The activity of the carbamoyl phosphate synthase 1 promoter in human liver‐derived cells is dependent on hepatocyte nuclear factor 3‐beta

Abstract Carbamoyl phosphate synthase 1 (CPS1) is the rate‐limiting enzyme in the first step of the urea cycle and an indispensable enzyme in the metabolism of human liver. However, CPS1 epigenetic regulation involves promoter analysis and the role of liver‐enriched transcription factors (LETFs), which is not fully elucidated. In this work, the promoter region of hCPS1 gene was cloned, and its activity was investigated. An LETF, hepatocyte nuclear factor 3‐beta (HNF3β), was found to promote the transcriptional expression of CPS1 in liver‐derived cell lines. In addition, dual‐luciferase reporter assay shows that the essential binding sites of the HNF3β may exist in the oligonucleotide −70 nt to +73 nt. Two putative binding sites are available for HNF3β. Mutation analysis results show that the binding site 2 of HNF3β was effective, and the transcriptional activity of CPS1 promoter significantly decreased after mutation. Electrophoretic mobile shift assay (EMSA) and ChIP assay confirmed that HNF3β can interact with the binding site in the CPS1 promoter region of −70 nt to +73 nt promoter region in vivo and in vitro to regulate the transcription of CPS1. Moreover, HNF3β overexpression enhanced the transcription of CPS1 and consequently improved the mRNA and protein levels of CPS1, whereas the knockdown of HNF3β showed the opposite effects. Finally, urea production in cells was measured, and ammonia detoxification improved significantly in cells after transfection with HNF3β. HNF3β plays a vital role in regulation of CPS1 gene and could promote the metabolism of ammonia by regulating CPS1 expression.


Introduction
CPS1 is a liver-specific, intramitochondrial, indispensable enzyme in the metabolism of human liver. CPS1 is the rate-limiting enzyme in the first step of the catalytic cycle of urea cycle. Previous research showed that CPS1 is at low expression levels in 75% of hepatocellular carcinoma (HCC) [1]. Gene mutation and epigenetic regulation may lead to abnormal gene regulation function, and numerous types of cancer cells present abnormal epigenetic patterns. Hypomethylation is a prerequisite for the activation of the DNA proximal promoter region, whereas hypermethylation leads to the silencing of genes [2]. Liu [3] speculated that the low CPS1 gene expression in HCC may be related to DNA methylation; in addition, Liu proved that the CpG island hypermethylation of CPS1 gene is a key mechanism for the low CPS1 enzyme expression in the liver cell carcinoma, and two CpG island hypermethylation sites were found in the vicinity of the promoter region. SIRT5, an NAD-dependent protein deacetylase, plays a pivotal role in the treatment and detoxification of ammonia by activating CPS1 expression [4].
CPS1 is particularly expressed in liver mitochondria and small intestine epithelial cells. CPS1 mutation can cause CPS1 deficiency (CPS1D), which causes hyperammonemia, neonatal death or mental retardation. CPS1D-associated mutations affect the function of a non-catalytic 20 kD C-terminal domain [5]. The domain is the allosteric region and combined with N-acetyl glutamate [6]; therefore, mutations disrupt the allosteric activation of N-acetyl glutamate on CPS1.
A recent study revealed that the ammonia removal of HepaRG cells through the urea cycle is limited. The low ammonia removal is related to the low expression of the urea cycle enzymes Arg1, OTC, and CPS1 (DMSO presence) [7]. Gene regulation of the hepatocyte lineage is mainly at the transcriptional level, and all kinds of LETFs or heterologous nuclear receptors play a vital role [8]. However, promoter studies on the low expression of key enzymes in the liver cell urea metabolic pathway are few. CPS1 epigenetic regulation involves promoter analysis, and the role of LETFs in the regulation is not yet fully elucidated. The promoter region of CPS1 gene was cloned, and the key transcription factor, HNF3b, was determined. Furthermore, we found an important HNF3b binding site (À21 nt to À10 nt, relative to the translation initiation site). Our study proved that the HNF3b performs a vital function in the regulation of CPS1 expression.

