Association between Hepatocyte Nuclear Factor 6 (HNF-6) and FoxA2 DNA Binding Domains Stimulates FoxA2 Transcriptional Activity but Inhibits HNF-6 DNA Binding

Expression plasmids, reporter plasmids, and adenovirus vectors.Expression vectors consisted of the cytomegalovirus (CMV) promoter driving expression of the mouse HNF-6 cDNA or the rat FoxA1, FoxA2, and FoxA3 cDNAs were described previously (35, 40). The pSP72 HNF6 L350A plasmid for in vitro transcription-translation and the pECE HNF6 L350A mammalian expression plasmid in which the simian virus 40 promoter drives expression of the mutant HNF-6 L350A cDNA were generous gifts from Frédéric Lemaigre (Brussels, Belgium) and were described previously (26).

The 6X HNF-6 TATA-luciferase reporter construct consisted of six copies of the HNF-6 binding site derived from the FoxA2 promoter (−141 to −127 bp) driving expression of the TATA box luciferase gene and was described previously (40). The 6X FoxA TATA-luciferase utilized six copies of the FoxA2 binding site (TTTGTTTGTTTG) from the cdx-2 promoter region, and this sequence was selective for FoxA2 binding (44, 53).

The glutathione S-transferase (GST) fusion protein with the FoxA2 (amino acids 144 to 279) and FoxA1 (amino acids 134 to 290) winged-helix DNA binding domains were described previously (34). GST fusion proteins with the FoxA1 and FoxA2 N-terminal regions (amino acids 7 to 103 and 7 to 94) and FoxA2 C-terminal region (amino acid 361 to 458) were described previously (17, 35).

Adenovirus vectors that used the CMV promoter to drive expression of either the mouse HNF-6 (AdHNF6) or rat FoxA2 (AdFoxA2; AdHNF3β) cDNA were described previously (48-50). For all the green fluorescent protein (GFP)-HNF-6 fusion proteins, the HNF-6 fragments were PCR generated, digested with EcoRI and BamHI, and then ligated into the appropriate GFP expression vector, except for the HNF-6 N-terminal fragment which was digested with EcoRI and XbaI. The PCR-amplified HNF-6 fragments were verified by DNA sequencing.

The following sense and antisense primers were used for PCR amplification of the designated HNF-6 coding regions: HNF-6 FL (1 to 465), 5′-CGCGAATTCATGAACGCACAGCTGACC-3′ and 5′-CGCGGATCCTGCTTTGGTACAAGTGCT-3′; HNF-6 Cut-Homeo (289 to 465), 5′-CGCGAATTCATGGAAGAGATCAATACC-3′ and 5′-CGCGGATCCTGCTTTGGTACAAGTGCT-3′; HNF-6 N-Term (1 to 288), 5′-CGCGAATTCATGAACGCACAGCTGACC-3′ and 5′-CGCTCTAGACTGCCCTGAATTACTTCC-3′; GFP HNF-6 ΔHomeo (1 to 384), 5′-CGCGAATTCATGAACGCACAGCTGACC-3′ and 5′-CGCGGATCCGGTGTTGCCTCTGTCCTT-3′; HNF-6 Cut (289 to 384), 5′-CGCGAATTCATGGAAGAGATCAATACC-3′ and 5′-CGCGGATCCGGTGTTGCCTCTGTCCTT-3′; and HNF-6 Homeo (385 to 465), 5′-CGCGAATTCCCCAAAAAGCCCAGGCTG-3′ and 5′-CGCGGATCCTGCTTTGGTACAAGTGCT-3′.

Cell culture and transient transfection.HepG2 cells were maintained in F12 medium supplemented with 10% fetal bovine serum, essential amino acids, penicillin-streptomycin, and insulin as previously described (44). For transient transfection, HepG2 cells were plated in six-well plates and transfected with Fugene 6 reagent (Roche) according to the manufacturer's protocol. Cells were transfected with 250 ng of CMV-HNF6 or deletion mutants, CMV-FoxA1, or -FoxA2, or FoxA3 with 1.6 μg of a 6X FoxA TATA-luciferase reporter with 30 ng of CMV-Renilla internal control.

We also performed these cotransfection experiments with either the −134 FoxA2 promoter luciferase construct (36) or the −188 Glut2 promoter luciferase construct (48). Total transfected DNA was kept constant with empty CMV vector. Twenty-four hours posttransfection, cells were prepared for dual luciferase assays (Promega). Luciferase activity was determined as fold induction over cells transfected with empty CMV vector, normalized to Renilla activity. Experiments were performed at least two to six times in triplicate unless otherwise stated. Statistical analysis was performed with Microsoft Excel tools.

