|Home | About | Journals | Submit | Contact Us | Français|
The upregulation of Src family kinases (SFKs) has been implicated in cancer progression, but the molecular mechanisms regulating their transforming potentials remain unclear. Here we show that the transforming ability of all SFK members is suppressed by being distributed to the cholesterol-enriched membrane microdomain. All SFKs could induce cell transformation when overexpressed in C-terminal Src kinase (Csk)-deficient fibroblasts. However, their transforming abilities varied depending on their affinity for the microdomain. c-Src and Blk, with a weak affinity for the microdomain due to a single myristate modification at the N terminus, could efficiently induce cell transformation, whereas SFKs with both myristate and palmitate modifications were preferentially distributed to the microdomain and required higher doses of protein expression to induce transformation. In contrast, disruption of the microdomain by depleting cholesterol could induce a robust transformation in Csk-deficient fibroblasts in which only a limited amount of activated SFKs was expressed. Conversely, the addition of cholesterol or recruitment of activated SFKs to the microdomain via a transmembrane adaptor, Cbp/PAG1, efficiently suppressed SFK-induced cell transformation. These findings suggest that the membrane microdomain spatially limits the transforming potential of SFKs by sequestering them away from the transforming pathways.
Src family kinases (SFKs) are membrane-associated, non-receptor protein tyrosine kinases involved in a variety of intracellular signaling pathways (5). SFKs are comprised of eight members in mammals: c-Src, Fyn, c-Yes, Lyn, Lck, Hck, c-Fgr and Blk. Among these, c-Src, Fyn, and c-Yes are ubiquitously expressed, whereas the others are relatively concentrated in hematopoietic cell lineages. The intracellular distribution of each SFK also varies depending on their unique N-terminal sequences and acyl modifications (5, 27). These distinctive features of SFKs suggest that each SFK member plays a unique role in particular tissues or cells. In contrast, it is also known that SFKs have redundant and pleiotropic functions in regulating critical cellular events, such as cell division, motility, adhesion, angiogenesis, and survival (26). In a variety of human cancers, protein levels and/or specific activities of c-Src and c-Yes are frequently upregulated (13, 35). Upregulation of Lyn, Lck, Hck, c-Fgr, or Blk is also observed in some leukemias and lymphomas (10, 16, 26). These observations imply a role for SFKs in cell transformation, tumorigenesis, and metastasis (31). However, because SFK genes are rarely mutated in human cancers (31), the mechanisms underlying their upregulation in these cancers remain unclear. Furthermore, the distinctive expression patterns and functional redundancy among SFK members have hampered concurrent analyses of their intrinsic transforming abilities and contribution to cancer progression.
In normal cells, the kinase activity of SFKs is negatively regulated by the phosphorylation of its C-terminal regulatory Tyr residue by C-terminal Src kinase (Csk) (21, 22). The cytoplasmic Csk requires Csk-binding scaffold proteins to gain efficient access to membrane-bound SFKs. Previously, we identified a transmembrane adaptor protein, Cbp (also known as PAG1), as a specific Csk-binding protein. Cbp/PAG1 is exclusively localized to a membrane microdomain enriched by cholesterol and sphingolipids and plays a scaffolding role for Cbp/PAG1 in Csk-mediated negative regulation of SFKs (3, 15). We also reported that expression of Cbp/PAG1 is noticeably downregulated by c-Src transformation and in some human cancer cells and that reexpression of Cbp/PAG1 can suppress c-Src-induced transformation and tumorigenesis (23). In addition, we showed that Cbp/PAG1 suppressed c-Src function independently of Csk by directly sequestering activated c-Src in the membrane microdomain. These findings suggest a potential role for Cbp/PAG1 as a suppressor for c-Src-mediated cancer progression. However, whether Cbp/PAG1 would serve as a suppressor for other SFK members and whether other microdomain adaptors, such as LIME (4, 11), would also contribute to the suppression of SFK-mediated transformation have yet to be examined.
