Up-regulation of per mRNA Expression by Parathyroid Hormone through a Protein Kinase A-CREB-dependent Mechanism in Chondrocytes*

In bone, clock genes are involved in the circadian oscillation of bone formation and extracellular matrix expression. However, to date little attention has been paid to circadian rhythm in association with expression of clock genes during chondrogenesis in cartilage. In this study, we investigated the functional expression of different clock genes by chondrocytes in the course of cartilage development. The mRNA expression of types I, II, and X collagens exhibited a 24-h rhythm with a peak at zeitgeber time 6, in addition to a 24-h rhythmicity of all the clock genes examined in mouse femurs in vivo. Marked expression of different clock genes was seen in both osteoblastic MC3T3-E1 and chondrogenic ATDC5 cells in vitro, whereas parathyroid hormone (PTH) transiently increased period 1 (per1) mRNA expression at 1 h in both cell lines. Similar increases were seen in the mRNA levels for both per1 and per2 in prehypertrophic chondrocytes in metatarsal organotypic cultures within 2 h of exposure to PTH. PTH significantly activated the mouse per1 (mper1) and mper2 promoters but not the mper3 promoter in a manner sensitive to both a protein kinase A inhibitor and deletion of the cAMP-responsive element sequence (CRE) in ATDC5 cells. In HEK293 cells, introduction of brain and muscle aryl hydrocarbon receptor nuclear translocator-like protein 1 (bmal1)/clock enhanced mouse type II collagen first intron reporter activity without affecting promoter activity, with reduction effected by either per1 or per2. These results suggest that PTH directly stimulates mper expression through a protein kinase A-CRE-binding protein signaling pathway for subsequent regulation of bmal1/clock-dependent extracellular matrix expression in cartilage.

In bone, clock genes are involved in the circadian oscillation of bone formation and extracellular matrix expression. However, to date little attention has been paid to circadian rhythm in association with expression of clock genes during chondrogenesis in cartilage. In this study, we investigated the functional expression of different clock genes by chondrocytes in the course of cartilage development. The mRNA expression of types I, II, and X collagens exhibited a 24-h rhythm with a peak at zeitgeber time 6, in addition to a 24-h rhythmicity of all the clock genes examined in mouse femurs in vivo. Marked expression of different clock genes was seen in both osteoblastic MC3T3-E1 and chondrogenic ATDC5 cells in vitro, whereas parathyroid hormone (PTH) transiently increased period 1 (per1) mRNA expression at 1 h in both cell lines. Similar increases were seen in the mRNA levels for both per1 and per2 in prehypertrophic chondrocytes in metatarsal organotypic cultures within 2 h of exposure to PTH. PTH significantly activated the mouse per1 (mper1) and mper2 promoters but not the mper3 promoter in a manner sensitive to both a protein kinase A inhibitor and deletion of the cAMP-responsive element sequence (CRE) in ATDC5 cells. In HEK293 cells, introduction of brain and muscle aryl hydrocarbon receptor nuclear translocator-like protein 1 (bmal1)/clock enhanced mouse type II collagen first intron reporter activity without affecting promoter activity, with reduction effected by either per1 or per2. These results suggest that PTH directly stimulates mper expression through a protein kinase A-CRE-binding protein signaling pathway for subsequent regulation of bmal1/clock-dependent extracellular matrix expression in cartilage.
Recent studies have revealed that the endogenous circadian rhythmicity generated at the cellular level by circadian core oscillators resides not only in the hypothalamic suprachiasmatic nucleus of the anterior hypothalamus (1,2), which is recognized as the mammalian central clock, but also in various peripheral tissues including heart (3), adipose tissue (4), pancreas (5), and liver (2). The suprachiasmatic nucleus is not essential for driving peripheral oscillations but rather acts as a synchronizer of peripheral oscillators, whereas the physiologi-cal rhythmicity may be under the direct control by their own local clock genes in peripheral tissues (7). In mice, the rhythmic transcription of two orthologs of the Drosophila Period (per) gene appears to be essential for circadian rhythms. Expression of mouse per (mper) genes is known to be positively regulated by other clock proteins belonging to the basic helix-loop-helix period/aryl hydrocarbon receptor nuclear translocator/single minded class, which are Clock and brain and muscle aryl hydrocarbon receptor nuclear translocator-like protein 1 (Bmal1), 2 respectively. In addition, mPer proteins constitute multimeric complexes with products of the cryptochrome (cry) genes, mcry1 and mcry2, which in turn negatively regulate the gene transcription mediated by Clock/Bmal1 (8 -11).
