SEARCH

SEARCH BY CITATION

Keywords:

  • GCF2;
  • dishevelled;
  • Wnt;
  • RhoA;
  • migration;
  • LRRFIP1

Abstract

  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. References

GC-binding factor 2 (GCF2), a transcriptional repressor that decreases the activity of several genes is capable of binding directly to the GC-rich sequence of the EGFR promoter and repressing the transcriptional activity of EGFR. In addition to its function as a transcriptional repressor, GCF2 can directly interact with other proteins such as flightless-1 (Fli-1). Many previous findings pertaining to the function of Fli-1 have suggested a role for fli-1 in providing a direct link between molecules involved in signal transduction pathways and the actin cytoskeleton. We hypothesized that GCF2, together with Fli-1, plays a role in regulating cytoskeleton function, cell migration, and/or morphology. In our study, we observed that GCF2 is crucial for the activation of RhoA, a small GTPase that plays a key role in the regulation of the actin cytoskeleton. RhoA was markedly inactivated as a result of the decreased expression of GCF2. Co-immunoprecipitations were subsequently performed to further investigate the mechanism for the repressive function. We identified dishevelled (Dvl), which is the key mediator for the Wnt pathway, as a binding partner with GCF2. These results strongly suggest that GCF2 plays a role in the Wnt-noncanonical planar cell polarity (PCP) signaling pathway. Consequently, GCF2 may regulate the cytoskeleton or migration via Dvls and RhoA.

The GC-rich binding factor 2/Leucine rich repeat in the Flightless1 interaction protein 1 (GCF2/LRRFIP1, GCF2) gene encodes a 752 amino acid protein with a detected molecular mass of 160 kDa.1, 2 A Northern blot analysis of human GCF2 mRNA, which is expressed as a 4.2 kbp mRNA, revealed that almost all tissues exhibit an abundant expression. GCF2 has been shown to have the capability to bind with the promoter region of several genes and repress their transcription activities. For example, GCF2 binds with the GC-rich region of the EGFR promoter and subsequently represses the transcription of EGFR.1, 2 Furthermore, GCF2 has been reported to repress the promoter activities of the PDGF A-chain,3 TNF-α,4 and the glutamine transporter EAAT2.5 Another study reported that in addition to its function as a transcriptional repressor of several genes, GCF2 can directly interact with other proteins such as flightless-I (FliI).6 FliI protein contains an actin-binding domain with homology to the gelsolin family and is likely to be involved in actin cytoskeletal rearrangement.7, 8 FliI is a cytoskeletal regulator which is essential for early embryonic development in both D. melanogaster and mammals.9 Because Flightless-1 protein (Fli-1) exhibits these crucial functions, we postulate that GCF2 may play a role in regulating the cytoskeleton function, cell migration and/or cell morphology.

Recent reports indicate that LRRFIP2, which has a high homology to GCF2/LRRFIP1, binds to Dishevelled (Dvl) and activates the canonical Wnt pathway by increasing the cellular abundance of β-catenin in both mammalian cells and in Xenopus embryos.10 In addition to its function as an essential component of the Wnt canonical signaling pathway,11, 12 Dvl also plays an indispensable role in the Wnt noncanonical planar cell polarity (PCP) pathway.13, 14 Dvl consists of three conserved elements, namely, the DIX (Dishevelled/Axin) domain, the PDZ (PSD-95, DLG, ZO1) domain, and the DEP (Dishevelled, EGL-10, Pleckstrin) domain.13, 15 The PDZ and DEP domains are used for the noncanonical PCP pathway, whereas DIX and PDZ are used in the canonical Wnt pathway, thus indicating that Dvl plays a significant role in determining the direction of these downstream cascades take.16–18 The PCP functions, which were analyzed and discovered in the study of convergent extension during Xenopus and zebrafish gastrulation,19 have also been identified in mammals.20–22 Because the molecules that constitute this pathway are conserved in many different tissues from flies to mammals, PCP signaling participates in various cellular functions, including the coordinated organization of cytoskeletal elements and migration of the cells.22–24 Although the characteristics of the signals which initiate the pathway in vertebrates have not yet been clarified, several members of the Wnt family are known to regulate this pathway.

