Nuclear Factor 1 Family Members Interact with Hepatocyte Nuclear Factor 1α to Synergistically Activate L-type Pyruvate Kinase Gene Transcription*

Transcription of hepatic L-type pyruvate kinase (L-PK) gene is cell type-specific and is under the control of various nutritional conditions. The L-PK gene contains multiple cis-regulatory elements located within a 170-bp upstream region necessary for these regulations. These elements can synergistically stimulate L-PK gene transcription, although their mechanisms are largely unknown. Because nuclear factor (NF) 1 family members bind to specific cis-regulatory elements known as L-IIA and L-IIB and hepatocyte nuclear factor (HNF) 1α binds to the adjacent element L-I, we examined the functional and physical interactions between these two transcription factors. Reporter gene assay showed that these two factors synergistically activated the L-PK promoter containing the 5′-flanking region up to –189. Although two NF1-binding sites are required for the maximum synergistic effect of NF1 family members with HNF1α, significant functional interaction between the two factors was observed in the L-PK promoter containing two mutated NF1-binding sites and also in the promoter containing only the HNF1α-binding site, raising the possibility that NF1 proteins function as HNF1α co-activators. Chromatin immunoprecipitation assay revealed that both NF1 proteins and HNF1α bound to the promoter region of the L-PK gene in vivo. In vitro binding assay confirmed that NF1 proteins directly interacted mainly with the homeodomain of HNF1α via their DNA-binding domains. This interaction enhanced HNF1α binding to the L-I element and was also observed in rat liver by co-immunoprecipitation assay. Thus, we conclude that cooperative interaction between NF1 family members and HNF1α plays an important role in hepatic L-PK transcription.

Pyruvate kinase (ATP, pyruvate 2-O-phosphotransferase; PK) 2 is an important regulatory enzyme in the glycolytic pathway. Four PK isozymes are known to exist in mammals and are designated as L, R, M 1 , and M 2 types (1). They are produced from two genes, PKL and PKM, by alternative use of two promoters and mutually exclusive alternative splicing, respectively (2)(3)(4). The L-type isozyme (L-PK) is expressed in a tissue-specific manner; this form is expressed primarily in hepatic parenchymal cells and is also present in pancreatic ␤ cells, kidney, and small intestine (1). Hepatic L-PK gene expression is also transcriptionally controlled positively by glucose in the presence of insulin and negatively by polyunsaturated fatty acids or glucagon via cyclic AMP (5)(6)(7)(8). These regulations of the L-PK gene transcription are mediated via the multiple cis-regulatory elements located within a 170-bp upstream region from the transcription start site, as shown in Fig. 1 (8 -10). We have named these elements L-I (Ϫ94 to Ϫ76), L-II (Ϫ149 to Ϫ126), and L-III (Ϫ170 to Ϫ150) (9). In addition, another group has identified a weak negative element (Ϫ116 to Ϫ99) between L-I and L-II (10). Although this element has been named L2, we will refer to this as L-IIA and L-II as L-IIB hereafter. Although L-IIB is considered to be a regulatory element responsible for polyunsaturated fatty acid suppression of L-PK gene transcription (8), L-III itself possesses carbohydrate responsiveness and is thus termed a carbohydrate response element (11,12). In the natural promoter context, however, L-III alone does not confer carbohydrate responsiveness to the promoter, and synergistic interaction between L-IIB and L-III is crucial for carbohydrate responsiveness (13,14). Therefore, L-IIB also functions as an accessory element for L-III. We have observed functional interactions between L-I and L-IIB, L-IIB and L-III, but not L-I and L-III (9). The inclusion of all three elements oriented in the same direction showed the maximum synergistic effect, indicating that these elements function as a unit. This unit possesses cell type specificity as well as carbohydrate responsiveness in vivo (15,16). Thus, the transcription factors that bind to these elements may interact with each other, and these interactions may be crucial to the further understanding of L-PK gene transcriptional regulation.
