NR2F1 disrupts synergistic activation of the MTTP gene transcription by HNF-4α and HNF-1α.

Regulation of microsomal triglyceride transfer protein (MTP) expression mainly occurs at the transcriptional level. We have previously shown that MTTP gene expression was repressed in nondifferentiated intestinal cells by nuclear receptor 2 family 1 (NR2F1). However, mechanisms involved in the repression of MTP by NR2F1 were not elucidated. Here, we show that MTP expression requires hepatic nuclear factor (HNF)-4α transcription factor. Different HNF-1 proteins synergistically enhance MTP promoter activity along with HNF-4α by binding to different cis elements. NR2F1 does not alter individual effects of HNF-4α and HNF-1 proteins on the MTTP gene promoter. However, NR2F1 suppresses synergistic activation of the MTP promoter by HNF-4α/HNF-1α by binding to a direct repeat 1 (DR1) element. This suppression is further enhanced in the presence of nuclear receptor corepressor 1. In short, these studies identified a novel mechanism of MTP repression that involves binding of NR2F1 to the DR1 element and recruitment of corepressors. In this mechanism, NR2F1 does not affect activities of individual transcription factors; instead, it abrogates synergistic activation by HNF-4α and HNF-1 proteins.

to the manufacturer's instructions. Viral titer was determined with a HIV-1 p24 Antigen ELISA kit (ZeptoMetrix; Buffalo, NY). Two shRNAs out of these viruses against NR2F1 (core sequence: 5 ′ -CG TCCGCAGGAACTTAACTTA-3 ′ or 5 ′ -CAGC TTCAACT GG-CC TTA CAT-3 ′ ) or NCOR1 (core sequence: 5 ′ -GCC CAC AGA T-GA TG AAGAAAT-3 ′ and 5 ′ -GCCTTAAATATCCCAAACAAA-3 ′ ), res pectively, were selected to knock down gene expressions in Caco-2 cells. One day before viral transduction, Caco-2 cells were plated at 40% confl uence in 60 mm dishes. Fresh growth media containing 8 g/ml polybrene (Sigma) and ‫ف‬ 1 × 10 6 viral particles were added the following morning. After overnight incubation, the virus-containing media was replaced with fresh growth media and the transduced cells were allowed to grow for 2 more days before collection for RNA or protein extractions.

Chromatin immunoprecipitation
Chromatin immunoprecipitation (ChIP) was used to study the binding of transcription factors to the MTTP promoter ( 6 ). In brief, Caco-2 cell cultures 4 and 14 days old were fi xed with 1% formaldehyde. After 15 min at room temperature, cells were harvested, and nuclear fractions were isolated and sonicated. Approximately 25 g of cross-linked genomic DNA in 1 ml 1% BSA/PBS was incubated overnight at 4°C with 10 l of antibodies against NR2F1, HNF-4 ␣ , HNF-1 ␣ , or a control rabbit IgG. The following day, 50 l of protein G beads (GE Life Sciences) were used to precipitate DNA. Then DNA was extracted by phenol-chloroform after reversal of cross-linkage by incubating at 65°C for 4 h. PCR (95°C, 3 min; 25 cycles of 95°C, 30 s; 58°C 30 s; 72°C 30 s; 72°C 7 min) was carried out using primers described previously ( 6 ) to determine binding of these factors to the MTTP promoter.

Statistics
Experiments were performed in triplicate and repeated at least twice. Values are presented as means ± SEM. Statistical signifi cance ( P < 0.05) was determined using Student's t test (Graph-Pad Prism; La Jolla, CA).

Three HNF-1 family members synergistically enhance HNF-4 ␣ -mediated activation of the MTTP promoter
The minimal 204 bp MTTP promoter contains cis elements (HNF-1, HNF-4, and DR1) that are critical for differentiation-dependent activation of MTP expression in Caco-2 cells ( 4,6,9 ). Putative binding of HNF-4 ␣ and different HNF-1 family members to various cis elements is We ( 6 ) and others ( 16 ) have shown that MTP expression is induced during differentiation of a human intestinal adenocarcinoma cell line, Caco-2 cells. During cellular differentiation, the binding of HNF-4 ␣ and HNF-1 ␣ to the MTP promoter did not change signifi cantly, but there was a significant increase in the binding of Pol II and acetylated histone 3 to the MTTP promoter (6). Therefore, induction of the MTTP gene expression was not related to increases in the binding of different transcription factors to the promoter. Instead, the induction was dependent on the reduced expression and binding of nuclear receptor 2 family 1 (NR2F1), a putative transcriptional repressor, to the MTTP promoter. However, the molecular mechanisms of how NR2F1 represses MTTP transcription were not illustrated. In this study, we studied the binding of NR2F1 to the cis elements in the minimal MTTP promoter, evaluated the activation potential of the two HNF-1 ␤ isoforms, and elucidated the mechanism by which NR2F1 represses MTTP gene expression.

