Regulation of Lipocalin-type Prostaglandin D Synthase Gene Expression by Hes-1 through E-box and Interleukin-1β via Two NF-κB Elements in Rat Leptomeningeal Cells

The promoter function of the rat lipocalin-type prostaglandin D synthase (L-PGDS) gene was characterized in primary cultures of leptomeningeal cells prepared from the neonatal rat brain. Luciferase reporter assays with deletion and site-directed mutation of the promoter region (−1250 to +77) showed that an AP-2 element at −109 was required for activation and an E-box at +57, for repression. Binding of nuclear factors to each of these cis-elements was demonstrated by an electrophoretic mobility shift assay. Several components of the Notch-Hes signaling pathway, Jagged,Notch1, Notch3, and Hes-1, were expressed in the leptomeningeal cells. Human Hes-1 co-expressed in the leptomeningeal cells bound to the E-box of the rat L-PGDS gene, and repressed the promoter activity of the rat L-PGDS gene in a dose-dependent manner. The L-PGDS gene expression was up-regulated slowly by interleukin-1β to the maximum level at 24 h. The reporter assay with deletion and mutation revealed that two NF-κB elements at −1106 and −291 were essential for this up-regulation. Binding of two NF-κB subunits, p65 and c-Rel, to these two NF-κB elements occurred after the interleukin-1β treatment. Therefore, the L-PGDS gene is the first gene identified as the target for the Notch-Hes signal through the E-box among a variety of genes involved in the prostanoid biosynthesis, classified to the lipocalin family, and expressed in the leptomeninges. Moreover, the L-PGDS gene is a unique gene that is activated slowly by the NF-κB system.

Lipocalin-type prostaglandin (PG) 1 D synthase (L-PGDS) was first identified in the rat brain as an enzyme that catalyzes the conversion to PGD 2 from PGH 2 , the latter being a common precursor of all prostanoids (1). Later, L-PGDS was found to be identical to ␤-trace (2, 3), which had been identified earlier (4) as the major protein in human cerebrospinal fluid (CSF). PGD 2 is a major prostanoid in the brain, known as the most potent endogenous somnogenic substance thus far identified (5), and is involved in various physiological events such as regulation of sleep and pain responses (5)(6)(7)(8). Because human L-PGDS-overexpressing transgenic mice (9) and L-PGDS-gene knock-out mice (10,11) showed abnormality in the regulation of non-rapid eye movement (NREM) sleep and pain responses, PGD 2 produced by L-PGDS is thought to play important roles in the regulation of NREM sleep and pain sensation in the central nervous system. Alternatively L-PGDS binds small lipophilic molecules such as retinal and retinoic acid (K d ϭ 70 -80 nM) (12) and biliverdin and bilirubin (K d ϭ 33-37 nM) (13) with affinities higher than those of other members of the lipocalin family. Therefore, L-PGDS is a unique bifunctional protein acting as a PGD 2 -producing enzyme and as a lipophilic molecule-binding protein (6 -8).
L-PGDS is expressed abundantly in the human heart (14), the male genital organs (15)(16)(17), and the central nervous system (18 -20) of various mammals. In the brain, this enzyme is expressed dominantly in the leptomeninges, chroid plexus, and oligodendrocytes but poorly in neurons, astrocytes, and microglial cells (18 -20). These distribution profiles suggest that L-PGDS gene expression is differentially regulated in a tissueand cell type-specific manner. García-Ferná ndez et al. (21,22) found that L-PGDS gene expression was regulated by the thyroid hormone during brain development in rats and demonstrated that the thyroid hormone responsive element was functional in COS-7 cells (23). White et al. (24) also demonstrated that L-PGDS gene expression was induced by the thyroid hormone through its responsive element in human TE671 cells derived from a medulloblastoma in the cerebellum. Moreover, L-PGDS gene expression was induced by dexamethasone in mouse neuronal GT1-7 cells (25). However, the regulatory mechanisms of L-PGDS gene expression have never been analyzed in the leptomeningeal cells, which show the highest expression of this gene (18 -20). On the other hand, the L-PGDS level in the rat CSF was increased by the administration of a proinflammatory cytokine, interleukin (IL)-1␤ (26). IL-1␤ and tumor necrosis factor-␣ activate the NF-B system (27) to elicit somnogenic activities (28,29). The sleep-inducing effects of IL-1␤ and tumor necrosis factor-␣ were proposed to be mediated by L-PGDS, because these effects were inhibited by administration of inhibitors of PG synthesis to animals (29). However, the transcriptional regulation of the L-PGDS gene by the NF-B system has never been clarified.
