Pygopus is a core component of the β-catenin/TCF (T-cell factor) transcriptional activation complex required for the expression of canonical Wnt target genes. Recent evidence suggests that Pygopus could interpret histone methylation associated with target genes and it was shown to be required for histone acetylation. The involvement of a specific acetyltransferase, however, was not determined. In this report, we demonstrate that Pygopus can interact with the HAT (histone acetyltransferase) CBP [CREB (cAMP-responsive-element-binding protein)-binding protein]. The interaction is via the NHD (N-terminal homology domain) of Pygopus, which binds to two regions in the vicinity of the HAT domain of CBP. Transfected and endogenous hPygo2 (human Pygopus2) and CBP proteins co-immunoprecipitate in HEK-293 (human embryonic kidney 293) cells and both proteins co-localize in SW480 colorectal cancer cells. The interaction with CBP also enhances both DNA-tethered and TCF/LEF1 (lymphoid enhancing factor 1)-dependent transcriptional activity of Pygopus. Furthermore, immunoprecipitated Pygopus protein complexes displayed CBP-dependent histone acetyltransferase activity. Our data support a model in which the NHD region of Pygopus is required to augment TCF/β-catenin-mediated transcriptional activation by a mechanism that includes both transcriptional activation and histone acetylation resulting from the recruitment of the CBP histone acetyltransferase.
- cAMP-responsive-element-binding protein (CREB)-binding protein (CBP)
- colorectal cancer
- histone acetylation
- Wnt signalling
The canonical Wnt/Wg (wingless) signal transduction pathway is required for embryonic development and maintenance of adult epithelia [1,2]. The core element responding to activation of the pathway is the armadillo repeat-containing protein β-catenin. In the absence of signalling, β-catenin is sequestered at the plasma membrane where it is involved in cell adhesion . Cytosolic levels of β-catenin are attenuated by a destruction complex consisting of the tumour suppressor APC (adenomatous polyposis coli), scaffolding proteins such as axin, and the kinases GSK3β (glycogen synthase kinase-3β) and casein kinase I. The destruction complex is inhibited upon Wnt stimulation, thus allowing β-catenin to enter the nucleus to participate in transcriptional activation of target genes [4,5]. Transmission of the Wnt signal to transcriptional activation requires the nuclear protein Pygopus, which is linked to β-catenin via the adaptor protein lgs/Bcl9 (legless/B-cell lymphoma 9) [6,7]. Pygopus family members share two conserved motifs, the NHD (N-terminal homology domain) and the C-terminal PHD (plant homeodomain) separated by a proline-rich linker region. The PHD is necessary for interaction with lgs/Bcl9 in the active β-catenin TCF (T-cell factor) complex. The NHD, on the other hand, has been proposed to be required for transcriptional activation by an unknown mechanism [8–14].
Upon nuclear entry, the β-catenin complex replaces the co-repressor, Gro/TLE (groucho/transducin-like enhancer), bound to TCF/LEF (lymphoid enhancing factor), at target gene enhancers. Recent evidence suggests that repressors, including APC, β-transducin repeat containing protein and Gro/TLE, in addition to the active β-catenin complex, are recruited to Wnt target gene enhancers in a dynamic equilibrium that shifts in response to signalling . The presence of repressors and activators on the same complex provides a rapid on/off mechanism to ensure tight regulation of genes involved in critical cell-cycle control. In the absence of negative regulation, such as in tumour cells with mutations in APC, the complex enters a mode of constitutive activation, leading to overexpression of oncogenic pathway components.
Chromatin remodelling events occur to accommodate entry of the basal transcriptional machinery to target gene promoters. The modification of N-terminal histone tails protruding from nucleosomal particles regulates the association of chromosomal DNA with the multimeric histone core, thereby promoting or repressing gene activation by relaxation or tightening of the nucleosome . In Wnt-mediated transcriptional activation, for example, β-catenin recruits members of the MLL/SET1 (mixed-lineage-leukaemia/SET1-type) family of histone methyltransferases to enhancer sites of target genes, resulting in trimethylation of core Histone 3 on Lys4 (H3K4Me3) . Recent crystallographic evidence indicates that a complex consisting of the PHD of Pygopus and the HD1 domain of Lgs/BCL9 makes direct contact with the methylated histone , suggesting a role of Pygopus to interpret the methylated histone code at sites of transcriptional activation.
