Bili Inhibits Wnt/β-Catenin Signaling by Regulating the Recruitment of Axin to LRP6

Background Insights into how the Frizzled/LRP6 receptor complex receives, transduces and terminates Wnt signals will enhance our understanding of the control of the Wnt/ß-catenin pathway. Methodology/Principal Findings In pursuit of such insights, we performed a genome-wide RNAi screen in Drosophila cells expressing an activated form of LRP6 and a β-catenin-responsive reporter. This screen resulted in the identification of Bili, a Band4.1-domain containing protein, as a negative regulator of Wnt/β-catenin signaling. We found that the expression of Bili in Drosophila embryos and larval imaginal discs significantly overlaps with the expression of Wingless (Wg), the Drosophila Wnt ortholog, which is consistent with a potential function for Bili in the Wg pathway. We then tested the functions of Bili in both invertebrate and vertebrate animal model systems. Loss-of-function studies in Drosophila and zebrafish embryos, as well as human cultured cells, demonstrate that Bili is an evolutionarily conserved antagonist of Wnt/β-catenin signaling. Mechanistically, we found that Bili exerts its antagonistic effects by inhibiting the recruitment of AXIN to LRP6 required during pathway activation. Conclusions These studies identify Bili as an evolutionarily conserved negative regulator of the Wnt/β-catenin pathway.


Introduction
Wnt signaling through b-catenin regulates the expression of genes involved in cell proliferation and cell fate during development, in adult homoeostasis, and in diverse diseases [1]. In the absence of a Wnt ligand, steady-state levels of b-catenin are maintained at a low level by a 'degradation complex' that promotes the phosphorylation, ubiquitination, and proteosomal degradation of b-catenin [2]. The binding of Wnt ligands to the transmembrane FZD (frizzled) and LRP5/6 (low-density-lipoprotein-related protein 5/6) co-receptors triggers an inhibition of the degradation complex resulting in the stabilization and nuclear accumulation of b-catenin, where it binds to transcription factors of the TCF/LEF family and regulates target gene expression.
Signaling events immediately following Wnt binding to the receptor complex are poorly understood. Previous studies have shown that in the presence of Wnt ligand, Dishevelled (DVL) facilitates phosphorylation of the cytoplasmic tail of LRP5/6 [3] by Casein Kinase 1c (CSNK1G1) and Glycogen Synthase Kinase-3b (GSK3B) [4,5]. Phosphorylation of LRP5/6 is followed by the recruitment of AXIN away from the degradation complex. As AXIN is likely a rate-limiting scaffold protein required for the degradation of b-catenin [6], the relocation of AXIN to the membrane away from the degradation complex likely promotes stabilization of b-catenin [7].
We performed a genome-wide RNAi screen in Drosophila cells to identify genes that regulate LRP6 mediated activation of a bcatenin reporter. Here we describe Bili (Band4.1 inhibitor LRP interactor), a previously uncharacterized FERM domain containing protein. Analyses of Bili function in Drosophila, zebrafish, and cultured human cells support the conclusion that Bili is an evolutionarily conserved antagonist of b-catenin signaling. Mechanistically, we show that Bili regulates the recruitment of AXIN to LRP6.

Results
RNAi screen in Drosophila identifies Bili as a negative regulator of Wg signaling Drosophila mutagenesis screens have resulted in the identification of several core Wg signaling proteins [8]. The potential for discovering additional contextually relevant molecular players in the pathway is enhanced by robust read-outs such as transcriptional reporter assays coupled with genome-wide RNAi libraries [9]. Our goal was to identify new proteins that regulate Wnt/bcatenin signaling at the level of the receptor complex. While Wnt and Frizzled proteins can signal independent of b-catenin, signaling through b-catenin is dependent on LRP5/6 co-receptors [7]. Therefore, we modified a Drosophila dsRNA screen [10] by using a constitutively active LRP6 mutant (DNLRP6) [11] to activate a fly-optimized b-catenin luciferase reporter, dTF12 [10]. In a high-throughput RNAi screen, clone 8 (cl8) Drosophila cells were transfected in a 384-well plate format with individual dsRNAs, dTF12, DNLRP6 and a Renilla luciferase control for cell viability and transfection efficiency. The cells were incubated for four days to allow knockdown of target RNAs and normalized reporter luminescence was then measured as an indicator of Wnt pathway activity. Approximately 0.5% of the library (,100 dsRNAs) modified the activation of the reporter by DNLRP6 in a manner validated by low throughput confirmation of the primary screen. dsRNA targeting eighteen genes with known human orthologs increased reporter activation two-fold or greater (Table S1). Seven of the eighteen amplicons that had potential off-target effects according to the publicly available Drosophila RNAi Screening Center (DRSC, flyrnai.org) database were not pursued further.