Immunofluorescence.HepG2 cells were plated on untreated glass coverslips and transfected with 500 ng of full-length GFP HNF-6 or mutant constructs with the Fugene 6 (Roche) reagent according to the manufacturer's protocol. Twenty-four hours posttransfection, the coverslips were washed three times in phosphate-buffered saline (PBS), and the cells were fixed in 3% paraformaldehyde for 15 min at room temperature. After several rinses in PBS, the coverslips were mounted on glass slides with mounting medium (15% [wt/vol] Vinol 205 polyvinyl alcohol, 33% [vol/vol] glycerol, 0.1% azide in PBS, pH 8.5). Cellular fluorescence was visualized with a Zeiss Axioplan microscope.

In vitro pulldown assay.Production of GST fusion proteins was described previously (34). For the binding reaction, 15 μl of ³⁵S-labeled in vitro transcribed and translated protein was incubated with 20 μl of glutathione-Sepharose beads (preadsorbed with 2.5 μg of various GST-FoxA fusion proteins) in 500 μl of binding buffer for 2 h at 4°C with constant rotation. After 2 h of binding, the beads were washed three times with 500 μl of wash buffer. The pelleted beads were then resuspended with sodium dodecyl sulfate (SDS) sample buffer plus 1% β-mercaptoethanol and boiled for 5 min, after which the supernatant was loaded on sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and processed for autoradiography. The binding buffer consisted of 85 mM KCl, 20 mM HEPES, pH 7.9, 1.5 mM MgCl₂, 0.2 mM EDTA, 0.5 mM dithiothreitol, 20% glycerol, 0.1% NP-40 detergent, 0.2 mM phenylmethylsulfonyl fluoride, and protease inhibitors (Roche). The wash buffer was identical to the binding buffer except that it contained 150 mM KCl.

Western and coimmunoprecipitation with transfected HepG2 nuclear extracts.HepG2 cells were cotransfected with 5 μg of full-length GFP-HNF6 (1 to 465) or GFP-ΔHomeo HNF-6 (1 to 384) with pcDNA3.1 FoxA1 or FoxA2 tagged with a C-terminal V5 epitope or CMV-CBP. Twenty-four hours posttransfection, cells were washed twice with cold PBS, collected, and prepared for nuclear extracts. Nuclear extracts were prepared as described with slight modifications (36, 45). Briefly, cells were resuspended with hypotonic buffer (10 mM HEPES, pH 7.9, 1.5 mM MgCl₂, 10 mM KCl, 0.5 mM dithiothreitol, and 0.2 mM phenylmethylsulfonyl fluoride supplemented with protease complete inhibitor tablets), incubated on ice for 20 min, and then ruptured by 30 passes in a Dounce homogenizer.

Nuclear pellets were washed with hypotonic buffer, and nuclear proteins were extracted by addition of high-salt buffer (hypotonic buffer containing 20% glycerol, 0.5 M KCl, 0.2 mM EDTA, and 0.1% NP-40) and incubated for 40 min at 4°C with occasional agitation. After centrifugation, the Bradford reagent (Bio-Rad) was used to determine protein concentrations of the nuclear extracts. For immunoprecipitation, 850 μg of nuclear extracts was diluted to a final salt concentration of 85 mM KCl and allowed to equilibrate for 45 min. The extracts were first cleared with protein A-Sepharose and then incubated with 1 μl of either V5 (Invitrogen) or CBP (C-1; Santa Cruz Biotech) monoclonal antibodies for 16 h at 4°C. The extracts were then immunoprecipitated by the addition of protein A-Sepharose. Beads were washed extensively with wash buffer (hypotonic buffer containing 150 mM KCl). The immunoprecipitates were eluted by the addition of SDS sample buffer and boiling for 5 min at 100°C. Eluted proteins were resolved by SDS-PAGE and transferred to nitrocellulose membranes for Western blot analysis. For Western blot analysis, a monoclonal GFP antibody (Clontech) was incubated at a dilution of 1:1,000 overnight at 4°C.