Histone acetylation, a prerequisite for chromatin relaxation , is also important for target gene activation [15,18]. Histone acetylation at Wg target genes upon Wg stimulation in Drosophila requires the activity of the HAT (histone acetyltransferase) CBP [CREB (cAMP-responsive-element-binding protein)-binding protein] . A number of chromatin remodelling proteins, including CBP, have been shown to be components of the β-catenin/TCF complex present at the promoters of mammalian Wnt-target genes [20–23].
Several studies indicate that Pygopus, and in particular its NHD domain, has transcriptional activity [8–14]. Furthermore, Pygopus was recently associated with Histone H3 acetylation in male germ cells . Consistent with these previous findings, we demonstrate in the present study, using HEK-293 (human embryonic kidney 293) cells, that the transcriptional activity of Pygopus is enhanced by the recruitment of CBP. In colorectal cancer cells, in which Pygopus is required for canonical Wnt signalling , we found that Pygopus and CBP form complexes in vivo and co-localize to nuclei. More specifically, we identified a direct interaction between CBP and the NHD region of Pygopus. This interaction was required to augment both DNA-tethered and TCF/LEF-dependent transcriptional activation. Our findings show that the recruitment of CBP by Pygopus is sufficient for histone acetylation, consistent with a model in which Pygopus/Bcl9 modifies the histone code by recruiting HAT activity.
MATERIALS AND METHODS
Cells and transfections
HEK-293 and SW480 colorectal cancer cells were obtained from A.T.C.C. HEK-293 cells were maintained in Dulbecco's modified Eagle's medium and SW480 cells were maintained in Liebovitz's L-15 medium. Media were supplemented with 10% fetal bovine serum and 10 units/ml penicillin/streptomycin. All transfections were performed using Lipofectamine™ and Plus Reagent (Invitrogen), as described in .
Plasmids and antibodies
GST (glutathione transferase)-, Gal4- and Myc-tagged human Pygopus2 constructs were PCR amplified and cloned in-frame into pGEX-4T1, pMG4 and pCS4+ respectively. pCS2+-hPygo2 (human Pygopus2) was previously described in . NPF-AAA mutant NHD hPygo2 constructs were prepared using the Site-Directed Mutagenesis kit (Stratagene). Full-length mouse CBP cDNA has been previously described in . FLAG- and HA (haemagglutinin)-tagged full-length, FLAG-RID-Bro (where RID is receptor interaction domain and Bro is bromodomain), FLAG-Bro-Q and HAT+1/2 CH3 CBP constructs were gifts from Dr Gary Paterno (Division of Biomedical Science, Memorial University, St. John's, Newfoundland, Canada) . FLAG-tagged CH3, Q-domain and HAT domain CBP constructs were PCR amplified and cloned in-frame into pCMV-Tag2 (Stratagene). The HAT-defective mutant of CBP (HAT mut) was constructed from full-length HA-tagged CBP using the Site-Directed Mutagenesis kit (Stratagene), resulting in amino acid substitutions at positions L1690K and C1691L . Human β-catenin was PCR amplified from HS-578Bst cDNA and cloned into pCS2+. All constructs were verified by sequencing. Primer sequences are given in Supplementary Table S1 (at http://www.BiochemJ.org/bj/422/bj4220493add.htm). Commercial antibodies were: hPygo2 (monoclonal and polyclonal), β-Catenin and CBP (monoclonal and polyclonal) (Santa Cruz Biotechnology), FLAG M2 (Sigma) and Myc 9E10 (Developmental Studies Hybridoma Bank). Polyclonal hPygo2 antiserum has been described previously in .