CPS1 promoter luciferase constructs
Genomic DNA from normal blood cells was extracted by use of Genomic DNA Purification Kit (Promega, Madison, WI, USA) and used as a template for polymerase chain reaction amplification. The plasmid pGL4-2086 with the CPS1 promoter driving firefly luciferase was constructed by ligation of the PCR-generated full-length CPS1 promoter (nucleotides À2086 to +73, relative to the translation initiation site) into the XhoI and HindIII (Thermo Scientific, Waltham, MA, USA) cleaved sites of the luciferase reporter plasmid pGL4-Basic (Promega). Various CPS1 promoter deletion constructs, which included pGL4-1408 (nucleotides À1408 to +73), pGL4-825 (nucleotides À825 to +73), pGL4-482 (nucleotides À482 to +73), pGL4-212 (nucleotides À212 to +73) and pGL4-70 (nucleotides À70 to +73), were also made by insertion of the corresponding PCR-generated fragment into the XhoΙ and HindIII sites of the pGL4-Basic plasmid. The primers used for amplification are shown in Table 1. The reverse primer À2086/+73R was used together with each of the forward primers.
The pGL4-70 plasmid was used as a template for construction of HNF3b binding sites mutants. The genetailor site-directed mutagenesis kit (Invitrogen, Carlsbad, CA, USA) was used following the manufacturer's instructions. The primers used for amplification are shown in Table 1. The resulting plasmids were pGL4-70-HNF3bmut1 and pGL4-70-HNF3bmut2. All constructs were confirmed by DNA sequencing.

Dual-luciferase reporter assay
Cells were transfected with the CPS1 promoter luciferase reporter constructs. The renilla luciferase expression vector pRL-TK (Promega) was used for normalization, and the promoterless vector pGL4-Basic served as the negative control. Cells were lysed 48 hrs after transfection. The Dual-Luciferase Reporter Assay System (Promega) was used to detect intracellular luciferase activity in 40 ll of cell lysates, following the manufacturers' recommendations. Luminescence measurement was carried out on an illuminometer (Orion II Microplate Luminometer, Berthold Detection Systems, Germany). Each transfection was performed in triplicate.

Western blotting
Protein (30 lg) was subjected to 10% sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE), and electrophoretic transfer to a polyvinylidene fluoride (PVDF) membrane (Merck Millipore, Darmstadt, Germany). Protein blots were incubated separately with a panel of specific antibodies that included anti-CPS1 (ab155083; Abcam, Cambridge, UK), anti-HNF3b (ab108422; Abcam) and anti-b-actin (sc-47778; Santa Cruz Biotechnology, Santa Cruz, CA, USA) antibodies at 1:1000 dilution. Immunoreactivity was detected using a chemiluminescence Western blot immunodetection kit (Invitrogen) according to the manufacturer's instructions and recorded on Hyperfine-ECL detection film. The amounts of each protein were semi-quantified as ratios to b-actin indicated on each gel.

RNA extraction and real-time RT-PCR analysis
Total RNA was extracted from HepG2 cells with TRIzol Reagent (Invitrogen). The reverse transcription reaction was carried out with 3 lg of RNA in a final volume of 20 ll by use of the Transcriptor First Strand cDNA Synthesis Kit (Roche Diagnostics, GmbH, Mannheim, Germany). Quantitative real-time PCR was performed with the ABI StepOne Real-Time PCR System (Applied Biosystems, Foster City, CA, USA) and the Fast Start Universal SYBR Green Master Mix (Roche). Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) served as an internal control. The primers used for GAPDH and CPS1 amplification are shown in Table 1. Each sample was analysed in triplicate. The relative expression level of CPS1 was calculated by normalization to the endogenous GAPDH mRNA expression prior to comparative analysis by use of the 2 ÀDDCt method as previously described [9]. All procedures followed the manufacturer's instructions.