It has been shown that hepatocyte nuclear factor (HNF) 1␣ binds to L-I (9, 10) and that nuclear factor (NF) 1 family members bind to L-IIA (10). Although several transcription factors are reported to interact with the L-III element (10,(17)(18)(19), recent studies revealed that carbohydrate response element-binding protein (ChREBP) is an important physiological factor (20 -23). ChREBP is reported to bind to L-III as a heterodimer with Max-like protein X. It has been widely accepted that HNF4␣ is involved in transcriptional activation of the L-PK gene, because reporter gene assays and binding assays showed that HNF4␣ stimulates the L-PK promoter activity by binding to the L-IIB element (10,13,24). However, we suspected an alternative mechanism, because NF1 family members also bind to the L-IIB element as well as the accessory element in the S 14 gene, whereas HNF4␣ does not bind to this element in the S 14 gene (25). The S 14 accessory element located in the adjacent region to carbohydrate response element is involved in carbohydrate responsiveness of the S 14 gene and therefore corresponds to L-IIB in the L-PK gene (14). In addition, expression of HNF4␣ and L-PK genes in the liver are regulated by hormones and diet in opposite ways (26,27). However, functional significance of NF1 family members in the L-PK gene regulation has never been clarified. We assumed that NF1 family proteins play important roles in transcriptional regulation of the L-PK gene by interactions with HNF1␣ and ChREBP.
HNF1␣ (also called LF-B1) is a homeodomain-containing transcription factor expressed in liver, kidney, pancreatic ␤ cells, and small intestine (28,29). HNF1␣ is implicated in hepatocyte differentiation and is required for liver-specific expression of several genes, including albumin, ␤-fibrinogen, and other genes for homeostasis and metabolisms of glucose, lipids, steroids, and amino acids. HNF1␣ consists of 628 amino acids and contains three functional domains, an N-terminal dimerization domain, a DNA-binding domain (a Pit1-Oct1-Unc86-like domain and an atypical homeodomain), and a C-terminal transactivation domain.
NF1 family proteins are ubiquitously expressed and regulate, not only ubiquitously expressed genes, but also hormonally and developmentally regulated genes (30 -32). There are four NF1 genes in mammals, NF1-A, -B, -C, and -X, from which numerous NF1 isoforms are generated through alternative splicing or promoter usage. NF1 proteins are highly conserved in the N-terminal DNA-binding/dimerization domains, whereas the C termini contain diverse transactivation/repression domains.
In this study, we focused our research on the role of NF1 family members in transcriptional regulation of the L-PK gene. We examined functional and physical interaction between HNF1␣ and NF1 family proteins that bind to adjacent regions of the L-PK promoter and found that cooperative interaction between HNF1␣ and NF1 family proteins plays an important role in L-PK transcription in the liver.

EXPERIMENTAL PROCEDURES
Materials-The dye terminator cycle sequencing kit was obtained from PE Applied Biosystems. The dual luciferase reporter gene assay kit, pGL3-Basic, pRL-SV40, and the T7 TNT quick-coupled transcription/ translation system were purchased from Promega. The Qiagen plasmid and Qiaex gel extraction kits and pQE-30 were purchased from Qiagen. Horseradish peroxidase-conjugated secondary antibodies, ECL plus reagent, Hyperfilm, pGEX vectors, protein A-Sepharose, glutathione-Sepharose, [␥-32 P]ATP (111 TBq/mmol), and [ 35 S]methionine (43.17 TBq/mmol) were obtained from Amersham Biosciences. Mutan-Super Express kit, T4 polynucleotide kinase and restriction endonucleases were purchased from Takara. Plasmid pSPORT 1 and Wako PURE system, an in vitro transcription/translation system, were obtained from Invitrogen and Wako, respectively.
Animals and Treatment-6 -8-week-old male Sprague-Dawley and Wister rats were used. To examine the effect of a high carbohydrate diet, rats were starved for 24 or 48 h and were then refed with a high carbohydrate diet containing 61% dextrin and 20% sucrose (5). All animal experiments were approved by the Animal Experiment Committee, Nagoya University, and rats were treated in accordance with the guidelines.
Mutagenesis-Site-directed mutagenesis was performed using the Mutan-Super Express kit according to the manufacturer's instructions. The mutations were introduced into pLluc189 using the following oligonucleotides: mL-IIA, Ϫ127 5Ј-GTACAAGGCTTCAGTACTCAA-GAGAGATGC-3Ј Ϫ98; mL-IIB, Ϫ156 5Ј-CCCGTGGTTCCTAGTA-CTTGCCAGCCAGTGTACAAG-3Ј Ϫ121 (the mutated nucleotides are underlined). All mutations were confirmed by DNA sequencing. Reporter Gene Assays-HeLa cells were co-transfected with 600 ng of reporter plasmid, 10 ng of pRL-SV40, and the indicated amount of RSV expression vectors (Figs. 2, 4, and 5). The total amount of DNA was adjusted by adding empty vector, when required. All plasmids used for transfection were prepared using the Qiagen plasmid Midi kit. The transfection conditions and luciferase reporter assays have been described previously (33).