Bone is formed through a well organized and highly regulated process (12). During embryogenesis mesenchymal precursor cells differentiate into skeletal elements by forming a cartilaginous model framework, which then induces the bone formation process known as endochondral ossification in the vertebral column and long bone (13). This endochondral ossification is responsible for the formation of the cartilaginous rudiment that is a tightly regulated region in both the differentiation and maturation of chondrocytes. Within this cartilaginous rudiment, for example, chondrocytes successively differentiate through a well organized mechanistic procedure with migration from resting cells at the rudiment edge to proliferating, hypertrophic and calcifying cells around the central region, which finally leads to mineralization of the cartilage matrix in the area of the hypertrophic chondrocytes. Shortly after the mineralization process takes place, however, most hypertrophic chondrocytes undergo sustained apoptosis. Upon death of chondrocytes after mineralization, circulating precursors of osteoblasts, osteoclasts, and capillaries begin to invade the cartilage matrix to produce new bones, leading to the growth of endochondral bones (12,13).
Previous reports have demonstrated that expression patterns of two major extracellular matrices in bone, type I collagen and osteocalcin, display circadian rhythms (14,15). Molecular clock components have been shown to mediate leptin-regulated bone formation (16), whereas to date no evidence is available for a role of the expression of circadian clock genes in the proliferation and/or differentiation of chondrocytes in the literature. Because parathyroid hormone (PTH) is well known to regulate a variety of cellular functions through PTH/PTH related protein (PTHrP) receptors (PPRs) in both osteoblasts and chondrocytes, it is conceivable that circadian clock genes may be functionally expressed to play a role in the mechanisms associated with PTH-induced modulation of cellular functions in chondrocytes as well as osteoblasts. In the present study, an attempt has been made to demonstrate the expression and functionality of different circadian clock genes in chondrocytes to elucidate circadian oscillation of cartilage development during endochondral ossification.

EXPERIMENTAL PROCEDURES
Animals-The protocol employed here meets the guidelines of the Japanese Society for Pharmacology and was approved by the Committee for Ethical Use of Experimental Animals at Kanazawa University. All efforts were made to minimize animal suffering, to reduce the number of animals used, and to utilize alternatives to in vivo techniques. Male ddY mice obtained at 4 weeks of age were maintained for 2 weeks under controlled temperature and humidity with a 12-h light (08:00 -20:00 h)/12-h dark (20:00 -08:00 h) cycle with access to standard laboratory food and water ad libitum. Animals were then decapitated to obtain total RNA from femurs at the following zeitgeber times (ZT): 0, 6, 12, and 18, where ZT 0 is defined as the time of lights on and ZT 12 lights off, respectively. Four animals were usually used for each time point on different days.
Primary Osteoclasts-Bone marrow was prepared from tibias and femurs of 4-week-old Std-ddY male mice and cultured for 24 h with macrophage colony-stimulating factor (M-CSF) (R&D Systems) at 10 ng/ml in ␣-minimal essential medium containing 10% FBS. After culturing for 24 h, nonadherent cells in supernatants were collected, followed by lamination on Ficoll gradient, and subsequent centrifugation at 500 ϫ g for 15 min. The monocyte/macrophage progenitor fraction was collected and suspended in minimal essential medium (Invitrogen) containing 10% FBS, M-CSF at 20 ng/ml, and receptor activator of nuclear factor-B ligand (RANKL) (R&D Systems) at 20 ng/ml. Cells were plated at a density of 1 ϫ 10 5 cells/cm 2 , followed by culturing at 37°C under 5% CO 2 for 5 days.
Embryonic Metatarsal Rudiment Organ Culture-Metatarsal organ culture was conducted as described previously (17). The three central metatarsal rudiments were isolated from ddY mouse embryos at 15.5 days post-gestation. Each of three metatarsals was placed in a well of a 24-well plate containing 1 ml of organ culture medium: minimal essential medium supplemented with 0.05 mg/ml ascorbic acid, 1 mM ␤-glycerophosphate, and 0.25% FBS. These explants were grown at 37°C in a humidified 5% CO 2 incubator for 5 days.