Adenovirus infection of HepG2 cells and electrophoretic mobility shift assays.HepG2 cells were infected with a constant amount of adenovirus expressing the mouse HNF-6 cDNA (AdHNF-6; 10 PFU per cell) with increasing amounts of AdFoxA2 (AdHNF-3β; 2, 10, or 50 PFU per cell) as described previously (48-50). HepG2 cells were also infected with a constant amount of AdFoxA2 (10 PFU per cell) with increasing amounts of AdHNF-6 (0.2-, 1-, or 5-fold excess). The AdLacZ adenovirus (CMV-LacZ) was used to keep the total amount of infecting adenovirus at 60 PFU per cell and used to eliminate secondary competition between CMV promoters. Nuclear extracts were prepared 24 h following infection and used for electrophoretic mobility shift assays with the HNF-6 DNA binding site from the Glut-2 promoter (48) or the FoxA DNA binding site from the cdx2 promoter as described previously (37, 53). We also included control lanes containing a 100-fold excess of unlabeled DNA competitor and HepG2 nuclear extracts infected with either AdHNF-6 or AdFoxA2. Western blot analysis of AdFoxA2 and AdHNF-6 HepG2 nuclear extracts with HNF-6 antibody was performed as described previously (42, 52).

HNF-6 and FoxA2 synergistically stimulated transcription of a FoxA-responsive promoter.In order to examine whether HNF-6 is capable of stimulating FoxA-dependent transcription, transient cotransfection assays were performed in human hepatoma HepG2 cells with a FoxA-dependent luciferase reporter plasmid (6X FoxA-TATA luciferase plasmid). This consisted of a TATA box sequence (35) adjacent to six copies of the FoxA (HNF-3) binding site (TTTGTTTGTTTT) derived from the mouse cdx-2 promoter (53) controlling expression of the luciferase reporter gene. This FoxA DNA recognition sequence was previously shown not to bind HNF-6, as it failed to compete for HNF-6 protein-DNA complex formation in electrophoretic mobility shift assays with liver nuclear extracts (44).

Consistent with this finding, in cotransfection assays with the 6X FoxA-TATA luciferase plasmid, the CMV-FoxA2 cDNA expression vector elicited transcriptional activation of the reporter gene, whereas cotransfection of the CMV-HNF-6 plasmid did not activate this promoter (Fig. 1B). Interestingly, combining the FoxA2 and HNF-6 expression vectors in HepG2 transient transcription assays caused a statistically significant increase in expression of the 6X FoxA-TATA luciferase gene (Fig. 1B, P = 0.0003). In contrast, this HNF-6 transcriptional synergy was not found with either the related FoxA1 (HNF-3α) or FoxA3 (HNF-3γ) expression vector (Fig. 1A). Interestingly, the HNF-6 Cut-Homeodomain (Fig. 1A, 289 to 465) sequences alone were sufficient to stimulate FoxA2 transcriptional activity to levels similar to full-length HNF-6 protein (Fig. 1B).

• Open in new tab
• Download powerpoint

FIG. 1.

HNF-6 Cut-Homeodomain sequences were sufficient for transcriptional synergy with FoxA2 in cotransfection assays. (A) Schematic representation of HNF-6 full-length, HNF-6 Cut-Homeodomain (Cut-Homeo 289 to 465), or HNF-6 containing a deletion of the homeodomain sequences (ΔHomeo 1 to 384). The numbers represent the HNF-6 amino acid position contained within each of the expression constructs. (B) The HNF-6 Cut-Homeodomain sequences were sufficient to stimulate FoxA2 transcriptional activity. HepG2 cells were transiently transfected with CMV FoxA2 expression vector and various CMV HNF-6 expression constructs (see A for coordinates) with the 6X FoxA TATA luciferase reporter. Twenty-four hours posttransfection, cells were collected and processed for dual luciferase assays. Transcriptional synergy with FoxA2 occurred with full-length HNF-6 (FL 1 to 465), and HNF-6 Cut-Homeodomain, but not with the HNF-6 containing a deletion of the Homeodomain, which could function as a dominant negative inhibitor of synergistic FoxA2 activation. Note that HNF-6 was unable to stimulate FoxA1 and FoxA3 transcriptional activity. (C) The HNF-6 Cut-Homeodomain sequences are transcriptionally inactive in cotransfection assays with an HNF-6 dependent reporter gene. HepG2 cells were transiently transfected with CMV HNF-6 expression constructs (A) and 6X HNF-6 TATA luciferase reporter and then analyzed for luciferase activity 24 h later. Transcriptional activity of the HNF-6 ΔHomeo protein was diminished by 73%, whereas the HNF-6 Cut-Homeo protein was transcriptionally inactive. (D) HNF-6 synergistically stimulated FoxA2 transcriptional activation of the −134 FoxA2 promoter luciferase construct. HepG2 cells were transiently transfected with the −134 FoxA2 promoter luciferase construct (36) and either CMV-FoxA2, CMV-HNF-6, or CMV-HNF-6 Cut-homeodomain expression vectors as indicated and then analyzed for luciferase activity 24 h later. Luciferase levels are represented as the fold increases over the expression of the reporter with empty CMV plasmid and are normalized to the CMV-Renilla internal control. The graphs represent the mean fold induction (±standard deviation) calculated from two independent transfections performed in triplicate.