Chemicals
Most chemicals, including DNA oligonucleotides, were purchased from Sigma (St. Louis, MO), whereas most primary antibodies were from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA), unless indicated otherwise.

Cell culture and transfection
HEK293 and Caco-2 cells were maintained in Dulbecco's modifi ed medium supplemented with 10% FBS, streptomycin, ampicillin, and glutamine in a humid 37°C cell culture incubator aired with 5% CO 2 . For transient transfection, cells were seeded at 40% confl uence in 24-well plates on the day before transfection. The Exgen 500 transfection reagent (Fermentas; Glen Burnie, MD) was used to introduce 0.5 µg of various pMTP-204 reporter plasmids ( 6 ), together with 5 ng of a control plasmid, pCMV-RL, into HEK293 or Caco-2 cells, according to the manufacturer's instructions. Cell lysates were then obtained after 48 h of incubation, and expression of fi refl y and Renilla luciferases was quantifi ed by using the dual luciferase assay kit (Promega; Madison, WI). The ratios of fi refl y to Renilla luciferase for each sample were calculated and used as the indices to compare different promoter activities.

Lentivirus-mediated shRNA knockdown in Caco-2 cells
shRNA sets in pLKO.1 clones against NR2F1 (RHS4533-NM-005654) and NCOR1 (RHS4533-NM-006311) were purchased from Open Biosystems, Inc. Subsequently, shRNA-coding lentiviral particles were synthesized with the Trans-Lentiviral TM Packaging System Kit with Arrest-In agent (Open Biosystems, Inc.) according that the two HNF-1 ␤ isoforms also synergistically activated the MTTP promoter in the presence of HNF-4 ␣ , although their effi ciency was slightly less than HNF-1 ␣ . These fi ndings suggest that MTTP promoter activity is illustrated in Fig. 1A . We compared MTTP promoter sequences from seven species using CLUSTAL W and found that these elements are conserved across vertebrates ( Fig.  1B ), as has been reported before ( 4,6,8,9 ). To identify transcription factors that bind to the three conserved cis elements, we cloned HNF-4 ␣ , HNF-1 ␣ , HNF-1 ␤ (a), and HNF-1 ␤ (b) and studied their expression in HEK293 cells that do not endogenously express these factors (data not shown). Transfection of HNF-1 ␣ expression plasmid resulted in the appearance of a protein of ‫ف‬ 80 kDa, whereas transfection of HNF-1 ␤ (a) and HNF-1 ␤ (b) plasmids revealed two protein bands with slightly different molecular mass around 60 kDa ( Fig. 2A ), consistent with a previous report ( 14 ) that HNF-1 ␤ (b) is 26 amino acids shorter than HNF-1 ␤ (a). As expected, transfection of HNF-4 ␣ yielded proteins of ‫ف‬ 55 kDa ( Fig. 2B ). These studies indicate successful cloning and expression of various HNF proteins.
Next, we studied the effect of these transcription factors on the MTTP promoter activity using a dual luciferase reporter assay. As shown in Fig. 2C , the minimal MTTP promoter activity was weak in HEK293 cells, probably due to the absence of critical transcription factors in these cells, but was robust in undifferentiated Caco-2 cells (see DISCUSSION). Expression of HNF-1 ␤ (b) had no effect on the MTP promoter; however, expression of HNF-1 ␣ or HNF1 ␤ (a) slightly ( ‫ف‬ 2-fold) increased this activity. By contrast, HNF-4 ␣ enhanced the promoter activity by ‫ف‬ 15-fold, consistent with mouse genetics and cell culture studies ( 8,10 ); this activity was comparable to that present in undifferentiated Caco-2 cells. Moreover, we coexpressed different HNF-1 isoforms with HNF-4 ␣ . HNF-1 ␣ synergistically ( ‫ف‬ 30-fold) activated MTTP promoter activity when expressed with HNF-4 ␣ , consistent with the studies reported in hepatoma cells by Sheena et al. ( 12 ). For the fi rst time, we observed Possible activators and repressors that could bind to these elements were identifi ed using online software MatInspector (Genomatix). MatInspector predicts that HNF-1 ␣ binds to the HNF site and HNF-4 ␣ binds to HNF-4 and DR1 elements. Binding of HNF-1 ␤ (a) and HNF-1 ␤ (b) to the HNF-1 site is based on the data presented in this study. B: The MTTP promoter sequences from seven species, including human (1), mouse (2), dog (3), rat (4), hamster (5), chicken (6) and zebrafi sh (7), were compared to identify conserved homologous sequences using CLUSTAL W (UCSD Biology Workbench). Consensus sequences for different cis elements is shown in bold underneath the original sequences, whereas mutated sequences are shown in italics above the original sequence. completely abolished the expression of MTP promoter activity by HNF-4 ␣ and its synergistic activation by HNF-1 ␣ ( Fig. 3B , HNF-4m+DR1m). We interpret these studies to suggest that HNF-4 ␣ is able to bind to both the HNF-4 and DR1 elements. The binding of HNF-1 ␣ to the HNF-1 element synergistically activates both of these activities.