In Situ Hybridization Analysis-In situ hybridization was carried out as described previously (17). In brief, mounted sections were fixed with 4% paraformaldehyde, followed by HCl, proteinase K, and triethanolamine/acetic anhydride treatment. After prehybridization, sections were covered with digoxigenin (DIG)-labeled cRNA probes at 65°C for 16 h. The slides were treated with RNase A, blocked with 1.5% blocking buffer, incubated with anti-DIG-AP-Fab fragments (Roche) at 4°C for 16 h, and treated with nitro blue tetrazolium/5-bromo-4-chloro-3indolyphosphate for different periods. Moreover, neonatal mice were subcutaneously injected with 400 g/kg PTH, followed by dissection of tibiae 1 h after administration and subsequent detection of Per1 mRNA expression using a DIG-labeled cRNA probe.
Luciferase Assay-Reporter vectors were co-transfected with a TK-Renilla luciferase construct into MC3T3-E1, ATDC5, or HEK293 cells using Lipofectamine and Plus reagent (Invitrogen). Two days after transfection, cells were lysed, and luciferase activity was determined using specific substrates in a luminometer according to the manufacturer's protocol (Promega). Transfection efficiency was normalized by determining the activity of Renilla luciferase.
RT-PCR-RT-PCR was conducted as described previously (18). In brief, cDNA was synthesized with the oligo(dT) primer and reverse transcriptase (Invitrogen) from extracted total RNA. PCR amplification was performed using specific primers (Table 1), and PCR products were subcloned into a TA cloning vector (Promega) for determination of DNA sequences. Although the results obtained by RT-PCR are by definition not quantitative, apparent semi-quantitative RT-PCR analysis was performed at cycles below 30 with relatively high linearity using primers for the housekeeping gene glyceraldehyde-3-phosphate dehydrogenase for comparative reference. PCR products were quantified by using a densitograph, followed by calculation of the ratios of expression of mRNA for each gene over that of glyceraldehyde-3-phosphate dehydrogenase.
Electrophoretic Mobility Shift Assay-Gel retardation electrophoresis was conducted as described previously (18). In brief, nuclear extracts were prepared by the method of Schreiber et al. (19) with minor modifications. Cells were collected in 10 mM HEPES-NaOH buffer (pH 7.9) containing 10 mM KCl, phosphatase inhibitors, and protease inhibitors. Following the addition of 10% Nonidet P-40 to make a final concentration of 0.6%, cells were centrifuged at 20,000 ϫ g for 5 min. Pellets were suspended in 50 mM Tris-HCl buffer (pH 7.5) containing 400 mM KCl and inhibitors for phosphatases and proteases. Suspensions were kept in an ice bath for 30 min, followed by centrifugation at 20,000 ϫ g for 5 min. The supernatants thus obtained were collected as nuclear extracts. Assays were carried out using double-stranded oligonucleotide containing consensus core sequence for the E-box from the Dec1 promoter (5Ј-CTA-GGTCCAACACGTGAGACTCTCGA-3Ј), type II collagen promoter (5Ј-ATCCATGGTTGCACGTGGCCCGTGCA-3Ј), and type II collagen first intron (5Ј-AATTTTCTTCCACGTG-TTTGGGCACG-3Ј), as a probe labeled by [␣-32 P]deoxy-ATP for detection of DNA binding activity. After determining protein contents, an aliquot of nuclear extracts was incubated at a fixed amount of 3 g of protein and analyzed by electrophoresis on a 6% polyacrylamide gel. Gels were fixed and dried, followed by exposure to x-ray films for different periods to obtain autoradiograms appropriate for subsequent quantification by a densitograph.
Western Blotting-Western blotting was conducted as described previously (18). In brief, nuclear extracts prepared from cultured osteoblasts were treated with SDS, and lysates were subjected to electrophoresis on a polyacrylamide gel containing SDS, followed by transference to a polyvinylidene fluoride membrane. The membranes were then subjected to blocking with skimmed milk, followed by reaction with the anti-phospho-CRE-binding protein (CREB) antibody or the anti-CREB antibody (New England Biolabs). Finally, membranes were reacted with the antirabbit immunoglobulin G antibody conjugated with peroxidase and detected with the aid of ECL detection reagents for subsequent exposure to x-ray films.