Electrophoretic Mobility Shift Assays (EMSAs)-Nuclear extracts were prepared from rat liver and HeLa cells as described previously (25,33). Rat liver nuclear extracts were heated at 60°C for 5 min and centrifuged at 15,000 revolutions/min. The resultant supernatant was used for EMSAs. HNF1␣ and NF1-C were synthesized in vitro from pT7-B1 using the T7 TNT quick-coupled transcription/translation system and from pT7-NF1-C using the Wako PURE system, respectively, and both were also used for EMSAs. The oligonucleotides used in EMSAs are listed in TABLE ONE. Double-stranded oligonucleotides were 5Ј-endlabeled with [␥-32 P]ATP and T4 polynucleotide kinase. 10 g of nuclear extract were incubated for 30 min at room temperature with the 32 Plabeled probe in a reaction mixture containing 12.5 mM HEPES-KOH (pH 7.9), 10% glycerol, 100 mM KCl, 1 mM EDTA, 1 mM dithiothreitol, and 1 g of poly(dI-dC). When in vitro synthesized HNF1␣ and NF1-C were used for EMSAs, the reaction mixture contained 10 mM Tris-HCl (pH 7.5), 50 mM NaCl, 0.5 mM dithiothreitol, 10% glycerol, 0.05% Nonidet P-40, 1 g of sonicated single-stranded DNA, and the 32 P-labeled probe. Competition experiments included a 200-fold molar excess of unlabeled competitor oligonucleotides. Supershift assays were carried out by preincubating the nuclear extract with anti-NF1 antibody or anti-HNF1␣ antibody (34) for 30 min at room temperature before the addition of labeled oligonucleotide. Anti-M 1 -PK antibody (35) was used as a control. Polyclonal anti-NF1 antibody 8199 was kindly provided by Dr. Naoko Tanese (New York University School of Medicine, New York, NY). After the binding reaction, the mixture was subjected to electrophoresis on a 4% polyacrylamide gel in 0.5ϫ Tris borate-EDTA buffer at 200 V for 1 h. The gel was dried and analyzed by a FUJIX BAS-2500 image analyzing system.
Chromatin Immunoprecipitation (ChIP) Assays-Rat hepatocytes were obtained from male Sprague-Dawley rats by the collagenase perfusion method (36). ChIP assays were performed essentially as previously described (37). Hepatocytes were treated with 1% formaldehyde at 37°C for 15 min. The cross-linking reaction was stopped by the addition of glycine to a final concentration of 0.125 M. After being washed with cold phosphate-buffered saline, the cells were suspended in lysis buffer containing 50 mM Tris-HCl (pH 8.1), 1% SDS and 10 mM EDTA. The lysate was sonicated to generate DNA fragments with a length of Ͻ1,000 bp. After removal of cell debris by centrifugation, the supernatant was diluted 10-fold with buffer containing 16.7 mM Tris-HCl (pH 8.0), 1.1% Triton X-100, 1.2 mM EDTA, 167 mM NaCl and then precleared for 1 h with protein A-Sepharose in the presence of normal rabbit IgG and 2 g of sonicated denatured DNA to reduce the nonspecific background. The supernatant was incubated with normal rabbit IgG, anti-NF1 antibody or anti-HNF1␣ antibody, or without any antibody overnight at 4°C and then incubated with protein A-Sepharose for 2 h at 4°C. The reaction mixture was centrifuged, and protein A-Sepharose was washed extensively. DNA was eluted from the precipitate, reverse cross-linked by incubating at 65°C for 6 h, and then treated by proteinase K. The purified DNA was dissolved in water and analyzed by PCR using primer sets for the promoter region (Ϫ246 to Ϫ23, 5Ј-TACAGACCTGATCT-GAGCCTTTG-3Ј and 5Ј-TTTATACTGCGCCTTTGTCAGTG-3Ј) and exon 12 (5Ј-GCTCTGGCTATACCAACATCATGC-3Ј and 5Ј-CGGATTGAGACAGGGATATACAAC-3Ј) of the L-PK gene. The input sample contained 1 ⁄ 250 of the supernatant used for immunoprecipitation. PCR products were separated by agarose gel electrophoresis and visualized by ethidium bromide staining.