NR2F1 represses gene transcription by disrupting synergistic activation of the MTTP promoter by HNF-4 ␣ /HNF-1 ␣
We have previously shown that overexpression of NR2F1 in Caco-2 cells repressed the MTTP promoter ( 6 ). Hence, attempts were made to fi nd out the molecular basis of MTTP gene repression by NR2F1, which has a molecular mass of ‫ف‬ 46 kDa when expressed in HEK293 cells ( Fig. 4A ). We hypothesized that NR2F1 might counteract the activation of the MTTP promoter by either HNF-4 ␣ or HNF-1 ␣ . Therefore, we fi rst studied the effect of NR2F1 on HNF-1 ␣ activity ( Fig. 4B ). Similar to that shown in Fig. 2C , HNF-1 ␣ increased MTTP promoter activity by about 2-fold. NR2F1 had no effect on the HNF-1 ␣ activity. Next, we studied the effect of NR2F1 on HNF-4 ␣ activity ( Fig. 4C ). HNF-4 ␣ signifi cantly enhanced the MTTP promoter activity, and again NR2F1 had no effect on the HNF-4 ␣ activity. These studies eliminated the possibility that NR2F1 might interact separately with these activators and repress their activities.
We then hypothesized that NR2F1 might suppress the synergistic activation of MTTP transcription by HNF-4 ␣ and HNF-1 ␣ . Consistent with data presented earlier ( Figs. 2, 3 ), coexpression of HNF-4 ␣ and HNF-1 ␣ synergistically activated the MTTP promoter activity ( Fig. 4D ). NR2F1 repressed this synergistic activation by 40% ( Fig. 4D ). Because we obtained only partial inhibition, we studied the effect of higher concentrations of NR2F1 ( Fig. 4E ). Transfections with higher amounts of plasmid DNA increased NR2F1 protein levels with no change in ␤ -actin. NR2F1 showed a dose-dependent reduction of synergistic activation of the MTP promoter by HNF-1 ␣ /HNF-4 ␣ at lower concentrations ( Fig. 4E ). But, again, at several higher concentrations, it only reduced the synergistic activation by ‫ف‬ 40%. These studies indicated that NR2F1 partially suppresses the synergistic activation by HNF-4 ␣ /HNF-1 ␣ .
Because NR2F1 did not completely repress, we hypothesized that corepressors might be required to abolish synergistic activation by HNF-4 ␣ /HNF-1 ␣ . Several lines of evidence have indicated that NCOR1 enhances the transrepressive potential of NR2F1 ( 17 ). Hence, we asked whether a corepressor is necessary for the repression of the MTTP promoter activity. Expression of NCOR1 resulted in the synthesis of a protein corresponding to ‫ف‬ 450 kDa ( Fig. 4F , arrow). When NCOR1 was expressed along with NR2F1, the MTP promoter was further suppressed ( Fig. 4D ). However, expression of NCOR1 in the absence of NR2F1 had no effect on synergistic activation of MTP promoter activity by HNF-1 ␣ /HNF-4 ␣ . These data suggest that NR2F1 represses MTTP promoter activity by disrupting the synergy between HNF-4 ␣ and HNF-1 ␣ and additional recruitment of NCOR1 augments this repression. synergistically activated by HNF-4 ␣ in combination with any one of the three HNF-1 family proteins.