Downstream of Dvl, the PCP signaling causes the activation of the small GTPases, Rho, Rac and cdc42.25, 26 The small GTPases have been clearly demonstrated to regulate the migration, morphology, polarity, microtubule dynamics and vesicle transport of the cells.27 ROCK (Rho-associated kinase), a downstream effector of Rho, re-organizes the cytoskeleton by mediating the formation of stress fibers and focal adhesions, thus resulting in cell migrations or changes in cell morphology.28 Actin rearrangement largely depends on the activation of Rho, which is postulated to correctly reflect the cytoskeletal features of cells. The role of PCP signaling is critical in multiple embryonic processes including embryo morphogenesis.29 It is noteworthy that embryonic development shares many similarities with cancer development and conserved signaling pathways involved in embryogenesis including Wnt, Hedgehog and Notch are dysregulated in tumorigenesis. Thus, it comes as no surprise that Wnt-PCP pathway is implicated in cancer cell adhesion, migration or invasion. Characterization of a variety of PCP components, including Wnt, Fzd, Dvl and RhoA, has shown their involvement in cancer cell invasion or metastasis.30

Our study was initiated by the finding that GCF2 and its family can interact with Dvl protein. Because of its characteristic expression pattern in normal tissues and its interaction with Fli-1, it has been suggested that GCF2 plays a role in regulating the morphology or migration of cells. We herein demonstrate that GCF2 acts as a component of the Wnt-PCP pathway by interacting with Dvl protein and positively controlling the activation of Rho A in cancer cells.

Material and Methods

  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. References

Cell culture and transfection

Human Hela cells and 293 cells were purchased from American Type culture Collection (ATCC) and cultured in MEM medium supplemented with 10% (v/v) FBS. Human DLD-1 cells, derived from human colon cancer cells, were cultured in RPMI 1640 medium supplemented with 10% (v/v) FBS. Transfections were performed with Lipofectamine 2000 reagent (Invitrogen Corp., Carlsbad, CA) according to the manufacturer's protocol.

Plasmid construction

Human GCF2 was cloned by RT-PCR from human ovary total RNA (Clontech Laboratories, San Diego, CA) and inserted into a Myc-tagged mammalian expression vector, pMyc CMV vectors (Sigma-Aldrich, St. Louis, MO). The GCF2 and mutant series were generated by PCR amplification and inserted into the same vector. The human Dvl3 was cloned by RT-PCR from human liver total RNA (Clontech) and inserted into the pFLAG CMV vectors (Sigma-Aldrich). The constructs for the Dvl3 mutants, DIX, ΔDIX, DEP, ΔDEP and PDZ, were generated by PCR amplification and cloned into the same vectors. The identities of all constructs were verified by DNA sequence analysis.

Small interference RNA

Human Stealth™ GCF2 siRNA was designed using the BLOCK-IT™ RNAi Designer web site (Invitrogen). The following two SiRNA GCF2 sequences were utilized. For the GCF2 SiRNA-1, sense: 5′GGA AAU CAA GGA CUC UCU AGC AGA A, antisense: 5′UUC UGC UAG AGA GUC CUU GAU UUC C. For the GCF2 SiRNA-2, sense: 5′CAG UAU ACU GCA AUU UCA GUU UGC U, antisense: 5′AGC AAA CUG AAA UUG CAG UAU ACU G.

Luciferase assays

Either the TOP flash (LEF/TCF-Luc) or FOP flash (mutant LEF/TCF-Luc) (Upstate Chemicon) was used as a reporter plasmid. Cells were harvested 24 hr after the transfection and cell extracts were prepared. The luciferase activities were quantified with a Lumat LB 9507 luminometer (Berthold Technologies, Bad Wildbad, Germany). All luciferase activities were normalized for protein concentration and transfection efficiency using B-galactosidase. All experiments were performed in triplicate and expressed as the mean ± standard deviation (SD).