Coimmunoprecipitation, immunoblotting and immunofluorescence
HEK-293 cells were transfected with hPygo2 and FLAG–CBP constructs and protein extracts were prepared 48 h after transfection. For endogenous interactions in SW480 cells, approx. 500 μg of whole cell extract was used. Proteins were extracted in 1× Triton buffer [1% Triton X-100, 150 mM NaCl, 10 mM Tris/HCl, pH 7.5, 10 mM EDTA, 1 μg/ml leupeptin, 1 μg/ml aprotinin, 5 μg/ml Tos-Lys-CH2Cl (tosyl-lysylchloromethane; ‘TLCK’) and 0.5 mM PMSF]. Extracts were cleared by centrifugation for 10 min at 4 °C. Supernatants were then incubated with control antibodies (anti-Gal4, anti-GST or preimmune serum) and anti-CBP or anti-hPygo2 overnight at 4 °C. Protein-A–Sepharose or Protein-G–Sepharose (GE Biosciences) beads were added to capture antibody complexes. Beads were washed extensively with 1× Triton buffer, followed by two washes with 150 mM NaCl. Immunoblotting was performed as described in . Immunofluorescence was performed as described in . Myc and Myc-hPygo2 were transiently transfected into SW480 cells and fixed 24 h after transfection. Cells were immunostained using anti-Myc/anti-mouse Cy5 (indodicarbocyanine; Jackson Laboratories) and anti-CBP/anti-rabbit FITC (GE Biosciences). Endogenous proteins were stained with anti-CBP/anti-mouse Cy5 and anti-hPygo2/anti-rabbit FITC. Nuclei were stained with DAPI (4′,6-diamidino-2-phenylindole) (0.1 μg/ml). Images were obtained using an Olympus FluoView FV1000 confocal microscope.
GST pulldowns and HAT assays
GST fusion proteins were prepared using the GST–gene fusion system (GE Biosciences) and radiolabelled proteins were prepared using the Coupled Transcription/Translation kit (Promega). GST pulldowns using labelled in vitro translated proteins were performed as described in . For pulldown experiments using HEK-293 cell extracts, approx. 1 μg of GST fusion proteins were purified on glutathione–Sepharose 4B beads. Approx. 50 μg of cell extract were diluted in 1× Triton buffer and incubated for 2 h at 4 °C. Beads were washed extensively in 1× Triton buffer followed by washes in 150 mM NaCl. Bound proteins were separated by SDS/PAGE, transferred to nitrocellulose membranes and analysed by immunoblotting. HAT assays were performed independently in triplicate as described  using hPygo2 complexes immunoprecipitated from HEK-293 cells or purified GST-fused domains of hPygo2 incubated with HEK-293 cell extract. Proteins were incubated at 30 °C for 30 min in the presence of [14C]acetyl CoA and biotinylated H4 peptide (Millipore).
Transcription assays were performed independently in triplicate exactly as described in . Reporter constructs used were pFR-LUC for the DNA-tethered hPygo2 transcription assays. pGL3-OT (wild-type TCF-binding sites) and pGL3-OF (mutant TCF-binding sites)  were used for the TCF (OT/OF) reporter assays. Luciferase values were normalized to β-galactosidase activity by co-transfection of pRSV-βGal.
The hPygo2 NHD region interacts with the conserved HAT/CH3 region of the CBP/p300 family of histone acetyltransferases
We identified a direct interaction between hPygo2 and CBP by performing in vitro GST pull-down analyses. These assays were carried out using various fragments of hPygo2 fused to GST (Figure 1A, top panel) to probe interactions with in vitro synthesized regions of CBP. As shown in Figure 1(A; bottom panel), full-length CBP directly interacted with full-length hPygo2, as well as fusion proteins containing the NHD (amino acids 1–95), but not the PHD.
The CBP/p300 family is a well-characterized group of HAT-containing proteins consisting of highly conserved binding and catalytic domains common to all family members , as illustrated in Figure 1(B). To determine which region of CBP was interacting with the NHD domain of Pygopus, we performed GST pulldown assays using GST–NHD (1–95) and different in vitro synthesized fragments of CBP. Binding was observed between the NHD region of Pygopus and two distinct regions of CBP: a fragment containing the CH3 domain (aa 1698–1900, large arrowhead) and another fragment containing the CH2/HAT domain (aa 1088–1625, small arrowhead) (Figure 1B).