Northern Blotting-Northern blotting was conducted as described previously (18). In brief, extracted RNA was resolved on 1% formaldehyde/agarose gel, and transferred onto positively charged nylon transfer membranes. After fixation of RNA to the blot by UV cross-linking, blotted membranes were prehybridized at 68°C for 1 h, and subsequently hybridized with a denatured DIG-labeled RNA probe of mPer1 at 68°C for 16 h. Membranes were then washed and incubated with anti-DIG-AP-Fab, followed by incubation with CDP-star and subsequent exposure to x-ray films for appropriate periods to detect chemiluminescence.
Chromatin Immunoprecipitation (ChIP) Assay-ChIP experiments were performed essentially according to the protocol provided with the ChIP assay kit (Upstate Biotechnology) using ATDC5 cells. ATDC5 cells were treated with formaldehyde for cross-linking, followed by sonication in lysis buffer containing protease inhibitors (1 mM phenylmethylsulfonyl fluoride, 1 g/ml aprotinin, and 1 g/ml pepstatin). Immunoprecipitation was performed with the anti-FLAG antibody (Sigma), followed by extraction of DNA with phenol/chloroform. PCR was performed using sequences found in the 5Ј-flanking region (Ϫ3215 to Ϫ2821) and first intron region (ϩ1042 to ϩ1365) of the mouse type II collagen gene, in addition to the mper promoter (Ϫ663 to Ϫ383), as primers.
Data Analysis-Results are all expressed as the mean Ϯ S.E. and statistical significance was determined by two-tailed and unpaired Students' t test or one-way analysis of variance with Bonferroni/Dunnett post-hoc test.

Rhythmic Expression of Clock Genes and Marker Genes in Bone
and Cartilage-To investigate whether mRNA expression for bone and cartilage marker genes exhibit daily rhythms in addition to circadian clock genes, total RNA was obtained from adult ddY mouse femurs every 6 h throughout a single 24-h period, followed by determination of the mRNA levels by semi-quantitative RT-   indeed expressed in bone cells, including osteoclasts, osteoblasts, and chondrocytes, RT-PCR analysis was conducted using specific primers for each clock gene. For the purpose of preparation of primary osteoclasts devoid of osteoblasts, monocyte/macrophage progenitor cells were cultured for 5 days in the presence of 20 ng/ml M-CSF and 50 ng/ml RANKL, followed by determination of expression profiles of each clock gene. RT-PCR analysis revealed that expression of mRNA was drastically increased for all the osteoclastic marker genes examined in primary cultured osteoclasts differentiated from hematopoietic progenitors in the presence of both M-CSF and RANKL ( Fig. 2A). Under these conditions, bmal1 mRNA was constitutively expressed at each of the differentiation stages, with transient expression of dec1 mRNA 1 day after differentiation induced by RANKL and M-CSF. However, no marked expression was seen for the other molecular clock genes, including clock, per1-3, cry1-2, dbp, and rev-erb␣, in cultured osteoclasts throughout the entire range of differentiation stages in primary osteoclasts (Fig. 2B). Moreover, osteoblastic MC3T3-E1 and chondrogenic ATDC5 cells were cultured for 7 days in individual differentiation media, followed by non-quantitative determination of expression of circadian clock genes. Expression of mRNA was seen for all clock genes examined in both MC3T3-E1 and ATDC5 cells cultured for 7 days, in addi-tion to adult mouse hypothalamus used as a positive control (Fig. 2C).
Expression Profile of mRNA for Clock Genes in ATDC5 Cells-We next examined the expression profiles of circadian clock genes in ATDC5 cells. ATDC5 cells are chondrogenic cells and gradually display a chondrocytic phenotype in the presence of insulin in proportion to the length of the culture period up to 28 days. Expression of mRNA was markedly increased for type X collagen during culture periods of from 7 to 14 days with a gradual increase thereafter up to 28 days, whereas a relatively steady level of mRNA expression was seen for both type II collagen and Runx2 for a period of up to 28 days (Fig. 3A). Under these conditions, expression of mRNA was drastically decreased for per2, bmal1, and clock from 7 to 14 days with a gradual decline thereafter up to 28 days, whereas mRNA levels were significantly decreased for per1 and cry2 at 28 days. By contrast, expression of dec2 mRNA was markedly increased from 7 to 14 days with a sustained increase up to 28 days. No significant alternation was seen in mRNA for per3, cry1, or dec1 on any day examined. We next examined DNA binding to the E-box sequence in ATDC5 cells cultured for different periods up to 28 days. Consistent with the mRNA expression for both clock and bmal1, DNA binding activity of the E-box was decreased in proportion to the duration of culture from 3 to 28 days (Fig. 3B, a). To verify the nuclear proteins involved in E-box binding, supershift analysis was carried out in nuclear extracts prepared from ATDC5 cells cultured for 14 days using the anti-clock antibody. The anti-clock antibody was effective in inducing an upward shift of the probe-protein complex on gels (Fig. 3B, b). Moreover, E-box binding was completely inhibited by excess unlabeled double-stranded oligonucleotide containing the E-box sequence, whereas oligonucleotide with mutations of the E-box core sequence did not markedly affect E-box binding even at a molar concentration ratio of 30-fold (Fig. 3B, c).