Nuclear extraction and EMSA
The 5 0 -biotin end-labelled oligonucleotides that corresponded with the HNF3b transcription factor recognition sequences within the CPS1 promoter region were synthesized (Viagene, Jiangsu, China) and used as probes. Unlabelled oligonucleotides, either wild-type (cold probes) or   Table 1. A Nuclear and Cytoplasmic Protein Extraction Kit (Beyotime, Jiangsu, China) was used to prepare nuclear proteins from HepG2 cells. An EMSA Gel Shift Kit (Viagene) was used to examine the interaction between HNF3b in the nuclear extract with the probe, in accordance with the manufacturer's instructions. Briefly, 20 lg of nuclear extracts were incubated with 0.5 lg of poly (dI-dC) and 13.8 nM labelled probe in a final volume of 18 ll. The competition assay was performed using 50-to 100-fold molar excess of cold probes or cold mutated probes preincubated with the reaction mixture before addition of biotin-labelled probes. In supershift assays, 4 lg of the specific antibodies (anti-HNF3b) was added to the mixtures of nuclear extracts and DNA probes (In here, the quality of poly (dI-dC) we use is 0.8 lg). The DNA-protein complexes were incubated at 25°C for 20 min., then subjected to electrophoresis on a 6.0% non-denaturing pre-made polyacrylamide gel (Invitrogen). Next, the complexes were transferred to a nylon Binding-membrane (Viagene) and fixed nearly 10 cm below the UV lamp for 30 min. cross-linking. The biotin endlabelled DNA was detected by addition of streptavidin-horseradish peroxidase conjugate and chemiluminescent substrate.

Chromatin immunoprecipitation assay (ChIP)
HepG2 cells (10 7 ) were cultured and cross-linked by adding formaldehyde to a final concentration of 1% at room temperature for 10 min. The crosslinking was stopped by addition of 125 mM glycine at room temperature for 5 min. Cells were rinsed twice and collected with ice-cold phosphatebuffered saline (PBS), then resuspended in 0.75 ml FA Lysis Buffer (Abcam, Cambridge, MA, USA) containing a protease inhibitor cocktail (Roche) and incubated on ice for 10 min. Cell lysates were sonicated for 20 min. with a Branson model 350 Sonifier (Branson Sonic Power Co., Danbury, CT, USA) at 20% duty cycle. This conditions yielded DNA fragments ranging from 200 to 500 base pair (bp) in size. Before immunoprecipitation, 50 ll of each lysate was removed for analysis of input chromatin DNA. Immuoprecipitation was conducted with 5 lg of specific antibodies (anti-HNF3b) followed by 80 ll of pre-blocked protein A/G. The normal rabbit IgG or pre-blocked protein A/G (no antibody control) was also performed for control purpose. Following elution and proteinase K digestion, DNA was recovered by phenol/chloroform extraction and ethanol precipitation in the presence of glycogen and dissolved in 100 ll distilled sterile water. Bound target DNA fractions were analysed by PCR. The primers used are shown in Table 1, which amplified a region that spanned nucleotides À70 to +73 (containing the HNF3b binding sites) of the CPS1 promoter. The resulting PCR products were electrophoresed on 2% agarose gels and stained with ethidium bromide.

RNA interference assay
The small interfering RNA (siRNA) targets for HNF3b (5 0 -GCCGUCCGA-CUGGAGCAGCUACUAUdTdT-3 0 ) were described by Tomaru [10]. All siRNA duplexes, which included a negative control that had no homology with known human genes, were synthesized chemically by the Gen-ePharma Company (Shanghai, China); 100 pmol of HNF3b-siRNA or the negative control was used for transfection. Total RNA and proteins were isolated at 48 hrs after transfection in accordance with the manufacturer's instructions.
Measurement of Urea production in cells 1 9 10 5 of cells were resuspended in 100 ll growth medium and added to a 96-well culture plate. Cells were transfected transiently with pFLAG-HNF3b and pFLAG-CMV-2 (control), respectively, when the cell density was about 80% and incubated at 37°C for 12 hrs, then the supernatant was replaced by 100 ll growth medium or STD buffer with 0-40 mM NH 4 Cl. After another 12 hrs, 50 ll supernatant was used for the detection assay. Twenty-five microlitres of 1:3 sulphuric acid/phosphoric acid (by volume) was added, and urea production was determined by adding 25 ll a-isonitrosopropiophenone (9% in absolute ethanol), incubating at 100°C in the dark for 15 min., and measuring absorbance at 490 nm. Meanwhile, the standard curve and regression equation for detection were produced using a series of urea standards with different concentration, which were used to calculate the urea concentration in the samples [15].

