Differential Modulation of TCF/LEF-1 Activity by the Soluble LRP6-ICD

The canonical Wnt/β-catenin (Wnt) pathway is a master transcriptional regulatory signaling pathway that controls numerous biological processes including proliferation and differentiation. As such, transcriptional activity of the Wnt pathway is tightly regulated and/or modulated by numerous proteins at the level of the membrane, cytosol and/or nucleus. In the nucleus, transcription of Wnt target genes by TCF/LEF-1 is repressed by the long Groucho/TLE co-repressor family. However, a truncated member of the Groucho/TLE family, amino terminal enhancer of Split (AES) can positively modulate TCF/LEF-1 activity by antagonizing long Groucho/TLE members in a dominant negative manner. We have previously shown the soluble intracellular domain of the LRP6 receptor, a receptor required for activation of the Wnt pathway, can positively regulate transcriptional activity within the Wnt pathway. In the current study, we show the soluble LRP6 intracellular domain (LRP6-ICD) can also translocate to the nucleus in CHO and HEK 293T cells and in contrast to cytosolic LRP6-ICD; nuclear LRP6-ICD represses TCF/LEF-1 activity. In agreement with previous reports, we show AES enhances TCF/LEF-1 mediated reporter transcription and further we demonstrate that AES activity is spatially regulated in HEK 293T cells. LRP6-ICD interacts with AES exclusively in the nucleus and represses AES mediated TCF/LEF-1 reporter transcription. These results suggest that LRP6-ICD can differentially modulate Wnt pathway transcriptional activity depending upon its subcellular localization and differential protein-protein interactions.


Introduction
The canonical Wnt/b-catenin (Wnt) pathway is a ubiquitous transcriptional regulatory pathway that affects the expression of over 1800 genes involved in at least 36 different pathways [1,2,3,4,5]. As a result, the Wnt pathway influences a diversity of biochemical functions including cell patterning, differentiation, embryonic development and apoptosis [1,2,3,5]. Defects at various levels of the pathway therefore result in numerous pathological conditions including various cancers and neurodegenerative disorders [3,6].
While b-catenin mediated activation of TCF/LEF-1 requires robust activation of the Wnt pathway [9], the transcriptional activity of the Wnt pathway (i.e., b-catenin stabilization and/or TCF/LEF-1 activity) can be modulated by various effector proteins at the level of the membrane, cytosol and/or nucleus [5,6,10,26,28,29,30,31,32,33,34,35,36]. We and others previously have shown the LRP6-ICD can function as an effector of the Wnt pathway by attenuating GSK3b mediated b-catenin phosphorylation thereby enhancing cytosolic b-catenin stabilization [35,36,37,38,39]. LRP6-ICD expression also enhances TCF/ LEF-1 activity [35,36,38], which may be an indirect result of enhanced b-catenin stabilization. However, the hypothesized mechanism has yet to be confirmed, which is one aim of the current study. It should also be noted that the LRP6-ICD is a GSK3b substrate, but it does not have to be phosphorylated by GSK3b to interact with [35,38] and attenuate GSK3b activity [35]. It has been shown that the soluble intracellular domain of various cell surface receptors, such as Fz and LRP1, can translocate to the nucleus and directly modulate gene expression [40,41,42,43,44]. While the cytosolic function of LRP6-ICD has been studied [35,36,37,38,39], it is unknown if LRP6-ICD can also translocate to the nucleus and directly modulate TCF/LEF-1 activity, which is the primary focus of the study.
Previously we reported that AES and LRP6-ICD interacted in a yeast 2 hybrid assay [39] but the interaction was not confirmed in situ. In the current study, we show the LRP6-ICD can translocate to the nucleus where it interacts with AES. In contrast to cytosolic LRP6-ICD, we show that nuclear localized LRP6-ICD represses transcription of a TCF/LEF-1 reporter and depresses AES mediated activation of the reporter. The current findings suggest LRP6-ICD can differentially modulate Wnt pathway transcriptional activity (i.e., b-catenin stabilization and/or TCF/LEF-1 activity) based upon its subcellular localization and protein-protein interactions.

