Pcyt2 (CTP:phosphoethanolamine cytidylyltransferase) is the rate-limiting enzyme in mammalian PE (phosphatidylethanolamine) biosynthesis. Previously, we reported that Pcyt2 mRNA levels increased in several types of cells after serum starvation, an effect that could be suppressed by supplementation with low-density lipoprotein or 25-HC (25-hydroxycholesterol). Transcription of Hmgcr, which encodes 3-hydroxy-3-methylglutaryl-CoA reductase, is also suppressed by 25-HC in the same dose-dependent manner. Nevertheless, a sterol-regulatory element was not detected in the Pcyt2 promoter region. The important element for transcriptional control of Pcyt2 by 25-HC (1.25 μM) was determined to reside between −56 and −36 on the basis of analysis with several Pcyt2 promoter deletion–luciferase reporters in NIH 3T3 cells. Using the yeast one-hybrid system, we found that NF-Y (nuclear factor-Y) binds at C−37CAAT−41 and YY1 (Yin Yang1) binds at C−42AT−40 in the Pcyt2 promoter. Endogenous NF-Y and YY1 bind clearly and competitively to these sites and are important for basal Pcyt2 transcription. Moreover, NF-Y binds to the Hmgcr promoter at C−14CA−12 in gel-shift analysis, and suppression of the basal luciferase activity of the Hmgcr promoter–reporter construct (−30/+61) by 25-HC was abolished when C−14CA−12 was mutated. Furthermore, transcriptional suppression of Pcyt2 by 25-HC was reduced following knockdown targeting of NF-YA or YY1. ChIP analysis revealed that 25-HC inhibited the interaction between NF-Y and RNA polymerase II on the Pcyt2 and Hmgcr promoters. On the basis of these results, we conclude that NF-Y and YY1 are important for the basal transcription of Pcyt2 and that NF-Y is involved in the inhibitory effects of 25-HC on Pcyt2 transcription.
Glycerophospholipids such as PE (phosphatidylethanolamine) and PC (phosphatidylcholine), as well as cholesterol, are major components of mammalian cell membranes, and the ratio of these lipids is almost constant. The distribution of these lipids in cell membranes is strictly restricted, with PE found mainly in the inner leaflet and PC in the outer leaflet . Cholesterol is found in both leaflets.
In the de novo biosynthesis of phospholipids, PE and PC are synthesized via the CDP-ethanolamine and CDP-choline pathways, which together are referred to as the Kennedy pathway. Each of these pathways involves a similar series of three sequential enzymatic reactions . First, ethanolamine/choline is phosphorylated by ethanolamine/choline kinase, producing phosphoethanolamine/phosphocholine. Secondly, Pcyt2 (CTP:phosphoethanolamine cytidylyltransferase) or Pcyt1 (CTP:phosphocholine cytidylyltransferase) catalyses the transfer of CTP to phosphoethanolamine or phosphocholine, generating CDP-ethanolamine or CDP-choline. This second step is the rate-limiting step in these pathways. Finally, ethanolamine/choline phosphotransferase transfers CDP-ethanolamine/CDP-choline to diacylglycerol, producing PE or PC. In mitochondria , and especially in the case of the absence of ethanolamine in vitro, the phosphatidylserine decarboxylation pathway predominantly contributes to PE biosynthesis . Thus, in normal cell and tissue contexts, the CDP-ethanolamine pathway is the major pathway of PE biosynthesis .
In mammalian tissues, four cDNA isoforms of Pcyt1 have been reported [5,6]. Transcription factors that are involved in cell growth and division are important for transcription of most active Pcyt1α [7,8]. We showed that the mRNA levels of Pcyt1a were not affected by treatment with LDL (low-density lipoprotein) or 25-HC (25-hydroxycholesterol) .
In humans, rats and mice, Pcyt2 is a cytosolic enzyme  and is encoded by a single gene, Pcyt2. Two alternatively spliced forms of Pcyt2 have been identified [11,12]. The human PCYT2 promoter is TATA-less and driven by a functional CAAT box, and is regulated by early growth response factor-1 and nuclear factor-κB . In C2C12 cells, C/EBP (CCAAT/enhancer-binding protein), Sp1 (specificity protein 1), Sp3 and MyoD are bound to the Pcyt2 promoter and enhance its transcription during muscle cell differentiation .
Cholesterol levels are controlled by a feedback system. The rate-limiting enzyme for cholesterol biosynthesis is HMGCR (3-hydroxy-3-methylglutaryl-CoA reductase), which is regulated at both the transcriptional and post-translational levels . Transcription of the Hmgcr gene is up-regulated when cellular cholesterol levels are low. Specifically, SCAP [SREBP (sterol-regulatory-element-binding protein) cleavage activation protein] binds to SREBPs and escorts them from the ER to the Golgi, and then active domains of SREBPs are released by two sequential cleavages. The released active domain stimulates Hmgcr transcription. By contrast, cholesterol inhibits release of SREBPs from ER . Cholesterol derivatives such as oxysterols also inhibit release of SREBPs mediated by the product of Insig (insulin-induced gene) .
Previously, we reported that the enhanced transcription of Pcyt2 in cells cultured in serum-starved medium can be suppressed by the addition of FBS to the culture medium. We identified LDL and oxysterols such as 25-HC as the substances responsible for this suppression, which is different from the regulation of Pcyt1a transcription. In a similar dose-dependent manner, LDL and 25-HC can also suppress Hmgcr transcription . These results suggest that oxysterols are important negative regulators for maintaining PE and cholesterol content in cell membranes by controlling the levels of Pcyt2 and Hmgcr mRNAs, which in turn affects levels of the Pcyt2 and HMGCR enzymes.
In the present study, we identified and characterized an element in the Pcyt2 promoter at position −56 to −36 that is regulated by 25-HC. NF-Y (nuclear factor Y) and YY1 (Yin Yang 1) transcription factors were identified as proteins that can bind to this element and regulate transcription. The data suggest that 25-HC suppresses the co-operative activities of NF-Y and YY1, interaction with RNA polymerase II at the Pcyt2 promoter and thus Pcyt2 transcription.