In this study, we characterized the rat L-PGDS gene promoter in the primary cultures of leptomeningeal (LM) cells prepared from neonatal rat brains and identified an AP-2 element for the transcriptional activation, an E-box element for the repression, and two NF-B elements for slow induction by IL-1␤. Among various genes involved in prostanoid biosynthesis, classified as members of the lipocalin family and expressed in the leptomeninges, the L-PGDS gene is the first one identified to be a target of Hes-1, a mammalian homolog of Drosophila Hairy and Enhancer of split of the Notch-Hes signal pathway. The L-PGDS gene is thus the second identified gene encoding an enzyme suppressed by the Notch-Hes signal, as the human acid ␣-glucosidase gene was the first one (30,31). Moreover, the L-PGDS gene is a novel example of a small number of genes slowly up-regulated through activation of NF-B subunits by IL-1␤, similar to the Mn-superoxide dismutase gene (32) but different from many of the IL-1␤-activated genes identified up to the present.
To ensure the transcription level of L-PGDS mRNA, we performed the competitive RT-PCR. Briefly, 200 ng of total RNA was converted to the first-strand cDNAs using RNase H Ϫ SuperScript II reverse transcriptase in the presence of serial diluted RNA standard (competitor) with random 9-mer primer (Takara Shuzo). PCR amplification was performed using sets of specific primers of L-PGDS-T and L-PGDS-C for L-PGDS or G3PDH-5Ј and G3PDH-3Ј for G3PDH, ExTaq DNA polymerase (Takara Shuzo), and the RT products. After the initial denaturation at 95°C for 5 min, PCR was carried out for 15 s at 94°C, 15 s at 55°C, and 20 s at 74°C for 28 cycles. The amplified products were analyzed by electrophoresis on 2% agarose gel followed by staining with ethidium bromide. The stained products were scanned and analyzed by ChemiImager system (Alpha Innotech, San Leandro, CA). The expression level of L-PGDS gene was expressed relative to that of the G3PDH gene.
Determination of Transcription Initiation Site-The transcription initiation site of the rat L-PGDS gene was determined by the Cap Site Hunting method (46) with rat brain CapSite cDNA (Nippon Gene, Toyama, Japan). Briefly, first-step PCR was performed by using LA Taq DNA polymerase (Takara Shuzo) with L-PGDS-E (5Ј-CGGAATTCTCT-CACCTGTGTTTACTC-3Ј) and 1RDT (5Ј-GATGCTAGCTGCGAGT-CAAGTC-3Ј; Nippon Gene) as the primers under the following amplification conditions: initial denaturation at 95°C for 5 min, followed by 94°C for 15 s, 55°C for 15 s, and 72°C for 1 min for 35 cycles. Nested PCR was carried out by using L-PGDS-C and 2RDT (5Ј-CGAGT-CAAGTCGACGAAGTGC-3Ј; Nippon Gene) as the primers and firststep PCR products as the template under the same condition as used for the first-step PCR. Resultant PCR products were cloned into pGEM-T Easy vector. Nucleotide sequences were determined by the dideoxy termination method with a Thermo Sequenase cycle sequencing kit (Amersham Biosciences) and a Long Tower Read DNA sequencer (Amersham Biosciences).
Construction of the Promoter-Luciferase Gene Plasmids-An ϳ1.3-kb fragment (Ϫ1250 to ϩ77) consisting of the rat L-PGDS gene promoter was amplified by PCR with the gene-specific primer set with XhoI (sense) or HindIII (antisense) sites at their respective 5Ј-end. PCR fragments doubly digested with XhoI and Hind III were cloned into the upstream region of the luciferase reporter gene of the pGL3-Enhancer vector (Promega). The luciferase-reporter construct carrying the promoter region from Ϫ1250 to ϩ77 was designated as Ϫ1250/ϩ77. A deletion series was constructed in the same manner. Site-directed mutations were introduced into the putative cis-elements by PCR using mutated primers. All constructs were subjected to nucleotide sequencing to verify their correct sequence and orientation.
Transient Transfection and Luciferase Assay-For transfection, LM cells or those of the three cell lines were seeded at a density of ϳ2 ϫ 10 5 cells/well in 24-well plates and co-transfected with each construct (0.4 g) and pRL-CMV (0.1 g; Promega), the latter carrying the Renilla luciferase gene under the control of the cytomegalovirus promoter as the transfection control, by use of Effectene (Qiagen, Hilden, Germany) according to the method prescribed by the manufacturer. The cells were cultured for 48 h. In the case of IL-1␤ stimulation, the cells were transfected in the same way as described above. After 48 h of transfection, the medium was replaced with fresh medium with or without IL-1␤ (1 ng/ml; R&D Systems, Minneapolis, MN), and the cells were then incubated further for 24 h. The luciferase activities were measured by using a dual luciferase reporter assay kit (Promega). The reporter activity was calculated relative to that of pGL3-Enhancer vector and was defined as 1. All data were obtained from at least three independent experiments, and each experiment was performed in duplicate. The relative promoter activities were depicted as the mean Ϯ S.D.