DNA Constructs
The GFP tagged wild type LRP6-ICD and LRP6-ICD65 m (Ser or Thr in all five PPP(S/T)P motifs were mutated to Ala and this construct is not phosphorylated by GSK3b) were described previously [35,39]. To make the GFP tagged LRP6-ICD constructs with a nuclear localization signal (NLS) or nuclear export signal (NES), wild type GFP-LRP6-ICD or GFP-LRP6-ICD65 m were used as templates for PCR amplification. The eGFP sequence from the pEGFP-C1 vector (BD Bioscience) was amplified as part of the LRP6-ICD PCR product using the following PCR primers: SalI-forward . The amplified GFP-LRP6-ICD products were digested with and subcloned into the SalI and NotI site of the pCMV/myc/nuc empty vector (EV-NLS; Invitrogen) or BamHI and NotI site of a pcDNA-NES empty vector to make GFP-LRP6-ICD-myc-NLS or GFP-LRP6-ICD-NES constructs. The pCMV/myc/nuc/GFP vector (EV-GFP-NLS; Invitrogen) was used only as a control for assays involving GFP-LRP6-ICD-NLS constructs. Using human AES cDNA as a template (accession # NM_001130, 197 amino acids), the following PCR primers, digestion and vector were used to make the AES-HA construct: EcoRI-forward (59-GCC-GGG-ATC-CAC-ATG-ATG-TTT-CCA-CAA-AGC-AG-39) and EcoRI-reverse (59-CCG-GCC-TCG-AGC-TAA-TCC-GAC-TTC-TCG-CCA-TC-39), digested with and subcloned into the EcoRI site of an HA expression vector [35]. To make the AES-myc-NLS or AES-HA-NES constructs using AES-HA as a template, the following PCR primers, digestion and vectors were used for each: NcoI-forward

Luciferase Assay
One day before transfection, 6.0610 4 HEK 293T cells were plated in a 12-well plate. Cells were transfected 24 hrs later with 0.2 mg LEF-1 (from Dr. R Grosschedl), 0.005 mg TK-Renilla internal control (Promega) and 0.2 mg TOPflash-luciferase reporter (TOP; Upstate) or the negative control reporter, Super8-XFOPflash (FOP; Addgene), whose TCF/LEF-1 binding sites have been mutated. Additionally, cells were transfected with AES-HA (0.5 mg), AES-myc-NLS (0.5 mg), AES-HA-NES (0.25 mg), GFP-LRP6-ICD (1 mg), GFP-LRP6-ICD-myc-NLS (1 mg) and/or GFP-LRP6-ICD-NES (0.25 mg). The empty NLS vectors (EV-NLS or EV-GFP-NLS) significantly activated the TOP reporter compared to lacZ or eGFP while the empty pcDNA-NES also had a slight activating effect on the TOP reporter. In contrast to the empty NLS vectors, the activating effect of pcDNA-NES was mitigated by using a lower cDNA concentration. Neither the empty NLS nor NES vectors affected the FOP reporter. However, to mitigate possible confounding variables, the NLS or NES tagged AES-HA and GFP-LRP6-ICD constructs are expressed as fold change relative to the empty NLS or NES vectors. Equal amount of cDNA for the empty NLS or NES vectors and the NLS or NES tagged AES-HA and/or GFP-LRP6-ICD constructs was used in all assays. All experiments were performed at least three times and each time the measurements were done in triplicate.