Statistical analysis
Statistical analyses were performed with SPSS software (SPSS13.0 version for windows) (SPSS Inc., Chicago, IL, USA) by use of analysis of variance (ANOVA) and t-test. All values were expressed as means AE standard deviation (SD) from duplicate experiments. A P-value <0.05 was considered to be statistically significant.
Next, the full length and a series of 5 0 -deletion constructs of the CPS1 gene promoter (pGL4-1408, pGL4-825, pGL4-482, pGL4-212 and pGL4-70) were transfected transiently into BEL-7404 cells. The plasmid pGL4-212 show full maximum luciferase activity relative to the full promoter region (Fig. 1C), suggesting that nucleotides À212 nt to +73 nt is crucial for the full activity of the CPS1 promoter. In addition, all recombinant promoter fragments were, respectively, cotransfected with HNF3b, demonstrating that the fluorescence activity of pGL4-70 was significantly higher than that of transient transfection after cotransfection with HNF3b (Fig. 1D). The experimental data indicate that the essential binding sites of the HNF3b may exist in the oligonucleotide À70 nt to +73 nt; therefore, the recombinant plasmid pGL4-70 was used for further studies.
Identification of HNF3b binding sites in À70 nt to +73 nt region of the CPS1 promoter TFSEARCH, TESS and TFBIND software were used to analyse the potential HNF3b binding sites in the nucleotides À70 nt to +73 nt, and a comprehensive evaluation of the optimal two binding sites 1 and 2 ( Fig. 2A) were selected for subsequent experiments. Then, sitedirected mutants of the putative HNF3b binding sites were generated and named pGL4-70-HNF3bmut1 and pGL4-70-HNF3bmut2, and the promoter activities of the corresponding constructs were measured. As shown in Figure 2B, mutations in HNF3b binding site 2 (pGL4-70-HNF3bmut2) led to a reduction in the promoter activity of approximately 52% to 57% compared with the control non-mutated construct (pGL4-70), not HNF3b binding site 1 (pGL4-70-HNF3bmut1). The results show that the binding site 2 of the transcription factor HNF3b was effective, and the transcriptional activity of the CPS1 promoter was significantly decreased after mutation. The CPS1 promoter displays the highest activities in liver-derived cell lines; therefore, we speculate that the promoter may be closely related to the high protein levels of endogenous HNF3b. As shown in Figure 2C, the HNF3b and CPS1 protein were measured in various cell lines. The protein levels of HNF3b in liver-derived cell lines HepG2, Huh7 and BEL-7404 were higher than those in the other cell lines. However, CPS1 protein was not detectable in HepG2 and Huh7 cells. This result is consistent with the study of Butler [16] and conforms to the view of Siddiqui [1], who demonstrated that the low level of CPS1 is expressed in 75% of human HCC tissues.
The role of HNF3b binding site was further verified by EMSA. The results demonstrate that the nuclear extract could 'shift' the position of the biotin-labelled probe because of a change in electrophoretic mobility caused by dsDNA/protein complex formation (Fig. 2D, lane  2); moreover, the DNA-transcription factor complexes could be 'super-shifted' by the addition of antibodies directed against human HNF3b (Fig. 2D, lane 6). In addition, a competition assay showed that preincubation with 50-or 100-fold molar excess of cold probe Fig. 2 Binding sites of HNF3b in the CPS1 promoter. (A), Nucleotide sequence of À70 nt to +73 nt region of CPS1 promoter is shown. The translation initiation site (+1) is indicated by the arrow. The putative HNF3b binding sites in the nucleotide region À38 nt~À27 nt (site 1) and À21 nt~À10 nt (site 2) are underlined. In the binding site 1, TT is mutation into AC, named mut1. Similarly, TA is mutation into AG named mut2 in the binding site 2. (B), Effect of mutation of the HNF3b binding sites on the activity of the CPS1 promoter. HNF3b binding sites were subjected to site-directed mutagenesis. A total of 1 lg of mutants together with 100 ng of pRL-TK was cotransfected into HepG2 cells. pGL4Basic served as the negative control. Luciferase activities were measured 48 hrs after transfection. The relative luciferase units (RLU) were obtained (right) by comparison with the wild-type of pGL4-70, which was set to 1. The total extracted DNA (Input DNA, 5%) prior to immunoprecipitation and the immunoprecipitated samples were PCR-amplified using primers specific to a region that spanned À70 nt to +73 nt (containing the HNF3b binding sites) of the CPS1 promoter. The normal rabbit IgG or no antibody control was also performed for control purpose. Band signals were quantified using the densitometric software, and the relative intensities to the input which was set to 1.00 were calculated. ND: None detected (*P < 0.05).  (Fig. 2D, lane 3 and lane 5), but not the cold mutated probes (Fig. 2D,  lane 4), evidently diminished the intensity of the bands.
To determine whether the HNF3b is associated with CPS1 promoter in vivo, we performed ChIP assay with specific antibody and polymerase chain reaction (PCR) using the primers against the regulatory elements of the CPS1 promoter. The results show that a 143bp DNA fragment covering HNF3b was amplified by chromatin immunoprecipitation with an anti-HNF3b antibody (Fig. 2E). The same band was obtained with the input DNA, whereas the normal IgG control and no antibody control did not result in immunoprecipitation of DNA fragments detectable by PCR amplification.
Taken together, our results confirm that HNF3b can interact with the binding site in the CPS1 promoter region of À70 nt to +73 nt promoter region in vivo and in vitro to regulate the transcriptional regulation of the CPS1 promoter.