The membrane microdomain has been regarded as a signaling platform that harbors various signaling molecules and positively transduces cell signaling evoked by activated receptors (29). This model has been best exemplified in immunoreceptor-mediated signaling (8). Moreover, it was reported that SFKs could function positively when bound to Cbp/PAG1 in the microdomain (30, 32). Such positive roles of the microdomain in cell signaling are apparently inconsistent with its suppressive role in Src-mediated transformation. However, this discrepancy rather raises the possibility that the membrane microdomain would function to segregate or protect the normal signaling pathway from the transforming pathways. To prove this hypothesis, more extensive analysis of the role of the membrane microdomain in controlling cell transformation remains to be performed (28).
To elucidate the role of the membrane microdomain in regulating the functions of SFKs, we first compared the transforming abilities of all SFK members using Csk-deficient cells, a reconstitution system in which wild-type SFKs can induce cell transformation (24), and we evaluated the relevance of the membrane distribution of SFKs to their transforming activities. We then investigated the role of the microdomain by disrupting or enhancing its function using methyl-β-cyclodextrin (MβCD) and a microdomain-specific adaptor, Cbp/PAG1, respectively. Our results show that the membrane microdomain and Cbp/PAG1 spatially limit the oncogenic potential of SFKs by sequestering them away from the transforming pathways.
Csk-deficient (Csk−/−) mouse embryonic fibroblasts (MEFs) and wild-type cells (Csk+/+) were kindly donated by Akira Imamoto (12) and were cultured as described previously (24). Csk−/− cells expressing mouse c-Src, c-Yes, Lyn, Lck, Hck, c-Fgr, or Blk or human Fyn, all of which were myc tagged at the C termini, were generated by retroviral gene transfer using pCX4pur vector as described previously (1). A constitutively active form of c-Src, c-Yes, Fyn, or Lck, which has a Tyr-to-Phe replacement at the C-terminal regulatory site, was generated by PCR and introduced into Csk+/+ cells using pCX4pur vector. Wild-type rat Cbp or mouse LIME was expressed using pCX4bleo vector. All cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum.
Cells were lysed in n-octyl-β-d-glucoside buffer (20 mM Tris-HCl [pH 7.4], 150 mM NaCl, 1 mM EDTA, 1 mM sodium orthovanadate, 20 mM NaF, 1% Nonidet P-40, 5% glycerol, 2% n-octyl-β-d-glucoside, and protease inhibitor cocktail), and immunoprecipitation (IP) and immunoblotting were carried out as described previously (23). Immunocytochemistry was performed as described previously (34). The following antibodies (sources indicated in parentheses) were used: antiphosphotyrosine (4G10; Upstate), anti-Src (Ab-1; Oncogene), anti-Src pY418 (Biosource), anti-Src pY529 (Biosource), anti-Lck (2102 from Santa Cruz for Western blotting and 3A5 from Upstate for IP), anti-myc (PL14; MBL), antiactin (Santa Cruz), anti-focal adhesion kinase (FAK; Santa Cruz), anti-FAK pY397 (Biosource), anti-extracellular signal-regulated kinase 1/2 (anti-ERK1/2; Cell Signaling), anti-ERK1/2 pT202/Y204 (Cell Signaling), anti-LIME (clone 6; Exbio), antivinculin (Santa Cruz), anti-β-tubulin (clone 2.1; Sigma), anti-transferrin receptor (Zymed), and Alexa Fluor 594-conjugated goat anti-rabbit immunoglobulin G (IgG) (Molecular Probes). Anti-Cbp was prepared as described previously (15). The chemicals used were Alexa Fluor 488-phalloidin (Molecular Probes), MβCD (Sigma), and polyethylene glycol (PEG)-cholesterol (Sigma).