Electrophoretic Mobility Shift Assay (EMSA)-Nuclear extracts were prepared as follows: cells were suspended in 500 l of ice-cold buffer comprising 10 mM HEPES-KOH, pH 7.5, 10 mM KCl, 0.1 mM EDTA, 1 mM dithiothreitol, 0.5 mM phenylmethylsulfonyl fluoride (PMSF), a mixture of pepstatin, leupeptin, and aprotinin at 2 g/ml each, and 0.1% (v/v) Nonidet P-40 and vortexed vigorously. After centrifugation at 1,000 ϫ g for 1 min at 4°C, the nuclear pellet was resuspended in 100 l of 50 mM HEPES-KOH, pH 7.5, containing 400 mM KCl, 0.1 mM EDTA, 1 mM dithiothreitol, 0.5 mM PMSF, and a mixture of pepstatin, leupeptin, and aprotinin at 2 g/ml each, and incubated for 1 h at 4°C with continuous agitation. Nuclear extracts were prepared by centrifugation at 12,000 ϫ g for 30 min at 4°C. Protein concentrations were determined according to the method of Bradford (47). EMSA was performed as described previously (48).
Co-expression Study with Hes-1-The expression vector for human Hes-1 was kindly provided by Dr. Michael Caudy (Cornell University) and contained the human Hes-1 coding region fused to a FLAG-tagged sequence (49). Rat LM cells were co-transfected with the human Hes-1 expression vector and the Ϫ1250/ϩ77 construct as described above. After 48 h of transfection, the cells were harvested, and the luciferase activities were then measured.
Chromatin Immunoprecipitation (ChIP) Assay-LM cells were crosslinked with 1% formaldehyde at 37°C for 10 min and subsequently washed with ice-cold phosphate-buffered saline containing 0.5 mM PMSF and a protease inhibitor mixture of pepstatin, leupeptin, and aprotinin at 2 g/ml each. The cells were then lysed in 200 l of buffer containing 1% (w/v) SDS, 10 mM EDTA, 50 mM Tris-Cl, pH 8.0, and 0.5 mM PMSF and a mixture of pepstatin, leupeptin, and aprotinin at 2 g/ml each and then sonicated. After centrifugation to remove the cell debris, the supernatant was diluted with 20 mM Tris-Cl, pH 8.0, buffer containing 150 mM NaCl, 2 mM EDTA, 1% (v/v) Triton X-100, 0.01% (w/v) SDS, and 0.5 mM PMSF and a mixture of pepstatin, leupeptin, and aprotinin at 2 g/ml each and incubated at 4°C for 1 h with 60 l of 50% slurry of protein G-Sepharose (Amersham Biosciences), 20 g of shared salmon sperm DNA (Sigma), and 60 g of bovine serum albumin (Sigma). After the protein G-Sepharose had been removed by centrifugation, the resulting supernatant was incubated with the specific antibody and then 50 l of protein G-Sepharose was added, and the incubation was continued at 4°C for 1 h. The Sepharose beads were washed five times with washing buffer containing 50 mM Tris-Cl, pH 8.0, 150 mM NaCl, and 0.05% (v/v) Triton X-100 and twice with 10 mM Tris-Cl, pH 8.0, and 1 mM EDTA. The immunoprecipitated DNA-protein complexes were eluted with 500 l of 1% (w/v) SDS and 0.1 M NaHCO 3 . DNA-protein cross-links were reversed at 65°C for 5 h, after which the DNA was purified with a Qiaquick PCR purification kit (Qiagen) and then used for PCR amplification with the following gene-specific primer sets: S1-F (5Ј-CTGTATAAAGCAGGAGAGTGC-3Ј) and S1-R (5Ј-GGCT-GATTTGGGAGTTTTCTCC-3Ј), S2-F (5Ј-TGTGAGAAGCAGGTCT-TAGCC-3Ј) and S2-R (5Ј-CATACCTCAGTGAAGAACGGG-3Ј), and H1-F (5Ј-CTCCAGTGGGCAGTCCTTGGG-3Ј) and H1-R (5Ј-GGGCTG-CACGGAGACCTGGGC-3Ј) as the sense and antisense primers for NF-B site 1, site 2, and E-box, respectively, of the rat L-PGDS promoter. PCR amplification was performed as follows: initial denaturation at 95°C for 3 min, followed by 94°C for 25 s, 55°C for 25 s, and 72°C for 30 s for 34 -40 cycles. The resultant PCR products were analyzed on 2% (w/v) agarose gels.
Western Blot Analysis-Nuclear extracts were prepared as described above and separated by SDS-PAGE and transferred onto polyvinylidene difluoride membranes (Immobilon P; Millipore Corp., Bedford, MA). The membranes were incubated with specific primary antibodies (Santa Cruz Biotechnology, Inc., Santa Cruz, CA), washed, and then incubated with second antibodies conjugated to horseradish peroxidase (Santa Cruz Biotechnology, Inc.). Immunoreactive proteins were detected by using the ECL Western blotting detection system according to the manufacturer's instructions (Amersham Biosciences).