Rho A activation assays

Wnt-3A recombinant protein (Millipore, Billerica, MA) was diluted to the indicated concentration in serum-free EMEM. Wnt3A transfected L cells and control L cells were purchased from ATCC (Manassas, VA). Wnt3A conditioned medium was prepared according to the previously described method.31

Rho A activation was demonstrated with a GST-RBD pull-down assay (Millipore), essentially as previously described.32

Immunoblotting and immunoprecipitation

Whole-cell lysates were pretreated with mouse or rabbit IgG and protein G sepharose (Amersham). Endogenous GCF2 was immunoprecipitated from supernatants by rabbit anti-GCF2 polyclonal antibody and protein G sepharose. Endogenous Dvl-2 or Dvl-3 was immunoprecipitated by mouse anti-Dvl-2 monoclonal antibody (Santa Cruz) or anti-Dvl-3 monoclonal antibody (Santa Cruz). Flag-tagged proteins were immunoprecipitated by mouse anti-Flag monoclonal antibody (M2, Sigma). Myc-tagged proteins were immunoprecipitated by rabbit antimyc polyclonal antibody (SantaCruz).

Immunofluorescence

Hela cells plated on glass coverslips were fixed with 4% paraformaldehyde in phosphate-buffered saline for 15 minutes, permeabilized with 0.1% Triton X-100, and blocked with 5% bovine serum albumin. Mouse anti-GCF2 antibody (BD pharmingen), rabbit anti-Dvl2 antibody (Millipore), and rabbit anti-Dvl3 antibody (Abcam, Cambridge, MA) were used as primary antibodies. To detect GCF2, the cells were incubated with Alexa Fluor® 568 (Invitrogen) conjugated antimouse IgG antibodies. To detect Dvl2 or Dvl3, the cells were incubated with Alexa Fluor® 488 (Invitrogen) conjugated antirabbit IgG antibodies.

Cell migration assays

Transwell assays were performed with a 24-well Cell Migration Chamber (8-μm pore size; Millipore). Si RNA Oligos transfected Hela and DLD-1 were transferred to the upper chamber (2 × 104 cells in 500 μl). EMEM containing 100 ng/ml of Wnt-3A recombinant protein was added to the lower chamber. After 3–6 hr incubation, the cells that had migrated were stained. The cells were subsequently extracted and detected on a microplate reader.

In vitro wound healing assays

The In vitro “scratch” wounds were created by scraping the confluent cell monolayer with a sterile pipette tip. The cells were incubated with Wnt-3A recombinant protein at a concentration of 100 ng/ml. After 24 hr, the distance of the wound closure (compared to the control at t = 0 hr) was measured in three independent wound sites per group. Relative cell motility was calculated as the percentage of the remaining cell-free area compared to the area of the initial wound. Values from at least three independent experiments were pooled and expressed as the mean ± standard deviation (SD).

Statistical analysis of data

Differences between data were analyzed with the Student's t-test for paired data. p values of < 0.05 were considered significant. Data correspond to the mean ± standard deviation (SD).

Results

  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. References

Interaction of GCF2 with the endogenous Dvl

Previous studies reported that LRRFIP2 can interact with Dvls, and the amino terminus of GCF2 is highly conserved with LRRFIP2.6, 10 To examine whether GCF2/LRRFIP1 interacts with Dvl, we performed co-immunoprecipitation assays with the endogenous proteins. Whole cell lysates from the Hela cells were immunoprecipitated with anti-Dvl2, anti-Dvl3 or control IgG antibodies. GCF2 was detected in both the anti-Dvl2 and Dvl3 immunoprecipitants but not with the control IgG (Fig. 1a). To confirm whether these interactions between GCF and Dvls were genuine, the lysates were further immunoprecipitated with anti-GCF2 antibody. Consistent with the first experiment, both Dvl2 and Dvl3 were observed in the anti-GCF2 immunoprecipitants but not in the control IgG (Fig. 1b).