Deletion of the HNF-6 Homeodomain motif (Fig. 1A, HNF-6 ΔHomeo 1 to 384) abrogated stimulation of FoxA2 transcription, and the HNF-6 ΔHomeo protein functioned as a dominant negative inhibitor of the HNF-6 transcriptional synergy with FoxA2 in cotransfection assays (Fig. 1B). Moreover, when an HNF-6-dependent reporter (6X HNF-6 TATA luciferase) was used, the HNF-6 ΔHomeo mutant exhibited a 75% reduction in transcriptional activation of the reporter gene and the Cut-Homeodomain sequences were transcriptionally inactive compared to full-length HNF-6 (Fig. 1C). These results suggested that the HNF-6 Homeodomain sequences were essential for potentiating transcriptional activity of both a FoxA-dependent and an HNF-6-dependent reporter gene. Furthermore, although the HNF-6 Cut-Homeodomain sequences mediated recruitment of the p300/CBP coactivator proteins (26, 40), our data suggested that they were unable to do so when bound to DNA. These data suggested that HNF-6 Cut-Homeodomain mediated stimulation of FoxA2-dependent transcription through protein-protein association.

To further assess the effects of FoxA2 and HNF-6 in the context of an intact promoter, we performed transcription assays with the −134 bp FoxA2 promoter luciferase reporter gene, which possesses a FoxA binding site but lacks a high-affinity HNF-6 binding site (36, 40, 44). Transfection with CMV HNF-6 mildly stimulated expression of the FoxA2 promoter (Fig. 1D), presumably through a potential weak-affinity HNF-6 DNA binding site found at −64 to −52 bp (ACctTgGATTTAA). In spite of the fact that the CMV-FoxA2 expression vector alone was unable to elicit strong transcriptional activation of the FoxA2 promoter, coexpression of both FoxA2 and HNF-6 proteins provided synergistic transcriptional activation of the −134 bp FoxA2 promoter region (Fig. 1D). Consistent with HNF-6 mediating synergistic transcriptional activity with FoxA2, the HNF-6 Cut-Homeodomain protein alone was unable to stimulate expression of the −134 bp FoxA2 promoter, but when the HNF-6 Cut-Homeodomain and FoxA2 expression vectors were combined, they provided transcriptional synergy of the FoxA2 promoter (Fig. 1D). These data suggested that HNF-6 mediated stimulation of FoxA2 transcriptional activation through FoxA specific DNA sites within an intact promoter region.

To determine whether the nuclear localization sequences were contained in the Cut-Homeodomain, we used mammalian expression vectors in which the GFP was fused to the N terminus of various HNF-6 protein regions (Fig. 2A). Transfection of HepG2 cells with either GFP HNF-6 full-length (HNF6 FL) or GFP-HNF-6 Cut-Homeodomain (HNF6 Cut-Homeo) fusion protein showed equivalent nuclear fluorescence (Fig. 2B to C). The GFP HNF-6 N-terminal fusion protein (HNF6 N-Term) displayed cytoplasmic fluorescence in HepG2 cell transfections (Fig. 2D), suggesting that the nuclear localization sequences were contained within the HNF-6 Cut-Homeodomain motif. Furthermore, predominant nuclear fluorescence was found with GFP fusion proteins that contained either the HNF-6 Cut-domain (Fig. 2E and F; HNF6 ΔHomeo and HNF6 Cut) or the HNF-6 Homeodomain (Fig. 2G, HNF6 Homeo) in HepG2 cell transfections. These data suggested that the inability of the HNF6 ΔHomeo expression construct to synergistically potentiate FoxA2 transcriptional activity in cotransfection assays was not due to a problem with nuclear localization.

• Open in new tab
• Download powerpoint

FIG. 2.