Western Blotting-The proteins were separated by electrophoresis using 9 or 10% SDS-polyacrylamide gels, followed by transfer to membranes. The membranes were incubated for 1 h at room temperature with the antibodies indicated in Figs. 8 and 9 and then washed three times with phosphate-buffered saline containing 0.1% Tween 20 before being probed with horseradish peroxidase-conjugated second antibodies. Blots were developed using ECL plus reagent and exposed to Hyperfilm, according to the manufacturer's protocols.
GST Pull-down Assays-35 S-labeled HNF1␣ was synthesized from pT7-B1 using the T7 TNT quick-coupled transcription/translation system in the presence of [ 35 S]methionine according to the manufacturer's protocol. Histidine-tagged NF1-C was expressed from pQE-NF1-C in E. coli. Preparation of GST fusion proteins containing NF1 family members or HNF1␣ and pull-down analysis were carried out as described previously (33). Finally, samples were analyzed by SDS-polyacrylamide gel electrophoresis followed by autoradiography or Western blotting.

Nucleotide sequences of oligonucleotides used in EMSAs
Mutated nucleotides are underlined.

JOURNAL OF BIOLOGICAL CHEMISTRY 39829
Co-immunoprecipitation-Rat liver nuclei were isolated as described previously (25), and nuclear extracts were prepared with immunoprecipitation buffer (40 mM Tris-HCl, pH 7.8, 0.3% Nonidet P-40, and 150 mM NaCl). The nuclear proteins (13.5 mg) were first precleared by incubation with 60 l of a 50% protein A-Sepharose slurry containing normal rabbit IgG for 1 h at 4°C and then incubated with either normal rabbit IgG or anti-NF1 antibody overnight at 4°C. Immune complexes were collected by incubation with 30 l of a 50% protein A-Sepharose slurry for 1 h followed by centrifugation. Immunoprecipitates were washed five times with immunoprecipitation buffer and subjected to Western blotting with anti-HNF1␣ antibody.

RESULTS
HNF1␣ and NF1 Family Members Synergistically Activate L-PK Promoter Activity-To examine the functional interaction between NF1 family members and HNF1␣, we used a non-hepatic cell line, HeLa, for the reporter gene assays, because this cell line does not express endogenous HNF1␣. The reporter gene used was pLluc189, containing the 5Ј-flanking region of the L-PK gene up to Ϫ189. This region contains L-I, -IIA and -IIB, and L-III elements that are bound by HNF1␣, NFI family members and ChREBP, respectively. Although co-transfection with NF1 family members resulted in only slight increases in the reporter activity (NF1-A, 1.7 Ϯ 0.26; NF1-B, 1.6 Ϯ 0.12; NF1-C, 5.5 Ϯ 2.3; NF1-X, 2.8 Ϯ 0.71), co-transfection with HNF1␣ expression vector caused ϳ50-fold enhancement in the activity when compared with cells that were not co-transfected with either transcription factor (Fig. 2). Co-transfection of both HNF1␣ and NF1 expression vectors resulted in an additional 8 -10-fold increase in the HNF1␣-mediated activation of the L-PK promoter. This synergistic effect was observed in all NF1 isoforms. These results indicate that NF1 family proteins weakly activated L-PK transcription in the absence of HNF1␣ but greatly enhanced the activity through functional interaction with HNF1␣.