HNF-1, HNF-4, and DR1 elements are important for synergistic activation of MTP expression
To identify cis elements necessary for the minimal and synergistic activation of MTP, we mutated these conserved elements individually or in combination and expressed them along with plasmids expressing HNF-4 ␣ alone or together with HNF-1 ␣ . Again, the MTTP promoter showed negligible activity when expressed alone in HEK293 cells ( Fig. 3A , control). Coexpression with HNF-4 ␣ enhanced the MTTP promoter activity by 15-fold. HNF-4 ␣ and HNF-1 ␣ synergistically activated the promoter activity by approximately 40-fold. Mutation of the HNF-1 element had no effect on the activation of the MTTP promoter by HNF-4 ␣ ; however, it signifi cantly curtailed the synergistic activation with HNF-1 ␣ ( Fig. 3B , HNF1m). Mutagenesis of the HNF-4 element signifi cantly reduced but did not abolish the HNF-4 ␣ -mediated enhancement of MTTP promoter activity ( Fig. 3B , HNF4m). Nevertheless, the residual activity was synergistically activated by HNF-1 ␣ ( Fig. 3B , HNF4m). Similar results were obtained with the mutant DR1 element ( Fig. 3B , DR1m). DR1m had signifi cantly low MTTP promoter activity in the presence of HNF-4 ␣ , and the residual activity was synergistically activated by HNF-1 ␣ . Double mutagenesis of the HNF-4 and DR1 elements plasmids were coexpressed with or without NR2F1-expressing plasmids. Synergistic activation of the MTP promoter by the HNF-4 ␣ /HNF-1 ␣ was normalized to 1. As before, NR2F1 partially reduced the synergistic activation of the MTP promoter [ Fig. 5A , wild type (WT)]. Similarly, NR2F1 repressed the synergistic activation when the HNF-4 site was mutated ( Fig. 5A , HNF-4m). In contrast, DR1m was resistant to NR2F1 repression ( Fig. 5A , DR1m). These data indicate that the DR1 element is required for MTTP gene repression by NR2F1.
Then we asked whether NR2F1 binds directly to the DR1 element on the MTTP promoter in Caco-2 cells. ChIP analysis showed that the binding of HNF-1 ␣ and HNF-4 ␣ to the MTP promoter was similar in both undifferentiated ( Fig. 5B , D4) and differentiated ( Fig. 5B , D14) Caco-2 cells. In contrast, NR2F1 and NCOR1 binding was observed only in undifferentiated Caco-2 cells. Experiments were then performed to determine whether this binding suppresses MTP expression in undifferentiated cells. To test this, we expressed shRNA against NR2F1 and/or NCOR1 in undifferentiated Caco-2 cells. As shown in Fig. 5C , shRNA against NR2F1 (NR2F1-KD) reduced NR2F1 but had no effect on NR2F2 and NCOR1 mRNA levels. Similarly, shRNA against NCOR1 (NCOR1-KD) knocked down only NCOR1 expression, with no effect on NR2F1 and NR2F2. Importantly, MTP mRNA levels were increased ‫ف‬ 1.5-to 2-fold after knockdown of either NR2F1 or NCOR1; their double knockdown synergistically enhanced MTP expression by ‫ف‬ 6-fold compared with the control cells. Consistent with its mRNA, MTP protein was also induced in the undifferentiated cells ( Fig. 5D ). These studies indicate that loss of NR2F1 and NCOR1 derepresses MTTP gene expression in undifferentiated Caco-2 cells.
On the basis of these observations, we propose that the MTTP promoter contains one binding site for HNF-1 ␣ , HNF-4 ␣ , and NR2F1 ( Fig. 5E , top). NR2F1 binds to the DR1 element and represses the synergistic activation of MTTP transcription by HNF-4 ␣ /HNF-1 ␣ with the help of NCOR1 in undifferentiated Caco-2 cells ( Fig. 5E , top). If NR2F1 levels are low, then MTP transcription is permitted due to derepression ( Fig. 5E , bottom).

DISCUSSION
We have previously shown that NR2F1 suppresses MTP expression in undifferentiated Caco-2 cells. Furthermore, there was an inverse relationship between the expression of NR2F1 and MTP at the jejunum-colon and villus-crypt axes in mouse intestines ( 6 ). Loss of NR2F1 after differentiation of Caco-2 cells coincided with increases in MTP expression. These studies were interpreted to suggest that high expression of NR2F1 in undifferentiated cells suppresses MTP expression ( 6 ). To understand how NR2F1 could inhibit MTP expression, we used HEK293 cells that are defi cient in several transcription factors important for MTP expression. These cells were transfected with various MTTP promoter reporter constructs along with different transcription factors. Furthermore, different cis elements in the MTP promoter were mutagenized to evaluate their