thumbnail image

Figure 1. GCF2 interacts with Dvl2 and Dvl3 at the endogenous level. (a) The cultured Hela cells were harvested for immunoprecipitation with anti-Dvl2 and anti-Dvl3 antibodies or a nonspecific control IgG. Dvl-associated GCF2 was revealed by anti-GCF2 immunoblotting (left and upper panels). The expression of GCF2 was confirmed by anti-GCF2 immunoblotting with the total cell lysates (left and lower panels). Immunoprecipitated tagged proteins are shown in the right two panels. (b) To confirm the interactions between GCF2 and Dvl2 or Dvl3, the whole cell lysates from Hela cells were immunoprecipitated with anti-GCF2 antibody or a nonspecific antibody (control IgG). GCF2-associated Dvl2 and Dvl3 were detected by the specific antibody (left and middle panel). The expression of Dvl2 and Dvl3 were confirmed by specific antibody immunoblotting (lower two panels). Data shown are representative of at least three similar experiments. Immunoprecipitated tagged proteins are shown in the right panel. Hela cells were stained for GCF2 with anti-GCF2 primary antibody and Alexa 568 conjugated antimouse IgG antibodies. The cells were sequentially stained for Dvl2 with anti-Dvl2 primary antibody and Alexa 488 conjugated antirabbit IgG antibodies (c) or for Dvl3 with anti-Dvl3 primary antibody and Alexa 488 conjugated antirabbit IgG antibodies (d). Scale bar, 10 μm.

Download figure to PowerPoint

To reveal the subcellular localization of GCF2, immunofluorescence studies were performed in Hela cells. While endogenous GCF2 was primarily present in the cytoplasm, a small amount was detected in the nuclear space. Although most of the proteins presented as fine granules, some punctate patterns were also observed in perinuclear region of the cells (Figs. 1c and 1d left panel). The subcellular localization of endogenous Dvl2 or Dvl3 was also examined. As shown, Dvl2 and Dvl3 displayed punctate patterns and diffuse distributions in both nuclear and cytoplasm (Figs. 1c and 1d middle panel). When the figures were merged, it was revealed that the GCF2 proteins were co-localized with Dvl2 or Dvl3 in the perinuclear region (Figs. 1c and 1d right panel). These data further support the notion that GCF2 physiologically binds to Dvl in the cell cytoplasm.

Identification of the region of GCF2 responsible for the interaction with Dvl

The region of GCF2 that binds to Dvl3 was determined using the binding assays described in Figure 2a. In the deletion mutant series from the carboxyl terminal region of the GCF2, the interactions were detected with 1/350-myc, 1/208-myc, 1/153-myc (Figs. 2b and 2c line 4, 5). However, it must be noted that the interactions with 1/153-myc were rather weak. Once the construct was deleted to 1/98, then no interaction was observed. When the constructs were deleted from the amino terminal region of the GCF2, then the interactions were not detected with the 142/752-myc or 338/752-myc, whereas the 42/752-myc represents the interaction between the Flag-Dvl3 (line 7, 8, 9). In addition, the interaction was detected with 42/208-myc (line 2), thus indicating that the 42–208 region of GCF2 is responsible for the GCF2-Dvl3 interaction.

thumbnail image

Figure 2. The GCF2-Dvl interaction is mediated by a serine-rich repeat region of GCF2. (a) Schematic representative of human GCF2 deletion mutants. The region from amino acids 42 to 208 of GCF2 contains a serine-rich region (Ser-rich). (b and c) Immunoprecipitation analyses to examine the interaction between the truncation mutants of GCF2 and Dvl3. HEK293T cells were transfected with the indicated plasmid. The cell lysates were immunoprecipitated using an anti-FLAG antibody (b) or an anti-Myc antibody, and the associated protein was detected by immunoblotting with an anti-Myc antibody (b) or an anti-FLAG antibody (c) (top blot panels). Myc- or FLAG-tagged proteins in the cell lysates are shown in middle blot panel. Immunoprecipitated tagged proteins are shown in the lower blot panels. Bands with the expected sizes are indicated with the asterisks (*). Data shown are representative of at least three similar experiments.