Both L-IIA and L-IIB Are Crucial for Maximum Synergistic Effect of NF1 Family Members-Previous studies suggested that NF1 family proteins bind to both L-IIA and L-IIB (10,25). To determine which site is necessary for synergistic interaction between NF1 proteins and HNF1␣, we introduced mutations into both sites of pLluc189 independently or simultaneously. As demonstrated by EMSAs, we confirmed that NF1 proteins bound to both L-IIA and L-IIB elements (Fig. 3). However, NF1 proteins failed to recognize mutated oligonucleotides of both L-IIA and L-IIB. We used NF1-C expression plasmid as a representative of NF1 family members for transfection experiments. When NF1-C alone was transfected as an expression plasmid, the promoter activity of the wild type reporter was only slightly stimulated (Fig. 4). Therefore, it was difficult to evaluate the effect of mutations on the activation of pLluc189 by NF1-C alone. However, when the pLluc189 reporter plasmid containing either mutated sequences of L-IIA or L-IIB was transfected with NF1-C and HNF1␣ expression plasmids into HeLa cells, synergistic activation by NF1-C and HNF1␣ was decreased by 44 or 35%, respectively. Interestingly, both mutations, which caused a further decrease in the activity (p Ͻ 0.05 compared with mL-IIA or mL-IIB), did not completely eliminate the synergistic effect of NF1-C with HNF1␣. Thus, although both NF1-binding sites are required for the maximum synergistic effect of NF1 family members with HNF1␣, NF1 proteins may be able to functionally interact with HNF1␣ without binding to promoter DNA to some extent. Alternatively, the reporter plasmid may contain additional unknown NF1-binding elements. The first possibility was further examined using p(LI) 2 luc62 as a reporter. This plasmid contains two copies of the L-I element (HNF1␣-binding site) of the L-PK gene but not the NF1-binding site. As shown in Fig. 5, this reporter was activated by HNF1␣ but not by NF1-C. However, co-transfection of HNF1␣ and NF1-C expression plasmids resulted in a further marked increase in activity, and this activation of NF1-C was dose-dependent. On the other hand, the activity of pLluc62, which contains neither the NF1-nor HNF1␣-binding sites, was not stimulated by transfection with HNF1␣ alone or HNF1␣ and NF1-C (data not shown). These results confirmed that synergistic effects between HNF1␣ and NF1 proteins do not necessarily require NF1 binding to the L-PK promoter. The second possibility was examined by EMSAs using rat liver nuclear extracts. The formation of complexes of labeled L-IIB was competitively inhibited by the addition of a 200-fold molar excess of unlabeled L-IIB and NF1-Ade but not by mL-IIB (Fig. 3). The addition of unlabeled L-A, mL-IIAЈ, and L-B oligonucleotides, which correspond to the regions between Ϫ187 and Ϫ147, Ϫ127 and Ϫ98, and Ϫ98 and Ϫ62, respectively, also failed to affect the formation of the complexes (data not shown). Thus, it is unlikely that the reporter plasmid pLluc189 contains other NF1-binding elements besides L-IIA and L-IIB.
NF1 Family Members and HNF1␣ Actually Associate with L-PK Promoter in Hepatocytes-To examine whether NF1 proteins and HNF1␣ associate with L-PK promoter in vivo under physiological conditions, hepatocytes were isolated from rats starved for 48 h or refed with a high carbohydrate diet and used for ChIP assays. A specific primer set for PCR was used to produce a 224-bp DNA fragment of the L-PK gene promoter region (Ϫ246 to Ϫ23) (4). This region included L-I, L-IIA, L-IIB, and L-III sites. Another primer set was used to amplify a 219-bp region in exon 12 of the L-PK gene as a control. A PCR product specific to the L-PK promoter region was obtained from immunoprecipitates with anti-HNF1␣ antibody or anti-NF1 antibody, and this product was observed in both the starved and refed states (Fig. 6). In contrast, there was no amplified band when DNA samples were obtained in the absence of antibody or with normal rabbit IgG or when primers specific for exon 12 of the L-PK gene were used for PCR. These results indicate that the L-PK promoter is occupied by HNF1␣ and NF1 family members in vivo under physiological conditions and that this binding is not under nutritional control.
NF1 Family Members Directly Bind to Mainly Homeodomain of HNF1␣ through Its N-terminal DNA-binding Domain-Our reporter gene assays indicate that HNF1␣ and NF1 family proteins cooperatively stimulate L-PK gene transcription. To examine physical interaction between HNF1␣ and NF1 family proteins, we carried out GST pulldown assays. 35 S-labeled HNF1␣ was synthesized by an in vitro tran- scription/translation system and incubated with GST, GST-NF1-C, or GST-NF1-X immobilized to glutathione-Sepharose beads. As shown in Fig. 7A, 35 S-labeled HNF1␣ interacted with GST-NF1-C and GST-NF1-X but not with GST alone. These data clearly demonstrate direct interaction between HNF1␣ and NF1-C or NF1-X. Because all of the NF1 family proteins can enhance HNF1␣-mediated activation of the L-PK promoter and N-terminal DNA-binding regions are highly conserved among NF1 family members, we hypothesized that these regions are responsible for physical binding to HNF1␣. To examine this hypothesis, GST fusion proteins with N-terminal regions of NF1 family proteins were prepared and incubated with 35 S-HNF1␣. As shown in Fig.  7B, the N-terminal regions of all NF1 family members specifically interacted with HNF1␣. Therefore, we conclude that the N-terminal DNAbinding regions of NF1 family members are involved in direct interaction with HNF1␣.