HNF6 DNA binding domain possessed a dual nuclear localization signal contained within the Cut-Homeodomain sequences. (A) Schematic representation of GFP-FL HNF-6 or various other GFP-HNF-6 fusion proteins used to determine the regions responsible for nuclear localization in HepG2 transfection assays. The numbers represent the HNF-6 amino acid position contained within each of the GFP fusion proteins. (B to G) The HNF-6 Cut and Homeodomain sequences contained independently functioning nuclear localization sequences. HepG2 cells were plated on glass coverslips and transiently transfected with 500 ng of various GFP-HNF-6 fusion constructs. At 24 h posttransfection, the cells were fixed and prepared for fluorescence as described in Materials and Methods.

HNF-6 associates with FoxA2: Cut-Homeodomain of HNF-6 and winged-helix DNA binding domain of FoxA2 are required for the interaction.Our observations that HNF-6 and FoxA2 synergistically activated a FoxA-dependent promoter led us to examine whether these two proteins would physically interact. To evaluate the association between HNF-6 and FoxA2 proteins, we performed GST pulldown experiments with ³⁵S-labeled full-length HNF-6 protein or deletion mutants (Fig. 3A and Fig. 2A for HNF-6 coordinates). We generated GST fusion proteins with the FoxA transcriptional activation domains (Fig. 3A), including the N-terminal region of FoxA2 (7 to 94) or FoxA1 (7 to 103) and the C-terminal region of FoxA2 (361 to 458) as well as with the winged-helix DNA binding domains of FoxA2 (144 to 279) or FoxA1 (134 to 290). The GST-FoxA1 and GST-FoxA2 fusion proteins were expressed in bacteria, immobilized on glutathione-Sepharose and then binding assays were performed with various labeled HNF-6 proteins (Fig. 3B) to determine the HNF-6 and FoxA sequences required for protein binding (see Materials and Methods).

• Open in new tab
• Download powerpoint

FIG. 3.

FoxA2 winged-helix DNA binding motif interacts with the HNF-6 Cut-Homeodomain sequences in vitro. (A) Schematic representation of the GST fusion constructs. Schematic representation of the FoxA2 protein containing N-terminal and C-terminal transactivation domains and a 100-amino-acid winged-helix DNA binding domain (4, 35, 39). The GST fusion constructs span the N-terminal activation domains (7 to 94 FoxA2 and 7 to 103 FoxA1), the winged-helix DNA binding domain (144 to 279 FoxA2 and 134 to 290 FoxA1), and the C-terminal activation domain (361 to 458 FoxA2). (B) Schematic representation of the HNF-6 proteins that were radioactively labeled and used for the GST pulldown assays. Note that the HNF-6 L350A full-length (FL) contains the indicated mutation of the LXXLL motif that is required for HNF-6 recruitment of the p300/CBP protein (26). (C) The FoxA2 winged-helix and HNF-6 Cut-Homeodomain sequences are sufficient for protein interaction. GST-FoxA fusion proteins (2.5 μg preadsorbed to 20 μl of glutathione-Sepharose) were incubated with in vitro transcribed and translated ³⁵S-labeled full-length HNF-6 or deletion mutants (Fig. 5B), then washed, and bound proteins were eluted by boiling in SDS sample buffer as detailed in Materials and Methods. Labeled bound HNF-6 proteins were resolved by SDS-PAGE and processed for autoradiography. Note that the full-length (FL) HNF-6 bound weakly and HNF-6 Cut-Homeodomain and HNF-6 L350A bound poorly to the FoxA1 winged-helix DNA binding domain (134 to 290). The HNF-6 N terminus lacking the Cut-Homeodomain sequence (1 to 288 N-term) did not associate with any of the FoxA1 or FoxA2 GST fusion proteins.

These studies demonstrated that both HNF-6 full-length and Cut-Homeodomain sequences interacted strongly with the GST-FoxA2 (144 to 279) winged-helix domain fusion protein (Fig. 3C), but bound less effectively with the GST-FoxA1 (134 to 290) winged-helix fusion protein (Fig. 3C). This finding is consistent with the inability of HNF-6 to produce transcriptional synergy with FoxA1 (Fig. 1A and see below). Furthermore, neither of these labeled HNF-6 proteins bound to the GST fusion proteins containing the FoxA1 or FoxA2 N-terminal sequences, the FoxA2 C-terminal sequence or the GST control (Fig. 3C). The labeled N-terminal HNF-6 protein (Fig. 3B, 1 to 288) was unable to bind to any of the GST-FoxA fusion proteins (Fig. 3C), suggesting that association with the FoxA2 winged-helix DNA binding domain required the HNF-6 Cut-Homeodomain sequences. Finally, the FoxA2 DNA binding domain displayed specific interaction with the HNF-6 L350A full-length protein (Fig. 3C), which contained a mutation of the HNF-6 Cut domain LXXLL sequence (Fig. 3B) that is essential for recruiting the p300/CBP protein (26). These results suggested that the LXXLL motif was not required for interaction between FoxA2 and HNF-6 proteins.