Download figure to PowerPoint

Identification of the domain of Dvl responsible for the interaction with GCF2

Dvl3 consists of three key domains, namely, the DIX, the PDZ and the DEP.11, 12 A series of truncation mutants of Dvl3 were constructed and coimmunoprecipitation experiments were performed. The extracts were subjected to immunoprecipitation with an antimyc antibody and evaluated by a Western blot analysis to detect the bound Flag-tagged Dvl3 proteins. Interactions were obtained with Flag-ΔDIX, Flag-ΔDEP and Flag-PDZ, whereas no interactions were detected with Flag-DIX or Flag-DEP (Fig. 3b). In addition, anti-Flag antibody co-immunoprecipitants were probed with antimyc antibody on Western blot analysis and similar results were obtained (Fig. 3c). These experiments suggest that the PDZ domain of the Dvl3 is responsible for the GCF2-Dvl3 interaction.

thumbnail image

Figure 3. The PDZ domain of Dvl3 is responsible for the GCF2-Dvl interaction. (a) Schematic representative of human Dvl3 deletion mutants. (b and c) Immunoprecipitation analyses to examine the interaction between the truncation mutants of Dvl3 and GCF2. HEK293T cells were transfected with the indicated plasmid. The cell lysates were immunoprecipitated by using an anti-Myc antibody (b) or an anti-FLAG antibody (c), and associated protein was detected by immunoblotting with an anti-FLAG antibody (b) or an anti-Myc antibody (c) (top blot panels). FLAG- or Myc-tagged proteins in the cell lysates are shown in the middle blot panel. Immunoprecipitated tagged proteins are shown in the lower blot panels. Bands with the expected sizes are indicated with the asterisks (*). Data shown are representative of at least three similar experiments.

Download figure to PowerPoint

GCF2 is able to activate the canonical Wnt pathway

To elucidate the role of GCF2 in the Wnt canonical pathway, LEF/TCF-mediated transcription was examined using the TOP flash or FOP flash reporter plasmids. Over-expression of GCF2 in 293T cells activated the TOP flash reporter activity in a dose-dependent manner and 2 μg of GCF2 expression plasmid resulted in a roughly 2.5-fold increase in luciferase activity. No activity was detected with GCF2 using the FOP flash (Fig. 4a). To investigate the functional interaction between GCF2 and Dvls, GCF2 and Dvl3 expression plasmids were co-transfected to the 293T cells and the TOP flash reporter activity was tested. The expressed GCF2 and Dvl3 synergistically activated the TOP flash reporter activity, thus suggesting that GCF2 could activate the Wnt canonical pathway via interaction with the Dvl proteins (Fig. 4c). To confirm the role of GCF2 in the Wnt canonical pathway, the activation was tested in GCF2 knockdown cells. An overexpression of Wnt3A robustly activated the TOP flash reporter activity in the control SiRNA transfected Hela cells, whereas little activation was observed in the GCF2 knockdown cells (Fig. 4d). As shown in Figure 2, the region from 42 to 208 amino acid sequence of GCF2 is responsible for the interaction with Dvl3. To confirm the interaction between GCF2 and Dvl3 in the activation of the Wnt canonical pathway, a TOP flash reporter assay was performed using the 208/752 expression plasmid. The activation of TOP flash was not observed with this deletion mutant form of GCF2 (Fig. 4b).

thumbnail image

Figure 4. GCF2 regulates the Wnt canonical pathway and TCF/LEF-mediated transfection. (a) HEK293T cells were transfected with 0.1 μg of a reporter plasmid and the indicated amount of GCF2 expression vectors. (b) The deletion mutant vector for GCF2, 208/752 (which lacks the serine-rich repeat region of GCF2) and the reporter plasmid were transfected into the HEK293T cells. (c) 2 μg of GCF2 expression plasmid was transfected to the HEK293T along with the indicated amount of Dvl3 expression vector and 0.1 μg of reporter plasmid. (d) GCF2 Si-1 or control Si transfected Hela were transfected with the Wnt3A vector and the TOP flash. Quantitative analysis was performed as described in Materials and methods. Data are expressed as the mean ± SD. (*p < 0.05).