Effect of PTH on Per mRNA Expression in MC3T3-E1 and ATDC5 Cells-To determine whether PTH induces per1 in osteoblasts and chondrocytes, osteoblastic MC3T3-E1 and chondrogenic ATDC5 cells were cultured for 7 days, followed by the addition of 100 nM PTH and subsequent collection of total RNA at different time points after the PTH addition for determination of the mRNA levels by Northern blotting. mPer1 mRNA levels were markedly increased 1 h after the addition of PTH-(1-34), with a recovery to the baseline within 2 h in both MC3T3-E1 and ATDC5 cells (Fig. 4A), whereas no marked expression was seen at any time point after the addition of PTH-(3-34) (data not shown). The addition of PTH-(1-34) stimulated phosphorylation of CREB at Ser-133 in ATDC5 cells with maximal phosphorylation at 1 h and a return to the basal level within 4 h, whereas no marked alteration was seen for the expression of CREB at any time point after PTH addition up to 24 h (Fig. 4B). We next determined the effects of different drugs known to affect chondrocytic functions on per expression, in addition to PTH. ATDC5 cells were transfected with a construct of mper promoter regions linked to the luciferase reporter gene, followed by treatment with different drugs for 48 h. Of the drugs tested, the adenylyl cyclase activator forskolin and PTH induced a significant elevation of mper1 promoter activity, with no significant alternations by other reagents. These included vitamin D 3 , glutamate, GABA, epidermal growth factor, insulin-like growth factor, fibroblast growth factor, transforming growth factor, BMP-2, and melatonin (Fig.  4C). In addition, only forskolin significantly elevated mper2 promoter activity in addition to PTH, whereas mper3 promoter activity was not markedly affected by any of the regents examined (data not shown).
Effect of PTH on mPer Promoter Activity-ATDC5 cells were transfected with either of the mper promoter constructs and treated with PTH at 10 nM for 48 h. Both mper1 and mper2 promoter activities were significantly increased after exposure to PTH, with no response for the mper3 promoter (Fig. 5A). To examine the possible signaling pathway responsible for mper1 and

. Effect of PTH on expression of mPer1 mRNA in MC3T3-E1 and ATDC5 cells.
A, MC3T3-E1 and ATDC5 cells cultured for 7 days were exposed to 100 nM PTH for a period up to 24 h, followed by isolation of total RNA at the indicated time points and subsequent Northern blot analysis using a probe for mPer1. mper2 transactivation by PTH, cells were transfected with the mper promoter, followed by treatment with the protein kinase A (PKA) inhibitor H89 and the protein kinase C (PKC) inhibitor Ro31-8220, respectively, 1 h before PTH stimulation. Inhibition of PKA by H89 almost completely blocked transactivation of both the mper1 and mper2 promoters by PTH, whereas the PKC inhibitor slightly but significantly inhibited the promoter activity of mper1 without significantly affecting that of mper2. Moreover, PTH significantly increased the activity of pCRE-luc containing three tandems of CRE. The inhibitor of PKA, but not of PKC, completely inhibited PTH-dependent CRE reporter activity. No marked alternation was found in any promoter activities, including mper1, mper2, and CRE, in the presence of either H89 or Ro31-8220 alone. To determine the element responsive to PTH within the mper promoter region and to identify the cis-and trans-acting elements responsible for transcriptional activation, mper1 and mper2 promoter 5Ј deletion constructs were generated from the full-length mPer1 promoter (Ϫ1803 to ϩ40) with one putative CRE and three E-box sequences. PTH significantly stimulated the reporter activity of full-length mper1 promoter containing CRE, without altering that of the other mutated mper1 promoters not containing CRE (Fig. 5B). These included a Ϫ1350 mutant with three E-boxes, a Ϫ900 mutant with two E-boxes, and a Ϫ450 mutant with one E-box. To further assess the importance of CRE in the responsiveness of the mper1 promoter to PTH, mutant constructs were made at the four nucleotides within CRE in the context of whole mper1 and mper2 promoters. This CRE mutation completely abolished the responsiveness of both the mper1 and mper2 promoters to PTH. In addition, MC3T3-E1 cells transfected with either of mper promoter constructs were similarly treated with PTH at 10 nM or forskolin at 10 M. mPer1 promoter activities were significantly increased by PTH and forskolin, whereas marked induction was observed by only forskolin for the mper2 promoter. In contrast, no response was seen for the mper3 promoter, as shown in ATDC5 cells. Moreover, the significant induction of mper1 promoter activity by PTH and forskolin was completely abolished when CRE mutant constructs of the mper1 promoter were used (data not shown).
