Human FasL Gene Is a Target of β-Catenin/T-Cell Factor Pathway and Complex FasL Haplotypes Alter Promoter Functions

FasL expression on human immune cells and cancer cells plays important roles in immune homeostasis and in cancer development. Our previous study suggests that polymorphisms in the FasL promoter can significantly affect the gene expression in human cells. In addition to the functional FasL SNP -844C>T (rs763110), three other SNPs (SNP -756A>G or rs2021837, SNP -478A>T or rs41309790, and SNP -205 C>G or rs74124371) exist in the proximal FasL promoter. In the current study, we established three major FasL hyplotypes in humans. Interestingly, a transcription motif search revealed that the FasL promoter possessed two consensus T-cell factor (TCF/LEF1) binding elements (TBEs), which is either polymorphic (SNP -205C>G) or close to the functional SNP -844C>T. Subsequently, we demonstrate that both FasL TBEs formed complexes with the TCF-4 and β-catenin transcription factors in vitro and in vivo. Co-transfection of LEF-1 and β-catenin transcription factors significantly increased FasL promoter activities, suggesting that FasL is a target gene of the β-catenin/T-cell factor pathway. More importantly, we found that the rare allele (-205G) of the polymorphic FasL TBE (SNP -205C>G) failed to bind the TCF-4 transcription factor and that SNP -205 C>G significantly affected the promoter activity. Furthermore, promoter reporter assays revealed that FasL SNP haplotypes influenced promoter activities in human colon cancer cells and in human T cells. Finally, β-catenin knockdown significantly decreased the FasL expression in human SW480 colon cancer cells. Collectively, our data suggest that β-catenin may be involved in FasL gene regulation and that FasL expression is influenced by FasL SNP haplotypes, which may have significant implications in immune response and tumorigenesis.


Introduction
FasL (Fas ligand or CD95 ligand) is a type II membrane protein and a member of the TNF ligand superfamily. FasL is mainly expressed in activated T cells, NK cells, macrophages, and various cancer cells. FasL triggers cell death and/or cell activation by binding and clustering Fas (CD95). Fas-mediated apoptosis or activation induced cell death (AICD) plays important roles in maintaining peripheral immune tolerance [1]. In mouse models, the gld and lpr mice containing the loss-of-function mutations in the FasL (gld mice) and Fas (lpr mice) genes develop spontaneous autoimmunity as a result of defective lymphocyte apoptosis [2,3]. The FasL/Fas system promotes immune tolerance through the deletion of auto-reactive T cells, B cells, and macrophages [4][5][6]. In humans, autoimmune lymphoproliferative syndrome (ALPS or Canale-Smith syndrome) is caused by the inherited loss-offunction mutations in Fas or FasL [7][8][9]. FasL also initiates cell activation and cell differentiation by engaging Fas, which promotes chronic inflammation and inflammatory responses [10,11]. Furthermore, FasL has a critical role in the pathogenesis of AIDS and in the induction of pulmonary silicosis [12][13][14].
The expression of FasL is tightly controlled in humans. Numerous transcription factors are involved the regulation of the FasL gene. NF-kB, NF-AT, Egr-3, and IRF-1 have been implicated in regulation of human FasL gene [37][38][39][40][41][42][43]. Previously, we reported that a polymorphic C/EBPb element in the FasL promoter region is involved in the regulation of FasL expression [44]. In the current study, we characterized two TCF/LEF-1 binding elements (TBEs) in the FasL promoter region. We demonstrate that FasL TBEs are involved in the regulation of promoter activity. Additionally, we show that FasL SNP haplotypes affected the promoter activities in human cells. Furthermore, our data support that b-catenin is involved in the regulation of FasL expression in cancer cells. Our study provides new insight into the genetics of human FasL and the mechanisms of FasL gene regulation in human immune cells and cancer cells.
Subsequently, we carried out transcription factor motif search to look for the potential transcription elements near the FasL polymorphic sites. Surprisingly, we found that a putative TCF/ LEF-1 binding element (TBE) is adjacent to the SNP -844C.T and that the second putative TBE sits on the FasL SNP -205C.G (Fig. 1C). For simplicity, we designated the putative TBE near the SNP -844C.T as TBE1 and the TBE containing the SNP -205C.G as TBE2 (Fig. 1C). Mutant FasL TBEs were generated by site-directed mutagenesis in the wild-type TBEs of the promoter reporter constructs (Fig. 1C).