NF1 Family Members Directly Associate with HNF1␣ in Vivo-To assess the physical interaction between HNF1␣ and NF1 family proteins in rat liver, we performed co-immunoprecipitation assay with the anti-NF1 antibody. The liver nuclear extracts were prepared from rats refed with a high carbohydrate diet for 16 h and were immunoprecipitated with anti-NF1 antibody or normal rabbit IgG. Immunoprecipitates were then analyzed by Western blotting using an anti-HNF1␣ antibody. As shown in Fig. 9, HNF1␣ was detected in the immunoprecipitates from rat liver nuclear extracts by the anti-NF1 antibody but not those by normal IgG. The same results were obtained when the liver nuclear extracts were obtained from rats starved for 24 h (data not shown). These results indicate that NF1 family members interact with HNF1␣ in rat liver and that this interaction is not under the control of nutritional conditions.
Binding of HNF1␣ to the L-PK Promoter Is Enhanced by Interaction with NF1 Family Members-DNA binding affinity of HNF1␣ may be increased by interaction with NF1 family members. To examine this hypothesis, we first performed EMSAs with nuclear extracts from HeLa cells transfected with HNF1␣ and/or NF1-C expression vectors. When labeled L-I oligonucleotide containing the HNF1␣-binding site was used as a probe, no specific complexes were detected with nuclear extracts from HeLa cells transfected with NF1-C or empty vector (Fig.  10A). However, transfection of HNF1␣ resulted in formation of the complexes that were competed by the addition of anti-HNF1␣ antibody. Co-transfection of HNF1␣ and NF1-C caused increases in intensities of these bands that were dependent on the amount of NF1-C transfected. A supershifted band was observed by the addition of anti-NF1 antibody to nuclear extracts. These results suggest that the DNA complexes contain both NF1-C and HNF1␣ and that DNA binding of HNF1␣ is stimulated by physical interaction with NF1-C. This was further confirmed by EMSAs using HNF1␣ and NF1-C that were synthesized in the in vitro transcription/translation system (Fig. 10B). We confirmed synthesis of both factors by Western blotting (data not shown). A DNA-protein complex was observed by incubation with HNF1␣ but not by NF1-C. The intensity of the complex was enhanced by the addition of increased NF1-C. This complex was reduced by the addition of a 200-fold molar excess of unlabeled L-I oligonucleotide or anti-HNF1␣ antibody and supershifted by the addition of anti-NF1 antibody. Thus, we conclude that NF1 family members enhance binding of HNF1␣ to the L-PK promoter through their direct interactions. Rats were starved for 48 h or refed with a high carbohydrate diet. Hepatocytes were isolated from the liver and subjected to formaldehyde cross-linking followed by immunoprecipitation with anti-NF1 antibody or anti-HNF1␣ antibody. Precipitated DNA was analyzed by PCR with a primer set specific for the promoter or exon 12 of the L-PK gene. Input corresponds to 0.4% of the amount of DNA used in the assay. Ab, antibody; Pro, promoter region.

DISCUSSION
Transcription factors not only bind to the cognate DNA sequences but also interact with one another to regulate gene transcription. In addition, these factors recruit non-DNA-binding proteins such as coactivators and co-repressors. They, in turn, form a ternary complex, interact with basic transcription machinery, and confer transcriptional activity to the gene promoter. Here we demonstrated for the first time that NF1 family members play an important role in the transcriptional activation of the L-PK gene by interacting with HNF1␣ in vivo under physiological conditions as follows. First, both NF1 family proteins and HNF1␣ bound to the promoter region containing L-I, L-IIA, L-IIB, and L-III of the L-PK gene in liver, as shown by ChIP assays. Second, NF1 family members directly interacted with HNF1␣ in liver as shown by co-immunoprecipitation assays, which resulted in increased binding of HNF1␣ to the L-I element and cooperative activation of the L-PK gene transcription. Although four family members of the NF1 gene are expressed in the liver (30,31), all of them can be involved in cooperative interaction with HNF1␣ through their DNA-binding domains, which are highly conserved among NF1 family members.
