difference in steady state levels and short term responses of signaling molecules. On the other hand, cell cycle arrest of fibroblasts lacking Crk and CrkL at G₁ points a new direction for the study of Crk/CrkL, although it is unclear how Crk and CrkL control the cell cycle. The existence of numerous Crk/CrkL-interacting proteins poses a great challenge in identifying specific proteins and pathways downstream of Crk/CrkL that drive cell proliferation. Potentially, Crk and CrkL may orchestrate multiple pathways for coordinated regulation of cell structure, motility, and cell cycle. Therefore, a systematic approach to identifying downstream mediators of Crk and CrkL functions may need to be developed to understand their complex functions. Our findings help provide a useful starting point for these approaches.

Formation of adherens junctions and subsequent cell cluster formation in addition to reduced cell motility in the absence of Crk and CrkL may provide important insights into biological functions of Crk and CrkL. Failure of migrating neurons to pass through earlier born neurons in the preplate region of the developing brain in reeler (53) and brain-specific Crk/CrkL knock-out mice (51) might have been caused by aberrant formation of cell-to-cell junctions among earlier born neurons in addition to defective cell motility of migrating neurons. On the other hand, it is likely that the inability of T cells lacking Crk and CrkL to cross an endothelial barrier in CD4-specific Crk/CrkL knock-out mice is mainly caused by defective chemotaxis and adhesion of migrating T cells but not by potential changes in formation of cell-to-cell junctions among endothelial cells (50). Recent reports about involvement of Crk and CrkL in epithelial-mesenchymal transition (EMT) (54–57) suggest that reduced cell-to-cell connections among tumor cells and among endothelial cells in addition to increased motility and invasiveness of tumor cells may contribute to metastasis and poor prognosis of human cancers with elevated levels of Crk and CrkL. Our studies bring significant insights into Crk/CrkL biology, but extensive and combined in vitro and in vivo studies on Crk and CrkL are still needed to uncover the mechanisms underlying various biological functions of Crk and CrkL in development, tumorigenesis, and metastasis.

Experimental Procedures

Generation of Immortalized Mouse Embryonic Fibroblasts and Electroporation of Cells with Synthetic mRNA

Mouse embryonic fibroblasts were prepared from single or double-floxed Crk/CrkL mouse embryos as previously reported (58). All mouse studies were carried out according to the protocols approved by the Institutional Animal Care and Use Committee at the Children's Hospital of Philadelphia Research Institute. To immortalize fibroblasts, they were either transfected with SV40 large T antigen (22) or passaged according to the 3T3 protocol. Generation of a variety of Crk and CrkL constructs with the epitope tag FLAG was previously described (21). All the FLAG-tagged constructs in the pLVX-IRES-mCherry vector (Clontech) were subcloned into the pcDNA3.1/myc-His B vector (Invitrogen) for in vitro transcription using the T7 promoter. Wild-type, constitutively active, and constitutively inactive JNK1 and JNK2 constructs with the FLAG tag deposited by Roger Davis were obtained from Addgene, and their gene names and plasmid numbers are as follows: Jnk1a1 (13798), Jnk2a2 (13755), MKK7B2Jnk1a1 (19726), MKK7B2Jnk2a2 (19727), Jnk1a1(apf) (13846), and Jnk2a2(apf) (13761). Synthetic messenger RNAs (synRNA) were synthesized in vitro using the Megascript T7 reaction (Invitrogen) together with pseudo-UTP and methyl-CTP and the EPAP poly-A tailing system (Ambion) to increase stability and translation (23, 24). For transfection, fibroblasts were harvested and counted, and 1 × 10⁵ cells were gently mixed with the indicated amounts (μg in 10 μl cell suspension) of synthetic mRNA or DNA, placed in a 10-μl Neon tip, and electroporated with a single pulse at 1350 V for 30 ms using the Neon transfection system (Life Technologies). Transfected cells were seeded onto 48-well plates, 8-well culture slides, 35-mm dishes, and 60-mm dishes for analyses of cell proliferation, immunocytochemistry, protein expression, and flow cytometry, respectively.

Cell Proliferation Assay Using WST-1

Cell proliferation was measured using cleavage of the tetrazolium salt WST-1 (Roche Applied Science) into a water-soluble formazan by cellular enzymes. Cells were seeded in triplicate at 3 × 10³ cells/well onto 48-well plates after transfection. To measure cell proliferation, cells were incubated for 2 h in a CO₂ incubator with 200 μl/well of fresh culture medium plus 20 μl/well of WST-1 stock solution provided by the manufacturer. As a background control, WST-1 was added to the medium without cells. At the end of the incubation, the conditioned medium was mixed for 1 min on a shaker, and absorbances were read using the Infinite M200 PRO NanoQuant microplate reader (Tecan) at 450 and 690 nm as measurement and reference wavelengths, respectively. The reference value was subtracted from the measurement value, and the A_450–690 value was multiplied by 1000 for easy presentation. Results were plotted on a graph using Microsoft Excel with the x axis for DPT and the y axis for A_450–690 in the logarithmic scale. An exponential trend line was obtained by using the Microsoft Excel program, and the slope of the trend line, which is the coefficient of x, is presented as the rate for exponential cell growth.

Crystal Violet Staining

To visualize cells that have grown in multiwell plates, cells were first washed with PBS and fixed with 4% (w/v) paraformaldehyde plus 4% (w/v) sucrose in PBS. Cells were then stained with 0.1% (w/v) crystal violet for 10 min at room temperature. Cells were washed twice with nanopure water and dried completely. Images of stained cells were captured using the Olympus MVX10 macroscope with a DP71 digital color camera and imported into Adobe Photoshop CS3 for analysis.