Effect of PTH on Per mRNA Expression in Metatarsals and Tibiae-To determine whether PTH induces mper ex vivo as seen in ATDC5 cells in vitro, mouse metatarsals were isolated before vascularization, followed by organotypic culture for 5 days and subsequent treatment with 100 nM PTH for in situ hybridization analysis. Small cells were labeled by a cRNA probe for type I collagen known to be specifically expressed by osteoblasts in the chondrocyte layer (data not shown). In cul- tured metatarsals, type II collagen mRNA was preferentially expressed by proliferating to prehypertrophic chondrocytes, whereas Indian hedgehog mRNA was selectively expressed by prehypertrophic chondrocytes with highly selective expression of type X collagen mRNA by hypertrophic chondrocytes, in addition to osteopontin mRNA expression by hypertrophic FIGURE 6. Effect of PTH on expression of mPer1 and mPer2 mRNA in chondrocytes. A, metatarsals were isolated from fetal mice before vascularization, followed by organotypic culture for 5 days and subsequent exposure to 100 nM PTH for different periods up to 4 h. Frozen sections were dissected after fixation with formalin in a cryostat, followed by in situ hybridization using cRNA probes for mPer1 and mPer2, in addition to different chondrocytic markers. B, neonatal mice were subcutaneously injected with 400 g/kg PTH, followed by dissection of tibiae 1 h after administration and subsequent in situ hybridization using cRNA probes for mPer1. Black arrows indicate osteoblasts in cancellous bone. Typical micrographic pictures are shown in this figure, whereas similar results were invariably obtained in at least five independent determinations. Ihh, Indian hedgehog; OP, osteopontin. chondrocytes (Fig. 6A). In metatarsals cultured for 5 days, PPR was predominantly expressed by prehypertrophic chondrocytes. Under the experimental conditions used here, mRNA levels peaked for both the per1 and per2 selectively expressed by prehypertrophic chondrocytes at 1-2 h with a return to basal levels within 4 h after the addition of PTH-(1-34). However, PTH-(3-34) failed to induce mper expression in cultured metatarsals at any time point after stimulation (data not shown). In sections treated with sense RNA probes, no marked expression was detected for any of the genes examined (data not shown).
To further demonstrate the induction of mper by PTH in vivo as seen in cultured ATDC5 cells in vitro and metatarsals ex vivo, neonatal mice were subcutaneously injected with 400 g/kg PTH, followed by dissection of tibiae and subsequent in situ hybridization analysis. In situ hybridization analysis revealed that up-regulation of mper1 mRNA expression was mainly detected in prehypertrophic chondrocytes, as seen in cultured metatarsals, 1 h after PTH administration (Fig. 6B). Interestingly, mper1 mRNA expression was also increased by the systemic administration of PTH in osteoblasts attached to cancellous bone in the tibiae.
Possible Involvement of mPER in Type II Collagen Transcription-To examine the possible role of mPer in the expression of chondrocytic marker genes, promoter regions were subjected to a Motif data base search. The mouse type II collagen promoter contains a putative E-box recognized by the Bmal1/ Clock heterodimer, whereas the first intron region of mouse type II collagen also has a putative E-box. In HEK293 cells transfected with a particular construct made up of different promoter regions, mper1 promoter activity was significantly enhanced by Bmal1/Clock expression, whereas no marked increase was found in cells transfected with Bmal1 or Clock alone (Fig. 7A). In contrast, mPer1, mPer2, and mCry1 significantly impaired the Bmal1/Clock-dependent enhancement of mper1 promoter activity. In particular, mCry1 almost completely inhibited the increase in mper1 promoter activity by Bmal1/Clock. Under these conditions, the mouse type II collagen promoter was neither activated by Bmal1/Clock nor inhibited by the clock genes tested (Fig. 7B). However, wild-type mouse type II collagen first intron reporter activity was significantly enhanced by Bmal1/Clock expression in a manner sensitive to inhibition by the introduction of mPer1, mPer2, or mCry1, as was also seen with mper1 promoter activity. In con- trast, Bmal1/Clock did not significantly induce the mouse type II collagen first intron reporter activity when the reporter with the mutated E-box core sequence was used in HEK293 cells (Fig. 7C).