Characterization of the distal TCF/LEF-1 binding element (TBE1)
The TCF/LEF-1 transcription factor members belong to the HMG (high mobility group) class of transcription regulators that have a highly conserved consensus recognition core motif, 59-CTTTG(A/T)-39 [45]. As shown in Fig. 1C, the nucleotide sequence between position -838 and -833 ( -838 ACAAAG -833 ) has a perfect match with the consensus T-cell factor (TCF/LEF-1) binding element  in the reverse orientation. Therefore, the putative FasL TBE1 (the distal TBE) is located next to the C/EBPb element and five nucleotides away from the SNP -844C.T. To determine whether the putative distal TBE (TBE1) could bind the specific transcription factors, we carried out electrophoretic mobility shift assays (EMSA). As shown in Fig. 2A, complexes were formed between the radio-labeled TBE1 probe and proteins in SW480 nuclear extracts (Lane 1, 2, and 5). The unlabeled TBE1 Specific Probe (cold SP) inhibited the complex formation between the radio-labeled probe and the nuclear extract ( Fig. 2A, Lane 3). The anti-TCF-4 antibody significantly decreased the specific complex formation between nuclear proteins and TBE1 probe ( Fig. 2A, Lane 4), providing evidence that TCF-4 is one of the transcription factors that binds the TBE1. In addition, cold Non-Specific probe (cold NP) failed to inhibit the complex formation between the labeled TBE1 probe and nuclear proteins ( Fig. 2A, Lane 5) and the mutations within the putative TBE1 binding core sequence abrogated the binding of the transcription factors (Lane 6).
In Wingless/Wnt signaling pathways, b-catenin could form complexes with nuclear TCFs to regulate gene expression [46]. To confirm whether the TBE1 probe is capable of forming super complexes with TCFs and b-catenin, we performed EMSAs using anti b-catenin antibody. Fig. 2B shows that radio-labeled TBE1 probe could form complexes with TCF-4 and b-catenin (Lane 1 and 4) and the unlabeled TBE1 probe inhibits the specific complex formation (Lane 2). Furthermore, anti b-catenin antibody dramatically decreased the formation of the probe-TCF-4-bcatenin complexes (b-Cat arrow-pointed band, Fig. 2B, Lane 3). Meanwhile, the probe-TCF-4 complex formation was notably increased (TCF-4 arrow-pointed band, Fig. 2B, Lane 3), confirming that super complexes could be formed between TBE1 probe and TCF-4 plus b-catenin. Mutations within the TBE1 core sequence disrupted the binding of both TCF-4 and bcatenin ( Fig. 2A, Lane 6 and Fig. 2B, Lane 5). Taken together, the putative TBE1 adjacent to the SNP -844 is indeed a TCF/LEF-1 transcription factor-binding element.
EMSAs provided direct evidence for the physical interaction between TCF-4 and the TBE1 probes in vitro. Next, we performed chromatin immunoprecipitation assays (ChIP) to confirm the in vivo interaction between the TBE1 and transcription factors (TCF-4 and b-catenin). As shown in Fig. 2C, anti TCF-4 antibody was able to specifically precipitate chromatin-DNA complexes containing TBE1 DNA fragment (Lane 5). In addition, anti b-catenin antibody was also able to precipitate chromatin-DNA complexes containing TBE1 DNA fragment (Fig. 2D, Lane 5). The identity of DNA fragments in the ChIP assay was further confirmed by direct DNA sequencing (data not shown). Our data confirmed that endogenous TCF-4 and b-catenin bound FasL TBE1 in vivo, consistent with the in vitro results in Fig. 2A and 2B.
Effect of SNP -205 on the binding affinities for TCF/LEF-1 transcription factors Multiple transcription factors (Egr-3, SP-1, and NF-AT) are located near FasL SNP -205C.G [38][39][40]42]. In addition, as FasL -205C allele is within the nucleotide sequence (59-C -205 TTTGA-39) that matches perfectly with the consensus TBE (59-CTTTGA-39) and therefore, we speculated that the SNP -205C.G may be within a functional TBE (Fig. 1C, TBE2). To examine whether the putative TBE2 is able to bind TCF-4, we carried out EMSAs using the double-stranded DNA probes. Fig. 3A shows that the labeled wild-type TBE2 probe (TBE2, -205C allele) formed complexes with the SW480 nuclear extract proteins (Fig. 3A, Lane 1, 2, 5, and 6). The complex formation was abrogated by the addition of unlabeled wild-type TBE2 probe (cold SP) (Fig. 3A, Lane 3), but the cold non-specific probes (cold NP) were unable to inhibit formation of the specific complexes (Fig. 3A, Lane 5 and 6), validating that the binding of nuclear proteins to the labeled probe was specific. Furthermore, the anti TCF-4 antibody dramatically reduced the complex formation (Fig. 3A, Lane 4), verifying that the transcription factor TCF-4 binds to the TBE2 probe. Interestingly, there was no complex formation between the labeled TBE2 mutant probe (mTBE2 or-205G allele) and TCF-4 ( Fig. 3A, Lane 7), suggesting that function of the putative FasL TBE2 may be affected by SNP -205 alleles.
We further determined whether the TBE2 probe can form supercomplexes with TCF-4 plus b-catenin. Besides the probe-TCF-4 complexes, we were able to detect probe-TCF-4-b-catenin supercomplexes (b-Cat arrow-pointed band, Fig. 3B, Lane 1 and 4). Again, the unlabeled TBE2 wild-type probe (cold SP) inhibited the specific complex formation (Fig. 3B, Lane 2). The anti b-catenin antibody was able to disrupt the specific super-complexes containing the labeled probe, TCF-4, and b-catenin (Fig. 3B, Lane 3). With the decrease of probe-TCF-4-b-catenin super-complexes, the probe-TCF-4 complexes increased dramatically (TCF-4 arrowpointed band, Fig. 3B, Lane 3), verifying that the wild-type TBE2 probe (-205C allele) could form super-complexes with TCF-4 and bcatenin. In contrast, the mutant TBE2 (mTBE2, -205G allele) probe almost lost the ability to bind TCF-4 and b-catenin (Fig. 3B, Lane 5). Taken together, our data indicate that SNP -205C.G is located in a bona fide TCF/LEF-1 transcription factor-binding element and the SNP -205C.G in the TEB2 affect the binding capacity for the transcription factors (TCF-4 and b-catenin). Subsequently, we used chromatin immunoprecipitation assays (ChIP) to examine whether TCF-4 and b-catenin bind to FasL TBE2 in vivo. As shown in Fig. 3C, anti TCF-4 antibody was able to specifically precipitate chromatin-DNA complexes containing TBE2 DNA fragment (Lane 5). In addition, anti b-catenin antibody was able to specifically precipitate chromatin-DNA complexes containing TBE2 DNA fragment (Fig. 3D, Lane 5). The identity of DNA fragments in the ChIP assay was further confirmed by direct DNA sequencing (data not shown). Our data confirmed that endogenous TCF-4 and bcatenine bound FasL TBE2 in vivo, which is in agreement with the in vitro results shown in Fig. 3A and 3B.