Overexpression of HNF3b and interference of HNF3b siRNA
To elucidate the transcriptional regulation of HNF3b on the CPS1 promoter, we cotransfected pGL4-70 with different doses of pFLAG-HNF3b and HNF3b siRNA in HepG2 and BEL-7404 cells. As shown in Fig. 3A and 3B, the fluorescence activity of CPS1 promoter increased with pFLAG-HNF3b overexpression, which was dependent on HNF3b concentration (Fig. 3A). Meanwhile, the fluorescence intensity of cotransfected HNF3b siRNA was significantly lower than that of the negative control (Fig. 3B). In addition, the levels of CPS1 in protein (BEL-7404 cell) and mRNA (HepG2 cell) also show the same trend ( Fig. 3C-3F). The results show that the overexpression of pFLAG-HNF3b promotes the transcription of the CPS1 promoter and then increases the level of CPS1 mRNA (Fig. 3E), ultimately resulting in the enhancement of CPS1 protein level (Fig. 3C). On the contrary, the interference of HNF3b siRNA reduced the transcription of the CPS1 promoter and inhibited the expression of CPS1 mRNA (Fig. 3F), and the expression of CPS1 protein level also decreased (Fig. 3D).

Comparison of ammonia detoxification by cells
The urea cycle is one of the metabolic pathways of ammonia. Thus, the production of urea is used to evaluate the capacity for ammonia detoxification in cells. Urea synthesis in HepG2 and BEL-7404 cells is shown in Fig. 4A and 4B. Urea products increased in different degrees with increasing ammonia concentration through transfection with pFLAG-HNF3b, although urea production in cells may be reduced after transient transfection with pFLAG-CMV-2 (control). The transcription factor HNF3b could promote the metabolism of ammonia by regulating the expression of CPS1.