RESULTS
Cell Type-specific Expression of the Rat L-PGDS Gene-The expression of the L-PGDS gene in the rat LM cells and rat cell lines was examined by RT-PCR. The rat L-PGDS mRNA was expressed abundantly in the LM cells (Fig. 1A, lane 1) and negligibly in F2408 (lane 2), RNB (lane 3), and C6 (lane 4) cell lines, whereas the expression of the mRNA for G3PDH as the internal control was almost identical in each cell (Fig. 1A). Transient transfection experiments with the Ϫ1250/ϩ77 construct in the rat LM cells and each of the three rat cell lines demonstrated that efficient reporter activity was detected only in the LM cells (Fig. 1B). This result is in agreement with the expression of the L-PGDS mRNA demonstrated by the RT-PCR analysis and shows that the promoter region from Ϫ1250 to ϩ77 contains the critical element(s) responsible for the cell type-specific expression of the rat L-PGDS gene. Therefore, we used this promoter region to study the transcriptional regulation mechanism involving the activation and/or suppression of L-PGDS gene expression in the LM cells.
Transcription Initiation Sites of the Rat L-PGDS Gene-Transcription initiation sites were determined by the CapSite Hunting method. The sequence analysis of 18 clones demonstrated that nine of them had a nucleotide C of 77 base pairs upstream (defined as ϩ1) of the A of the translation initiation codon, ATG, and that the remaining nine had a C of 75 base pairs upstream (Fig. 2), indicating that rat L-PGDS gene expression was initiated predominantly from both of these transcription initiation sites in the brain. Surrounding both transcription initiation sites, a TATA-like box (ATAAATA) was found at a position from Ϫ28 to Ϫ22, but no typical CCAAT box was observed. Through a TRANSFAC search of the promoter region, several putative cis-elements such as NF-B elements at Ϫ1106 and Ϫ291, an Sp-1 element at Ϫ926, an AP-2 element at Ϫ109, and an E-box at ϩ57 were found. To elucidate the evolutional conservation of the promoter sequences between the human and rat L-PGDS genes, we aligned the 5Ј-flanking regions of both sequences (Fig. 2). The overall homology (61.9%) between them was observed in a region of ϳ400 bp around the transcriptional initiation site. One of the transcription initiation sites of the rat L-PGDS gene showed a onenucleotide difference from that of the human L-PGDS gene (24). Furthermore, several cis-elements were conserved between them, including the TATA-like box (Ϫ28 to Ϫ22 for rat and human), the AP-2 element (Ϫ109 to Ϫ101 and Ϫ98 to Ϫ90 for rat and human, respectively), and the E-or N-box (ϩ57 to ϩ62 and Ϫ334 to Ϫ329 for rat and human, respectively). These results suggest that the human and rat L-PGDS genes are regulated in a similar manner.
Promoter Activity of the Rat L-PGDS Gene-To determine the promoter activity and the important cis-elements involved in the regulation of L-PGDS gene expression, we introduced deletion and site-directed mutation into the promoter region and then conducted reporter analyses using rat LM cells. When the region from Ϫ1250 to ϩ77 was used for the luciferasereporter assay, efficient reporter activity was detected, indicating that this region contained cis-element(s) responsible for the regulation of gene expression (Fig. 1B). By deletion of the region from Ϫ1250 to Ϫ281, which contained two NF-B elements at Ϫ1106 and Ϫ291, and the Sp-1 element at Ϫ926, the promoter activity was decreased by ϳ30% compared with that of the Ϫ1250/ϩ77 construct (Fig. 3), indicating that this region contained weak activation activity for the expression.
A deletion from Ϫ280 to Ϫ100 resulted in a drastic decrease in the promoter activity by nearly 80% as compared with that of the Ϫ1250/ϩ77 construct. This region contained the AP-2 element at Ϫ109. By introduction of the site-directed mutagenesis into the AP-2 element, the luciferase reporter activity was decreased to almost the same level as when the AP-2 element had been deleted. These results indicate that the AP-2 element is an important cis-element for the transcriptional activation of the rat L-PGDS gene. Furthermore, when the Ϫ1250/Ϫ75 construct was used for transfection, measurable reporter activity was no longer detected, indicating that the region from Ϫ74 to ϩ1, containing the TATA-like box at Ϫ28 was essential for the basal expression of the L-PGDS gene in the LM cells.
In contrast, the deletion of the region from ϩ40 to ϩ77 in exon 1 caused the reporter activity to increase significantly by ϳ160% compared with that of the Ϫ1250/ϩ77 construct, indicating that a negative regulator binds to this region from ϩ40 to ϩ77. This region contained the E-box at ϩ57 involved in the Notch-Hes signaling pathway. Introduction of a mutation into the E-box increased the reporter activity by ϳ130% compared with that of the wild-type Ϫ1250/ϩ77 construct. These results clearly indicate that L-PGDS gene expression was suppressed through the E-box in the LM cells.