To further verify the possible involvement of clock genes in type II collagen transcription in chondrocytic cells, mouse type II collagen first intron reporter (ϩ959 to ϩ2954) containing both the chondrocyte-specific cis-acting elements and Bmal1/ Clock recognition sites was used in ATDC5 cells. In ATDC5 cells, mouse type II collagen first intron reporter (ϩ959 to ϩ2954) activity was significantly enhanced by Bmal1/Clock expression, whereas no marked increase was found in cells transfected with Bmal1 or Clock alone (Fig. 7D).
To examine the possible association of Bmal1/Clock with the intron E-box of type II collagen in vivo, we next performed ChIP assays using ATDC5 cells expressing FLAGtagged Bmal1/Clock. As shown in Fig. 8A, the association of Clock was detected in ATDC5 cells transfected with Bmal1/ Clock within the type II collagen first intron, in addition to the mper1 promoter, but not the type II collagen promoter. In contrast, no association of Clock was found in ATDC5 cells transfected with empty vectors alone within the type II collagen first intron in addition to the mper promoter.
The binding of Bmal1/Clcok to the E-box found in the promoter and first intron region of mouse type II collagen was evaluated by electrophoretic mobility shift assay in HEK293 cells transfected with Bmal1/Clock. DNA binding was detected in nuclear extracts of HEK293 cells transiently expressing Bmal1/Clock in the E-box located in the first intron, but not in the promoter, of mouse type II collagen (Fig. 8B).

DISCUSSION
The essential importance of the present findings is that PTH stimulates expression of both mper1 and mper2 mRNA through transactivation mediated by a PKA-CREB signaling pathway in both cultured chondrocytes and metatarsals, with inhibition by mPer of Bmal1/Clock-dependent mouse type II collagen first intron reporter activity but not type II collagen promoter reporter activity. Circadian rhythm was found for the expression of particular osteoblastic and chondrocytic marker genes in adult mouse femurs, and in addition, a subcutaneous injection of PTH markedly elevated mper1 mRNA expression predominantly in prehypertrophic chondrocytes in the tibia of neonatal mice in vivo. To our knowledge, this is the first direct demonstration of the expression of molecular clock genes by chondrocytes in terms of direct modulation of cellular functions. Although several previous studies have already demonstrated the functional expression of molecular clock genes in peripheral tissues, including bone, no direct evidence for a role of these clock genes in the cellular functionality of chondrocytes has been previously reported.
The prevailing view is that PPR is a member of the B superfamily of G protein-coupled receptors with seven transmembrane domains in association with at least three different isoforms of the trimeric G protein. These include G s protein for the activation of adenylyl cyclase/PKA, G q /G 11 protein for the activation of phospholipase C/PKC, and G i protein for the inhibition of adenylyl cyclase (20,21). Previous studies have demonstrated that PPR may differentially regulate the expression of various genes on the basis of intracellular signaling pathways in target cells. The PKC signaling pathway appears to regulate the expression of type X collagen, Bcl-2, and vitamin D 3 receptors, for example, whereas the cAMP/PKA signaling pathway plays a pivotal role in controlling the mineralization and synthesis of glycosaminoglycans, in addition to inducing alkaline phosphatase and Indian hedgehog, in primary chondrocytes and chondrogenic cells such as ATDC5 cells (22)(23)(24). However, it is widely accepted that the roles of PKA versus PKC signaling activated by PTH/PTH-related protein vary according to experimental conditions, such as cell densities and differentiation stages, in terms of regulating proliferation and differentiation of chondrocytes (25,26). Although PKC signaling is associated with the enhancement of cell proliferation and DNA synthesis in addition to calcium signaling (27), cAMP-PKA signaling appears to be the dominant pathway for the actions of PTH on osteoblasts (28). Indeed, in osteoblasts PTH transiently activates various immediate early genes, including c-fos, c-jun, c-myc, interleukin-6, and leukemia inhibitory factor (29 -31), through the cAMP-PKA signaling pathway.