Discussion
We demonstrated the relationship between LETFs and CPS1 expression in liver-derived cell lines. Our data showed that HNF3b played an important role in the regulation of CPS1 gene, and a potent HNF3b binding site in CPS1 promoter was found.
The transcription of gene is tissue-specific and is mainly controlled by the interaction between transcription factors and their corresponding cis-acting elements [17,18]. After the online software prediction of the existence of LETF binding sites in À2086 nt to +73 nt, we selected the HNF1a, HNF3b, HNF4a, HNF6, C/EBPa and C/EBPb as the research object. Our study shows that the activity of CPS1 significantly enhanced after cotransfection with HNF3b, whereas cotransfection with C/EBPa and C/EBPb does not enhance transcription. Our findings are not consistent with the study of Howell [19] and Lagace [20] that multiple C/EBP binding sequences were found in the promoter of rat CPS1. A minimum of two possibilities could explain this paradox. First, other transcription factors may do the opposite of C/EBP in the human CPS1 promoter. However, the more plausible explanation is that despite the function of the C/EBP binding site in the rat CPS1 promoter, whether this function is true in humans was unknown. We performed luciferase activity on all truncated fragments, and transient cotransfection with C/EBPb was not enhanced (data not shown). The exact mechanism needs further study. In addition, the À212 nt to +73 nt sequence retains full maximum luciferase activity of the CPS1 promoter, thereby indicating that the region is the core promoter region of the CPS1 gene. Moreover, we identified two potential HNF3b binding sites (À38 nt to À27 nt and À21 nt to À10 nt) in the À70 nt to +73 nt region. However, due to the limitations of the prediction software, HNF3b binding sites are not shown in À212 nt to À70 nt. We speculate that other important transcription factors may be involved.
The important role that HNF3b plays in the regulation of CPS1 expression is supported by the following experiments. First, sitedirected mutagenesis was used to construct the mutant recombinant plasmid according to the prediction of the binding site, and binding site 2 was found to exert a definite effect on the transcription of CPS1. Second, probes were synthesized on the basis of the binding site to carry out EMSA and ChIP assay by use of HNF3b specific  antibody, confirming that HNF3b can interact with the CPS1 promoter in vivo and in vitro. In addition, the overexpression of HNF3b and interference of HNF3b-siRNA strengthened the conclusion that the transcription of CPS1 gene is regulated by HNF3b. The overexpression of HNF3b promotes the transcription of CPS1 and then increases the level of CPS1 mRNA and protein expression. Conversely, the expression levels of CPS1 mRNA and protein are significantly reduced after HNF3b interference. Finally, urea production in HepG2 and HepG2/HNF3b cells was measured, and ammonia detoxification was improved more significantly in HepG2/HNF3b cells compared with HepG2 cells.
The HNF3 gene family consists of HNF3a, HNF3b and HNF3c, and the family structure is characterized by a helix/fork DNA binding region containing approximately 110 amino acids. This area is required for nuclear localization and transcriptional activation, which can identify the same DNA sequence. The area shows a high degree of continuity in the evolution and existence of their family members from yeast to humans [21]. The HNF3 gene family is involved in important metabolic and detoxification processes. HNF3a functions in regulating the activation of glucagon. The expressions of tyrosine aminotransferase and phosphoenolpyruvate carboxykinase were significantly reduced in the liver cells of HNF3c gene deletion, and the serum glucose level was decreased, as well [22]. The excessive expression of HNF3b may lead to the decrease in glycogen synthesis ability [23]. Therefore, the HNF3 binding site in the promoter of CPS1 was associated with the regulation of serum glucose. CPS1 gene expression was closely related to the glucocorticoid and glucagon. The CPS1 distal enhancer element (GRU [24,25]), proximal enhancer element (GAGA-box [26], GRE [27]) and the core promoter region are indispensable for the glucocorticoid cascade response element. CREB stimulates the transcription of CPS1 and ASS based on glucagon signals [28,29]. The transcription of partial enzymes in the urea cycle is also regulated by glucocorticoid and the glucagon signalling pathway, but the regulation mechanism is different [30][31][32].
In summary, CPS1 gene is regulated by the transcriptional regulation of the LETF HNF3b. CPS1 gene expression is closely linked with numerous kinds of liver diseases. We hope that the research on the CPS1 transcription presents a certain guiding significance for the treatment of disease. Odom [33] proposed that a large number of liver-specific transcription factors binding to the promoter indicated a high possibility of expression of the gene in liver. Moreover, a positive correlation was shown between the number of transcription factors binding to the promoter and the level of gene expression. The regulation of liver-specific gene is more likely to be accomplished by a number of LETFs and heterologous nuclear receptors, and a single LETF or heterologous nuclear receptor cannot correct the complex intracellular regulation. Therefore, further studies are needed to elucidate other potential trans-acting factors interacting with CPS1 gene, as well as the regulation mechanism of their common components. Finally, this work will provide appropriate intervention targets for our study on constructing novel liver cells with a better solution to address ammonia toxicity.