The in vitro interaction assays indicated the potential for interaction of hPygo2 with CBP. To confirm that hPygo2 and CBP interact in vivo, full-length constructs encoding both proteins were transiently transfected into HEK-293 cells (Figure 2A). Co-immunoprecipitation assays were performed using control, hPygo2 or CBP antibodies. Blots were then probed to detect the presence of both proteins. As expected, immunoprecipitated hPygo2 interacted with CBP (Figure 2A, left panel). In the reciprocal experiment, immunoprecipitated CBP interacted with hPygo2 (Figure 2A, right panel). To confirm that the Bro-Q region of CBP interacts with hPygo2 in vivo, FLAG-tagged fragments of CBP were transiently transfected along with full-length hPygo2 into HEK-293 cells, hPygo2 complexes were immunoprecipitated and antibodies against FLAG were used to detect the presence of CBP protein in the complexes (Figure 2B). Pre-immune serum, as expected, did not precipitate complexes containing FLAG–CBP (Figure 2B, lane 1). hPygo2 antiserum, however, precipitated complexes containing FLAG-tagged full-length CBP (Figure 2B, lane 3) as well as a FLAG-tagged fragment of CBP consisting of a region spanning the bromodomain to the Q domain (Figure 2B, lane 5), but not a region spanning the RID to the bromodomain (FLAG-RID-Bro, lane 4). Consistent with the GST-pulldown analyses (Figure 1), these results indicated that hPygo2 could interact with full-length CBP as well as the fragment of CBP containing the CH3/HAT domains in vivo.
The human colorectal cancer cell line, SW480, was used to investigate whether endogenous hPygo2 and CBP proteins co-localize and associate. Initially, we tested whether a Myc-tagged hPygo2 could co-localize with endogenous CBP (Figure 3A). Cells were transiently transfected with either Myc alone (Figure 3A, upper panel) or with Myc-hPygo2 (Figure 3A, lower panel) and cells were stained with antibodies against Myc and CBP, as well as DAPI to visualize nuclei. Myc staining was observed throughout the entire cell, whereas Myc-hPygo2 staining was strictly nuclear and showed significant co-localization with endogenous CBP. Next, we stained SW480 cells for endogenous hPygo2 and CBP (Figure 3B). Consistent with Myc–hPygo2, endogenous hPygo2 showed significant overlap with CBP within the nuclei of SW480 cells. To more specifically demonstrate the endogenous interaction of hPygo2 and CBP in SW480 cells, both proteins were subjected to immunoprecipitation analysis (Figure 3C). When endogenous protein complexes containing hPygo2 were immunoprecipitated, CBP was detected as an interacting protein (left panel). Likewise, when endogenous CBP was immunoprecipitated, hPygo2 was detected as an interacting protein (right panel). These analyses indicate that both endogenous proteins exist in a complex in SW480 cells, consistent with the in vitro pull-downs and in vivo interactions in HEK-293 cells.
hPygo2 is required to augment DNA-tethered and TCF-dependent transcription by the recruitment of CBP
Although the above experiments confirmed that CBP interacts with the Pygopus NHD region, we wanted to test whether this interaction played a functional role in transcription. Previous studies indicated that, when fused to either the DNA binding domain of TCF or Gal4, the NHD of Pygopus displayed transcriptional activity from TOPFLASH or Gal4-upstream activating sequence reporters respectively [9,11]. Since CBP can recruit basal transcription factors, including RNA polymerase II , we hypothesized that co-expression of CBP might augment DNA-tethered NHD reporter transcription above its basal level. Gal4–NHD was therefore transfected with increasing amounts of full-length CBP into HEK-293 cells (Figure 4A). As a negative control, Gal4 alone was co-expressed with the reporter with increasing amounts of CBP, and little reporter activation was observed. Co-expression of the reporter with the Gal4–NHD fusion protein resulted in a three-fold activation over Gal4 alone, consistent with previous results . When Gal4–NHD was co-expressed with increasing amounts of CBP, a significant dose-dependent increase in reporter activation to a maximum of 17-fold was observed, indicating that CBP could augment the transcriptional activity of the NHD.