Bili is a conserved FERM-domain protein
Of the 11 remaining candidate negative regulators, we chose to focus on CG11848 (dBili) due to its predicted plasma membrane localization and the specificity of its effect on the Wg pathway. More specifically, of the 12 other screens in the DRSC database, dBili was identified only as a weak hit in a MAPK kinase screen (Friedman et. al., 2007). Bili, annotated as CG11848 (fly) and FRMD8/FKSG44 (human), also has orthologs in worms, zebrafish and mice (Fig. S1A). In human tissue samples, we found Bili mRNA expression highest in heart and spleen (Fig. S1B). Bili is a predicted 464 amino acid protein with a Band4.1/FERM domain in its N-terminus (amino-acids, 26-272) (Fig. S1C). FERM domains (4.1 Ezrin Radixin Meosin) are highly conserved protein domains important for directing plasma membrane association and, in some cases, linking integral membrane proteins to the actin cytoskeleton and modulating signaling [12]. Given that no prior Band4.1 domain proteins have been implicated in Wnt signaling, we focused on Bili to determine whether it functioned in a unique capacity in the Wnt/b-catenin pathway.
Bili dsRNA validates in low throughput assays dsRNA targeting either dBili or Ran, a previously identified negative regulator of Wg signaling [13], enhanced the activation of dTF12 by either DNLRP6 or Wg (Fig. 1A). As expected, dBili dsRNA decreased endogenous dBili transcripts compared to control dsRNA (Fig. S3A). A second independent dsRNA targeting dBili, dsBili_2, also decreased dBili transcript levels and enhanced DNLRP6-activated reporter activity ( Fig. S3A and B). Additionally, overexpressing dBili in S2R+ or Clone8 Drosophila cell lines inhibits Wg-dependent reporter activation ( Fig. S3C and D). These data support the conclusion that dBili is a novel negative regulator of Wg signaling.

Bili is coexpressed with Wg during Drosophila development
If dBili is a component of the Wg pathway in Drosophila, then it should be co-expressed with other genes involved in Wg signaling. To test this hypothesis we performed in situ hybridization of dBili in Drosophila embryos (Fig. 1B) and larval imaginal discs (Fig. 1C). At early stages (stages 5-6) dBili RNA is uniformly expressed in the early embryonic ectoderm at low levels (data not shown), and more highly expressed in the neuroectoderm (Fig. 1B, a and b). Additionally, dBili is expressed in cells adjacent to the ventral midline just prior to initiation of the invaginating ventral furrow (black arrows, Fig. 1B, a and b). Interestingly, at later stages (stages [13][14][15][16][17], the expression of dBili is evident in the developing central nervous system (CNS, Fig. 1B, c and e) and the ventral epidermis in stripes (black arrows, Fig. 1B, e and e') similar to the expression of Wg (Riggleman et al. 1990). At the end of embryogenesis (stage 16/17), the majority of dBili expression is restricted to the CNS (Fig. 1B, d).
dBili and Wg expression patterns also overlap in wing, leg, and the eye-antennal imaginal discs in third instar larvae (Fig 1C). In the wing imaginal disc, dBili is expressed broadly in the wing pouch (black arrows, Fig. 1C, panel a) and in the notum (red arrow, Fig. 1C, panel a). The expression of dBili in the notum and that of Wingless (Wg) protein (blue arrow, Fig. 1C, panel b) appears to be non-overlapping, however they do abut each other. In the leg disc, dBili is expressed in a small discrete domain in the posterior leg disc (red arrow, Fig. 1C, panel c) as well as in the ventral wedge (black arrows, Fig. 1C, panel c) in a pattern slightly broader yet nearly identical to Wg expression (white arrow, Fig. 1C, panel b). In both dorsal and ventral compartments of the eye disc, dBili expression (black arrows, Fig. 1C, panel d) partially overlaps with the lateral edge expression of Wg (white arrows, Fig. 1C, panel e). In the antennal disc, dBili is expressed as a dorsal wedge (red arrow, Fig. 1C, panel d) which also overlaps with that of Wg (blue arrow, Fig. 1C, panel e). We conclude that the overlapping and adjacent expression of dBili and Wg are consistent with dBili functioning in the Wg pathway.