Fractionation of membrane compartments on a sucrose gradient was performed as described previously (23). Cells were lysed with a buffer (50 mM Tris-HCl [pH 7.4], 150 mM NaCl, 1 mM EDTA, 1 mM sodium orthovanadate, 20 mM NaF, and protease inhibitor cocktail) containing 0.25% Triton X-100 and separated on a discontinuous sucrose gradient (5 to 35 to 40%) by ultracentrifugation at 40,000 × g for 6 h at 4°C. Twelve fractions (1 ml) were collected from the top of the gradient. Separation of detergent-resistant membrane fractions (DRMs) and non-DRMs was confirmed by immunoblotting with the transferrin receptor and GM1 ganglioside (B subunit of cholera toxin; Sigma) as markers of non-DRMs and DRMs, respectively (data not shown) (23). To measure the cholesterol concentration, 50 μl of each fraction was analyzed with the Amplex Red cholesterol assay kit (Molecular Probes) according to the manufacturer's instructions. The protein concentration was determined using the bicinchoninic acid protein assay kit (Pierce).
For reverse transcription-PCR (RT-PCR), total RNA was prepared using Sepasol (Nacalai Tesque) and reverse transcribed using SuperScript II reverse transcriptase (Invitrogen). The expression of the glyceraldehyde-3-phosphate dehydrogenase (GAPDH) housekeeping gene gapdh was used to normalize the amount of total RNA. The following primers were used: mouse Cbp, forward, 5′-TACTGAGCAGTGGGCAGATG-3′, and reverse, 5′-AGGTTGGCATTCTCATCCAG-3′; mouse GAPDH, forward, 5′-ACTCCACTCACGGCAAATTC-3′, and reverse, 5′-CCCTGTTGCTGTAGCCGTAT-3′; and mouse LIME, forward, 5′-GCCCACTCAGTGAAAGAAGC-3′, and reverse, 5′-ACTTGCAGATCTTGCCCACT-3′. PCR products were electrophoresed on a 1.0% agarose gel and visualized by staining with SYBR gold (Molecular Probes).
Single-cell suspensions of 4 × 104 cells were plated onto 60-mm culture dishes in 3 ml of Dulbecco's modified Eagle's medium containing 10% fetal bovine serum and 0.36% agar on a layer of 5 ml of the same medium containing 0.7% agar. For MβCD or cholesterol treatment, MβCD or PEG-cholesterol was added to the medium. Colonies were stained with 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) 11 days after plating, and photographs of the stained colonies were taken; in some cases, the numbers of stained colonies were counted.
We first compared the transforming potentials of all SFK members. To examine the functions of wild-type SFKs, we employed an experimental system using Csk−/− cells. Although endogenous SFKs are activated in these cells, the cells are not transformed because of the substantial degradation of activated SFKs via the ubiquitin-proteasome system (7). However, exogenous expression of c-Src, at more than twice the endogenous level, can efficiently induce cell transformation (24). We expressed myc-tagged versions of all SFK members and isolated clones expressing nearly comparable levels of each SFK (Fig. (Fig.1A,1A, myc blots). All SFKs were activated to a level similar to that of c-Src, which had been shown to be activated about fourfold higher than endogenous c-Src (24), although there were some differences in the patterns of tyrosine-phosphorylated cellular proteins (Fig. (Fig.1A,1A, pY blots).
The impact of the expression of each SFK on cell transformation was examined by morphological study and colony formation assay in soft agar. As shown in Fig. Fig.1B,1B, the expression of Fyn or c-Yes did not induce significant changes in cell morphology or cytoskeletal organization. However, other SFKs (c-Src, Lyn, Lck, Hck, c-Fgr, and Blk) did induce apparent morphological changes characteristic of the transformed cells. Anchorage-independent growth in soft-agar was also conferred by the expression of all SFKs, except for Fyn and c-Yes (Fig. (Fig.1C),1C), indicating that these SFKs can induce cell transformation. Consistently, tyrosine phosphorylation of major Src substrates, such as FAK, cortactin, and annexin II, was elevated in cells expressing the transforming SFKs (Fig. (Fig.1A)1A) (data not shown). These results suggest that all SFKs, except for Fyn and c-Yes, exert transforming activity under the conditions used here. Furthermore, we analyzed the transforming ability of constitutively active forms of SFKs with Tyr-to-Phe replacements at the C-terminal regulatory sites using Csk+/+ cells (Fig. (Fig.1D).1D). Under these conditions, Fyn and c-Yes mutants also exhibited less transforming activity than c-Src and Lck mutants. This indicates that the weaker transforming activity of Fyn and c-Yes than those of other SFKs is not due to the difference in the C-terminal regulatory function.