Up-regulation of FasL promoter activities by human LEF-1 and b-catenin
The high mobility group (HMG) domain of TCF/LEF-1 transcription factor family members is responsible for the binding to target genes in a sequence specific fashion while b-catenin provides a domain for transcription activation. The transcriptional activation of target genes occurs only when TCF/LEF-1 transcription factors are associated with b-catenin in cell nuclei. Nuclei of human APC -/colon cancer cells contain stable bcatenin-TCF-4 complexes for target gene activation [47]. To examine whether FasL TBEs are involved in the regulation of FasL promoter activities, we carried out promoter reporter assays in human colon cancer SW480 cells, which contain constitutively high levels of active b-catenin in the nuclei [47]. As shown in Fig. 4A, co-transfection of LEF-1 significantly increased promoter activities of the FasL reporter construct containing two wild-type TBEs in SW480 cells as compared with the vector control. To further confirm that FasL promoter activities could be affected by LEF-1 and b-catenin, we carried out the promoter reporter assays

Role of FasL TBEs in human colon cancer cells
Because the proximal FasL TBE (TBE2) contains the SNP -205C.G, next we examined whether the SNP could affect promoter activities. We carried out the promoter reporter assays with the constructs containing either the triplicate wild-type TBE2 (3XTBE2W, -205C allele) or the triplicate mutant TBE2 (3XTBE2M, -205G allele). As shown in Fig. 4C, triplicate wildtype TBE2 (-205C allele) increased reporter promoter activities four folds in SW620 cells (APC -/colon cancer cells) and twelve folds in SW480 cells (APC -/colon cancer cells) over the vector controls respectively. In contrast, triplicate mutant TBE2 (-205G allele) enhanced reporter promoter activities only by 1.3 folds in SW620 cells and by 2.9 folds in SW480 cells over the vector controls. Therefore, wild-type TBE2 (-205C allele) serves as a much better enhancer for FasL gene expression and that the SNP -205C.G may affect the FasL promoter activities in human colon cancer cells.
APC -/b-catenin-TCF signaling pathways specifically enhance expression in genes containing the TCF/LEF-1 transcription factor elements. To examine the role of TBEs within the FasL promoter, we generated four promoter reporter constructs. As shown in Fig. 4D, simultaneous mutations in both TBE1 and TBE2 (TBE1M/2M) significantly decreased FasL promoter activities in SW480 cells. However, FasL promoter constructs containing either one mutant TBE (single mutant TBE1 = TBE1M/2W or single mutant TBE2 = TBE1W/2M) failed to affect promoter activities in SW480 cells (Fig. 4D), suggesting that a single TBE (TBE1 or TBE2) may be necessary and sufficient to support FasL promoter function in APC -/colon cancer cells.

Effect of TBEs on FasL promoter activities in human T cells
b-catenin is critical in the TCR signaling pathway which mediates TCR2CD3-driven signals necessary for T cell differen-tiation [48]. We speculated that TBE may be one of critical transcription elements affecting FasL promoter function in human T cells. In deed, as shown in Fig. 5A, FasL promoter carrying mutant TBE1 and wild-type TBE2 (TBE1M/2W) had significantly lower activities than that carrying both wild-type TBEs (TBE1W/2W) in T cells at all conditions (P,0.01). Surprisingly, FasL promoter carrying the wild-type TBE1 and mutant TBE2 (TBE1M/2W) failed to affect the FasL promoter activities in T cells (Fig. 5A). Furthermore, although mutations in both TBE1 and TBE2 (TBE1M/2M) significantly reduced FasL promoter activities in T cells, yet similar promoter activities were observed between the FasL promoter carrying mutant TBE1 with wild-type TBE2 (TBE1M/2W) and the FasL promoter carrying mutant TBE1 with mutant TBE2 (TBE1M/2M). Collectively, our data demonstrate that TBE1 is one of critical elements in FasL promoter. On the other hand, FasL TBE2 may not be functionally important in T cells (Fig. 5A).

Differential effect of FasL SNP haplotpes in colon cancer cells and T cells
Single TBE1 mutation dramatically reduced FasL promoter activities in human T cells and but not in human colon cancer cells ( Fig. 5A and Fig. 4D). Because lymphocytes (T cells) and epithelial cells (colon cancer cells) express completely different sets of transcription factors, we speculated that FasL promoter SNP haplotypes may function differently in T cells and colon cancer cells. In deed, as shown in Fig. 5B, FasL-3 haplotype (-844T/-756G/-478A/-205G) had significantly reduced promoter activities compared to FasL-1 (-844C/-756A/-478A/-205C) in colon cancer cells (Fig. 5B, P,0.05). In contrast, the promoter activities were not significantly different between FasL-3 and FasL-1 haplotypes in T cells (Fig. 5C). On the other hand, FasL-1 (-844C/-756A/-478A/-205C) always drives the highest promoter activities and FasL-2 (-844T/-756A/-478A/-205C) the lowest among the three major haplotypes in colon cancer cells and in T cells (Fig. 5B and 5C). Our data indicate that SNP haplotypes might have different effects on FasL expression in different cell populations.

Role of b-catenin in the regulation FasL expression in SW480 cells
To further examine whether b-catenin is involved in FasL expression, we utilized b-catenin siRNA to knockdown expression of b-catenin in SW480 cells. As shown in Fig. 6, b-catenin mRNA expression decreased by 88% in SW480 cells transfected with bcatenin specific siRNA as compared to the cells transfected with scrambled control siRNA. Concomitantly, the expression of FasL mRNA decreased by 42% in the cells along with the decreased bcatenin mRNA. These data indicate that b-catenin promotes expression of FasL.