Immunocytochemistry and Measurement of Cytoplasmic and Nuclear Areas

Immunostaining of fibroblasts cultured on 8-well culture slides was carried out as previously reported (21). Images of immunostained cells were captured using a Nikon 90i microscope equipped with Roper EZ monochrome and DS-Fi1 color cameras. Quantitative assessments of synRNA transfection efficiency and cell morphology were performed by the fluorescence imaging-based method using the Object Count function of the Nikon NIS element program as we described previously (21). To quantify the efficiency of synRNA transfection, both GFP and DAPI images of cells were taken throughout the entire area of the well. Images of a minimum of 24 different fields of view were taken for analysis of each sample. Both DAPI-positive objects and DAPI/GFP-positive objects were counted and used to calculate the percentage of GFP-positive cells. To quantify areas of the cytoplasm and nucleus, GFP and DAPI images of cells were taken throughout the entire area of the well. Areas and numbers of DAPI-positive objects were calculated to obtain the average size of the nucleus for each field of view. Then the total area of cells with GFP signals was divided by the number of DAPI-positive objects to obtain the average size of the cytoplasm for each field of view. The values for all the fields of view (n = 24 for control cells; n = 17 for synCre-transfected cells) were combined to calculate the overall sizes of the nucleus and cytoplasm. High magnification images of cells were captured using an Olympus IX70 inverted microscope equipped with Deltavision deconvolution software at the Cell and Developmental Biology Microscopy Core, Perelman School of Medicine at the University of Pennsylvania. Images were processed using FiJi (59) followed by the Nikon NIS element program.

Scratch Wound Healing Assay

Fibroblasts were transfected with synRNA, plated on culture slides, and cultured for 2 days. Then a wound was created using a micropipette tip to scratch through the middle of the cell monolayer, and migration of cells into the gap at the wound site was monitored for 24 h. Cells were fixed and stained with a heat shock protein 90 (HSP90) antibody and DAPI to visualize the cytoplasm and nucleus of migrated cells. After fluorescence images were taken using a Nikon 90i microscope, a ROI of 2230 × 1000 μm was drawn as the original wound gap using the Nikon NIS element program. DAPI-stained objects were counted using the program for the whole image and for the ROI, and the percentage of cells that migrated to the ROI was calculated for each image.

Western Blot Analysis

Preparation of lysates of fibroblasts and Western blot analyses using the Odyssey infrared imaging system (Li-Cor Biosciences) were described previously (21). Protein bands from at least three Western analyses were quantified using the Odyssey system, and their mean ± S.D. values are shown in the graph. In addition to the previously described antibodies, the following antibodies were used: antibodies for phosphorylated MAP kinases, p-ERK (pTEpY, V8031), p-JNK (pTPpY, V7931), and p-p38 (pTGpY, V1211) were from Promega; anti-JNK1 antibody (554286) was from BD Biosciences; anti-JNK2 antibody (sc-7345) was from Santa Cruz Biotechnology.

Stimulation of Cells with Serum and Growth Factors

Fibroblasts were electroporated with synRNA and plated onto 48-well culture plates at low densities (3000 cells/well) in DMEM plus 10% FBS. One day later the medium was replaced with DMEM plus 0.5% FBS supplemented with growth factors (100 ng/ml EGF (Gibco, PHG0314), 50 ng/ml basic FGF (Gibco, 13256-029), 50 ng/ml PDGF (Millipore, GF149), or 100 ng/ml IGF-I (Gibco, PHG0078)). Cell proliferation was measured every day for 4 days using WST-1 (Roche Applied Science). To examine activation of signal transduction pathways, fibroblasts were electroporated with synRNA, plated onto 60-mm dishes, and incubated with DMEM plus 10% FBS for 2 days. Then the medium was removed, and cells were incubated with DMEM plus 0.5% FBS for 24 h before treatment with 20% FBS or 50 ng/ml basic FGF in DMEM for 5 min or 1 h. Cell lysates were prepared for Western blot analyses using the Odyssey infrared imaging system (Li-Cor Biosciences).

Flow Cytometry

Crk/CrkL double-floxed fibroblasts were electroporated with synCre and plated onto 60-mm culture dishes. Subconfluent fibroblasts were harvested at the indicated time points by trypsin-EDTA treatment and centrifugation at 1600 × g for 5 min and resuspended in cold Dulbecco's PBS. For cell cycle analysis, cells were fixed by adding cold 70% ethanol dropwise to the cell pellet while gently vortexing and incubating on ice for 15 min. The cell pellet was washed twice with Dulbecco's PBS and treated with DNase-free RNase (Roche Applied Science) at room temperature for 10 min. Then, propidium iodide (Calbiochem) was added to the cell suspension to a final concentration of 10 μg/ml and cells were analyzed by flow cytometry using Accuri C6 (BD Biosciences). For cell death analysis, cells were harvested, resuspended in annexin V binding buffer (BD Biosciences), and stained with allophycocyanin (APC)-annexin V (BD Biosciences) and propidium iodide (a final concentration of 0.05 μg/ml) for flow cytometry with Accuri C6 (BD Biosciences).

Statistical Analysis

All quantitative data were presented as mean ± S.D. Statistical analyses of data were carried out using unpaired two-tailed Student's t test for comparison between two experimental groups. Differences were considered to be significant when probability (p) values were <0.05.

Author Contributions

T. P. and T. C. designed the study, analyzed the results, and wrote the paper. T. P. conducted all the experiments. M. K. and T. P. conducted flow cytometry experiments and interpreted the results. All authors analyzed the results and approved the final version of the manuscript.