Analysis of Factor Binding to Regulatory Elements by EMSA-To demonstrate the nuclear factor(s) specific for binding to the putative cis-element(s) identified by the promoter luciferase assay, we conducted an EMSA using 32 P-labeled double-stranded oligonucleotides containing the putative ciselement, shown in Table I, and nuclear extracts prepared from the rat LM cells. As shown in Fig. 4A, when the doublestranded rL3 oligonucleotide carrying the AP-2 element was incubated with the nuclear extracts, two major DNA-protein complexes were found (Fig. 4A, lane 1). The formation of these complexes was inhibited by the addition of a 10-or 50-fold excess amount of the unlabeled rL3 oligonucleotide (lanes 2 and 3) or the authentic AP-2 wild-type (WT) DNA probe (lane 6). In contrast, a 100-fold excess amount of the nonspecific pUC19 DNA (lane 4) or the inactive AP-2 mutant (mu) DNA probe (lane 5) did not prevent the complex formation. These results demonstrate the specific binding of nuclear factors to the AP-2 element.
Furthermore, 32 P-labeled rL4 oligonucleotide containing the E-box bound a nuclear factor (Fig. 4B, lane 1). The formation of this DNA-protein complex was decreased in a concentrationdependent manner by the addition of a 10-or 50-fold excess amount of the unlabeled rL4 oligonucleotide (lanes 2 and 3) but was not inhibited by a 100-fold excess amount of the nonspecific pUC19 DNA (lane 4) or the inactive E-box (mu) mutant oligonucleotide (lane 5). Thus, the nuclear factor prepared from the rat LM cells bound to the E-box in a nucleotide sequencespecific manner.

Function of the E-box in the Regulation of L-PGDS Gene Expression-The E-box is well known as the cis-element for the
Hes protein that is involved in the Notch-Hes signaling pathway. Thus, we examined the expression of the genes involved in the Notch-Hes signaling pathway by RT-PCR and found that Jagged, a ligand for Notch receptors, Notch1 and Notch3, and Hes-1 were expressed in the rat LM cells (Fig. 5A).
Co-expression of the Ϫ1250/ϩ77 construct with the human Hes-1 protein was performed in the rat LM cells. The recombinant human Hes-1 protein fused with a FLAG tag was detected with anti-FLAG antibody in the nuclear extracts prepared from the transfected LM cells (Fig. 5B). The recombinant human Hes-1 protein level in the nucleus of the transfected LM cells was dependent on the amount of the expression vector used for the transfection (Fig. 5B). Furthermore, the promoter activity was decreased in a concentration-dependent manner by the Hes-1 expression vector (Fig. 5C). On the other hand, Renilla luciferase activities as the internal control for transfection were almost the same in all experiments, even when the Hes-1 protein was overexpressed in the LM cells (Fig. 5C).
We then conducted EMSA by incubating the rL4 oligonucleo-tide containing the E-box with nuclear extracts prepared from rat LM cells transiently expressing the human Hes-1 protein.
Evidence of formation of a DNA-protein complex was observed (Fig. 5D, lane 2), but no complex was found in the absence of the nuclear extracts (lane 1). This complex failed to form in the presence of 10-or 50-fold excess amount of the unlabeled rL4 oligonucleotide (lanes 3 and 4). The formation of the complex was not inhibited by a 100-fold excess amount of nonspecific pUC18 DNA (lane 5). In the supershift assay, this DNA-protein complex migrated more slowly when anti-FLAG monoclonal antibody had been added (lane 6). In the non-transfected LM cells, the corresponding shifted band (*) was observed in both in the presence or absence of the FLAG antibody (lanes 7 and  8). In contrast, no corresponding supershifted band (**) was observed even when FLAG antibody was added (lane 8). Furthermore, we confirmed the in vivo binding of the recombinant Hes-1 protein to the E-box of the rat L-PGDS gene by using the ChIP assay. The Hes-1 protein bound to the E-box. The amount of DNA-protein complex was increased in a concentration-dependent manner by the transfection with the Hes-1 vector (Fig. 5E). In the negative control experiment, when S1-F and S1-R as primers for NF-B site 1 of L-PGDS gene were used for the subsequent PCR amplification, the amplification of its region containing no E-box consensus sequence was not detected (N.C. in Fig. 5E). Therefore, these results, taken together, indicate that the Hes-1 protein binds specifically to the promoter region of the L-PGDS gene in vivo, as well as in vitro, and thereby suppresses the L-PGDS gene a These double-stranded oligonucleotides were purchased from Santa Cruz Biotechnology, Inc. Introduced mutations are underlined.

Regulation of Rat L-PGDS Gene Promoter
expression through the E-box in the rat LM cells. This is the first proof that the L-PGDS gene, which is involved in prostanoid biosynthesis, classified as a member of the lipocalin gene family, and expressed in the leptomeninges, is regulated transcriptionally by the Notch-Hes signal pathway. Enhancement of Transcription by IL-1␤-We examined the change of L-PGDS mRNA level in the rat LM cells treated with or without IL-1␤ (1 ng/ml) by the competitive RT-PCR assay. When the LM cells were treated with IL-1␤, L-PGDS mRNA level was increased slowly with a peak at 24 h after the start of treatment (ϳ2.5-fold over the control) (Fig. 6A). We also confirmed its induction by Northern blot analysis. L-PGDS gene expression was enhanced to a level of ϳ2.5-fold over the control after 24 h of IL-1␤ treatment (Fig. 6B, upper panel), whereas G3PDH gene expression was not altered by the IL-1␤ treatment (Fig. 6B, lower panel). In the nuclear fraction of the LM cells, NF-B subunits, p65 and c-Rel, but not p50, were detected by Western blot analysis, and their nuclear accumulations was enhanced by the treatment with IL-1␤ (Fig. 6C).