We next prepared nuclear extracts from HepG2 cells cotransfected with expression vectors containing the GFP protein fused to either HNF-6 FL (1 to 465) or HNF-6 ΔHomeo (1 to 384) (see Fig. 1A) and a V5 epitope-tagged FoxA2 or FoxA1 cDNA. These nuclear extracts were subjected to immunoprecipitation with a V5 monoclonal antibody followed by Western blot analysis with a monoclonal GFP antibody (Fig. 4A). All of the transfected GFP-HNF-6 fusion proteins were expressed as determined by Western blot analysis of the transfected HepG2 nuclear extracts with the GFP monoclonal antibody (Fig. 4A, left panel). Both GFP-HNF-6 full-length and GFP HNF-6 ΔHomeo proteins were able to coimmunoprecipate with the FoxA2 protein (Fig. 4A, right panel), suggesting that the HNF-6 Homeodomain sequence was dispensable for FoxA2 protein association. Interestingly, we were unable to coimmunoprecipitate FoxA1 and GFP full-length HNF-6 from transfected HepG2 nuclear extracts (Fig. 4A, right panel), a finding consistent with the inability of HNF-6 to enhance FoxA1 transcriptional activity in cotransfection assays (Fig. 2A). These data demonstrate that the HNF-6 protein associates with FoxA2 but not FoxA1 in nuclear extracts prepared from transfected HepG2 cells.

• Open in new tab
• Download powerpoint

FIG. 4.

HNF-6 interacts with FoxA2 and CBP in vivo. (A) HNF-6 interacts with FoxA2 but not with FoxA1 in vivo. HepG2 cells were cotransfected with GFP-HNF6 FL (1 to 465) or GFP ΔHomeo (1 to 384) with a C-terminal V5 epitope-tagged FoxA2 or FoxA1 or empty vector controls. Nuclear extracts were prepared and coimmunoprecipitated with V5 antibody, followed by Western blot analysis with GFP monoclonal antibody as detailed in Materials and Methods. To examine expression levels of each of the GFP-HNF-6 proteins, 1/10th of the input nuclear extract (85 μg) used for coimmunoprecipitation was analyzed by Western blot analysis. Both GFP-HNF-6 full-length (FL) and GFP HNF-6 ΔHomeo fusion proteins were coimmunoprecipitated with FoxA2 protein with the antibody to the V5 epitope tag, whereas GFP FL-HNF-6 protein was not coimmunoprecipitated with FoxA1 under similar conditions. Note that the CMV V5 epitope-tagged FoxA1 expression vector produced functional FoxA1 protein, as evidenced by transcriptional activation of the FoxA reporter gene in cotransfection assays (data not shown). (B) HNF-6 interacts with CBP in vivo. HepG2 cells were cotransfected with GFP-HNF6 FL (1 to 465) or GFP ΔHomeo (1 to 384) with the CBP expression vector. Nuclear extracts were prepared and coimmunoprecipitated with CBP antibody followed by Western blot analysis with GFP monoclonal antibody as detailed in Materials and Methods. GFP-HNF-6 full-length (FL) fusion protein was coimmunoprecipitated with CBP protein, but the GFP HNF-6 ΔHomeo GFP protein was less effectively coimmunoprecipitated with the CBP antibody.

To confirm that HNF-6 interacts with CBP in vivo, we prepared nuclear extracts from HepG2 cells cotransfected with expression vectors containing the GFP protein fused to either HNF-6 FL (1 to 465) or HNF-6 ΔHomeo (1 to 384) and the CMV-CBP expression vector. These nuclear extracts were subjected to immunoprecipitation with CBP antibody followed by Western blot analysis with a monoclonal GFP antibody (Fig. 4B). Only the GFP-HNF-6 full-length protein was able to coimmunoprecipate with the CBP protein (Fig. 4B), suggesting that the HNF-6 Homeodomain was required for interaction with the CBP protein, a finding consistent with the inability of HNF-6 ΔHomeo protein to synergistically stimulate FoxA2 transcription.