Matchmaker Yeast One-hybrid System for screening DNA binding proteins and the 11-day mouse embryo Matchmaker cDNA library in the pACT2 vector were purchased from Clontech. Saccharomyces cerevisiae strain BY5444 MATa was provided by the NBRP (National Bio-Resource Project) of the MEXT (Ministry of Education, Culture, Sports, Science and Technology) (Osaka, Japan). 3-AT (3-amino-1,2,4-triazole) was purchased from Sigma–Aldrich.
The promoterless Photinus pyralis luciferase vector pGL4.24, the control Renilla reniformis luciferase vector pRL-CMV and the Dual-Luciferase Reporter Assay System were obtained from Promega. We purchased Fugene 6™ transfection reagent and Lipofectamine™ 2000 from Roche Applied Science and Invitrogen respectively.
The NIH 3T3 and Hepa1 cell lines were supplied by the RIKEN Cell Bank (Tsukuba, Japan). High-glucose DMEM (Dulbecco's modified Eagle's medium) (Wako Chemical) and FBS (Invitrogen) were used for cell culture. Cholesterol, 25-HC, T0901317, GW3965 and 3-methylcholanthrene were obtained from Sigma–Aldrich, cycloheximide was purchased from Merck Calbiochem, and 24-HC (24-hydroxycholesterol), 27-HC (27-hydroxycholesterol) and 24(S),25-epoxycholesterol were obtained from Enzo Life Sciences.
Anti-V5 antibody was obtained from Invitrogen, and anti-YY1, anti-NF-YA, anti-NF-YC and rabbit IgG antibodies were purchased from Santa Cruz Biotechnology. Anti-RNA polymerase II antibody was obtained from Millipore.
Construction of plasmids and yeast strains containing the Eh sequence
As reported previously , the triple repeat of Eh (5′-TTGGCGCACGCCATTGGCTGCG-3′) (−53/−32) of the mouse Pcyt2 promoter, flanked by an EcoRI site at the 5′-end and by StuI and XbaI sites at the 3′-end (5′-GAATTC-TTGGCGCACGCCATTGGCTGCGTTGGCGCACGCCATTG-GCTGCGTTGGCGCACGCCATTGGCTGCGAGGCCTCTC-GAGTCTAGA-3′) (restriction sites are underlined), and its complementary strand (500 pmol each) were annealed. The annealed sequence (3Eh) was digested and inserted into the EcoRI and XbaI sites of the yeast pHIS vector (Clontech), resulting in pHIS-Eh. pHIS-Eh were transformed into the yeast host strain BY5444, and stable integrants with pHIS-Eh (YHIS-Eh yeast) were isolated. To obtain cDNAs encoding proteins that can bind to 3Eh, YHIS-Eh yeast were transformed with the pACT2 mouse embryo cDNA library and incubated in SD/−His/−Leu (synthetic dropout lacking histidine and leucine) selective medium containing 15–60 mM 3-AT.
Construction of luciferase reporter vectors
To construct the mouse Pcyt2 promoter–luciferase reporters, mouse genomic DNA was obtained from mouse blood using a DNA Extractor WB Kit (Wako). ET1-4 (−2362/+56) was prepared by PCR using a forward primer (ET1) with a HindIII site (5′-AAAAAGCTTGCCTTCCTTGTCTTTAGTGGACC-3′) and a reverse primer (ET4) with a HindIII site (5′-AAAAAGCTTCGCAAATCCTGGCGGCT-3′) (restriction site is underlined). The Pcyt2 promoter region (−2362/+56) was cloned into the HindIII site of pGL4.24. ET1-4 was digested with NheI and re-ligated to obtain a clone containing (−1395/+56) of the Pcyt2 promoter sequence (ET-Nhe). The vector series comprising ETe (−464/+56), ETd (−86/+56), ETf (−56/+56), ETg (−36/+56) and ETa (−15/+56) was obtained from ET1-4 by PCR using the following HindIII site (underlined)-containing forward primers, ETe (5′-AAAAAGCTTAAACATCTCTCAATTG-3′), ETd (5′-AAAAAGCTTGGGGGTGTGGCCCGGCTG-3′), ETf (5′-AAAAAGCTTTGGTTGGCGCACGCCATTG-3′), ETg (5′-AAAAAGCTTCTGCGCCAAGCGGGTGAG-3′) and ETa (5′-AAAAAGCTTCGCGTGAGCACGCGCAAG-3′), and a reverse primer (ET4).
Site-directed mutagenesis was performed using the QuikChange™ Site-Directed Mutagenesis Kit (Stratagene) as described previously . To introduce mutations into the XRE (xenobiotic-response element) (XREm; see Figure 1A), PCR was performed using the oligonucleotide primers 5′-GCCCTGGTTGGCGTTTGCCATTGGCTGCGC-3′ (mutated residues are underlined) and a complementary primer. The primer 5′-TTGGCGCACGCCATGTTCTGCGCCAAGCGG-3′ (mutated residues are underlined) was used to introduce a mutation in the CCAAT binding consensus site (CCAATm). To introduce mutations into both XRE- and CCAAT-binding consensus sites (XRE/CCAATdm), PCR was performed using 5′-GCCCTGGTTGGCGTTTGCCATGTTCTGCGC-3′ (mutated residues are underlined) and its complementary primer using ETd (CCAATm) as template. To introduce mutations into the NF-Y- and YY1-binding consensus site (NF-Y/YY1m) in ETe and ET-Nhe, PCR was performed using 5′-TTGGCGCACGCGGGTGGCTGCGCCAAGCGG-3′ (mutated residues are underlined) and its complementary primer. The sequences of the constructs were verified by direct DNA sequencing (ABI Prism 377-XL).