dBili negatively regulates Wg signaling during Drosophila embryogenesis
We next investigated the function of dBili during Drosophila embryogenesis. Wg has a well-established role in patterning the ventral epidermis and has been shown to be involved in cell fate determination [14,15]. Epidermal cells with active Wg signaling secrete naked-cuticle, whereas cells lacking Wg signaling activity secrete denticle belts. If dBili is a negative regulator of Wg signaling, reduced Bili function should resemble Wg gain of function phenotypes including ectopic naked cuticle.
dBili levels were reduced during embryogenesis by expressing a short-hairpin RNA (shdBili) in the paired domain (prdGA-L4.UAS-shdBili). Quantitative real time-PCR (qRT-PCR) analysis following expression of dBili shRNA under the control of prd-GAL4 in the Drosophila embryo revealed .75% knockdown of dBili mRNA levels compared to prd-GAL4 control embryos (Fig.  S2A). As a control for the specificity of shdBili we also quantified levels of armadillo mRNA and found no change (Fig. S2A). Importantly, the expression of shdBili led to the expansion of Wg signaling activity as demonstrated by the secretion of mostly naked cuticle with very few cells secreting cuticle with trichomes (denticles) ( Fig. 2A and 2B). Furthermore, driving dBili shRNA expression with even skipped stripe 3/7-GAL4, disrupted denticle stripe 2/3 (Fig. S2B). Thus, depleting dBili in the Drosophila embryo phenocopies hyperactivated Wg signaling, suggesting that dBili is a negative regulator of the Wg pathway in Drosophila embryos.
Next we assessed the molecular consequences of depleting dBili in the paired domain. During embryonic segmentation, Wg activates and maintains the expression of engrailed (En) and defines para-segment boundaries [16]. Consistent with Wg gain-offunction, immunohistochemistry for En protein revealed a marked and uniform expansion of its expression in shdBili embryos (  Bili is conserved in vertebrates and negatively regulates Wnt/b-catenin signaling in zebrafish The Wg signaling pathway is highly conserved throughout evolution. Therefore, we next asked if Bili negatively regulates Wnt signaling in vertebrates. To this end, we asked if Bili function is conserved in zebrafish. Ectopic activation of Wnt/b-catenin signaling in early zebrafish development causes dose-dependent anterior truncations and mesodermal defects (Fig. 3A) [17]. If zfBili functions as a negative regulator of Wnt/b-catenin signaling during zebrafish development, then silencing zfBili expression should exacerbate these phenotypes. Overexpression of Wnt8 results in anterior truncation phenotypes of varying severity (Fig. 3A). Silencing Bili with either of two non-overlapping antisense morpholinos targeting zfBili (bilimo or bilimo2) shifted the Wnt8 induced phenotypes toward increasing severity (Fig. 3C,  top panel, Fig. S4E). As controls, we first confirmed that zfBili is expressed throughout early zebrafish development (Fig. S4A-D) and second verified that bilimo, but not a control morpholino (como), repressed translation of zfBili (Fig. 3B).
We next investigated the effect of zfBili gain-of-function on zebrafish development. Embryos injected with zfBili RNA, but not control RNA, displayed a slight expansion of anterior structures, which is consistent with diminished Wnt/b-catenin signaling (data not shown) [18]. Importantly, zfBili overexpression rescued Wnt8 gain-of-function phenotypes (Fig. 3C, bottom). Together, these results demonstrate that Bili is an evolutionarily conserved negative regulator of Wnt/b-catenin signaling.
Bili negatively regulates Wnt/b-catenin signaling in human cultured cells We next tested whether Bili regulates Wnt signaling in human cells. We performed siRNA knockdown or cDNA overexpression of hBili in RKO colorectal carcinoma cells or human embryonic kidney (HEK293T) cells expressing a b-catenin responsive luciferase reporter and Renilla luciferase normalization control. Activation of the reporter in RKO cells following treatment with WNT3A conditioned media was enhanced by transfection of either hBili or AXIN1 siRNA and repressed by b-catenin siRNA (Fig. 4A). Similarly, activation of the reporter in HEK293T cells by transfection of WNT1, a pathway activator, was enhanced by hBili siRNA and repressed by b-catenin siRNA (Fig. S5A). Consistent with these loss-of-function studies, overexpression of hBili in HEK293T cells inhibited WNT3A-mediated activation of the reporter (Fig. 4B).
A series of additional controls confirm these results and the conclusion that Bili is a negative regulator of the Wnt/b-catenin pathway in human cells. First, hBili siRNA effectively reduced both hBili mRNA and tagged protein levels. By qRT-PCR, hBili siRNA depleted hBili transcript levels by 50% (Fig. S5B), and reduced levels of overexpressed hBili protein (Fig. S5C). Second, four additional independent siRNAs designed against hBili had a similar or greater effect on Wnt-mediated activation of the b-catenin responsive reporter (Fig. S5D and E and data not shown) ruling out the formal possibility that the activity of the first siRNA was due to off-target effects. Third, hBili siRNA synergized with WNT3A conditioned media to induce transcriptional upregulation of the endogenous b-catenin target gene, AXIN2 (Fig. S5F). This effect on expression of an endogenous b-catenin target gene validates the conclusions of the synthetic reporters. Fourth, knockdown or overexpression of hBili had no effect on control b-catenin unresponsive reporters (Fig. S6A andB) or a CREB reporter that was activated with forskolin (Fig. S6C). These data collectively support the conclusion that Bili is a conserved inhibitor of Wnt/b-catenin signaling in diverse species.