Previously, we observed that Fyn, which is preferentially distributed to the membrane microdomain, was much less active in cell transformation than the non-microdomain-localized c-Src, and that a c-Src mutant carrying a microdomain localization signal became defective in transformation (23). These findings suggested that the transforming activities of SFKs are commonly suppressed by being distributed to the microdomain. To investigate this possibility, we assessed the membrane distribution of all SFKs by separating the DRMs, in which major components of the microdomain are concentrated (17). As observed previously, c-Src was preferentially distributed to non-DRMs, whereas Fyn was mainly recovered in DRMs (Fig. (Fig.1E).1E). We also found that c-Yes, which did not induce transformation, was highly concentrated in DRMs in a manner similar to Fyn. In contrast, the transforming SFKs Lyn, Lck, Hck, and c-Fgr were more widely distributed to both DRMs and non-DRMs. The transforming SFK Blk, which has only a single myristate modification at the N terminus like c-Src (Fig. (Fig.1F),1F), was mainly distributed to non-DRMs. These results suggest that the transforming abilities of SFKs are associated with their distributions to the nonmicrodomain compartments.
To further confirm the relationship between the membrane distribution of SFKs and their transforming abilities, we reevaluated the transforming activities of Fyn, Lck, and Lyn, which have both myristate and palmitate modifications (Fig. (Fig.1F),1F), in Csk−/− clones expressing different levels of these SFKs (Fig. (Fig.2A).2A). Clones Fyn#2 and Fyn#1, which contained activated Fyn at a level equivalent to that of c-Src in the transformed cells, were not transformed, whereas clones Fyn#6 and Fyn#3, which expressed greater amounts of Fyn, exhibited significant transformed phenotypes (Fig. (Fig.2B).2B). Titration analysis of Fyn protein levels using anti-Fyn antibody showed that more than a 20-fold increase in Fyn levels was required for inducing cell transformation (Fig. (Fig.2C).2C). Furthermore, analysis of c-Yes clones expressing different levels of c-Yes showed that, like Fyn, c-Yes requires more than a 20-fold increase in its expression for cell transformation (Fig. (Fig.2D).2D). These results indicate that Fyn and c-Yes are much less transforming than c-Src, which can efficiently transform Csk−/− cells when expressed at only twice the endogenous level (24).
DRM separation analysis showed that activated Fyn was highly concentrated in DRMs in the nontransformed Fyn#2 and Fyn#1 clones, whereas it was distributed to both DRMs and non-DRMs in the transformed Fyn#6 and Fyn#3 clones (Fig. (Fig.2E).2E). These results raise the possibility that Fyn is capable of inducing cell transformation when a substantial amount of activated Fyn is distributed to the nonmicrodomain compartments. It should be noted that the amounts of activated Fyn in DRMs are almost equal among these cell lines, suggesting that there is a limitation in the distribution of SFKs to the microdomain. Similar results were obtained for Lck and Lyn. Clones expressing Lck or Lyn at lower levels than that used in Fig. Fig.11 (Lck#19′, Lyn#2, and Lyn#21′) showed dominant distribution of each SFK to DRMs, and these cells were not significantly transformed (Fig. 2A, B, and E). In contrast, other clones expressing higher levels of Lck or Lyn, in which each SFK was substantially distributed to non-DRMs, were readily transformed. These results indicate that the distribution of activated SFKs to the nonmicrodomain compartments is tightly associated with the ability of cell transformation and suggest that the membrane microdomain can limit the transforming potential of SFKs.