Proximal promoter of mouse Pcyt2 and effects of 25-HC on activities of Pcyt2 promoter–luciferase constructs
Mouse genomic DNA was used to construct the mouse Hmgcr promoter–luciferase reporters. HR (−1000/+61) was prepared by PCR using a forward primer (3F) containing an NheI site (underlined) (5′-AAAAGCTAGCGACAAGTCATAGACTATTAC-3′) and a reverse primer (R2) containing a HindIII site (underlined) (5′-AAAAAAGCTTAAGGAACTGCGCTTACGC-3′). The Hmgcr promoter region (−1000/+61) was cloned into the NheI and HindIII sites of pGL4.24. HR (−30/+61) was generated from HR (−1000/+61) by PCR using the following NheI site (underlined)-containing forward primer: 5′-AAAAGCTAGCGCAGCCGCCCGACGATTG-3′ with a reverse primer (R2). To mutate the CCAAT-binding consensus site (C−12CAAT−15) in HR (−1000/+61) and HR (−30/+61), PCR was performed with the oligonucleotide primer 5′-CTGAGCAGCCGCCCGACGATGTTCTAGGGG-3′ (mutated residues are underlined) and a complementary primer, resulting in the constructs HR (−1000/+61) CCAATm and HR (−30/+61) CCAATm (see Figure 6A).
Tissue culture and transfection
NIH 3T3 and Hepa1 cells were maintained in DMEM with 10% (v/v) FBS, then treated with trypsin/EDTA (Invitrogen) for cell dispersion.
For luciferase analysis, NIH 3T3 and Hepa1 cells (2.5×104) were plated on 24-well plates (BD) and grown overnight. Next, FuGENE 6™ was mixed with Pcyt2 or Hmgcr promoter–luciferase vector and the pRL-CMV Renilla vector. Alternatively, the Pcyt2 or Hmgcr promoter–luciferase vector and pRL-CMV Renilla vector were diluted with Opti-MEM (Invitrogen), mixed with diluted Lipofectamine™ 2000, and incubated for 20 min. Transfection was initiated by dropwise addition of the DNA suspension to cells followed by 12 h of incubation. Next, 25-HC (1.25 μM) was added to the culture medium and cells were collected 24 h later.
For RT (reverse transcription)–PCR analysis, NIH 3T3 cells (105) were dispensed into 35-mm-diameter plates (BD) and cultured overnight. To analyse the effect of LXR (liver X receptor) agonists T0901317 and GW3965, and cycloheximide or 3-methylcholanthrene on the transcription of Pcyt2, cells were cultured in serum-starved medium (0.5% FBS) for 48 h before adding one of the compounds with or without 25-HC (1.25 μM). Cells were collected 12 h later.
To analyse the effects of NF-YA or YY1 knockdown, 100 pmol or 250 pmol of Stealth Select RNAi (Invitrogen) targeting NF-YA [NF-YA-si1 (NfyaMSS207041), NF-YA-si2 (NfyaMSS207042) or NF-YA-si3 (NfyaMSS207043)], YY1 [YY1-si4 (YY1MSS279064) or YY1-si5 (YY1MSS279065)] or an siRNA control were suspended in 250 μl of Opti-MEM, mixed with diluted Lipofectamine™ 2000, and transfection was initiated by dropwise addition of the siRNA. For immunoblot analysis, after 36 h, total cell lysates were prepared in lysis buffer as described previously . For RT–PCR analysis, after 12 h, 1.25 μM 25-HC was added to the culture medium and cells were collected 24 h later. Total RNA was collected using RNeasy mini kits (Qiagen).
For electromobility gel-shift assays, 2×105 NIH 3T3 cells were dispensed into 60-mm-diameter plates and grown overnight.
For ChIP analysis, NIH 3T3 and Hepa1 cells (5×104) were dispensed into 100-mm-diameter dishes and grown overnight. Next, the cells were cultured in serum-starved medium (0.5% FBS) for 48 h before adding one of the compounds with or without 25-HC (1.25 μM). Cells were collected 8 h later.
Cells were harvested and lysed in 100 μl of passive lysis buffer (Promega). A 10 μl volume of the cell lysate was used in a dual-luciferase reporter assay and luciferase activity was measured using a BLR-301 luminescence reader (Aloka). The results were normalized to account for differences in transfection efficiency by calculating the ratio of the activity obtained with the Pcyt2 or Hmgcr promoter deletion and mutant constructs to that obtained with the pRL-CMV construct.
To quantify the mRNA levels of Pcyt2, Pcyt1a and Gapdh (glyceraldehyde-3-phosphate dehydrogenase) in NIH 3T3 cells, 0.7–2 μg of total RNA was reverse-transcribed at 37°C for 120 min using a high-capacity reverse transcription kit (Applied Biosystems). Reverse-transcribed samples were then subjected to 40 cycles of amplification using the FastStart Universal SYBR Green Master (ROX) for RT–PCR (Roche Applied Science) or THUNDERBIRD™ SYBR® qPCR (Toyobo) using a 7300 Real-Time PCR System (Applied Biosystems). Oligonucleotide primers used were 5′-CTCATCGTGGGTGTGCATACGG-3′ and 5′-CCTGTACCATCTTGTACCTCTC-3′ for mouse Pcyt2, 5′-GAATTCCTTCCAAAGTGCAG-3′ and 5′-ATAGGGCTTACTAAAGTCAAC-3′ for mouse Pcyt1a and MA050371-(F)(R) (Gapdh) (TaKaRa) for mouse Gapdh. TaqMan primer for mouse Pcyt2, Mm00470327_m1, was used to certify all results.