Bili functions upstream of b-catenin stabilization and associates with LRP6
We next carried out epistasis experiments in HEK293T cells expressing a b-catenin responsive luciferase reporter and Renilla luciferase normalization control. hBili or GFP were overexpressed in the presence of BIO (6-bromoindirubin-39-oxime), a GSK3 inhibitor [19], or co-expressed with other pathway activators including WNT1, DNLRP6, Dishevelled (DVL) or b-catenin. Bili gain-of-function reduced reporter activation by WNT1, DNLRP6, and DVL (Fig. 4C). In contrast, hBili overexpression did not affect reporter activation by overexpression of b-catenin or BIO treatment (Fig. 4C). These results place the function of hBili between the Wnt receptor complex and the b-catenin degradation complex. Consistent with these epistasis studies, and the observed membrane associated localization of other FERM domain containing proteins, we found that dBili localizes in a concentric ring, adjacent to Arm protein at the plasma membrane in fly cells ( Fig. S3E and S3E').
Based on these results we predicted that Bili might regulate signaling from the Wnt co-receptors LRP5/6 or Frizzled. In direct support of this prediction, tagged hBili colocalizes with tagged LRP6 as detected by immunocytochemistry (ICC) (Fig. 4D). Furthermore, hBili-FLAG pulls down tagged-LRP6 in co-immunoprecipitation (co-IP) assays (Fig. 5A). The association of Bili and LRP6 was also confirmed in Drosophila S2R+ cells where immunoprecipitated tagged-DNLRP6 pulled down tagged-dBili (Fig. 5B). Collectively, these data suggest that Bili functions upstream of b-catenin stabilization and associates with the LRP6 receptor complex.

Bili antagonizes the recruitment of Axin to LRP6
During Wnt signaling, LRP6 is phosphorylated in a Wntdependent manner by CSNK1G1 and GSK3B, which then facilitates the binding of AXIN to LRP6 [4,5]. Consistent with this model, overexpression of GSK3B consistently and robustly mimics the ability of WNT to induce the binding of AXIN to LRP6 ( [20,21]. Our observations that Bili negatively regulates Wnt/bcatenin signaling upstream of the destruction complex and associates with LRP6 lead us to hypothesize that Bili may regulate the recruitment of AXIN to LRP6. To address this hypothesis we employed co-immunoprecipitation to monitor the association of AXIN and LRP6 following gain-offunction (Fig. 5C) or loss-of-function (Fig. 5D) of Bili. HEK293T cells stably expressing LRP6-EGFP were transiently transfected with AXIN-MYC. As expected LRP6 co-immunoprecipitated AXIN in the presence of overexpressed GSK3B (Fig. 5C & 5D, lane 1 compared to lane 3). This interaction of LRP6 and AXIN was attenuated approximately 2-fold in the presence of overexpressed hBili (Fig. 5C, lane 2 compared to lane 1) and enhanced more than 2-fold when hBili levels were knocked down ( Figure 5D, lane 2 compared to lane 1). Furthermore, even in the absence of GSK3B, knockdown of hBili enhanced the association of AXIN and LRP6 approximately 4-fold (Fig. 5D, lane 4 compared to lane 3). Together these data suggest that Bili negatively regulates the recruitment of AXIN to LRP6 thereby providing a mechanism for how Bili negatively regulates Wnt/b-catenin signaling.