To address the role of the microdomain in limiting the transforming potential of SFKs, we examined the effects of disruption of the microdomain on the transforming abilities of endogenous SFKs. For this experiment, we employed nontransformed cell lines, MEFs, and Csk−/− and Csk−/− cells expressing Csk (Csk−/−/Csk) (Fig. (Fig.3A).3A). We disrupted the microdomain of these cells by treatment with MβCD, which efficiently depleted cholesterol from DRMs (Fig. (Fig.3B),3B), and examined its effect on the anchorage-independent growth of the cells. The MβCD treatment did not significantly affect the growth of MEFs, whereas it dramatically enhanced the anchorage-independent growth of Csk−/− in a dose-dependent manner (Fig. 3C and D). The MβCD-induced cell growth was significantly suppressed by the expression of Csk, indicating that the effects are dependent on the activity of SFKs. Conversely, the addition of water-soluble cholesterol (PEG-cholesterol) suppressed the anchorage-independent growth of Src-transformed Csk−/− cells (Csk−/−/Src). This further supports the limiting role of the microdomain in cell transformation. Immunocytochemical analysis showed that MβCD-treated Csk−/− cells gained a characteristic transformed phenotype; actin fibers were dramatically rearranged and the activated SFKs (pY418) were relocated to and highly concentrated in the tips of the cells, where focal contacts were formed (Fig. (Fig.4A,4A, left panels). The accumulation of activated SFK in focal contacts was confirmed by staining for vinculin as a marker of focal contacts (Fig. (Fig.4A,4A, right panels). These results clearly show that disruption of the membrane microdomain induces SFK-dependent cell transformation in Csk−/− cells.
To address the mechanisms by which MβCD-treatment induced cell transformation, we examined the phosphorylation and activity status of some signaling molecules (Fig. (Fig.4B).4B). Surprisingly, MβCD treatment did not induce significant change in the activity of SFKs (left panels) or tyrosine-phosphorylated cellular proteins (middle panel). However, we detected a significant, but subtle (~10%), increase in the phosphorylation of FAK Y397 (an autophosphorylation site) and an apparent activation of ERK1/2 by treatment with MβCD (right panels). These results suggest that disruption of the microdomain induces relocation of activated SFKs to focal contacts, where SFKs activate the FAK-mediated ERK pathway leading to cell transformation.
Moreover, we confirmed the role of the membrane microdomain by introducing Cbp/PAG1, a microdomain-specific anchoring protein for activated c-Src (23). Previously, we found that expression of Cbp/PAG1 was strongly downregulated in c-Src-transformed cells (23), thereby liberating activated c-Src from the microdomain. Thus, we first confirmed Cbp expression in Csk−/− cells expressing comparable amounts of each SFK (Fig. (Fig.1).1). Consistent with the case for c-Src, Cbp/PAG1 proteins and transcripts were dramatically reduced in parallel with the transforming activity of each SFK (Fig. 5A and B). This indicates that the expression of Cbp/PAG1 is commonly regulated by the pathway downstream of activated SFKs. In these cells, we ectopically expressed Cbp/PAG1 (Fig. (Fig.5C)5C) and examined the effects on the anchorage-independent growth in soft agar. Although the expression of Cbp/PAG1 did not affect nontransformed cells, it could efficiently suppress the anchorage-independent growth of the SFK-transformed cells (Fig. (Fig.5D)5D) and reverted the cell morphology to a nearly normal shape (data not shown). These results suggest that Cbp/PAG1 is functionally associated with all SFKs and can generally serve as a suppressor for SFK-mediated cell transformation.
We then examined the mechanisms for the Cbp/PAG1-mediated suppression of cell transformation. Consistent with the effects of MβCD, the expression of Cbp/PAG1 did not significantly affect the activity of all SFKs (Fig. (Fig.5C;5C; pY418). DRM fractionation analyses showed that Cbp/PAG1 expression induced relocations of activated c-Src, Lyn, and Lck from non-DRMs to DRMs (Fig. (Fig.6A).6A). The effect of Cbp/PAG1 expression on Fyn clone was not observed under these conditions (Fig. (Fig.6A).6A). However, when the transformed Fyn clones (Fyn#6 and Fyn#3) expressing greater amounts of Fyn were used, the Cbp/PAG1 expression relocated activated Fyn from non-DRMs to DRMs (Fig. (Fig.6B)6B) and readily suppressed cell transformation (Fig. (Fig.6C).6C). This further supports that transforming activity of SFK is not due to total SFK expression but is correlated with the SFK contents in the nonmicrodomain membrane. IP assays revealed that Cbp/PAG1 could interact with all activated SFKs examined here (Fig. (Fig.6D).6D). These observations demonstrate that the recruitment of activated SFKs to the microdomain is mediated by direct interaction with Cbp/PAG1 and is critical for suppressing SFK-mediated cell transformation.