Download figure to PowerPoint

GCF2 is necessary for Wnt-3A-induced RhoA activation

After recognizing the role of GCF2 in the wnt-induced RhoA activation, we examined the ability of Wnt-3A to stimulate RhoA activation in GCF2 knockdown cells. Using the SiRNA technique, the expression of GCF2 was efficiently knocked down in both Hela and DLD-1 cells (data not shown). After 48 hr of serum starvation, GCF2 SiRNA transfected or control cells were treated with mouse Wnt-3A recombinant protein at concentrations of 0, 20, 100 or 200 ng/ml (Figs. 5a and 5d). The activation of Rho A was followed by a roughly two-fold increase in the control Hela cells at the concentrations of 100 and 200 ng/ml, whereas no increase was observed in the GCF2 Si-1 nor GCF2 Si-2 transfected cells (Fig. 5a). Similarly, the RhoA activity did not increase in the GCF2 Si-1 transfected DLD-1 cells at any concentration in contrast to the control cells (3.0 to 4.0-fold at 100 and 200 ng/ml, respectively) (Fig. 5d). Thereafter, the cells were treated with 100 ng/ml of Wnt-3A for 0, 0.5, 1 or 2 hr and the level of RhoA activation was evaluated. As shown in Figures 5b and 5e, a smaller, less sustained increase in activated RhoA was detected in the GCF2 Si-1 transfected Hela or DLD-1 cells, whereas a time-dependent increase in RhoA activation was observed in the control cells. These data demonstrated that the Wnt-3A induced Rho A activation was strongly suppressed in the GCF2 knockdown Hela or DLD1 cells. Similar results were also obtained with the Wnt-3A conditioned medium (Wnt-3A CM) (Figs. 5c and 5f). The above data leads us to conclude that GCF2 plays an indispensable role in the Wnt-3A-induced RhoA activation.

thumbnail image

Figure 5. GCF2 is an important factor in Wnt3A-induced RhoA activation. GCF2 Si-1, GCF2 Si-2 or Control Si RNA were transfected to the Hela or DLD-1 cells. (a and d) 48 hr after the transfection, GCF2 Si knocked down or control cells were incubated for 2 hr with Wnt3A recombinant protein at the indicated concentrations (a; Hela, d; DLD-1). (b and e) cells were incubated with 100 ng/ml of Wnt3A recombinant protein for 0.5–2 hr (b; Hela, e; DLD-1). (c and f) cells were incubated with Wnt3A conditioned medium for 0.5 – 2 hr (c; Hela, f; DLD-1). Quantitative analysis was performed as described in Materials and methods. Data are expressed as the mean ± SD. (*p < 0.05).

Download figure to PowerPoint

Wnt-3A induced cell migration is suppressed in the GCF2 knockdown cells

To identify the role of GCF2 in cell migration, standard transwell assays were carried out with Hela cells (Fig. 6a). Cell migration was strongly repressed among both GCF2 Si-1 and GCF2 Si-2 transfected Hela cells, and the number of cells that passed through the membrane within 3 hr were less than 50% that of the control cells. After 6 hr, very few GCF2 knockdown Hela cells had migrated to the bottom of the filter and the number was about one-third of the number of control cells. A similar result was observed with the DLD-1 cells. To confirm the effect of GCF2 on the cell migration, in vitro wound healing assays were performed with Hela cells (Fig. 6b). In the control Hela cells, observable wound healing was observed after 24 hr of incubation in 100 ng/ml of Wnt-3A recombinant protein. No obvious wound healing was not clearly observed in the GCF2 knockdown cells treated with Wnt-3A recombinant for 24 hr. These results indicate that GCF2 protein plays an important role in Wnt-3A induced cell migration.