Discussion
Many aspects of Wnt/b-catenin signal transduction remain poorly understood. One aspect in particular is the transduction events occurring at the plasma membrane following Wnt activation of the Frizzled/LRP6 coreceptor complex. Studies over the last several years have shed light on these events by uncovering several kinases that phosphorylate LRP6 resulting in its recruitment of AXIN [4,5]. As AXIN recruitment to LRP6 appears to be a key step in the propagation of a WNT signal, it is likely a hotspot for pathway regulation. Our identification of Bili as a negative regulator of Wnt/b-catenin signaling and inhibitor of AXIN recruitment to LRP6 is the first report of negative regulation at this step in pathway transduction. The specific details of how Bili affects AXIN recruitment will be the focus of future studies. One proposed mechanism is that Bili sterically hinders the recruitment of the kinases required for LRP6 phosphorylation and subsequent AXIN recruitment. A second possibility is that Bili inhibits the formation of LRP6 aggregates or 'signalosomes' which have been shown to be the LRP6 population associated with AXIN (Bilic et al. 2007).
We propose that inhibitors of the Wnt/b-catenin pathway fall into two groups: 'constitutive inhibitors' that keep basal levels of bcatenin low, such as members of the degradation complex; and 'inducible inhibitors' that act outside of the degradation complex and may function only when the pathway has been activated. The 'constitutive inhibitors' would include APC and AXIN, while 'inducible inhibitors' would include NKD and Bili. Consistent with this categorization, NKD and Bili both antagonize a WNT signal between receptor activation and the degradation machinery as well as having similar phenotypes in both flies [22] and fish [23]. In concert, constitutive and inducible inhibitors allow for rapid pathway activation in response to WNT ligand and tight regulation of the strength and duration of the WNT signal. Continued characterization and identification of both constitutive and inducible inhibitors is key to our understanding of the regulation of Wnt/b-catenin signaling in development and disease.

Fly screen
The screen was performed as previously described [10].

Drosophila in situ hybridization and immuno-staining
In situ hybridization and immuno-staining in the fly embryos and larval imaginal discs were performed using standard protocols described in the laboratory manual ''Drosophila Protocols'' by William Sullivan et al., CHSL press. qRT-PCR in Drosophila embryos RNA was isolated from Drosophila embryos using Trizol and was further purified using the RNA ''clean-up'' protocol with the Qiagen RNAeasy kit. This was followed by reverse transcription

Constructs
MGC clone for human Bili (MHS1010-9204707) was purchased (Open biosystems). Bili was sub-cloned into pCS2+ using the EcoR1 and Xho1 sites. Forward primer sequence 59 GTAGTAAAGCTTGCCACCATGGCCCTGAGGATG-GACGGG 39 and reverse primer sequence 59 TACTACGTC-GACTCAGCCCTGCTCCAGGCTGT 39. Zebrafish Bili was cloned out of a mixed embryonic library (6 hpf-24 hpf) using forward primer 59 TCTCCAGCTCAGGATTTGTTGGTG 39 and reverse primer 59 TACTACCTCGAGTCAACTTTCAGT-CAC 39 based on an ENSEMBLE BLAST prediction using Bili. Since the ATG was missing, a 59 RACE kit (Invitrogen) was used with a nested primer 59 GCCACTGACGACACAGCTTGTAA 39 that yielded the full-length clone.

Co-immunoprecipitation
HEK293T cells were made to stably express LRP6-EGFP by transfecting pCS2+LRP-EGFP and pPUR (Clontech) into cells and treating with 2 mg/ml of puroMYCin. Stable cells were seeded into 6-well plates at 50% confluency and transfected with AXIN-MYC (300 ng), GSK3B (50 ng) and Bili-FLAG (200 ng). Cells were lysed the next day in a buffer containing 1% Triton-X, 50 mM Tris-HCL pH 7, 150 mM NaCl, protease and phosphatase inhibitors. LRP6-EGFP was immunoprecipitated using polyclonal GFP and protein G beads for 2 hours. The beads were washed three times with the lysis buffer that contained 0.6 M NaCl instead of 150 mM NaCl. Proteins that bound were eluted using SDS gel loading buffer and western blot analysis was used to detect GFP, MYC, FLAG and GSK3B. Quantification was done using NIH image.