Nuclear Fractionation
Separation of cytosolic and nuclear fractionation was carried as previously described [45]. Briefly, cells were washed twice with and harvested in ice-cold PBS and a small aliquot of the total lysate was collected before centrifugation and resuspended in lysis buffer (represents whole cell lysate). The remaining lysate was spun at 800 x g for 5 min at 4uC and the cell pellet was resuspended in lysis buffer by triturating followed by centrifugation at 380 6 g for 5 min at 4uC. The supernatants were collected and used as the cytosolic fractions. The pellets were washed once in lysis buffer and twice in wash buffer (30 mM sucrose, 10 mM Pipes, pH 6.8, 3 mM MgCl 2 , 1 mM EGTA, 25 mM NaCl with phosphatase and protease inhibitors). The crude nuclei were overlaid on top of 1 m sucrose and spun at 1200 6 g for 10 min at 4uC. The pellets were collected and resuspended in buffer B (300 mM sucrose, 10 mM Pipes, pH 6.8, 3 mM MgCl 2 , 1 mM EGTA, 25 mM NaCl, 0.5% Triton X-100 with phosphatase and protease inhibitors) and used as the nuclear fractions.
Immunoprecipitation HEK 293T cells were co-transfected with 4 mg of AES-HA along with 5 mg of eGFP-C1 (eGFP), wild type GFP-LRP6-ICD or GFP-LRP6-ICD65 m. Forty-eight hours post transfection, the AES-HA/eGFP cells were collected with immunoprecipitation buffer (0.5% Nonidet P-40, 150 mM NaCl, 10 mM Tris-Cl (pH 7.4), 1 mM EGTA and 1 mM EDTA with phosphatase and protease inhibitors) while the AES-HA/GFP-LRP6-ICD cell's were fractionated into cytosolic and nuclear fractions. For immunoprecipitation, cytosolic or nuclear fractions (50 mg) were incubated at 4uC for 3 hrs on a rotational shaker with M-280 sheep anti-mouse magnetic IgG beads (Invitrogen/Dynal Biotech) preconjugated to a monoclonal anti-HA antibody (Sigma) in the presence of immunoprecipitation buffer. Cytosolic or nuclear precipitates (and their respective controls) were washed 66 with PBS containing 350 mM NaCl (high salt) and 0.2% Triton X-100 or 26/26with high salt PBS/regular PBS. The beads were boiled for 10 min in 26stop buffer (2% sodium dodecyl sulfate, 250 mM Tris-Cl (pH 7.4), 10% glycerol, 5 mM EDTA and 5 mM EGTA) and subjected to SDS-PAGE and immunoblot analysis.

Immunofluorescence
Cells were plated onto poly-D-lysine/laminin coated glass coverslips (20/5 mg/ml) and transfected with the indicated constructs for 48 hrs. Cells were fixed with 4% PBS-paraformaldehyde for 15 min, permeabilized with ice cold methanol for 5 min and stained with anti-GFP or anti-HA antibody at 4uC overnight. Cells were then washed with PBS and incubated with 5% FBS containing a Texas red-or FITC -conjugated secondary antibody (Jackson ImmunoResearch) and DAPI (Vectashield, Vector Labs) at room temperature for 1 hr. Slides were analyzed using a Zeiss Axio Observer inverted microscope with a 636 oil immersion objective. Images were captured with an Axio-CamMR3 camera (Zeiss Corp) and exported using the AxioVision software (Zeiss Corp.).

Data Analysis
Data were analyzed using ANOVA between individual groups and considered significantly different when p,0.05. Results were expressed as mean 6 S.E.

LRP6-ICD translocates to the nucleus
It has been shown exogenous LRP6-ICD localizes to the cytosol [36] but nuclear localization of the ,20 kDa (,50 kDa with GFP tag) protein has not been examined. Figure 2 clearly shows exogenous wild type LRP6-ICD and a phospho-mutant LRP6-ICD that can not be phosphorylated by GSK3b (LRP6-ICD65 m) also localize to the nucleus in CHO cells. Although the presence of nuclear LRP6-ICD was evident, the level of nuclear LRP6-ICD relative to cytosolic LRP6-ICD ( Fig. 2A), suggests that the majority of LRP6-ICD remains in the cytosol. Immunocytochemistry (Fig. 2B) also show that both wild type LRP6-ICD and LRP6-ICD65 m localize within DAPI stained nuclei which confirms the fractionation data showing nuclear localization of LRP6-ICD does not require GSK3b phosphorylation. Interestingly, nuclear LRP6-ICD exhibits punctuate nuclear staining whereas cytosolic LRP6-ICD exhibits a more diffuse pattern similar to eGFP. Both LRP6-ICD constructs also exhibited similar nuclear localization in HEK 293T cells (Fig. 1a and Fig.  S1). It should also be noted that the subcellular distribution of LRP6-ICD was not significantly altered by co-expression with LEF-1 and/or transcriptional activation of the Wnt pathway by LiCl [7] (data not shown).