NR2F1 represses synergistic activation by binding to the DR1 element
Next, attempts were made to identify the cis element(s) crucial for NR2F1-mediated MTTP repression. MTP promoter luciferase construct and HNF-4 ␣ /HNF-1 ␣ expression repressors. These transcription factors homodimerize or heterodimerize with other nuclear receptors and regulate gene expression involving four different mechanisms ( 18,19 ). First, they bind to a variety of direct repeats and directly compete for the binding of different transcription factors to these repeats. Second, they interact with RXR, thereby interfering with its binding to other receptors. Third, they can actively repress gene expression by interacting with corepressors. Fourth, NR2Fs can transrepress gene expression by interacting with the ligand binding domains of various nuclear hormone receptors. Our studies indicate that the repression of MTP expression by NR2F1 involves a novel mechanism that includes a combination of the fi rst and third mechanisms. In this mechanism, NR2F1 directly interacts with the DR1 element and recruits corepressors to cause gene silencing. NR2F1/NCOR1 repressed MTTP promoter activity to levels seen in the presence of HNF-4 ␣ alone in transfected cells ( Fig. 4E ) but did not completely silence it. It is known that NCOR1 interacts with histone deacetylase 3 (HDAC3) and this activity is essential for its repressor activity ( 20 ). Additionally, the NCOR1 complex contains chromatin remodeling enzymes and CpG methylation machinery ( 20,21 ). We do not know whether these enzymes are present in HEK cells used in this study. Therefore, one possibility is that absence of these enzymes might have avoided complete silencing of the MTP promoter activity. Alternatively, these enzymes may not be recruited to and may not silence a promoter activity that is present extrachromosomally, as role in NR2F1-mediated suppression of MTP expression. Additionally, direct binding of NR2F1 to the MTP promoter was studied by ChIP assays. Using these approaches, we show that NR2F1 directly binds to the DR1 element in the MTP promoter. Binding of NR2F1 to this element does not affect the individual activation potentials of HNF-4 ␣ and HNF-1 ␣ . However, it signifi cantly diminishes the synergistic activation of MTP promoter by HNF-4 ␣ /HNF-1 ␣ . Moreover, NR2F1 in the presence of the corepressor NCOR1 further suppresses the synergistic activity of HNF-4 ␣ /HNF-1 ␣ . We then extended these studies to Caco-2 cells and showed that the binding of NR2F1 and NCOR1 is high in undifferentiated Caco-2 cells. After differentiation, NR2F1 and NCOR1 did not interact with MTP, suggesting that MTP suppression in undifferentiated Caco-2 cells involves binding of NR2F1 to the DR1 element and recruitment of cosuppressors to the MTP promoter. To further substantiate that NR2F1/NCOR1 suppress MTP expression in undifferentiated Caco-2 cells, we reduced their expression using shRNA. Knockdown of NR2F1 and NCOR1 signifi cantly increased MTP expression in differentiated Caco-2 cells. These studies show that MTP expression is suppressed in undifferentiated cells due to the binding of NR2F1 to the DR1 element and recruitment of NCOR1 to this site.
The NR2F family consists of two homologous members, NR2F1 (also known as COUP-TF1 or EAR3) and NR2F2 (COUP-TF2 or ARP-1). Although they were identifi ed as activators, it is now generally believed that they act as with HNF-4 ␣ has been examined for the ␣ 1 -antitrypsin gene. In this case, HNF-1 ␤ does not substitute for HNF-1 ␣ in synergistically activating ␣ 1 -antitrypsin promoter with HNF-4 ␣ ( 24 ). The molecular bases for differences between MTP and ␣ 1 -antitrypsin promoters are unknown. However, it is worth noting that the positions of the HNF-1 ␣ and HNF-4 ␣ binding sites from the transcription start sites in the MTP and ␣ 1 -antitrypsin promoter are reversed. In the MTP promoter, the HNF-4 ␣ binding site is proximal to the transcription start site, whereas in the ␣ 1 -antitrypsin gene, the HNF-1 ␣ binding site is proximal to the transcription start site. The importance of the positional occupancy of HNF-1 ␣ and HNF-1 ␤ in the synergistic activation of HNF-4 ␣ remains to be determined.