Zebrafish in situ hybridization and microinjection
In situ hybridization and microinjections were carried out as previously described on zebrafish online resource, ZFIN. ATG blocking morpholino sequences (Gene Tools) GGAAGTCGCCAT-CATCTCCCTCCAT (bilimo) and GAGACACTCTCCTTC-GATTCAGAAG (bilimo2). A 10 mM stock was diluted to 1.5 mM. bilimo, bilimo2, or control morpholino at this concentration were then coinjected into 1-cell embryos at three doses of wnt8 RNA (stock concentrations 2 ng/ml, 4 ng/ml and 8 ng/ml). The same wnt doses were used for zfBili overexpression coinjection experiments with Renilla as a control at 50 ng/ml stock concentration. Injection drop size 0.5-1 nl.

QT-RTPCR
RNA was isolated following treatment using Qiagen RNAeasy kit. This was followed by reverse transcription (Invitrogen ThermoScript or SuperScript) and quantitative PCR (sybr green) using Roche Light cycler 2.0. PCR primers Bili forward 59CCAAGCAGGCC-GAACT39, Bili reverse 59CCTTGCCGTCCTCCACGTA39, GAPDH forward 59CCA CCC ATG GCA AAT TCC ATG GCA39, GAPDH reverse 59TCTAGACGGCAGGTCAGGTC-CACC39. Axin 2 forward 59CTCCCCACCTTGAATGAAGA39 Axin 2 reverse 59 TGGCTGGTGCAAAGACATAG39 Gene accession numbers. NM_143085 (Drosophila), XM_682750.2 (Zebrafish), NM_031904 (Human).   Figure S2 (A) qRT-PCR analysis of dBili knockdown. Expression of short-hairpin (shRNA) dbili (#3M) under the control of prd-GAL4 in the Drosophila embryo results in .75% knockdown in message level (Red bar) as compared to prd-GAL4 control embryos (Blue bar). As a control, Armadillo mRNA was not affected in embryos expressing dBili shRNA. (B) Embryos expressing dBili shRNA under the control of even skipped stripe 3/7-GAL4: These embryos display a partial lack/disruption of the 2nd or 3rd denticle belt which coincides with the eve-stripe 3 expression. This phenotype is consistent with a localized increase in Wingless signaling activity in the region around eve-stripe 3. The same however was not observed for stripe 7 (posterior end of the embryo), perhaps due to differential expression of GAL4 in stripe 3 versus stripe 7. cDNA and a b-catenin responsive reporter and assayed the following day. (B) hBili siRNA effectively knocked down hBili transcripts in HEK293T as measured by qRT-PCR. (C) hBili siRNA effectively decreased the expression of a hBili-Venus fusion protein as assayed by fluorescence measurement. hBili siRNA had no effect on Venus expression. (D) A second siRNA (Bili #2) targeting hBili enhanced WNT1 mediated activation of a b-catenin responsive reporter. (E) Bili #2 siRNA also effectively decrease expression of a hBili-Venus fusion protein with no effect on Venus alone. (F) hBili knockdown in HEK293T cells with hBili siRNA enhanced the WNT3A mediated transcriptional induction of endogenous AXIN2 transcripts as assayed by qRT-PCR. Error bars represent STDEV (*p,0.05, **p,0.005, ***p,0.0005). Found at: doi:10.1371/journal.pone.0006129.s006 (1.48 MB EPS) Figure S6 Bili does not affect control reporters or CREB responsive reporters. (A) HEK293T cells were transfected with FUBAR, a control reporter not responsive to Wnt/b-catenin signaling, renilla luciferase normalization, and control or hBili siRNA. FUBAR was not responsive to WNT3A conditioned media (right two bars) and hBili siRNA had no effect. (B) HEK293T cells were transfected with FOPFlash, another control reporter not responsive to Wnt/b-catenin signaling, renilla luciferase normalization, and GFP cDNA or increasing doses of hBili cDNA or control siRNA or hBili siRNA. Neither hBili cDNA control nor hBili siRNA had an effect on the reporter. (C) hBili siRNA HEK293T cells stably expressing a CREB responsive reporter were transfected with renilla luciferase normalization, and control, hBili, or hBili#2 siRNA and treated with DMSO, 1 uM forskolin, or 10 uM forskolin. Error bars represent STDEV. Found at: doi:10.1371/journal.pone.0006129.s007 (1.14 MB EPS)