Preparation of nuclear extracts and electromobility gel-shift assays
Nuclear extracts were prepared using nuclear and cytoplasmic extraction reagents (Thermo Scientific). For the gel-shift assay for Pcyt2 promoter analysis, 500 pmol each of the following oligonucleotides and their reverse complement strands, pEh (−53/-24) (5′-TTGGCGCACGCCATTGGCTG-CGCCAAGCGG-3′), XREm (5′-TTGGCGTTTGCCATTGGC-TGCGCCAAGCGG-3′), CCAATm (5′-TTGGCGCACGCC-ATGTTCTGCGCCAAGCGG-3′), XRE/CCAATdm (5′-TTGG-CGTTTGCCATGTTCTGCGCCAAGCGG-3′) or NF-Y/YY1m (5′-TTGGCGCACGCGGGTGGCTGCGCCAAGCGG-3′), and for Hmgcr promoter analysis, pHr (5′-C−34TGAGCA-GCCGCCCGACGATTGGCTAGGGG−5-3′) or pHr (CCAATm) (5′-C−34TGAGCAGCCGCCCGACGATGTTCTAGGGG−5-3′) were annealed at 70°C for 10 min in 100 μl of 25 mM Tris/HCl (pH 8.0), 0.5 mM MgCl2 and 25 mM NaCl, and then cooled to room temperature (see Figures 1A and 6A). The double-stranded oligonucleotides (10 pmol) were 5′-end-labelled with [32P] ATP (PerkinElmer) and T4 polynucleotide kinase (TaKaRa), and purified using a Sephadex G-50 column (GE Healthcare). DNA–protein-binding reactions were performed for 30 min at room temperature in 40 μl of binding buffer (40 mM Tris/HCl, pH 7.9, 4 mM MgCl2, 2 mM EDTA, 100 mM NaCl, 2 mM DTT, 200 μg/ml BSA, 20% glycerol and 0.2% Nonidet P-40) containing 4 μg of poly(dI-dC) (Sigma), 1 μl of the radiolabelled probe (30000–80000 c.p.m.) and NIH 3T3 cell nuclear extract (2.5 μg). In some cases, anti-YY1, anti-NF-YA or anti-NF-YC antibody was included in the incubation mixture. BSA was added to normalize the protein content in each lane. Labelled probes were separated from DNA–protein complexes by electrophoresis on a 6% non-denaturing polyacrylamide gel in Tris/borate/EDTA buffer (44.5 mM Tris/HCl, pH 8.3, 44.5 mM boric acid and 1 mM EDTA) at 4°C. The gels were then dried and autoradiography was performed via exposure of the gel to an imaging plate for 30 min–24 h. The images were analysed using a Fuji BAS-2000 instrument.
Total cell lysates, prepared in lysis buffer as described previously , were subjected to immunoblot analysis. SDS/PAGE was performed using the Laemmli method, with 20 μg of protein separated in 10% (w/v) acrylamide. Proteins were transferred on to a nitrocellulose membrane using a semi-dry electroblotter, and the membrane was then treated for 2 h at 4°C with 5% (w/v) non-fat dried skimmed milk powder in 20 mM Tris/HCl (pH 8.0) containing 0.15 M NaCl. Next, the membrane was washed, incubated overnight with a 1:200 dilution of mouse anti-YY1 or anti-NF-YA antibody or 1:5000 dilution of anti-β-actin (Sigma), and washed extensively with 20 mM Tris/HCl (pH 8.0) containing 0.15 M NaCl. The membrane was then treated with a 1:5000 dilution of anti-mouse (for detection of YY1) or anti-rabbit (for detection of NF-YA or β-actin) IgG–horseradish peroxidase complex for 1 h. Immunoproteins were visualized using a ChemiDoc imaging system (Bio-Rad Laboratories) after exposure to the SuperSignal West Dura Extended Duration Substrate (Thermo Fisher Scientific).
The ChIP assay was performed following the protocol of the ChIP Assay Kit (Millipore) with some modifications. Briefly, formaldehyde was added to the cell culture medium to a final concentration of 1% before incubation for 10 min at 37°C. The cells were harvested and lysed on ice for 10 min, then sonicated to generate DNA fragments 200–500 bp in length. Pre-cleared chromatin was incubated with 2 μg of anti-RNA polymerase II or anti-NF-YA, or anti-rabbit IgG antibody as control, at 4°C overnight. Immune complexes were collected on 30 μl of Dynabeads–Protein G (Life Technologies). DNA–protein cross-linking was reversed by incubation with 0.4 M NaCl at 65°C for 4 h followed by proteinase K treatment. DNA was recovered by phenol/chloroform extraction and ethanol precipitation. PCR was carried out with 4 μl (eluate) or 2 μl (input) of sample using the following primers: mPcyt2-F (5′-CTGGGCGGAGGGGGGTGTGG-3′) and mPcyt2-R (5′-GCGGCTCCCGCGACACACAG-3′) for the −96/+44 fragment of the Pcyt2 promoter, mGapdh-F (5′-CTCTCTGCTCCTCCCTGTTC-3′) and mGapdh-R (5′-TCCCTAGACCCGTACAGTGC-3′) for the −21/+144 fragment of the Gapdh promoter, and mHmgcr-F (5′-CAGTGGGCGGTTGTTAGGG-3′) and mHmgcr-R (5′-AAGGAACTGCGCTTACGC-3′) for the −80/+61 fragment of the Hmgcr promoter. PCR was followed by analysis on 1% agarose gels.
All values are expressed as means±S.D. Group means were compared by using Student's t test after ANOVA to determine the significance of the differences between individual mean values. P<0.05 was considered statistically significant.
Analysis of the mouse Pcyt2 promoter
The transcriptional initiation site of Pcyt2 was determined previously . The proximal promoter of the mouse Pcyt2 gene is shown in Figure 1(A). Perhaps surprisingly, typical SREs (sterol-regulatory elements) and LXR-binding elements were not detected in the region from −1 to −2000 of Pcyt2, which would typically contain the promoter, using computational analysis, even though oxysterols dramatically suppress Pcyt2 mRNA levels in NIH 3T3 cells . To establish which element(s) in the Pcyt2 promoter is responsible for basal activity and suppression by oxysterols, we cloned a series of 5′-deleted promoters of Pcyt2 into the luciferase reporter vector pGL4.24.
As shown in Figure 1(B), the reporter activities did not change much in a set of sequential 5′-deletion constructs from −1395 bp (ET-Nhe) to −56 bp (ETf) when they were transfected into NIH 3T3 cells. As reported previously, the mRNA levels of Pcyt2 were clearly decreased when NIH 3T3 cells cultured in 10% FBS were incubated with 25-HC . We were therefore interested to determine the effects of 25-HC on the luciferase activities of Pcyt2 promoter–luciferase constructs. As expected, the reporter activities of these constructs were clearly suppressed by 25-HC, but not by cholesterol. These results suggest that the reporters recapitulate the effects of 25-HC on endogenous Pcyt2 mRNA levels. Moreover, the results suggest that element(s) in the Pcyt2 promoter responsible for the suppression of 25-HC treatment might be identified through further analysis using the Pcyt2 promoter–luciferase constructs. When the 5′-end 20 bp was removed from construct ETf (−56 bp), the luciferase activity of the resulting construct ETg (−36 bp) decreased to ∼30%. In addition, the negative response to 25-HC was reduced. Thus we reasoned that the 20-bp element from −37 to −56 is important for the response to 25-HC.