thumbnail image

Figure 6. GCF2 regulates Wnt3A-induced cell migration. (a and b) GCF2 Si-1, GCF2 Si-2 or Control SiRNA oligos transfected Hela (a) and DLD-1 (b) were transferred to the transwell migration chamber (8-μm pore size). EMEM containing Wnt-3A recombinant protein at the concentration of 100 ng/ml was added to the lower chamber. The motility behavior of Wnt3A-treated cells was analyzed in an in vitro wound model (c). Quantitative analysis was performed as described in Materials and methods. Data are expressed as the mean ± SD. (*p < 0.05).

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. References

In our study, we identified GCF2 as a novel binding partner of Dvl2 and Dvl3. The interactions between GCF2 and Dvl2 or Dvl3 were verified by co-immunoprecipitation experiments. As shown by the immunofluorescence study, endogenous GCF2 co-localized with Dvl2 or Dvl3 in the cytoplasm of the cells. Utilizing the deletion mutant expression vector series for GCF2 and Dvl3, we determined that the 42 to 208 region of GCF2 and the PDZ domain of Dvl were responsible for the interactions. A recent study reported that LRRFIP2 binds to Dvl and activates the canonical Wnt pathway by increasing the cellular abundance of β-catenin in both mammalian cells and Xenopus embryos.10 The interaction between GCF2 and Dvl proteins was hypothesized because a database analysis revealed that the 42 to 208 region of the GCF2 amino acid sequence has a high homology with the N-terminus region of LRRFIP2. Myc-tagged deletion mutant constructs for the GCF2 were prepared and the interactions with Dvl3 were investigated. We observed that the amino acid sequence from 42 to 208 was responsible for this interaction (Fig. 2). These findings led us to presume that the 42–208 amino acid sequence plays a significant role in the interaction of GCF2 with other molecules such as Dvls and is a key domain for its function.

Using the deletion mutant vector for Dvl3, we demonstrated that GCF2 interacts with the PDZ domain, which is the central conserved domain in Dvl proteins. The PDZ domain is conserved in many proteins that are involved in submembranous receptor and signal transduction complex clustering.33 Whereas the requirement of the PDZ domain for Dvl activity is varied and remains unclear, many reports indicate that the PDZ domain is involved in both canonical Wnt signaling and the noncanonical PCP pathway by interacting with many Dishevelled-associated proteins.15 GCF2 may therefore be involved in the Wnt pathway by regulating the Dvl activity.

To investigate the role of GCF2 in the Wnt canonical pathway, TOP flash reporter assays were performed. The level of TOP flash reporter activity was used as an indicator of TCF/LEF-dependent transcriptional activity and, indirectly, β-catenin levels.34 An overexpression of GCF2 in HEK293T cells resulted in the activation of TOP flash reporter activity in a dose-dependent manner, whereas no activation was observed with the negative control FOP flash. The over-expression of Wnt3A did not induce TOP flash activity in GCF2 Si knockdown cells. In addition, when the 1–208 amino acid sequence of GCF2 was deleted, the over-expression of mutant GCF2 did not induce TOP flash activity. The 1–208 amino acid sequence was revealed as a key domain in the interaction with Dvls (Fig. 2). These data suggest that GCF2 activates the Wnt canonical pathway via interaction with Dvl proteins.