Discussion
TCF/LEF-1 transcription factors are expressed in lymphoid cells and cancer cells. TCF/LEF-1 transcription factor family has four major members; TCF-1, LEF-1, TCF-3, and TCF-4, which share homology in their DNA binding domain with members of the HMG box transcription factors family [49]. TCF/LEF-1 family members were originally identified as lymphoid-specific DNA-binding proteins that recognize the nucleotide sequence 59-CTTTG(A/T)-39 (or in reverse orientation, 59-(A/T)CAAAG-39) [50][51][52]. TCF/LEF-1 factors bind their target DNA sequences within gene promoters through the HMG box DNA binding domain. The HMG box not only mediates DNA sequence recognition, but also induces a dramatic bend in the DNA to facilitate assembly of functional nucleoprotein structures [53,54]. DNA binding by TCF/LEF-1 alone is not sufficient to cause transcription activation. Promoter activation is accomplished only after a functional bipartite transcription factor is created thorough complex formation between TCF/LEF-1 transcription factor and b-catenin [55]. Within the functional complex, TCF/LEF-1 contributes the DNA binding and b-catenin confers the transcription activation [56]. Therefore, TCF/LEF-1 family factors serve as repressor to suppress target gene expression in the absence of bcatenin binding [57]. The functional transcription complexes of TCF/LEF-1 transcription factors and b-catenin are formed as a result of extracellular signal transduction in which TCF/LEF-1 factors act as molecular switches for the gene expression.
TCF/LEF-1 binds a set of target genes and programs their transcriptional competence to respond to specific signaling for bcatenin, a key down stream component of the Wnt signaling pathway. In the presence of Wnt signals, the signaling cascade leads to translocation of b-catenin into nuclei, where b-catenin interacts with TCF/LEF-1 to generate transcriptionally active multi-protein complexes [46,49]. b-catenin is also a target of TCR signaling pathway as TCF/LEF-1 transcription factor members play critical roles in the T cell development [51,58].
Colon cancers typically express high levels of FasL, which presumably serves as a potent mediator of immune privilege [28]. Expression of FasL potentially enables colon tumors to counterattack Fas-sensitive anti-tumor immune effector cells through Fasmediated death. The high prevalence of expression of FasL in various tumor cells suggests that FasL may be a general perhaps essential factor in the inhibition of anti-tumor responses by cancer cells [19][20][21][22][23][24][25][26][27][28][29]36]. Despite the tremendous attention to define FasL expression in cancer cells, it remains unclear how non-lymphoid cancer cells gain ability to express FasL and whether genetic variations affect FasL expression in cancer cells. It is well established that mutations in the adenomatous polyposis coli (APC) tumor suppressor gene initiate the majority of colorectal cancers. One consequence of this inactivation is constitutive activation of b-catenin/TCF-mediated transcription [47,[59][60][61]. Wnt pathway targets many genes critical for cell survival and proliferation (list of target genes can be found at www.stanford. edu/˜rnusse/pathways/targets.html). The c-myc oncogene has been identified as a critical target gene in the Wnt signal pathway and the expression of c-myc was regulated by two TCF/LEF-1 binding elements (TBEs) in c-myc promoter [62]. In the current study, we identified two TCF/LEF-1 binding elements (TBE1 and TBE2) in human FasL promoter. We demonstrated that both TBE1 and TBE2 were capable of forming complexes with TCF-4 and/or TCF-4-b-catenin in vitro and in vivo. Additionally, the FasL TCF/LEF-1 binding element could serve as an enhancer in colon cancer cells carrying APC mutations. Mutational analysis and cotransfection assays further confirmed that TBE1 and TBE2 were involved in the regulation of FasL promoter activities in human cells. Furthermore, b-catenin knockdown significantly decreased FasL expression. Taken together, we propose that FasL is a target gene in the APC pathway and that the FasL TBEs are the critical transcriptional elements controlling FasL expression in cancer cells. Accordingly, one can speculate that the constitutively active b-catenin-TCF-4 pathways would either enhance FasL promoter activities in human cancer cells or enable cancer cells to express FasL through binding to FasL TBEs.