As computational analysis suggested that the promoter element that is responsive to 25-HC includes the XRE that is the recognition motif for binding of the AhR (aryl hydrocarbon receptor)/Arnt (aryl hydrocarbon receptor nuclear translocator), as well as CCAAT transcription factor-binding consensus sequences, we prepared a set of mutant promoter constructs in which we mutated XRE (XREm) (−47/−45), the CCAAT consensus sequence (CCAATm) (−39/−37) or both (XRE/CCAATdm) in the parent construct ETd (−86/+56) (Figure 1A). As shown in Figure 2(A), the promoter–luciferase activity of XREm did not significantly change compared with wild-type ETd. However, the reporter activity of CCAATm or the double-mutant construct XRE/CCAATdm was significantly decreased and did not respond well to 25-HC. Thus the normal Pcyt2 transcriptional response might reflect co-operation of two or more factors that bind to the Pcyt2 promoter, i.e. at −47/−45 and −39/−37.
Effects of mutations in the murine Pcyt2 promoter, LXR and AhR agonists, and cycloheximide on Pcyt2 transcription
As shown in Supplementary Figure S1(A), other oxysterols, such as 24-HC, 27-HC and 24(S),25-epoxycholesterol also clearly suppressed the luciferase activities of ETd transfected into NIH 3T3 cells. The luciferase activities of Hepa1 cells transfected with wild-type and mutated ETd showed a similar response to that of NIH 3T3 cells (Supplementary Figure S1B). We reported previously that these oxysterols also suppressed Pcyt2 mRNA levels in NIH 3T3 cells and 25-HC suppressed Pcyt2 mRNA levels in NIH 3T3, Hepa1 and HeLa cells . From these results, the suppression of Pcyt2 mRNA levels in cultured cells by oxysterols may be general conclusions.
Effects of LXR and AhR agonists, and cycloheximide on Pcyt2 mRNA levels in NIH 3T3 cells
Oxysterol is known to activate LXR. However, there does not appear to be a typical LXR-binding consensus in the Pcyt2 promoter region. To explore the potential role of LXR in Pcyt2 mRNA transcription, synthetic LXR agonists, T0901317 or GW3965, were added to the culture medium of NIH 3T3 cells after 48 h of serum-starvation. As shown in Figure 2(B), Pcyt2 mRNA levels were not changed by treatment with either chemically synthesized LXR agonist despite the fact that, as in previous studies, 25-HC clearly suppressed Pcyt2 mRNA levels. As shown in Supplementary Figure S1(C), a higher dose of T0901317 (20 μM) suppressed Pcyt2 mRNA levels as reported previously , but the dose is higher than in common usage.
We next determined the potential role of the putative XRE-binding consensus sites, which were detected computationally in the Pcyt2 promoter region, on Pcyt2 mRNA transcription. For this purpose, we treated cells with an AhR agonist, 3-methylcholanthrene, and tested the effects on Pcyt2 mRNA levels. However, Pcyt2 mRNA levels were not changed following 3-methylcholanthrene treatment (Figure 2C). Thus LXR and XRE may not be involved in the suppression of Pcyt2 transcription by 25-HC. Instead, these data suggest that another novel mechanism explains Pcyt2 mRNA suppression by 25-HC.
To explore this mechanism further, we next tested the effects of cycloheximide on suppression of Pcyt2 mRNA by 25-HC. As shown in Figure 2(D), pre-treatment with cycloheximide did not exert a significant effect on the suppression of Pcyt2 mRNA levels by 25-HC, although the basal mRNA level of Pcyt2 and the effect of 25-HC were reduced. These results suggest that Pcyt2 mRNA suppression by 25-HC is not predominantly mediated by newly produced proteins. Instead, 25-HC or its derivatives might be directly involved in suppression via an unknown mechanism.
Transcription factors that bind to the Pcyt2 promoter region and are important for response to 25-HC
To identify transcription factors that bind to the Pcyt2 promoter region between −56 and −36, we performed a gel-shift analysis using radiolabelled pEh oligonucleotides (−53/−24) as a probe and nuclear extracts from NIH 3T3 cells. As shown in Figure 3(A), faint doublet bands (indicated by arrowhead a) and a clear upper band (indicated by arrowhead b) were detected in lanes 2 and 3. To test which sequence(s) are important for the formation of DNA–protein complexes in pEh, three mutant versions of the probe, XREm, CCAATm and the double mutant XRE/CCAATdm, were prepared as shown in Figure 1(A). With the XREm probe, a clear upper band (arrowhead b) was detected, but faint lower bands were not observed (lanes 5 and 6). The intensity of the upper band was reduced with the CCAATm probe and two faint, but more discrete, lower bands (arrowhead a) were clearly detected (lanes 8 and 9). None of these bands was present with the XRE/CCAATdm probe (lanes 11 and 12). These results suggest that at least two proteins bind competitively to the pEh probe. The strong bands below the mobility region indicated by arrowhead a, which appear in all lanes, are artefacts.
Gel-shift analysis of wild-type and mutant pEh promoter fragments and NF-Y and YY1 in the presence of an NIH 3T3 cell nuclear extract
Yeast one-hybrid system identification of putative regulatory proteins, NF-Y and YY1, that bind the Eh element (−53/−32) of the Pcyt2 promoter
On the basis of the Pcyt2 promoter–luciferase reporter and gel-shift results, we reasoned that the region in the Pcyt2 promoter that was important for the response to 25-HC resided between −53 and −32. We therefore tried to identify transcription factors that bind to this site as shown in Figure 3(A) by using a yeast one-hybrid system. A yeast integrant (YHIS-Eh) carrying three repeats of Eh (22 bp) (−53/−32) upstream of the his3 coding region was constructed and used to screen for Eh-binding proteins and promoter activation using the one-hybrid expression system. The yeast transformation efficiency was 103 colony-forming units/μg as determined on SD/−Leu (synthetic dropout lacking leucine) selective medium.