MatInspector predicts that HNF-4 ␣ can bind to HNF-4 and DR1 elements. Promoter-reporter expression studies presented here ( Fig. 3 ) and by others ( 12 ) indicate that the DR1 site can be occupied by HNF-4 ␣ . However, our previous ( 6 ) and present ( Fig. 5 ) ChIP studies in undifferentiated and differentiated Caco-2 cells indicate that the MTTP promoter binds similar amounts of HNF-4 ␣ in both of these cells ( 6 ), indicating that the DR1 site may not be occupied by HNF-4 ␣ in differentiated cells. At this time, it is unknown whether any other transcription factor(s) binds to the DR1 element in differentiated Caco-2 cells. The DR1 element has been shown to be occupied by transcription activators such as RXR/RAR heterodimers in liver-derived cell lines ( 15 ). Our preliminary studies do not support the hypothesis that the DR1 element is occupied by RXR/PPAR in differentiated Caco-2 cells (data not shown). Hence, we speculate that a different, yetunidentifi ed activator/enhancer might bind to the DR1 element in differentiated Caco-2 cells to augment MTP expression during differentiation.
In short, this study provides a mechanistic explanation for the silencing of the MTTP gene in undifferentiated intestinal cells. We show that NR2F1 binds to the DR1 element in the MTTP promoter and abolishes the synergistic activation of the promoter by HNF-1 transcription factors and HNF-4 ␣ by recruiting NCOR1. We speculate that NCOR1 brings histone-modifying and chromatin-remodeling enzymes to silence the MTTP gene in undifferentiated intestinal cells.
in promoter-reporter plasmids. This is supported by the observation that the expression of promoter-reported constructs in undifferentiated Caco-2 cells results in significant expression of the promoter activity despite lack of endogenous gene expression. These studies suggest that endogenous gene silencing might involve recruitment of deacetylases and chromatin remodeling enzymes by NCOR1. In contrast, extra-chromosomal plasmids are not subjected to chromatin remodeling to achieve maximum suppression. We propose that repression of the MTTP gene by NR2F1 in undifferentiated Caco-2 cells might involve two mechanisms. First, NR2F1 binds to the DR1 element and abrogates synergistic activation by HNF-4 ␣ and HNF-1 ␣ . Second, it recruits corepressor complexes involving NCOR1/HDAC3 to deacetylate and remodel chromatin structure to silence the gene.
NCOR1 is now well recognized as transcriptional repressor of multiple genes. Recently, Doyon et al. ( 22 ) showed that NCOR1 protein levels are high in intestinal crypt cells and that its protein levels are low in the uppermost villus cells. We have previously ( 6 ) shown that MTP expression is low in crypt cells and high in villus cells. Thus, there is a reciprocal relationship between MTP expression and NCOR1 along the villus-crypt axis. This relationship is consistent with the idea that NCOR1 represses MTP expression. Doyon et al. ( 22 ) also showed that NCOR1 expression is high in proliferating Caco-2 cells and that its protein levels decrease during differentiation. Our observation that NCOR1 binds less to the MTTP promoter in differentiated cells might be related to decreases in its protein levels. Thus, reduced levels of NCOR1 might play a role in the expression of MTP during differentiation of Caco-2 cells and possibly during the differentiation of crypt cells into enterocytes.
It has been shown that HNF-1 ␣ synergistically activates HNF-4 ␣ in cell culture studies ( 12 ); however, MTP expression is not altered in HNF-1 ␣ knockout mice ( 13 ). This could either be due to the fact that HNF-1 ␣ is not involved in MTP expression or another transcription factor could substitute for its activity. To delineate between these two possibilities, we tested the hypothesis that HNF-1 ␤ might compensate for the loss of HNF-1 ␣ by cloning two different isoforms of HNF-1 ␤ . Individually, both the isoforms had very little activation potential for MTP promoter, similar to HNF-1 ␣ . Analogous to HNF-1 ␣ , both HNF-1 ␤ isoforms were able to synergistically activate MTP expression. These studies indicate that in the absence of HNF-1 ␣ , either of the two HNF-1 ␤ isoforms can bind to the HNF-1 cis element in the MTTP promoter and synergistically enhance HNF-4 ␣ activity and provide an explanation for the lack of effect of HNF-1 ␣ defi ciency on MTP expression in HNF-1 ␣ knockout mice.
Functional interactions between HNF-1 ␣ and HNF-4 ␣ have been described for several genes involving different mechanisms [( 23 ) and references therein]. In general, when HNF-1 ␣ and HNF-4 ␣ proteins interact with two different cis elements, they synergistically enhance promoter activity. The possibility that HNF-1 ␤ could substitute for HNF-1 ␣ for the synergistic activation of promoter activity