Cbp/PAG1 is a member of transmembrane adaptor proteins (TRAPs) (18) consisting of LAT (36, 37), LIME (4, 11), NTAL/LAB/LAT2 (2, 14), TRIM (6), SIT (19), and LAX (38). To examine whether the action of Cbp/PAG1 is specific or not, we analyzed the function of the microdomain-anchored LIME because it was also shown to serve as a negative regulatory adaptor through binding to Lck and Csk in a manner similar to Cbp/PAG1 (4). RT-PCR analysis revealed that the expression of LIME was not significantly downregulated by c-Src-induced cell transformation (Fig. (Fig.7A),7A), suggesting no functional link between LIME and cell transformation. We then compared the functions of LIME and Cbp/PAG1 by ectopically expressing these TRAPs in c-Src- or Lck-transformed cells. Both Cbp and LIME were efficiently tyrosine phosphorylated by c-Src or Lck (Fig. (Fig.7B).7B). However, LIME did not affect cell morphology (Fig. (Fig.7C)7C) or the anchorage-independent growth (Fig. (Fig.7D),7D), whereas Cbp/PAG1 could restore the normal cell morphology (Fig. (Fig.7C)7C) and suppressed the anchorage-independent growth in soft agar (Fig. (Fig.7D).7D). Furthermore, LIME failed to suppress cell transformation induced by other members of SFKs (Fig. (Fig.7E)7E) and could not relocate activated c-Src of Lck from non-DRMs to DRMs. These results indicate that LIME lacks the potential to suppress SFK-induced cell transformation and in turn highlight the specific role of Cbp/PAG1 in limiting the transforming potential of SFKs by sequestering activated SFKs in the membrane microdomain.
Accumulated evidence has shown that the upregulation of various SFKs is linked to cell transformation, tumorigenesis, and metastasis (31), but concurrent analyses of their oncogenic potentials and regulatory mechanisms have not yet been achieved. In this study, we first systematically compared the transforming activities of all members of SFKs using Csk−/− cells (23) and examined the contribution of the membrane microdomain to the regulation of their transforming activities. We showed that (i) all SFK members have the ability to induce cell transformation, but with distinctive potency depending on the affinity for the microdomain; (ii) disruption of the microdomain can induce transformation in Csk−/− cells that contain a limited amount of activated SFKs; and (iii) the addition of cholesterol or recruitment of activated SFKs to the microdomain via an adaptor, Cbp/PAG1, specifically suppresses SFK-induced cell transformation. These results indicate that the membrane microdomain and Cbp/PAG1 are crucial for controlling the transforming potential of all SFK members.
We observed that all SFKs could induce cell transformation when expressed over certain thresholds. However, under particular conditions in which SFKs were modestly expressed (Fig. (Fig.1),1), we were able to compare the transforming ability of each SFK. Under such conditions, Fyn and c-Yes, which were predominantly distributed to the microdomain, failed to induce transformations, whereas other SFKs, many of which were distributed outside the microdomain, efficiently exerted transforming activities. The less-transforming activity of Fyn and c-Yes was also observed when constitutively active forms of SFKs were expressed in Csk+/+ cells, indicating that the functional difference in SFKs is not due to the difference in the C-terminal regulatory function. Furthermore, the expression of Fyn proteins over the threshold could induce cell transformation but was suppressed when activated Fyn was depleted from nonmicrodomains by the expression of Cbp/PAG1. These observations indicate that the transforming ability of SFKs can be defined by the distribution to the non-microdomain compartments.