After screening 1.3×106 colonies, we obtained 105 positive clones that were able to grow on SD/−His/−Leu selective medium containing 15–60 mM 3-AT. Several of the cDNAs obtained from positive yeast clones encoded transcription factors. These factors were identified as YY1, NF-YC, homeobox protein Hox-A11 and chromodomain-helicase DNA-binding protein1. A YY1-binding consensus region 5′-(C/g/a)(G/t)(C/t/a)CATN(T/a)(T/g/c)-3′  was found in the Eh promoter region (−45/−37); we therefore focused on YY1 and analysed its role in Pcyt2 transcription. The cDNA obtained in the yeast screen encoded only the C-terminus of YY1 (amino acids 156–414). The important domains in YY1 have been well characterized. A Zn2+ finger-type DNA-binding domain is located between amino acids 295 and 414 . In addition, a transcriptional activation domain is located in the N-terminus, and a transcriptional repression domain is located centrally.
A cDNA encoding full-length NF-YC was also obtained in the yeast one-hybrid screen. The CCAAT consensus core element (−37/−41) for NF-Y binding was also found in Eh. An overlapping YY1/NF-Y-binding site important for an enhancer-silencing effect has been reported previously . We therefore were also interested in the possible significance of NF-Y for Pcyt2 transcription.
DNA-binding properties of endogenous mammalian NF-Y and YY1 in NIH 3T3 cells
We next determined whether NF-YC might be the transcription factor responsible for the slow-migrating band in the above gel-shift assays (Figure 3A, arrowhead b). Since NF-Y complexes that include NF-YC are known to bind the CCAAT region, NF-YC was considered to be the most likely candidate for this transcription factor. For this analysis, we isolated nuclear extracts from NIH 3T3 cells, incubated them with or without an anti-NF-YC antibody and then performed gel-shift analysis using the intact pEh probe. When the nuclear extracts from the NIH 3T3 cells were incubated with the NF-YC-specific antibody, a supershift was detected (Figure 3B, lane 2, arrowhead d), and the lower band seen in the absence of the antibody (arrowhead b) disappeared. A supershift was also detected when nuclear extracts from NIH 3T3 cells were incubated with an NF-YA-specific antibody (Figure 3B, lane 3, arrowhead d). On the basis of these results, we concluded that the slow-migrating band indicated by arrowhead b corresponds to an NF-Y–DNA complex.
To examine whether or not endogenous YY1 can bind to the pEh promoter element, we performed gel-shift analysis using the intact pEh probe or the mutant probe CCAATm. Nuclear extracts from NIH 3T3 cells produced rapidly migrating doublet bands (arrowhead a in Figure 3C). Because these were faint doublet bands, a longer exposure to the imaging plate was necessary in order to be observed (Figure 3C, left-hand panel, lane 2). To confirm that the rapidly migrating band was an endogenous YY1–DNA complex, nuclear extracts from NIH 3T3 cells were incubated with YY1-specific antibody. As shown in Figure 3(C), lane 3, the faint band clearly disappears under these conditions, supporting the idea that endogenous YY1 can bind to the pEh probe. When the CCAATm probe was used, the intensity of the fast-migrating band increased (Figure 3C, lane 5), and a supershift (Figure 3C, arrowhead c) was also clearly detectable when the nuclear extract was incubated with an anti-YY1 antibody (Figure 3C, lane 6). The higher band (Figure 3C, arrowhead b) was not present when the CCAATm probe was used (Figures 3A and 3C). On the basis of these results, we concluded that the fast-migrating bands indicated by arrowhead a correspond to YY1–DNA complexes. These results also suggest that a protein that binds to C−37CAAT−41 might interfere with the binding of YY1 to −47/−45 in the pEh probe. Thus the YY1-binding site, in the region −47 and −37, could be an overlapping NF-Y/YY1-binding site important for the suppressive effect of 25-HC on Pcyt2 mRNA levels.
Importance of NF-Y and YY1 for suppression of Pcyt2 mRNA levels by 25-HC
We next looked at the effects of siRNA knockdown of NF-YA in NIH 3T3 cells. As shown in Figure 4(A), upper panel, protein levels of NF-YA were reduced following treatment with siRNA against NF-YA, NF-YA-si1 and NF-YA-si3. Although NF-YA knockdown was not complete, basal transcription and transcriptional suppression of Pcyt2 by 25-HC treatment was reduced following NF-YA knockdown via NF-YA-si3 (Figure 4B).
Effects of NF-YA or YY1 knockdown on Pcyt2 and Pcyt1a transcription
We next examined the expression of Pcyt2 in NIH 3T3 cells, as well as the effects of 25-HC on endogenous Pcyt2 mRNA levels, following siRNA knockdown of YY1. As shown in Figure 4(A), lower panel, the levels of the YY1 protein were clearly reduced following pre-treatment with YY1-si4 or YY1-si5. The expression of Pcyt2 mRNA decreased following YY1 knockdown via YY1-si5, and transcriptional suppression of Pcyt2 by 25-HC treatment was reduced in this context (Figure 4C). These results suggest that basal Pcyt2 transcription and its suppression by 25-HC are regulated co-operatively by both NF-Y and YY1, despite the fact that these factors can bind competitively to the Pcyt2 promoter. Expression of Pcyt1a did not change following siRNA knockdown of YY1 (Figure 4D).
As NF-Y and YY1 may bind competitively to the Pcyt2 promoter region between −47 and −37, we prepared probes in which the C−42AT−40 sequence was mutated (NF-Y/YY1m; Figure 1A). We reasoned that this core site might be important for both NF-Y and YY1 binding [20,22]. As expected, bands indicating YY1 and NF-Y protein–DNA binding disappeared completely when a probe in which the NF-Y/YY1-binding consensus site was mutated was used (Figure 5A).