Through the data base search, we found two NF-B elements in the promoter region of the rat L-PGDS gene and termed them NF-B site 1 at Ϫ1106 and site 2 at Ϫ291 (see Fig. 3B and Fig. 6D). Progressive deletion analysis of the L-PGDS promoter region of the Ϫ1250/ϩ77 construct demonstrated that two regions containing the NF-B element from Ϫ1250 to Ϫ950 and from Ϫ450 to Ϫ280 were required for the IL-1␤-mediated induction of L-PGDS gene expression (Fig. 6E). By introducing mutation into each of these two NF-B sites, we found the luciferase reporter activity in the presence of IL-1␤ to have decreased to almost the same extent as when the region containing the NF-B site had been deleted (Fig. 6D).
Furthermore, we confirmed that nuclear factors bound to each of the NF-B elements by conducting EMSA using nuclear extracts prepared from the rat LM cells treated with IL-1␤ and specific oligonucleotides, rL1 and rL2 containing NF-B site 1 and 2, respectively. Fig. 7A shows that nuclear factor-DNA complexes (␣ and ␤) were observed in both NF-B sites (lanes 2 and 6). In addition, a weak signal (␥) was often observed only in the site 2 (lane 6). Each signal was absent when an excess amount of the corresponding DNA probes was added (lanes 3, 4,  7, and 8). To confirm the binding of NF-B subunits to each of the two NF-B sites, we carried out the ChIP assay using specific antibodies for p50, p65, and c-Rel of NF-B subunits. No binding of NF-B subunits was observed at either NF-B site when the LM cells were not treated with IL-1␤ (Fig. 7B). When the LM cells were treated with IL-1␤, however, p65 and c-Rel predominantly bound to both NF-B site 1 and site 2.
These results indicate that L-PGDS gene expression was induced by IL-1␤ through these two NF-B elements in the LM cells and that this induction occurred through the binding of p65 and c-Rel to both NF-B site 1 and site 2. DISCUSSION We earlier reported that the L-PGDS gene was expressed predominantly in the leptomeninges in the rat brain (18,19). Indeed, its expression in the rat LM cells was higher than that  Table I. M indicates the 100-base pair ladder marker. The predicted sizes of the PCR products for Jagged, Jagged2, Notch1, Notch2, Notch3, Hes-1, Hes-2, Hes-3, and Hes-5 were 563, 523, 512, 560, 502, 373, 474, 528, and 501 base pairs, respectively. Amplified signals are indicated by arrows. B, recombinant Hes-1 protein in the nuclear fraction of the LM cells transfected with the indicated amount of Hes-1 expression vector. Recombinant Hes-1 protein fused with FLAG tag in the nucleus was detected with anti-FLAG monoclonal antibody. C, suppression of L-PGDS gene expression by Hes-1. Rat LM cells were co-transfected with the Ϫ1250/ϩ77 construct and 0, 0.1, 1, or 5 g of the human Hes-1 expression vector. The luciferase activities were measured 48 h after transfection. When no expression vector was used, the reporter activity was defined as 100%. Firefly and Renilla luciferase activities are indicated by closed and open columns, respectively. D, double-stranded rL4 oligonucleotide containing the E-box labeled with [␥-32 P]ATP was incubated with nuclear extracts prepared from the rat LM cells transfected with the human Hes-1 expression vector, and the mixtures were then subjected to EMSA. Unlabeled rL4 oligonucleotide was added as the competitor. The shifted band is indicated by an asterisk. Supershift assay was conducted by using anti-FLAG monoclonal antibody, and the supershifted signal is indicated by a double asterisk. E, in vivo binding of the Hes-1 protein to the E-box of the rat L-PGDS gene. The LM cells transfected with the Hes-1 expression vector were treated with formaldehyde to cross-link DNA and proteins. The DNA-protein complexes were immunoprecipitated by the anti-FLAG antibody, and the extracted DNA was used as the template for PCR amplification (IP). A small aliquot before immunoprecipitation was used for PCR amplification as the input control (Input). N.C. means the negative control for this assay using a primer set of S1-F and S1-R for the NF-B site 1, whose expected amplified region does not contain the E-box consensus sequence. Results are representative of two independent experiments. in other cell lines such as astrocytes and other glial cells (Fig.  1). These results indicate that rat L-PGDS gene expression is strictly regulated in a cell type-specific manner. In this study, we examined the promoter function of the rat L-PGDS gene in the LM cells. Fig. 8 shows a schematic representation of our proposed regulatory mechanism governing the transcription of the rat L-PGDS gene in the LM cells.