The identification of TCF/LEF-1 binding elements in the FasL promoter will provide valuable information for the understanding of FasL expression in human cancer cells. Interestingly, both TCF/LEF-1 binding elements in the FasL gene are either overlapped with or adjacent to known transcription factor binding sites. The TBE1 is overlapped with C/EBPb binding motif while TBE2 is situated close to an Egr3 binding element (59-G -215 TGGTGT -207 -39) (Fig. 1C). FasL promoter region around SNP -205C.G contains several critical transcription elements that are essential for the optimal responsiveness to TCR-mediated activation and confer cyclosporin A sensitivity [38][39][40]42]. Therefore, the SNP -205C.G may affect binding of multiple transcription factors to FasL promoter and influence the promoter activities. In the current study, we provided the direct evidence that the SNP -205C.G (in the context of haplotype) significantly affect the promoter activities in cancer cells. Additionally, SNP -205C.G has been reported to significantly associate with hemolytic anemia in lupus patients [63], suggesting that SNP -205C.G is a functional polymorphism.
Because the ability of cancer cells to express FasL may play important role in tumorigenesis, we also expanded the functional study for polymorphic C/EBPb element (SNP -844C.T) in human colon cancer cells. We observed that the SNP in C/EBPb also significantly altered FasL promoter activity in colon cancer cells in the context of SNP haplotypes (Fig. 5B). One can imagine that the differences in promoter activities of SNPs and/or SNP haplotypes may have a significant impact in the cancer development because the levels of FasL may affect cancer cells to establish immune privilege and metastasis. Hence, the polymorphism of C/EBPb (-844T/C) of FasL promoter might have significant implication in the prognosis of cancers that rely on FasL for metastasis. In deed, SNP in FasL promoter was reported to be significantly associated with susceptibility to esophageal squamous-cell carcinoma and cervical cancer, underscoring the functional importance of FasL SNPs in the cancer development [64,65].
Cross-linking TCR (T cell receptor) on T cells is a potent signal to upregulate FasL expression, which plays a critical role in AICD of T cells [66,67]. It is possible that TCR signaling may upregulate FasL expression through b-catenin/TCF pathway in view of the fact that TCR signaling drives nuclear accumulation of b-catenin in human T Cells [68]. Interestingly, b-catenin nuclear accumulation also down-modulates the inhibitory isoforms of TCFs (TCF1 and LEF-1) and alters the TCF isoform balance in favor of the stimulatory TCFs that are capable of binding b-catenin in T cells [68,69]. Consequently, TCR signaling may result in the upregulation of b-catenin/TCF target genes, which is consistent with the observations of increased FasL expression after TCR stimulation [66,67] and consistent with the assumption that FasL is b-catenin/TCF target gene. Activation-induced cell death (AICD) is a major mechanism to maintain immune homeostasis [66,67,70]. Consequently, the inducibility and level of expression of FasL under basal and induction conditions could be very crucial in maintaining immune tolerance and homeostasis. The biological importance of FasL is well established in TCR mediated AICD in lymphocyte homeostasis, in the maintenance of immune privilege within certain tissues, and in inflammations and infections. In this study, we characterized the function of three major FasL haplotypes, which affect the FasL promoter activities in human cancer cells and T cells. The functional characterization of FasL SNP haplotypes will certainly facilitate our understanding of the FasL expression in various human diseases.