Nuclear AES positively modulates TCF/LEF-1 dependent transcription
AES is a ubiquitously expressed ,25 kDa cytosolic and/or nuclear protein [10,12,14,17,22]. Although AES lacks a putative NLS [11,15] its subcellular distribution is cell type and possibly context dependent [12,14,17,22,46]. Fractionation of HEK 293T cells expressing a non-targeted AES-HA construct (i.e. AES) demonstrates AES is a nucleocytoplasmic protein that exhibits a high cytosol:nuclear localization ratio ( Fig. 1b and  Fig. S1). It has been shown xenopus and mouse AES (,90% and 99% protein sequence identity to human AES) can enhance TCF/LEF-1 dependent TOP transcription [10,12] by antagonizing the repressor function of long Groucho/TLE members  [10,11,12,15,17,18,20,23]. Figure 4a confirms AES acts as a positive modulator of the Wnt pathway as transient transfection of AES amplified TCF/LEF-1 dependent TOP transcription by ,105%. It is speculated that AES functionally modulates transcription in the nucleus, but AES has also been shown to function in the cytosol [22]. To determine if AES mediated TOP transcription is spatially regulated, the TOP assay was utilized in HEK 293T cells transiently transfected with NES or NLS tagged AES constructs. Figure 1b shows the NES or NLS tagged AES constructs localize almost exclusively to the cytosol or nucleus. Whereas cytosolic AES (i.e. AES-NES) does not affect TOP transcription (Fig. 4b), nuclear targeted AES (i.e. AES-NLS) significantly enhances TOP transcription (Fig. 4c). Transcriptional modulation of TOP was specific for TCF/LEF-1 as none of the expression constructs significantly affected FOP or the transfection control reporter (TK-Renilla; data not shown). Neither AES nor cytosolic AES affected stabilization of endogenous cytosolic b-catenin (Fig. 4d) which was not unexpected since AES does not interact with b-catenin [8]. The results clearly show AES positively modulates TCF/LEF-1 activity and that AES functional activity is spatially regulated in HEK 293T cells. Similar to LRP6-ICD, AES subcellular distribution was not significantly altered by co-expression with LEF-1 and/or transcriptional activation of the Wnt pathway by LiCl [7] (data not shown).

LRP6-ICD interacts with AES in the nucleus
Previously we reported the LRP6-ICD interacted with AES in a yeast 2 hybrid assay [39] and AES is moderately expressed in HEK 293T cells [10,22]. Therefore, interaction with a nuclear protein that positively modulates TCF/LEF-1 activity, such as AES (Fig. 4c), is a possible mechanism by which nuclear targeted LRP6-ICD negatively modulates TCF/LEF-1 activity (Fig. 3c).
To determine a possible interaction, HEK 293T cells were cotransfected with AES and LRP6-ICD, fractionated and a coimmunoprecipitation assay (CoIP) was carried out on the cytosolic and nuclear fractions. The CoIP confirms the yeast 2 hybrid data that LRP6-ICD interacts with AES in situ (Fig. 5a) and that the interaction is exclusively nuclear (Fig. 5a; GFP blot, lane 4). The data also shows the AES/LRP6-ICD interaction does not require GSK3b mediated phosphorylation of LRP6-ICD as the nuclear phospho-mutant LRP6-ICD65 m interacted with AES ( Fig. 5a; GFP blot, lane 5). The control samples confirm the absence of a non-specific interaction between the constructs, anti-HA antibody and/or magnetic beads ( Fig. 5a; GFP blot, lanes 1-3).