In a previous study, the transcription initiation site for the rat enzyme was determined at ϩ39 by primer extension analysis (50). However, recent efforts at expressed sequence tags sequencing gave several rat expressed sequence tags clones showing high homology to L-PGDS and carrying the upstream region of the transcription initiation site decided previously. We determined the novel transcription initiation sites of the rat L-PGDS gene by the 5Ј-oligo capping method to be located 77 or 75 base pairs upstream from the translational initiation ATG codon (Fig. 2). The upstream transcription initiation site showed only a 1-base difference from that for the human L-PGDS gene (24) or the mouse gene. 2 Furthermore, the transcription of the mouse gene was initiated from one other position 2 that was identical to the one decided previously at ϩ39 in the rat (50). Therefore, the initiation site identified previously at ϩ39 in the rat gene is considered to also be one of the transcriptional initiation sites of the rat gene. A TATA-like sequence ( Ϫ28 ATAAATA Ϫ22 ) was found around the transcription initiation site in the rat gene. The deletion of this TATAlike element abolished completely the promoter activity, indicating that the TATA-like element was essential for the basal expression of the rat L-PGDS gene. The sequences around the TATA-like element and transcription initiation sites of human and rat genes showed a high degree of conservation, indicating that the transcriptional regulation of these L-PGDS genes is similar among various animal species.
We analyzed the 5Ј-flanking region of the rat L-PGDS gene in the LM cells. The AP-2 element at Ϫ109, GCCGCCCGC, whose consensus sequence is GCCNNNGCC (51), is a cis-element that binds AP-2 proteins, which are involved in the regulation of numerous cellular events and are responsible for transmitting the signal from cAMP or phorbol ester (52,53). Deletion of the AP-2 element caused a significant decrease in the promoter activity, but the basal promoter activity still remained intact (see Figs. 3 and 6). EMSA demonstrated that a nuclear factor bound to the AP-2 element (Fig. 4B). These findings indicate that AP-2 activates rat L-PGDS gene expression but is not associated with the basal promoter activity, unlike the case of the human hematopoietic PGDS gene, where it is associated with both (48).
The rat L-PGDS promoter carries the thyroid hormone re-2 K. Fujimori and Y. Urade, unpublished data.
FIG. 6. Identification of the cis-elements responsible for the effect of IL-1␤ treatment. A, rat LM cells were cultured in the presence or absence of IL-1␤ (1 ng/ml) for the indicated time and then collected for the isolation of RNA and the subsequent competitive RT-PCR assay. Open and closed bars indicate notreatment (control) and IL-1␤ treatment. The data represent the mean Ϯ S.D. from three independent assays. B, induction of L-PGDS gene expression by IL-1␤. The cells were cultured for 24 h with or without 1 ng/ml IL-1␤. Total RNA (2 g) was utilized for the Northern blot analysis with rat L-PGDS cDNA (upper panel) (35) or G3PDH cDNA (lower panel) (36) as the probes. C, induction of the NF-B proteins by the treatment of LM cells with IL-1␤. NF-B proteins, p50, p65, or c-Rel were detected with each specific antibody in the nuclear fraction prepared from LM cells incubated with (ϩ) or without (Ϫ) IL-1␤ (1 ng/ml) for 24 h. D, a site-directed mutation was introduced into each of the two NF-B elements, and the nucleotide changes are indicated. E, each region-deleted or mutated NF-B element was connected to the luciferase reporter gene, and LM cells were transfected with each construct along with the Renilla luciferase expression plasmid. After 48 h, IL-1␤ (1 ng/ml) was added (closed columns) or not (open columns), and the cells were then cultured for 24 h. Data are presented as the means Ϯ S.D. from at least three independent experiments. sponsive element at Ϫ596, and it was shown to function in COS-7 cells (23). On the contrary, this element was not functional in the LM cells under our experimental conditions (Fig.  3). In hyperthyroidism, L-PGDS gene expression was decreased in the interlaminar meninges but not in the external leptomeninges surrounding the cerebellum, although its expression was strong in both membranes in normal rats (22). Thus L-PGDS gene expression in the leptomeninges is not affected by thyroid hormone and is regulated by a cell typespecific mechanism.