The N terminus of SFKs is modified by myristate or by both myristate and palmitate (Fig. (Fig.1G).1G). The myristate modification is required for membrane association, and the palmitate modification facilitates the localization of SFKs to the membrane microdomain (27). Blk and c-Src have a single myristate modification and are dominantly distributed to the non-microdomain compartments. Thus, the strong transforming ability of these SFKs could be accounted for by their reduced distribution to the microdomain (Fig. (Fig.1).1). This notion is supported by our previous observation that a c-Src mutant with a microdomain localization signal had lower transforming activity (23). In contrast, the weak transforming ability of Fyn and c-Yes could be associated with their preferential distribution to the microdomain via their dual N-terminal acyl modification (27). Indeed, an Fyn mutant lacking a palmitoylation site was liberated from the microdomain and gained stronger transforming activity (23). Other SFKs also have dual modification and, thus, were supposed to be distributed preferentially to the microdomain. However, when these kinases were expressed over a certain threshold (Fig. (Fig.1E1E and and2E),2E), a significant fraction was distributed to the nonmicrodomain compartments, which might be caused by an overflow from the microdomain or by an insufficient acyl modification. Thus, the transforming activities of these overloaded SFKs might be due to their presence in the non-microdomain compartments. From these observations, we postulate that the membrane distribution of SFKs, which is determined by their N-terminal acyl modifications, contributes to the regulation of their functions. In this context, the most frequent contribution of upregulated c-Src to human cancers could be explained by its unstable distribution to the microdomain. Because palmitoylation of signaling molecules is regulated dynamically (33), it would be possible that the normal function as well as transforming potential of SFKs is regulated through their N-terminal modification. More detailed analysis of the acylation status of SFKs in cancer cells will be necessary to address this interesting regulatory mechanism.
In normal cells, it is generally accepted that the microdomain-anchored SFKs, such as Lck and Lyn, positively transduce cell signaling evoked by receptor activation (9, 25, 32) and that the microdomain functions as signaling platforms to control normal cellular responses by compartmentalizing signaling components. In contrast, our observations suggest that the microdomain plays a suppressive role with respect to SFK-mediated transformation. These are apparently contradictory, given that SFKs share the same signaling pathway for normal functions and cell transformation. In this study, we observed that, when the microdomain in Csk−/− cells was disrupted by treatment with MβCD, activated SFKs were relocated to focal contacts and selectively activated FAK and ERKs required for cell transformation (20). Furthermore, the upregulation of SFKs over the capacity of the microdomain or the downregulation of Cbp/PAG1 resulted in liberation of activated SFKs from the microdomain to allow interaction with the transforming pathway. These observations suggest that the microdomain plays a crucial role in spatially sequestering or protecting normal SFK functions from the transforming pathways. Analysis of the molecular basis for such a shielding or scaffolding function of the microdomain will provide a new clue to the mechanisms for spatial regulation of the complex intracellular signaling pathways.
When we overexpressed SFKs at different levels (Fig. (Fig.2),2), we recognized that there is a limitation in the capacity of the microdomain and that SFKs are not transforming as long as they are expressed under such a limitation. We also showed that Cbp/PAG1 is substantially downregulated by SFK-mediated cell transformation. Considering that Cbp/PAG1 serves as a specific anchor for the activated SFKs in the microdomain, the downregulation of Cbp/PAG1 would further reduce the limitation of the capacity of the microdomain, thereby contributing to the upregulation of SFKs in some cancer cells. Thus, it is likely that, under normal conditions, the microdomain and Cbp/PAG1 coordinately function to maintain cell homeostasis by limiting or suppressing the transforming potential of SFKs. Further in vivo analysis of the function of these microdomain components would help identify new therapeutic targets that can control upregulation of SFKs in some human cancers.
We thank A. Imamoto, T. Akagi, and T. Yamamoto for their generous gifts of reagents.
This work was supported by a grant-aid for Scientific Research of Priority Areas, Cancer, and for Young Scientists from the Ministry of Education, Culture, Sports, Science and Technology of Japan.
Published ahead of print on 12 October 2009.