Recruitment of the p300/CBP proteins is required for HNF-6 transcriptional synergy with FoxA2.Because we found that the HNF-6 protein could immunoprecipitate with the CBP protein from HepG2 nuclear extracts, we next examined whether HNF-6 mediates FoxA2 transcriptional activation by recruiting the p300/CBP coactivator proteins. To test this hypothesis, we cotransfected the 6X FoxA TATA luciferase reporter plasmid with the FoxA2 and mutant HNF-6 L350A expression constructs, the latter of which can interact with FoxA2 (Fig. 3C) but can no longer recruit the p300/CEBP protein (26). These experiments demonstrated that the mutant HNF-6 L350A was unable to stimulate FoxA2 transcription and functioned as a dominant negative protein to inhibit FoxA2 and HNF-6 transcriptional synergy (Fig. 5A).

• Open in new tab
• Download powerpoint

FIG. 5.

Inhibition of histone acetyltransferase activity abrogated HNF-6 synergistic activation of FoxA2 transcription. (A) HNF-6 recruitment of the p300/CBP is required for transcriptional synergy with FoxA2. HepG2 cells were transiently transfected with the indicated combinations of HNF-6 FL FoxA2, or HNF-6 L350A and 6X FoxA TATA luciferase reporter constructs. Similar transfection experiments were also performed with either CMV E1A (S12) or CMV CBP HAT (−) (histone acetyltransferase deficient) expression constructs that inhibit p300/CBP activity. Transfections were performed and normalized as described in the Fig. 1 legend. The graphs represent the mean induction (±standard deviation) calculated from two independent transfections performed in triplicate. (B) Inhibition of p300/CBP histone acetyltransferase activity does not influence HNF-6 transcriptional activation of an HNF-6 dependent reporter gene. HepG2 cells were transiently transfected with full-length HNF-6 and the 6X HNF6 TATA luciferase reporter construct in combination with or without either the CMV E1A (S12) or CMV CBP HAT (−) expression construct.

These results suggest that HNF-6 mediates transcriptional synergy with FoxA2 by serving to recruit the p300/CBP coactivator proteins. To further confirm this finding, we performed cotransfection experiments with expression plasmids containing either a dominant negative CBP protein deficient in histone acetyltransferase activity [CBP HAT (−)] (46) or the adenoviral E1A protein, the latter of which binds to the coactivator CBP and inhibits histone acetyltransferase activity (2). Cotransfection of FoxA2 and either CMV E1A or CMV CBP HAT (−) did not influence the ability of FoxA2 to activate 6X FoxA-TATA luciferase expression (Fig. 5A). In contrast, HNF-6 and FoxA2 cotransfection assays in combination with either E1A or CBP HAT (−) expression vectors inhibited HNF-6 mediated FoxA2 activation and produced reporter expression levels identical to those found with FoxA2 transfected alone (Fig. 5A).

These data confirm that the HNF-6 protein functions to recruit the p300/CBP coactivators, which mediate synergistic activation of a FoxA-dependent promoter expression. Interestingly, neither E1A nor CBP HAT (−) expression vectors inhibited HNF-6 mediated activation of the 6X FoxA-TATA luciferase reporter gene (Fig. 5B). These results suggest that when HNF-6 protein binds to its target DNA sequence, the HNF-6 Cut-Homeodomain sequences are unable to recruit the p300/CBP protein and HNF-6 therefore mediates transcriptional activation through its N-terminal domain (26).

FoxA2 inhibits HNF-6 stimulation of the Glut-2 promoter by inhibiting HNF-6 DNA binding.Previous cotransfection and DNA binding studies demonstrated that the Glut-2 promoter is transcriptionally activated by HNF-6 but not by the FoxA (HNF-3) proteins (48). Consistent with these studies, HNF-6 cotransfection stimulated expression of the −188 Glut-2 promoter, whereas FoxA2 did not (Fig. 6). Interestingly, cotransfection of equal amounts of FoxA2 and HNF-6 expression vectors caused a significant reduction in HNF-6 activation of the Glut-2 promoter (Fig. 6).

• Open in new tab
• Download powerpoint

FIG. 6.