Although maximum synergistic effect between NF1 family members and HNF1␣ on the L-PK gene required binding of NF1 family proteins to both the L-IIA and L-IIB elements, significant functional interaction was still observed, even in the absence of NF1 binding to the two elements because of their mutations. It is unlikely that the reporter plasmid pLluc189 contains additional NF1-binding elements, because EMSAs showed that no other sequences than L-IIA and L-IIB in the L-PK promoter region between Ϫ62 and Ϫ187 interacted with NF1 proteins and because another reporter plasmid of pLluc62 containing the promoter region up to Ϫ62 did not respond to NF1-C with or without HNF1␣. In addition, it has been reported that no other nuclear protein-binding sequences than the four elements described above were observed in the promoter region up to Ϫ183 (10). Moreover, NF1 proteins activated the promoter containing only the HNF1␣-binding site. Thus, we consider the interesting possibility that NF1 family members function as coactivators of HNF1␣, although this issue must be further addressed in future studies. There are co-activators with histone acetyltransferase activity that can physically interact with HNF1␣ (38,39). They include p300/cAMP-response element-binding protein, p300/cAMP-response element-binding protein-associated factor, receptor-associated co-activator 3, and steroid receptor co-activator-1. We examined whether p300 is involved in the synergistic transcriptional effect of NF1 proteins and HNF1␣ on the L-PK gene using the reporter gene assay. The results indicated that p300 did not enhance NF1-and HNF1␣-mediated activation of the L-PK promoter. 3 The reason for this is unclear, because p300 has been reported to interact with the activation domain of HNF1␣ (39). However, it is possible that other co-activators are involved in cooperative activation of the L-PK gene. In addition, it has been reported that HNF4␣ functions as a co-activator of HNF1␣ and cooperates with p300 to achieve the highest HNF1␣-mediated transcription (40). However, its physiological significance remains to be determined.
Although the L-PK promoter is under the control of nutritional conditions, occupancy of NF1 proteins and HNF1␣ on the L-PK promoter was not altered by nutritional manipulations. Physical interaction between the two transcription factors was also observed in starved and refed states, suggesting that this interaction is also not regulated by dietary carbohydrate. Thus, ChREBP/Max-like protein X binding to the L-III element of the L-PK promoter is critical for carbohydrate regulation of the L-PK gene as reported (20 -23). Nevertheless, cooperative interaction between NF1 proteins and HNF1␣ plays an important role in carbohydrate regulation as well as cell type-specific transcription of the L-PK gene in liver, because maximum transcriptional activity of this gene is achieved through functional interactions among all cis-acting elements. Furthermore, it is possible that NF1 family members are also involved in the transcriptional regulation of the L-PK gene by interacting with ChREBP. However, further studies are required to solve this problem. 3 S. Satoh and T. Noguchi, unpublished results. FIGURE 9. Co-immunoprecipitation analysis of NF1 family members and HNF1␣ in rat liver. Rats were refed with a high carbohydrate diet for 16 h. The liver nuclear extracts were prepared and subjected to immunoprecipitation with anti-NF1 antibody or normal rabbit IgG. The precipitates were then analyzed by immunoblotting with anti-HNF1␣ antibody. Input corresponds to 0.23% of the nuclear extracts used in the assay. FIGURE 10. NF1-C enhances HNF1␣ binding to the L-I element. A, HeLa cells were transfected with the indicated amounts of pRSV-HNF1␣ and/or pRSV-NF1-C. Nuclear extracts were prepared from cells after 48 h and incubated with anti-HNF1␣ antibody or anti-NF1 antibody where indicated. EMSAs were performed using 32 P-labeled L-I oligonucleotide as a probe. B, HNF1␣ and NF1-C were synthesized using an in vitro transcription/translation system. The indicated amounts of products were used for EMSAs as described under "Experimental Procedures." The total amount of reticulocyte lysates was adjusted by adding unprogrammed lysates. Oligonucleotide was added as a competitor at 200-fold molar excess where indicated. SS, supershifted complex. DECEMBER