Transcriptional regulation of Pcyt2 by NF-Y and YY1, and the effect of 25-HC by ChIP analysis
To test whether NF-Y and YY1 affect suppression of Pcyt2 mRNA levels by 25-HC, we performed reporter analysis with the luciferase reporters containing ETe or ET-Nhe Pyct2 promoter fragments (Figure 1B) and a version of ETe or ET-Nhe that was mutated at the NF-Y/YY1-binding consensus site (ETe (NF-Y/YY1m) or ET-Nhe (NF-Y/YY1m). As shown in Figure 5(B), ETe (NF-Y/YY1m) and ET-Nhe (NF-Y/YY1m) reporter activities were reduced compared with the non-mutated reporters. The suppressive effects of 25-HC on the luciferase activities of ETe and ET-Nhe were reduced when the NF-Y/YY1-binding consensus was mutated in both constructs. Thus binding of both NF-Y and YY1 appears to be important for basal transcription and for the suppressive effects of 25-HC on Pcyt2 transcription.
The NF-Y complex is composed of three subunits, NF-YA and NF-YB/NF-YC and is important for RNA polymerase II recruitment for initiation of transcription . ChIP analysis showed that RNA polymerase II and NF-Y clearly bind to the Pcyt2 promoter (Figures 5C and 5D), as expected from gel-shift analysis (Figure 3B). Addition of 25-HC to NIH 3T3 cells suppressed RNA polymerase II recruitment to the promoter; however, the binding of NF-Y to the Pcyt2 promoter and recruitment of RNA polymerase II and NF-Y to the Gapdh promoter were not affected. ChIP analysis with Hepa1 cells also showed similar results to those in NIH 3T3 cells (Supplementary Figure S1D). These results suggest that 25-HC may specifically inhibit the recruitment of RNA polymerase II to the NF-Y complex on the Pcyt2 promoter and thereby suppress Pcyt2 transcription.
Identification of the promoter element and binding factor important for the response of Hmgcr to 25-HC
25-HC can also suppress Hmgcr transcription, in a dose-dependent manner , similar to its effect on Pcyt2 transcription. We were interested in which transcriptional factors are important for Hmgcr transcription. The mouse proximal Hmgcr promoter sequence is shown in Figure 6(A) and includes an SRE consensus sequence at G−111TGCGGTG−104 , as well as NF-Y-  and YY1-  binding consensus sequences as reported previously.
Effects of 25-HC on activities of murine Hmgcr promoter deletion-luciferase constructs, NF-Y binding to the regulated promoter element, and the effect of 25-HC by ChIP analysis
As reported previously , the C−12CAAT −16 element is included in the Hmgcr proximal promoter, suggesting that the NF-Y complex might bind to this region. To identify transcription factors that bind to this region, we performed a gel-shift analysis using radiolabelled pHr oligonucleotide (−34/−5) as a probe. As shown in Figure 6(B), a clear band (indicated by arrowhead b) was detected in lane 2, and a corresponding supershift was clearly recognized (arrowhead d, lane 3) by incubation with the anti-NF-YC antibody. When the pHrCCAATm probe (C−12CAAT−16 mutated to A−12ACAT−16) was used for the analysis, the previously clear upper band disappeared. These results showed that NF-Y binds to the C−12CAAT−16 element in the Hmgcr promoter.
To establish which element(s) in the Hmgcr promoter is important for basal activity and response to oxysterols, we cloned 5′-deleted promoters of Hmgcr in the luciferase reporter vector pGL4.24 and compared their activities. As shown in Figure 6(C), reporter activities of constructs −1000 and −30 were clearly suppressed by 25-HC, although basal transcriptional activity of the −30 construct was reduced by only ∼30% as much as that observed for the other construct. To confirm the importance of NF-Y for Hmgcr transcription, the CCAATm mutation (C−12CAAT−16 to A−12ACAT−16) was introduced into the Hmgcr promoter–luciferase reporters, resulting in HR (−1000/+61) CCAATm and HR (−30/+61) CCAATm. When HR (−30/+61) CCAATm was transfected into NIH 3T3 cells, the basal reporter activity of HR (−30/+61) CCAATm was significantly decreased compared with wild-type, and the response to 25-HC was abolished (Figure 6C). On the basis of these results, we conclude that, as for Pcyt2 transcription, NF-Y is also important for both basal Hmgcr transcription and for the Hmgcr response to 25-HC.
ChIP analysis showed that RNA polymerase II and NF-Y clearly bind to the Hmgcr promoter (Figures 6D and 6E), as expected from gel-shift analysis (Figure 6B). Addition of 25-HC to NIH 3T3 cells suppressed recruitment of RNA polymerase II to the Hmgcr promoter, similar to its effect on the Pcyt2 promoter. However, the binding of NF-Y to the Hmgcr promoter and recruitment of RNA polymerase II and NF-Y to the Gapdh promoter were not affected. ChIP analysis with Hepa1 cells also showed similar results to those in NIH 3T3 cells (Supplementary Figure S1D). These results suggest that the NF-Y complex on the Pcyt2 and Hmgcr promoters may be specifically recognized by 25-HC, and that 25-HC inhibits the recruitment of RNA polymerase II to the NF-Y complex and thereby suppresses Pcyt2 and Hmgcr transcription. However, as reported previously , the SRE is very important for the Hmgcr transcriptional response to 25-HC.
In the present study, we identified a region in the Pcyt2 promoter, located between −56 and −36, that is important for transcriptional suppression by 25-HC (Figure 1B). Using the yeast one-hybrid system, we identified NF-YC and YY1 as proteins that possibly bind to this region, and we confirmed the binding of these factors to this promoter region by gel-shift analysis (Figure 3). NF-Y is a ubiquitous transcription factor that binds to the CCAAT box, which is present in 30% of eukaryotic promoters. NF-Y is composed of three subunits, NF-YA and NF-YB/NF-YC, all of which are necessary for binding to the CCAAT box. These three subunits are well conserved evolutionarily , and are important for RNA polymerase II recruitment for transcription initiation .