FoxA2 diminished HNF-6-mediated stimulation of the Glut-2 promoter in cotransfection assays. HepG2 cells were transiently transfected with CMV FoxA2 or CMV HNF-6 expression vectors either alone or together with −188 Glut2 promoter luciferase construct. Twenty-four hours posttransfection, cells were collected and processed for dual luciferase assays. As previously described (48), HNF-6 stimulated expression of the −188 Glut-2 promoter, whereas FoxA2 did not. Cotransfection of HNF-6 and FoxA2 expression vectors demonstrated that FoxA2 diminished HNF-6 stimulation of the −188 Glut-2 promoter. luciferase levels are represented as the increases over the expression of the reporter with empty CMV plasmid and are normalized to the CMV-Renilla internal control. The graphs represent the mean induction (±standard deviation) calculated from two independent transfections performed in triplicate.

Because interaction between the FoxA2 and HNF-6 proteins was mediated by sequences within their DNA binding domains, we examined whether FoxA2 inhibited HNF-6 DNA binding. HepG2 cells were infected with constant amounts of adenovirus expressing HNF-6 (AdHNF-6) and increasing amounts of AdFoxA2 (from 0.2- to 5-fold increase). Nuclear extracts were prepared 24 h postinfection and then used for a electrophoretic mobility shift assay with the HNF-6 site from the Glut2 promoter. The electrophoretic mobility shift assay showed that increasing FoxA2 protein levels caused diminished HNF-6 protein complex formation (Fig. 7A), but elevated FoxA2 infection did not influence expression or stability of the HNF-6 protein as evidenced by Western blot analysis (Fig. 7B). This result suggested that the interaction between FoxA2 and HNF-6 proteins inhibited HNF-6 DNA binding activity.

• Open in new tab
• Download powerpoint

FIG. 7.

Increasing amounts of FoxA2 diminished HNF-6 DNA binding activity but not HNF-6 expression levels. Increased FoxA2 levels inhibit HNF-6 DNA binding activity (A) but increased HNF-6 levels do not prevent FoxA2 DNA binding activity (C). HepG2 cells were infected with a constant amount of AdHNF6 (50) and infected with increasing amounts of AdFoxA2 (0.2-, 1-, or 5-fold excess; panel A) as described previously (48). HepG2 cells were infected with a constant amount of AdFoxA2 and infected with increasing amounts of AdHNF-6 (0.2-, 1-, or 5-fold excess; panel C). Nuclear extracts were prepared 24 h following infection and used for electrophoretic mobility shift assays with HNF-6 DNA binding site from the Glut-2 promoter (48) or the FoxA DNA binding site from the Cdx-2 promoter as described previously (37, 53). We also included control lanes containing a 100-fold excess of unlabeled DNA competitor (C) and HepG2 nuclear extracts either infected with AdHNF-6 or AdFoxA2 (−). (B) Western blot analysis of AdFoxA2 and AdHNF-6 HepG2 nuclear extracts with HNF-6 antibody (42) demonstrated that AdFoxA2 infection does not reduce HNF-6 protein levels. (D) Including a 100-fold molar excess of unlabeled FoxA (Cdx2 promoter) binding sequence (indicated by +) did not reduce FoxA2's ability to inhibit HNF-6 binding activity in electrophoretic mobility shift assay. We used the same AdHNF-6- and AdFoxA2-infected nuclear extracts for an electrophoretic mobility shift assay as described for panel A.

In order to examine whether HNF-6 would, in turn, inhibit FoxA2 DNA binding, we prepared nuclear extracts from HepG2 cells infected with constant amounts of AdFoxA2 while increasing the amount of AdHNF-6 used in the infection. These electrophoretic mobility shift assay studies demonstrated that increasing HNF-6 protein levels did not influence FoxA2 protein-DNA complex formation (Fig. 7C), a finding consistent with the ability of HNF-6 to enhance FoxA2 transcription through a FoxA dependent promoter. Furthermore, addition of a 100-fold excess of unlabeled FoxA DNA binding oligonucleotide did not influence the ability of FoxA2 to inhibit HNF-6 DNA binding activity (Fig. 7D), suggesting that HNF-6 can interact with FoxA2 protein when FoxA2 is bound to DNA.

Taken together, our studies suggested that association between FoxA2 and HNF-6 DNA binding domains disrupted HNF-6 DNA binding, thereby inhibiting transcriptional activation of its target genes. Conversely, FoxA2 protein association with the HNF-6 protein synergistically activated FoxA2 transcription through HNF-6 mediated recruitment of the p300/CBP coactivator proteins.