Donors
Anti-coagulated peripheral blood was obtained from healthy normal volunteers. The human studies were reviewed and approved by the institution review board of the University of Alabama at Birmingham and all donors provided written informed consent.

Nucleic acid isolation
Genomic DNA was isolated using the Puregene DNA isolation kit (Gentra Systems, Minneapolis, MN).

Sequencing and amplification of FasL promoter region
The FasL promoter region was amplified from position -1032 to +33 with the sense primer 59-TTA TGC CTA TAA TCC CAG CTA CTC A-39 annealing to nucleotide position from -1032 to -1008, and anti-sense primer 59-CTG GGG ATA TGG GTA ATT GAA G-39 annealing to position from +12 to +33 (+1 site corresponds to the A of the ATG translation start codon). The PCR reaction was performed in a 9700 PCR System with 500 ng of DNA, 300 nM of each primer, 200 mM of dNTPs, 1.5 mM of MgCl 2 , and 2.5 U of Taq polymerase in a 50-ml reaction volume starting with 95uC for 5 min, 35 cycles of denaturing at 94uC for 30 s, annealing at 58uC for 45 s, extension at 72uC for 1 min with a final extension at 72uC for 7 min. All the PCR products (1065 bp) were separated on 2.5% agarose gels and purified with the QIAquick Gel Extraction Kit (QIAGEN, Valencia, CA). The purified PCR products were sequenced from both directions using BigDye terminator sequence kit on an ABI 377 Sequencer (Applied Biosystems, Inc., Foster City, CA). All new data have been deposited in GenBank SNP database as rs763110 for (SNP -844C.T), rs2021837 (SNP -756A.G), rs41309790 (SNP -478A.T), and rs74124371 (SNP -205 C.G).