Nuclear LRP6-ICD represses AES mediated activation of TCF/LEF-1 dependent transcription
By interacting with AES in the nucleus, it is possible that nuclear targeted LRP6-ICD repressed TCF/LEF-1 dependent TOP transcription (Fig. 3d) by repressing AES mediated  (Fig. 4a). To determine this, TCF/LEF-1 activity was measured in HEK 293T cells co-transfected with AES along with LRP6-ICD, LRP6-ICD-NES or LRP6-ICD-NLS. As shown above, AES overexpression induced TOP transcription (Fig. 4a) but co-expression with LRP6-ICD, which exhibits nuclear localization, significantly repressed AES mediated TOP transcription by ,65% (Fig. 6a). When LRP6-ICD was excluded from the nucleus, AES transcriptional activity was not repressed (Fig. 6b), but repression was re-established when LRP6-ICD was targeted to the nucleus (,137.6%; Fig. 6c). Repression of AES mediated TOP transcription by nuclear targeted LRP6-ICD and not cytosolic LRP6-ICD combined with the CoIP/interaction data (Fig. 5a) demonstrate LRP6-ICD and AES functionally interact exclusively in the nucleus.
Through various protein-protein interactions and post-translational modifications, Wnt pathway transcriptional activity is commonly modulated by influencing the stability and/or activity of cytosolic/nuclear b-catenin [2,3,5,52,53,54,55,56] as well as the transcriptional activity of TCF/LEF-1 [3,8,9,10,13,30,31]. In this study, we report a novel modulatory function for the soluble LRP6 intracellular domain (LRP6-ICD), the proteolytic product of LRP6, in that it can positively or negatively modulate Wnt pathway transcriptional activity (i.e., b-catenin stabilization and/ or TCF/LEF-1 activity) depending upon its subcellular localization and differential protein-protein interactions. Cytosolic LRP6-ICD positively modulates TCF/LEF-1 activity indirectly through b-catenin. Further, we show the LRP6-ICD can translocate to the nucleus where it negatively modulates TCF/LEF-1 activity by interacting with and repressing nuclear AES, a positive TCF/LEF-1 modulatory protein [10,12]. We also demonstrate the functional relationship between LRP6-ICD and AES is exclusively nuclear.
Cleavage, release, nuclear translocation and subsequent transcriptional modulation by the soluble intracellular domain (ICD) of a transmembrane receptor is a novel signaling mechanism that has been demonstrated for various cell surface receptors including Fz and the LRP6 family member, LRP1 [40,41,42,43,44]. As a result, such receptors (e.g., LRP -1 and -6 and Fz) can modulate transcription through multiple mechanisms. In some cases both the full length receptor and corresponding soluble ICD similarly affect transcription (e.g., Fz) [40,44] whereas certain ICD's differentially affect transcription compared to their full length receptor (e.g., PC1, LRP1) [43,57]. In this respect and similar to LRP1/LRP1-ICD, activated full length LRP6 positively affects transcription whereas the soluble nuclear LRP6-ICD negatively modulates TCF/LEF-1 activity.
While modulation of TCF/LEF-1 activity by LRP6-ICD or AES was significant, the scale of modulation in relation to overexpression clearly indicates that both proteins act as modulators and not primary effectors of Wnt pathway transcriptional activity. Although expression of endogenous Wnt target genes has been shown for primary activators of the Wnt pathway [2], no significant changes could be detected for LRP6-ICD or AES (data not shown). This was not unexpected as robust activation of the pathway, as shown for overexpression of primary activators, is required to initiate expression of Wnt target genes [8,9]. It is also possible that LRP6-ICD and/or AES are capable of modulating only a specific subset of TCF/LEF-1 dependent Wnt target genes that were not included in our analysis.