As shown in Figure 3, the NF-Y- and YY1-binding sites in the Pcyt2 promoter overlapped within −42/−37. The band density of NF-Y binding to this element in gel-shift assays is much higher than that of YY1 despite the evidence suggesting that these transcription factors bind competitively to this region. ETd reporter activities and the effect of 25-HC on these activities decreased when the CCAAT-binding consensus, which is thought to be important for NF-Y binding, was mutated. However, reporter activity was not changed if only the XRE binding sequence for YY1 was mutated (Figure 2A). On the basis of these results, we propose that NF-Y is more important than YY1 for basal Pcyt2 mRNA expression and transcriptional suppression in response to 25-HC. As shown in Figure 5(C), addition of 25-HC to NIH 3T3 cells suppressed recruitment of RNA polymerase II to the transcription-initiating NF-Y complex on the Pcyt2 promoter. However, 25-HC did not affect the binding of NF-Y to the Pcyt2 promoter or recruitment of RNA polymerase II to the NF-Y complex on the Gapdh promoter region. These results suggest that NF-Y complexes on specified promoters could be recognized by 25-HC or its metabolites by an unknown mechanism and that 25-HC may inhibit the recruitment of RNA polymerase II to the NF-Y complex and transcription initiation (Figure 7).
Schematic view of transcriptional suppression of Pcyt2 by 25-HC
YY1 was initially cloned and characterized in 1991 by two independent groups, Shi et al.  and Park and Atchison . YY1 behaves as both a transcriptional activator and a repressor. Indeed, the full name given to the protein, Yin Yang1, was meant to reflect its dual transcriptional activities. The results of early studies showed that E1A mediates activation of the adeno-associated virus P5 promoter, resulting from relief of YY1 repression via induction of the p300–YY1 complex. Overlapping YY1/NF-Y-binding sites in HOXB7 (homeobox B7) and embryonic β-type globin gene promoters important for the enhancer-silencing effect have been reported previously [30,31]. YY1 sites are also known to overlap and compete with serum-response factor sites in the c-fos serum-response element, skeletal and smooth muscle α-actin muscle regulatory elements, and the muscle creatine kinase CArG motif. In most cases, the overlapping sites have been shown to result in competition between the two factors for occupancy . Pcyt2 mRNA levels, as well as the inhibitory effects of 25-HC on Pcyt2 transcription, were decreased following NF-YA siRNA treatment (Figure 4B) and in YY1-knockdown cells (Figure 4C). On the basis of these results, we propose that NF-Y and YY1 co-operate to regulate Pcyt2 expression in response to 25-HC. Moreover, YY1 appears to be able to compensate for the effect of NF-Y on basal transcription and transcriptional suppression by 25-HC when NF-YA is depleted.
We were also interested in the transcriptional regulation of HMGCR by 25-HC because enhanced mRNA levels of both Pcyt2 and Hmgcr, which were induced by serum starvation in cultured cells, decreased following treatment with 25-HC in the same dose-dependent manner . As reported previously , an NF-Y-binding site (−12/−16) was detected by gel-shift analysis in the Hmgcr promoter near the SRE (−111/−104) (Figure 6B). SREBP transcription factors themselves are relatively weak activators of gene expression, and commonly require co-operation with the cofactors Sp1 and NF-Y . The reporter activities of both (−1000/+61) and (−30/+61) Hmgcr promoters were clearly decreased by mutation of the NF-Y-binding consensus sequence (−12/−16), and there was almost no response to 25-HC with the −30/+61 promoter (Figure 6C). These results suggest that, as for Pcyt2 transcription, NF-Y is also important for basal Hmgcr transcription and transcriptional suppression in response to 25-HC. As shown in Figure 6(D), recruitment of RNA polymerase II to the NF-Y complex on the Hmgcr promoter as well as on the Pcyt2 promoter was suppressed when NIH 3T3 cells were treated with 25-HC. NF-Y may be an important transcription factor for the response to 25-HC, which in turn is important for maintaining the relative amounts of PE and cholesterol in cellular membranes.
Cellular oxysterol levels might be regulated via incorporation of oxysterols from the outside of cells, especially from LDL , as well as via enzymatic production of oxysterols from cholesterol within cells [33,34]. It has been proposed that NF-Y facilitates transcription factor binding  and modulates transcription via histone acetylation because of its interaction with the histone acetyltransferase p300/CBP [CREB (cAMP-response-element-binding protein)-binding protein] and GCN5 (general control non-derepressible 5)/PCAF (p300/CBP-associated factor) . Thus it might be possible that 25-HC levels in cells directly regulate the activity of NF-Y by regulating its interaction with general transcription factors and regulators. By screening using Genome Browser (http://genome.ucsc.edu), hypermethylated histone (H3K4m3) at the Pcyt2 promoter region in several cell lines was reported, but histone acetylation was not.
PE is a major component of mammalian cell membranes as are PC and cholesterol, and can be metabolized to PC by PE methyltransferase in liver. Therefore understanding the machinery of controlling PE biosynthesis by Pcyt2 could be related to most phospholipid metabolisms in cells and organisms. Our combined data show that NF-Y can bind to the Pcyt2 and Hmgcr promoters and maintain basal transcription of these genes by recruitment of RNA polymerase II. Moreover, increased levels of 25-HC in cells may suppress the normal transcriptional enhancing effect of NF-Y and YY1, by suppressing the recruitment of RNA polymerase II to the NF-Y complex, thus repressing transcription of Pcyt2 and Hmgcr. These results suggest that oxysterols are important for major lipid homoeostasis in cell membranes. The precise mechanisms by which 25-HC exerts an effect on these transcription factors is unclear and merits further study.
The project strategy was devised by, and most of the experiments were performed by, Hiromi Ando and Hiroyuki Sugimoto. Kohei Hosaka also advised on the project strategy. Constructions of plasmids were partly performed by Chieko Aoyama, Yasuhiro Horibata and Satomi Mitsuhashi. siRNA experiments were partly performed by Motoyasu Satou and Masahiko Itoh.
We thank Dr Takashi Namatame of the Clinical Research Center for DNA sequencing.
This work was supported by the Japan Society for the Promotion of Science (JSPS) [KAKENHI grant number 24590358 and 15K08284], and Promotion and Mutual Aid Corporation for Private Schools of Japan.
aryl hydrocarbon receptor
CREB (cAMP-response-element-binding protein)-binding protein
Dulbecco's modified Eagle's medium
liver X receptor
nuclear factor Y
synthetic dropout lacking histidine and leucine
Yin Yang 1