We also tested whether this increase in transcriptional activity was independent of the HAT activity of CBP (Figure 4B). When a plasmid was introduced encoding CBP harbouring a mutation in the HAT domain, thereby rendering it catalytically inactive, a 2-fold increase in transcriptional activity over hPygo2 alone was observed. This observation is consistent with a previous finding in which a HAT-defective p300 was shown to have a positive effect on a naked Wnt reporter construct . This suggests that the HAT domain is not required for the transcriptional activity observed by the recruitment of CBP by the NHD of hPygo2.
A conserved NPF (Asn-Pro-Phe) tri-peptide present within the NHD domain at amino acids 76 to 78 was suggested to be a motif required for the transcriptional activity of Pygopus proteins . We wanted to test, therefore, whether the ability of CBP to augment NHD-specific transcription was dependent on this NPF motif. Initially, using GST pulldown analysis, we tested the binding of CBP to a GST-tagged fragment encoding the hPygo2 NHD region with the NPF motif mutated to alanines (NHDNPF−AAA). As shown in Figure 4(C), CBP could readily bind to both the wild-type NHD and NHDNPF-AAA. This result suggested that the NPF motif was not required for CBP binding. Thus the interaction between CBP and Pygopus, similar to that recently shown between TAF4 (TATA box-associated factor)/TFIID and Pygopus , must occur at a site different from the NPF.
We next asked whether the increased transcriptional activity we observed by the recruitment of CBP to the hPygo2 NHD was dependent upon the NPF motif. Expression of Gal4-NHDNPF−AAA lacked transcriptional activity, as expected , and was no different than the values obtained with Gal4 alone (Figure 4D). Co-expression of CBP with wild-type Gal4–NHD, however, resulted in a significant increase in transcription, whereas co-expression of CBP with Gal4–NHDNPF−AAA showed no significant increase over Gal4 alone. These results suggested that the observed increase in transcription by the recruitment of CBP to the NHD was dependent on the NPF motif. Thus, the binding of CBP to Pygopus is not dependent on the NPF, but is dependent on the NPF for its ability to increase the transcriptional activity of the NHD.
A hallmark of canonical Wnt signal transduction is the recruitment of the β-catenin transcriptional activation complex to the TCF/LEF DNA-binding protein. Pygopus is considered to be a distal link in a chain of transcription factor adaptors which connect the β-catenin–TCF complex to the basal transcription machinery . We therefore determined the ability of CBP to facilitate the transcriptional activity of hPygo2 from a TCF/LEF-dependent promoter in the presence and absence of β-catenin. For these experiments, we used the transcriptional reporters which contained wild-type TCF/LEF (pOT) or mutated TCF/LEF (pOF) binding sites, that have minimal background activation .
Without β-catenin, neither hPygo2 nor CBP alone or in combination could stimulate TCF-dependent transcription above background levels, as determined by transfection of cells with empty vector (Figure 5A). In contrast, β-catenin alone significantly stimulated transcription over background levels as expected. Co-transfection of β-catenin with either CBP or hPygo2 resulted in an approx. 1.6- to 1.8-fold increase respectively over β-catenin alone. These data are consistent with previous observations that Pygopus increased TCF-dependent transcription in HEK-293 cells by, at best, 2-fold . When hPygo2 and CBP were co-transfected with β-catenin, however, reporter activation increased to 2.5-fold, suggesting that hPygo2 and CBP can act co-operatively to increase β-catenin mediated TCF-dependent transcription.
Removal of the HAT domain of CBP caused an increase in transcription over wild-type CBP (Figure 5B). Consistent with the activation observed with GAL–NHD (Figure 4B), CBP with a catalytically inactive HAT domain (CBP HAT mut) increased transcription from the pGL3 OT reporter above that obtained with wild-type CBP (Figure 5A). When β-catenin and hPygo2 were co-expressed with the CBP HAT mutant, an even higher increase in TCF-dependent transcription was observed. These results suggest that the HAT domain of CBP is not required for the increase in TCF-dependent transcription observed when β-catenin and hPygo2 are present.