In addition to its function as an essential component of the Wnt canonical signaling pathway,11, 12 Dvl protein is also an important factor in the Wnt noncanonical PCP pathway.13, 14 Downstream of Dvl, the PCP signal activates the small GTPases Rho, Rac, and cdc42.25, 26 The small GTPases regulate cellular migration, morphology, polarity, microtubule dynamics, and vesicle transport.27 RhoA activation is one of the molecular mechanisms implicated in cell motility by promoting the formation of contractile actin bundles and large adhesive structures that interact with the extracellular substrate. GCF2 can also directly interact with flightless-I (FliI).6 FliI was identified as a cytoskeletal regulator that is essential for early embryonic development in both D. melanogaster and mammals.9, 35 Therefore, GCF2 should be involved in mammalian cell actin reconstruction and migration. To investigate the GCF2 involvement in Wnt-induced RhoA activation, RhoA activation assays were performed. As shown in Figure 5, Wnt3A induced RhoA activation in concentration-dependent and time-dependent manners in control Hela or DLD-1 cells, whereas no activation was observed in GCF2 SiRNA transfected Hela or DLD-1 cells. These results were confirmed using Wnt3A conditioned medium. Although Wnt11/ silberblick and Wnt5 are known to specifically regulate this pathway,16, 36, 37 recent studies suggest that Wnt3A (together with Dvl) has the capability to regulate the noncanonical PCP pathway by activating RhoA in cultured mammalian cells and results in directed cell migration.31, 38–40 We demonstrated that Wnt3A was able to activate RhoA in human carcinoma cell lines. RhoA, with a molecular mass of 21 kDa, is the most extensively studied member of the Rho GTPase family, which belongs to the Ras superfamily of small G proteins. RhoA has been reported to regulate many biological activities, including the formation of stress fibers and focal adhesions.41 RhoA acts as a molecular switch in cells and regulates signal transduction from cell surface receptors to intracellular target molecules; furthermore, it is involved in a variety of biological processes, including cell morphology, motility, and cytokinesis.42 The role of GCF2 in Wnt3A-induced cell migration was investigated in Hela and DLD-1 cells. Transwell migration assays and wound healing assays showed that cell migration was notably suppressed in GCF2 SiRNA treated cells (Fig. 6). This suppression was apparently due to the inactivation of RhoA in the GCF2 knockdown cells. It was suggested that GCF2 plays an important role in Wnt3A-induced cancer cell migration by activating the small GTPase RhoA. An increasing amount of evidence indicates that RhoA is an important player in the proliferation, invasion, and metastasis of carcinoma cells.43, 44

In our study, Hela and DLD-1 carcinoma cell lines were used in the RhoA activation assays and cell migration assays. DLD-1 cells were derived from human colorectal cancer and are characterized by an APC mutation lacking the Axin binding domain and the continuous inhibition of β-catenin degradation. The effect of GCF2 on the Wnt-induced RhoA activation or cell migration was observed not only in the Hela cells but also in APC-truncated DLD-1 cells. These findings indicated that GCF2, via the interaction with Dvls, regulates RhoA activation and cell migration independently from the Wnt canonical pathway. Recently, the metastasis of nonsmall cell lung cancer was suggested to be mediated by not only Dvl1 via the canonical Wnt pathway but also by Dvl3 via in a β-catenin independent manner.45 The Wnt β-catenin independent PCP pathway is now thought to be closely related to the tumor invasion or metastasis because of the fact that its components are highly associated with the cell adhesion or metastasis.30, 46 Thus, GCF2 is also supposed to have role in cancer cell invasion or metastasis in the β-catenin independent manner.

This report identified GCF2 as a novel binding partner for Dvls, which is a key molecule in the Wnt signaling pathways. In addition, GCF2 can interact with the PDZ domain of Dvl protein and is an important factor in both the Wnt canonical pathway and the noncanonical PCP pathway. It was demonstrated that GCF2 regulates the Wnt canonical pathway via interaction with Dvl proteins. Furthermore, GCF2 was shown to play an important role in Wnt-induced RhoA activation and migration in human carcinoma cell lines. Although further study is necessary to clarify the molecular mechanism of Rho GTPase regulation of GCF2 in the Wnt signaling pathway, there is strong evidence that GCF2 plays an important role in cancer biology.

References

  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. References