Reagents
Rabbit anti human TCF-4 polyclonal IgG was from Santa Cruz Biotechnology (Santa Cruz, CA). Mouse anti b-catenin mAb IgG was obtained from Transduction Laboratories (Lexington, KY). Transfection reagents DMRIE-C and Lipofectamine 2000 were from Invitrogen. Reagents and vectors for luciferase assays were from Promega (Madison, WI). Protease inhibitor cocktail was obtained from Roche Diagnostics (GmbH, D-68305 Mannheim, Germany).

FasL promoter reporter constructs
The

Generation of the human LEF-1 and b-catenin expression constructs
The human lymphoid enhancer factor-1 (LEF-1) expression construct was generated by cloning the full length LEF-1 coding region cDNA (1200 bps) into the pcDNA3. The human leukemic T cell line (Jurkat clone E6-1, ATCC) was maintained in RPMI with 10% fetal calf serum, penicillin (1000 units/ml), streptomycin (1000 units/ml), and glutamine (2 mM). Each transient transfection experiment was carried out with 2610 6 Jurkat cells, 2 mg of reporter construct plasmid DNA plus 0.5 mg of pCMV.SPORT-b-gal plasmid DNA, and 4 ml of DMRIE-C reagent. Transfected cells were then cultured separately without treatment (no treatment), with treatment of 50 ng/ml PMA plus 1 mM ionomycin (PMA+Iono), or with anti-CD3 treatment (aCD3) for 18 hrs as previously described [44]. The harvested cells were washed twice with PBS (pH 7.4) before being lysed in 300 ml of 16 lysis buffer. Cell debris was removed by centrifugation and the supernatants were used in the luciferase reporter assay [44]. Relative luciferase light units, standardized to b-galactosidase activities in luciferase assays or standardized to Renilla luciferase activities in dual luciferase reporter assays, are reported as the mean of triplicate samples.

Nuclear extract preparations
SW480 cancer cells were grown to monolayer and harvested by treating with 0.25% trypsin and 0.02% EDTA. Jurkat cells (3.5610 6 /ml) were either stimulated with 10 mg/ml of LPS for two hours or cultured on anti-CD3 mAb coated plates for 3.5 hours before nuclear proteins were extracted as previously described [44].

Electrophoretic mobility shift assays (EMSAs)
For each binding reaction, 8 mg of nuclear extract was incubated in 16 binding buffer (4% glycerol, 1 mM MgCl 2 , 0.5 mM EDTA, 0.5 mM DTT, 50 mM NaCl, 10 mM Tris-HCl, pH 7.5, and 50 mg/ml poly(dI-dC) . poly(dI-dC)) with 32 P labeled probes in a volume of 10 ml. Binding reactions were carried out at room temperature for 30 min with 50,000 cpm (0.1-0.5 ng) of double-stranded oligonucleotides end-labeled with [c-32 P]ATP using T4 polynucleotide kinase. Unlabeled specific (SP) or nonspecific (NP) competitor probes were used at 200-fold excess. Protein/DNA complexes and unbound DNA probe were then resolved on 5% non-denaturing polyacrylamide gel and visualized by autoradiography. The following double-stranded oligonucleotide probes were used in these experiments: 1) wild-type TBE1