Although differential modulation of transcription has been shown for non-ICD proteins, such as Hipk1 which differentially modulates b-catenin:TCF/LEF-1 transcription [32], there are no reports to our knowledge showing a particular ICD can differentially modulate transcription [40,41,42,43,44,57,58]. We demonstrate here for the first time [41,42], to our knowledge, that a released ICD has dual functions as both a positive and negative modulator of transcription whose functional activity is dictated by a combination of localization and differential protein-protein interactions. Cytosolic LRP6-ICD positively modulates Wnt pathway activity by interacting with [35,38,39] and attenuating GSK3b [35,37,38,39], a negative Wnt pathway regulator, which indirectly up-regulates TCF/LEF-1 activity through enhanced bcatenin stabilization. However, nuclear LRP6-ICD negatively modulates TCF/LEF-1 activity by interacting with AES, which functions in a dominant negative manner over the TCF/LEF-1 repressive activities of long Groucho/TLE members. By interacting with and possibly sequestering AES, nuclear LRP6-ICD may attenuate the dominant negative activity of AES, thereby allowing long Groucho/TLE members to reengage in full repression of TCF/LEF-1.
We show that the functional activity of LRP6-ICD and AES as well as their functional relationship is regulated in part by localization. Both proteins lack a putative NES or NLS and therefore regulation by the importin or exportin family of proteins seems unlikely [59]. Although the size of LRP6-ICD (,20 kDa) and AES (,25 kDa) allow for nuclear diffusion [60], it is reasonable to assume the transcriptional activity of the Wnt pathway is not modulated by the random diffusion of effector proteins. Subcellular distribution of LRP6-ICD does not appear to be affected by cell type as we showed it is primarily cytosolic in CHO, HEK 293T and COS (data not shown) cells which suggests LRP6-ICD functions primarily as a positive modulator within the Wnt pathway. Perhaps similar to the soluble Fz-ICD [40], LRP6-ICD nuclear localization is regulated through interaction with other proteins in a context-dependent manner. In contrast, AES subcellular distribution appears to be influenced by cell type as it is exclusively nuclear in Gonadotropin releasing neurons [17], cytosolic in COS cells [12,14] and here we show nucleocytoplasmic distribution in HEK 293T cells. Although AES is primarily cytosolic in HEK 293T cells, we show nuclear AES functionally modulates TCF/LEF-1 dependent transcription. This suggests AES activity is also spatially regulated in HEK 293T cells. It should be noted that transcriptional activation of the Wnt pathway by LiCl [7] and/or co-expression of LEF-1 did not significantly alter the subcellular distribution of LRP6-ICD or AES in HEK 293T cells (data not shown). Therefore, we were unable to determine the regulatory mechanism(s) and/or proteins that mediate their subcellular distribution. The assembly of transcriptionally active or repressive complexes depends on the concentrations of long Groucho/TLE and bcatenin protein(s) in the nucleus and their affinities for TCF/LEF-1 proteins [8]. This delicate balance is important as Wnt pathway transcriptional activity influences a broad range of cellular functions. For example, the balance between long Groucho/ TLE members (via HDAC recruitment) and b-catenin on TCF transcriptional activity regulates oligodendrocyte differentiation, timing of differentiation and maturation [13]. Therefore, effector proteins such as AES and/or LRP6-ICD which are capable of modulating TCF/LEF-1 activity through long Groucho/TLE and/or b-catenin, have the potential to influence Wnt pathway mediated cellular functions such as proliferation/differentiation. However, the influence modulatory proteins have on cellular function can be difficult to quantify for it can be influenced by environmental cues, cell type, specific cellular function analyzed and/or the magnitude by which Wnt pathway transcriptional activity is modulated, just to name of few.
In summary, we show the soluble LRP6-ICD differentially modulates Wnt pathway transcriptional activity depending upon its subcellular localization and differential protein-protein interactions. In future studies it will be interesting to further examine how the LRP6-ICD/AES interaction affects cellular function mediated by the Wnt pathway and the cellular cues that regulate their functional localization.