The Notch-Hes signaling contributed to the regulation of rat L-PGDS gene expression in the LM cells. Hes is a member of the basic helix-loop-helix protein family and acts as a transcriptional repressor by binding to the E-or N-box of target genes (39,54). The Notch-Hes signaling pathway is important for the determination of the cell fate or the embryonic development by inhibiting cellular differentiation in a variety of cells (55,56). The rat L-PGDS gene expression was repressed through the E-box, and the binding of the Hes-1 protein to the E-box was proved in vitro and in vivo by EMSA and ChIP assay. Further evidence for the role of Hes-1 as a transcriptional repressor was provided by the observation that co-expression of the human Hes-1 protein decreased the reporter activity of the Ϫ1250/ϩ77 construct in the LM cells (Fig. 5C). In the human L-PGDS promoter, an N-box (CACCAG) was found at Ϫ334, and the lack of the region containing the N-box increased significantly the reporter activity, indicating that this N-box might be functional for transcriptional repression and probably that its regulation is the same as in the case of the rat gene (24). However, such action was not discussed with respect to the human gene (24). Major component proteins involved in the Notch-Hes signaling pathway such as Jagged, Notch1, Notch3, and Hes-1 were expressed in the rat LM cells (Fig. 5A). Thus it is possible that the Notch-Hes signal is functionally involved in the regulation of L-PGDS gene expression through the binding of the Hes-1 to the E-box in the rat LM cells. Recently, presenilin-1 was found to digest the intracellular domain of the Notch receptor to activate the Notch-Hes signal (57). Furthermore, presenilin-1 is involved in the development of the central nervous system. Its mRNA is expressed in the leptomeninges in the mouse brain, and, moreover, in presenilin-1-deficient mice, the leptomeninges appeared to be thicker (58). Thus, the L-PGDS gene, highly expressed in the leptomeninges, might have potential roles mediated by the signal of the Notch-Hes pathway during development. Our present data provide the first proof that the Notch-Hes signal regulates transcriptionally a gene involved in prostanoid biosynthesis, classified as a member of the lipocalin gene family, and expressed dominantly in the leptomeninges. Roles of L-PGDS under the control of the Notch-Hes signal need to be further elucidated.
The region from Ϫ1250 to Ϫ280 contained three putative cis-elements, two NF-B elements at Ϫ1106 and Ϫ291 and the Sp-1 element at Ϫ926; however, deletion of this region did not result in a significant decrease in the promoter activity. On the other hand, when the cells were treated with the proinflammatory cytokine IL-1␤, L-PGDS gene expression was enhanced ϳ2.5-fold, but it was slowly induced to its maximal level, which was attained after 24 h of treatment. Such a slow induction by proinflammatory cytokines resembles the case of Mn-superoxide dismutase gene, in which there was a peak at 24 h after the start of IL-1␤ treatment (59). These two NF-B sites (GGGGGTTTCA at Ϫ1106 and GGGCTTTGCT at Ϫ291) of the L-PGDS gene are well in line with the NF-B consensus sequence, GGGNNTYYCC (60). Site-directed mutation analysis and EMSA demonstrated that these two NF-B sequences were required for IL-1␤-mediated promoter activity of the L-PGDS gene. Moreover, the ChIP assay with antibodies against NF-B subunits demonstrated that each of the two NF-B sites was recognized and bound in vivo by the NF-B p65 and c-Rel as the homodimeric or heterodimeric complex. These results indicate the activation mechanism of the L-PGDS gene by proinflammatory cytokine IL-1␤ acting through these two NF-B elements in the LM cells. Moreover, IL-1␤ induces the L-PGDS level in the CSF (26), in which case its induction may occur FIG. 8. Schematic representation of the proposed mechanism for transcriptional regulation of the rat L-PGDS gene in LM cells.

FIG. 7. Binding of nuclear factors to the two NF-B elements.
A, double-stranded rL1 or rL2 oligonucleotide containing NF-B site 1 or site 2 labeled with [␥-32 P]ATP was incubated with nuclear extracts prepared from the rat LM cells that had been treated with 1 ng/ml IL-1␤ for 24 h, and the mixtures were then subjected to EMSA. Unlabeled rL1 or rL2 oligonucleotide was added as the competitor. Shifted bands are indicated by arrows. B, ChIP assay to detect the in vivo binding between NF-B proteins and L-PGDS promoter. The LM cells were treated (ϩ) or not (Ϫ) with IL-1␤ for 24 h and then fixed with formaldehyde. Chromatin prepared from the formaldehyde-fixed cells was sonicated and then immunoprecipitated with anti-p50, anti-p65, or anti-c-Rel antibodies. Cross-linkage of the co-precipitated DNA-protein complexes was reversed, and DNA was used as the template for PCR amplification with specific primers for NF-B site 1 or site 2 of the L-PGDS promoter (IP). The aliquot prior to immunoprecipitation was also amplified as the input control (Input). Results are representative of two or three independent experiments. through enhancement of L-PGDS gene expression via activation of NF-B subunits. Cyclooxygenase-2, an upstream enzyme of PGDS, is responsible for the inflammation signal and is also transcriptionally activated by NF-B (61,62). However, IL-1␤-mediated induction of the cyclooxygenase-2 gene is rapid and completely different from the slow induction of the L-PGDS gene. The molecular mechanism and physiological roles of such slow induction of L-PGDS gene expression are currently under investigation.
L-PGDS gene is involved in the regulation of NREM sleep (9), and IL-1␤ is one of the somnogenic substances and is involved in NREM sleep induction through the activation of NF-B (27)(28)(29). Furthermore, IL-1␤ receptor was detected in high concentration in the leptomeninges (63,64). Our present findings, together with previous results, indicate that the IL-1␤-mediated NREM sleep may, at least in part, occur by enhancement of L-PGDS gene expression in the leptomeninges.