Confirmation of the TBE/DNA bindability by the chromatin immunoprecipitation (ChIP) assay
Chromatin immunoprecipitation (ChIP) assays were performed with Chromatin Immunoprecipitation Assay Kit as described by the vendor (Upstate, Lake Placid, NY). Briefly, Jurkat T cells (1610 7 ) were washed twice with ice cold PBS. The cells were fixed and cross-linked in 1% formaldehyde. The cells were then centrifuged and washed twice with ice cold PBS containing 16 proteinase inhibitor cocktail (Roche Diagnostics GmbH, D-68305 Mannheim, Germany). Finally, the cells were lysed in SDS Lysis Buffer for 10 minutes on ice. The lysates were sonicated to shear DNA to lengths between 200 and 600 bps. The sonicated cell supernatant was diluted 10 folds in ChIP dilution buffer and precleared with salmon sperm DNA/Protein A agarose beads for 30 minutes at 4uC with agitation. Anti human TCF-4 antibody was added to the pre-cleared supernatant. Salmon sperm DNA/ Protein A agarose beads were used to collect the antibody/ transcription complexes. The DNA extracted from the precipitated complexes was used for PCR amplification by following the vendor's instruction. For TBE1 ChIP assay, upper primer 59-AAT AAA TAA ACT GGG CAA ACA-39 (annealing from position -883 to -863) and lower primer 59-AAC TAC CAT TTA CCC TGA CCT-39 (annealing from position -741 to -721) were used. For TBE2 ChIP assay, upper primer 59-CAG AAA ATT GTG GGC GGA AAC TT-39 (annealing from position -291 to -269) and lower primer 59-CGG GAC CCT GTT GCT GAC TG-39 (annealing from position -97 to -78) were used. The ChIP assay PCR reactions were set up with 300 nM of each primer, 200 mM of dNTPs, 1.5 mM of MgCl 2 , 2.5 U of Taq polymerase, and 2 ml of the DNA template in a 50-ml reaction volume. The cycle reaction was started with 95uC for 5 min, 40 cycles of denaturing at 94uC for 30 s, annealing at 56uC for 45 s, extension at 72uC for 1 min with a final extension at 72uC for 7 min. The specific ChIP DNA products were separated on 2.5% agarose gels along with the 100 bp DNA molecular weight marker (Invitrogen) and the identities of PCR products of ChIP assays were further confirmed with direct DNA sequencing.

Haplotype determination
Haplotypes were deduced from homozygous donors and by using the PHASE 2.0.1 program for haplotype reconstruction [71,72]. Further confirmation of the FasL SNP haplotypes was carried out by cloning FasL promoter genomic DNA (1065 bps) from six heterozygous donors. The FasL promoter DNA fragment was cloned into pGEM-T Easy Vector (Promega Cor., Madison, WI) following vendor's manual. At least ten clones from each donor were sequenced on an ABI 377 Sequencer with ABI Dye Terminator Cycle Sequencing Kit.

b-catenin knockdown and quantitative real-time RT-PCR
The human colon cancer cell line SW480 was maintained in DMEM supplemented with 10% fetal calf serum and L-glutamine (2 mM). Cells (5610 5 ) were plated in each well of 12-well plates. Cells were transfected with 50 nM (final concentration in medium) of siRNA corresponding to either b-catenin or a Scrambled control siRNA (Dharmacon, SMARTpool siRNA, ThermoFischer Lafeyette, CO) using Lipofectamine 2000 transfection reagent (Invitrogen, Carlsbad, CA) after they reached 75% confluence (approximately 24 hours after plating). Cells were harvested 48 hours after transfection and RNA was isolated using the Qiagen RNAeasy isolation system (Qiagen, MD) according to manufacturer's instructions. Purified RNA samples were subsequently treated with RNAse-free DNAse I (Invitrogen) before reverse transcription with Quantas qScript cDNA Supermix according to manufacturer's instructions (Quantas Biosciences, MD). Real-time RT-PCR was performed using Ssofast EvaGreen Supermix with low Rox (Biorad, CA) on a 7500 Real Time PCR System with 7500 Software v2.0.1 (Applied Biosytems, Foster City, CA). Melting curve analysis was performed to ensure that the primers amplified the desired amplicons and that primer dimers were absent. The following primers were used for our assay: bcatenin-Forward (59-TCT TGC CCT TTG TCC CGC AAA  TCA-

Data analysis
TESS (Transcription Element Search Software; http://www. cbil.upenn.edu/tess/) and MatInspector (http://www.gsf.de/ biodv/matinspector.html) were used to search for candidate transcription factors around the SNP sites. Differences in FasL promoter activities of various constructs were analyzed by Student's t test. The null hypothesis was rejected at the 95% confidence level (P,0.05).