Recruitment of CBP by hPygo2 is sufficient for histone acetylation in vitro
CBP is a large multifunctional protein, which in addition to having significant HAT activity, is able to directly recruit the basal transcriptional machinery. This property of CBP would explain why we observed an increase in reporter gene expression with hPygo2 and CBP (Figures 4 and 5). The HAT family of proteins promote transcription by acetylating specific residues on N-terminal histone tails, resulting in nucleosomal relaxation, thereby allowing access to transcription factors and the basal transcription machinery . We therefore tested the possibility that the interaction between Pygopus and CBP may also promote histone acetylation.
To test whether hPygo2 recruits CBP-dependent HAT activity, we co-transfected the HAT domain-containing region of CBP (Bro-Q) with and without full-length hPygo2 into HEK-293 cells. Protein complexes were then immunoprecipitated using either pre-immune or anti-hPygo2 antiserum, and their ability to acetylate H4 peptide in vitro (HAT activity) was then quantified. Minimal background HAT activity was observed with the pre-immune control (Figure 6A; Pre). In contrast, a 6-fold increase in HAT activity was observed by immunoprecipitation of endogenous hPygo2 (pCS2+), and a 10-fold increase in HAT activity was observed by immunoprecipitation of overexpressed hPygo2. These results suggest that Pygopus may function to recruit CBP-dependent HAT activity required for histone acetylation in vitro.
We next determined more precisely the ability of different regions of hPygo2 to recruit CBP to acetylate histones. GST pulldown assays were performed using GST, NHD (GST–NHD), the intervening region between the NHD and PHD (GST–Δ) and the PHD (GST–PHD). GST proteins were incubated with cell extracts transfected with either the catalytic Bro-Q domain or the non-catalytic RID-Bro domain lacking HAT activity and were then tested for their ability to acetylate Histone H4 (Figure 6B).
GST fusion proteins incubated with cell extracts containing the non-catalytic RID-Bro region of CBP displayed low levels of histone acetylation, indicating that Pygopus on its own does not contain endogenous HAT activity. GST fusion proteins incubated with cell extracts containing the HAT domain of CBP, on the other hand, resulted in a significant increase in HAT activity. GST–NHD showed an approx. 8.6-fold increase in HAT activity compared with background obtained with GST alone. GST-Δ showed no significant increase in activity above background levels. GST-PHD showed an approx. 2-fold increase over background levels, which may be partly explained by the fact that the PHD may recruit β-catenin via Bcl-9, which was shown to display CBP-dependent HAT activity . Figure 6(C) indicates that the Bro-Q region of CBP is recruited by the NHD of hPygo2. These results strongly suggest that the hPygo2 NHD domain is able to functionally recruit the HAT activity of CBP.
Pygopus was proposed to represent the last component in a chain of adaptors that connect the β-catenin/TCF activation complex to the basal transcription machinery [9,11]. CBP was demonstrated to interact with β-catenin [22,23], and has been proposed to act, in addition to its HAT activity, as a bridging protein to stabilize transcription complexes in conjunction with other transcriptional regulators. Our data now suggest that CBP makes contact with Pygopus in this complex. Our observed TCF/LEF-dependent reporter activation appeared to be dependent on β-catenin, which is consistent with the proposed role for Pygopus in Wnt signalling [6,7]. The presence of CBP at the promoter of Wnt target genes suggests that it is required, in conjunction with both Pygopus and β-catenin, for augmenting transcriptional activation.
Through the use of in vitro pulldown assays, we have identified two distinct regions of CBP that interact with hPygo2 (Figure 1B). It has been shown that, similar to hPygo2, β-catenin can also bind to two distinct sites of CBP [22,23]. It is therefore plausible that both hPygo2 and β-catenin could either bind co-operatively or sequentially to CBP, or that multiple protein interactions with CBP may be required for the activation of a Wnt target gene. It is also plausible that the interaction of hPygo2 at each distinct site on CBP may be required to carry out an individual function. Whether it is transcription or histone acetylation in the activation of a Wnt target gene remains to be determined.