The present study clearly demonstrates the involvement of cAMP-PKA signaling in the regulation by PTH of mper gene expression. For instance, pretreatment with the PKA inhibitor, H89, almost completely prevented the PTH-dependent enhancement of both the mper1 and mper2 reporter activities, with the PKC inhibitor, Ro-31-8220, being rather ineffective. Forskolin strongly potentiated the mper reporter activity to an extent similar to PTH, whereas the PTH analog PTH- , which retains the ability to signal through PKC and calcium FIGURE 8. DNA binding of circadian clock genes to mouse type II collagen first intron. A, ATDC5 cells were transiently transfected with Bmal1/Clock or empty vector (EV), followed by incubation for 48 h and subsequent ChIP assay using the anti-FLAG antibody. Typical pictures are shown in this figure, whereas similar results were invariably obtained in at least three independent determinations. B, HEK293 cells were transiently transfected with Bmal1/ Clock, followed by incubation for 48 h and subsequent electrophoretic mobility shift assay using the E-box probe on mouse type II collagen promoter and the first intron region. Typical autoradiograms are shown in the figure, whereas similar results were invariably obtained in at least three independent determinations. pathways without PKA signaling activity (32), had no effect on mper1 mRNA expression or mper1 promoter activity (PTH- ). Moreover, the cAMP-PKA signaling pathway has been shown to stimulate the phosphorylation of CREB on serine 133, which is required for the recruitment of the CREB binding protein/p300 co-activator for gene transcription (33). Taken together, it appears that PTH activates mper1 gene expression through a mechanism related to cAMP-PKA-CREB signaling in both osteoblasts and chondrocytes.
A data base analysis revealed a putative E-box, which would be able to bind a Bmal1/Clock heterodimer, exists within 3 kb of the 5Ј-flanking region and the first intron of the type II collagen gene. The present findings from reporter and ChIP assays, however, suggest that the Bmal1/Clock heterodimer binds to the E-box located in the first intron but not that in the 3 kb of the 5Ј-flanking region. In addition to the mouse type II collagen gene described above, the E-box element is present within the first intron region but not in the 5Ј-flanking region of the type X collagen gene. Moreover, the type I collagen gene, an osteoblastic marker, contains an E-box element in the 5Ј-flanking region. Several independent lines of evidence indicate that effective transactivation would be mediated by the E-box element located in the intron regions but not that in the 5Ј-flanking regions with regard to Bmal1/Clock target genes. E-box elements located in the first and second introns of Dbp are critical for circadian transactivation by the Clock protein (34), in fact, the Bmal1/Clock heterodimer transactivates the peroxisome proliferator-activated receptor ␣ gene through an E-box-rich region located in the second intron (6). As the Per protein would be translocated into the nucleus to inhibit the activity of the Bmal1/Clock heterodimer (8), it may be speculated that the mPer transiently up-regulated by PTH negatively regulates the transactivation of chondrocytic marker genes by Bmal1/Clock through the E-box located in the intron region in chondrocytes.
In this study, we conducted reporter activity assays using HEK293 cells, which are a human embryonic kidney cell line with high transfection efficiency. The mouse type II collagen first intron reporter construct used in HEK293 cells does not contain chondrocyte specific cis-acting elements reported in the literature. However, the chondrocyte-specific type II collagen first intron construct containing the Bmal1/Clock recognition sites was up-regulated by overexpression of Bmal1/Clock in ATDC5 cells. It is therefore conceivable that Bmal1/Clock binds the E-box located in the mouse type II collagen first intron for subsequent positive regulation of transcription, and mPer proteins in turn negatively regulate the transcription mediated by Bmal1/ Clock even in chondrogenic cells, as described previously (8 -11). The exact molecular mechanism as well as the functional significance of differential expression profiles of clock genes during in vitro maturation of chondrogenic ATDC5 cells, however, remains to be elucidated in future studies.
It appears circadian clock genes are functionally expressed by chondrocytes to regulate extracellular matrix in cartilage. Molecular clock signaling is thus a potential target for the development of a drug useful for the treatment and therapy of a variety of bone and cartilage diseases relevant to abnormal development and maturation of chondrocytes as well as osteoblasts in human beings.