One hypothesized role of Pygopus is to target and anchor β-catenin in the nucleus . This function might be fulfilled by the recruitment of β-catenin through a CBP complex, thereby directing β-catenin to the sites of active transcription through the interaction with methylated Histones . Another possible role of the Pygopus NHD domain could be to capture transactivating complexes that sequentially bind to β-catenin . Consistent with this hypothesis, there is evidence indicating that mediator proteins could possibly bind sequentially to the NHD domain of Pygopus and β-catenin [8,39]. Pygopus and β-catenin may therefore act co-ordinately to capture CBP as well as other transcription factors such as MED12/13 during target gene activation.
Other studies indicated that the NHD of Pygopus could bind directly to the general transcription factor TAF4/TFIID, and that this interaction is independent of the NPF motif, suggesting a second important motif or protein docking site . Our observed activation of a naked DNA reporter suggests a possible role of the ability of Pygopus to recruit the general transcription machinery indirectly, through contact with CBP. The recruitment of CBP appears to be independent of the conserved NPF motif, but requires the NPF for its transcriptional activity. This emphasizes the importance of the NPF in the transcriptional activity of the NHD. Whether the NPF is required to recruit an unknown factor or if it is structurally important for transcriptional activity needs to be investigated.
A previous model suggested that members of the MLL/SET1 histone methyltransferase family bind to the C-terminal region of the Armadillo repeat segment of β-catenin. Furthermore, chromatin immunoprecipitation assays revealed that CBP was present at Wnt gene promoters along with β-catenin and Pygopus . The PHD of Pygopus, like that of many other PHD-containing proteins, recognizes methylated histone tails which it could use as a docking site . These interactions place Pygopus in a unique position within the β-catenin transcription complex to integrate methylation on specific histones, such as H3K4Me3, with acetylation events carried out by chromatin remodelling proteins, such as CBP. Indeed, recent evidence has emerged for a tissue-specific requirement for Pygopus2 in Histone H3 acetylation . The HAT family member that was required for their observed acetylation, however, was not determined. It is therefore possible that CBP, identified to interact with the NHD of Pygopus in this study, is responsible for the tissue-specific histone acetylation by Pygopus2. We propose that a possible function of Pygopus in Wnt signalling is to recruit or capture CBP, which could be required for both transcriptional activation and histone acetylation (Figure 7).
Phillip Andrews designed and performed most of the experiments and wrote the first draft of the manuscript. Zhijian He performed the co-localization and interaction assays in SW480 cells. Cathy Popadiuk and Kenneth Kao supervised this research and participated in writing the manuscript.
This research was supported by grants from CIHR (Canadian Institutes of Health Research) [grant number MOP-14983] and CBCF (Canadian Breast Cancer Foundation) (Atlantic).
We thank Dr Gary Paterno, Dr Laura Gillespie and Tina Blackmore for technical assistance, as well as for CBP constructs; Dr Bert Vogelstein (Sidney Kimmel Cancer Centre, Johns Hopkins University School of Medicine, Baltimore, MD, U.S.A.) for the pOT/OF reporter constructs; and Mark Kennedy for helpful comments on the manuscript.
Abbreviations: APC, adenomatous polyposis coli; Bro, bromodomain; CBP, CREB (cAMP-response-element-binding protein)-binding protein; Cy5, indodicarbocyanine; DAPI, 4′,6-diamidino-2-phenylindole; Gro/TLE, groucho/transducin-like enhancer; GST, glutathione transferase; H3K4Me3, Histone H3 trimethyled on Lys4; HA, haemagglutinin; HAT, histone acetyltransferase; HEK-293 cell, human embryonic kidney 293 cell; hPygo2, human Pygopus2; LEF, lymphoid enhancing factor; lgs/Bcl9, legless/B-cell lymphoma 9; MLL/SET1, mixed-lineage-leukaemia/SET1-type; NHD, N-terminal homology domain; PHD, plant homeodomain; RID, receptor interaction domain; TAF, TATA box-associated factor; TCF, T-cell factor; Wg, wingless
- © The Authors Journal compilation © 2009 Biochemical Society