Regulated Nuclear-Cytoplasmic Localization of CCAAT/ Enhancer-binding Protein d in Osteoblasts*

Insulin-like growth factor I (IGF-I) plays a central role in skeletal growth by promoting bone cell replication and differentiation. Prostaglandin E 2 (PGE 2 ) and para- thyroid hormone enhance cAMP production in cultured rat osteoblasts and stimulate IGF-I expression through a transcriptional mechanism mediated by cAMP-depend-ent protein kinase (PKA). We previously showed that PGE 2 activated the transcription factor CCAAT/enhanc- er-binding protein d (C/EBP d ) in osteoblasts and induced its binding to a DNA element within the IGF-I promoter. We report here that a PKA-dependent pathway stimulates nuclear translocation of C/EBP d . Under basal conditions, C/EBP d was cytoplasmic but rapidly accumulated in the nucleus after PGE 2 treatment ( t 1 ⁄ 2 < 30 min). Nuclear translocation occurred without con-current protein synthesis and was maintained in the presence of hormone. Nuclear localization required PKA and was blocked by a dominant-interfering regulatory subunit of the enzyme, even though C/EBP d was not a PKA substrate. Upon removal of hormonal stimulus, C/EBP d quickly exited the nucleus ( t 1 ⁄ 2 < 12 min) through a pathway blocked by leptomycin B. Mutagenesis

Insulin-like growth factor-I (IGF-I), 1 a 70-amino acid secreted protein, plays a central role in regulating growth and development in mammals and other vertebrates (1,2). IGF-I promotes the survival, proliferation, and differentiation of many cell types and tissues, including bone, where it enhances osteoblast replication and type I collagen synthesis, among other actions (3,4). IGF-I is produced by various cells within the skeleton, including osteoblasts (5), and its synthesis is enhanced by systemic and locally produced hormones that regulate skeletal function, such as parathyroid hormone and prostaglandin E 2 (PGE 2 ) (5,6). The increase in IGF-I induced by these hormones may explain their anabolic actions within the skeleton, and IGF-I may serve as a coupling factor to balance the remodeling sequence of resorption and new bone formation (5,7,8).
In cultured bone cells, both PGE 2 and parathyroid hormone stimulate IGF-I gene and protein expression by a transcriptional mechanism (9 -11). These effects on IGF-I gene transcription are mediated by hormone-induced increases in cAMP and subsequent activation of cAMP-dependent protein kinase (PKA) (5,6,12). As evidence for this pathway, the major IGF-I promoter can be induced in transient transfection experiments in osteoblasts by a co-transfected catalytic subunit of PKA to the level seen with PGE 2 treatment (10). Furthermore, a dominant-interfering mutant regulatory subunit of PKA that does not bind cAMP blocks hormone-activated gene expression (10). In past studies, we mapped a functional cAMP response element to the 5Ј-untranslated region of IGF-I exon 1 within a previously footprinted site termed HS3D (10,13) and showed that this sequence was required for full hormonal responsiveness of the IGF-I promoter in osteoblasts (13). More recently, we identified CCAAT/enhancer-binding protein ␦ (C/EBP␦) as the critical hormone-regulated transcription factor responsible for PKA-stimulated IGF-I gene transcription through the HS3D sequence (14,15) and showed that hormones that activate PKA induce binding of C/EBP␦ to this site (14). C/EBP␦ belongs to a family of transcriptional regulators that function in tissue differentiation, metabolism, healing, and immune responses (16). Members of the C/EBP family are related structurally, each consisting of an NH 2 -terminal transactivation region, a central basic DNA-binding domain, and a COOH-terminal dimerization interface termed the leucine zipper segment (16). C/EBP proteins share similarities in the latter two domains with a larger group of basic-leucine zipper transcription factors (16,17). The first C/EBP proteins to be characterized, C/EBP␣ and C/EBP␤, have key roles in adipocyte differentiation and in gene expression in the liver and other tissues (16, 18 -21). C/EBP␦ has been implicated in control of adipogenesis and in mediating the acute phase response to inflammatory stimuli (16,18,19). In addition, our previous work indicated a regulatory role for this protein in IGF-I gene expression in bone cells (14,15).
The current experiments were designed to assess mechanisms of activation of C/EBP␦ in osteoblasts. We now find that a PKA-dependent pathway stimulates the rapid nuclear translocation of C/EBP␦ in the absence of ongoing protein synthesis. Continual PKA activity is required for nuclear retention of C/EBP␦, because C/EBP␦ is quickly removed from the nucleus through an exportin-mediated pathway upon cessation of hormone action. Mutagenesis studies indicate that the basic domain of C/EBP␦ is necessary for nuclear localization and that the leucine zipper region permits full nuclear accumulation. In the aggregate, this report defines a pathway for hormonemediated activation of C/EBP␦ through its regulated nuclear import.

EXPERIMENTAL PROCEDURES
Materials-Timed-pregnant Sprague-Dawley rats were purchased from Harlan Sprague-Dawley (Indianapolis, IN). Collagenase types 1 and 2 were obtained from Worthington Biochemical Corporation (Lakewood, NJ). Leptomycin B was a gift from Dr. Minoru Yoshida (University of Tokyo, Tokyo, Japan). PGE 2 , forskolin, and cycloheximide were purchased from Sigma. H-89 and KT 5720 were from Calbiochem (San Diego, CA); [␥-32 P]ATP was obtained from PerkinElmer Life Sciences. LY294002 was purchased from Biomol Research Laboratories (Plymouth Meeting, PA), and UO126 was from Promega Corporation (Madison, WI). The long lasting IGF-I analogue, R 3 IGF-I, was from Gropep (Adelaide, Australia), and PDGF-BB was from Life Technologies, Inc. PGE 2 was reconstituted at a concentration of 1 mM in ethanol, R 3 IGF-I was resuspended in 10 mM HCl, and PDGF-BB was resuspended in 11.1 mM acetic acid with 10% BSA. All other drugs were dissolved in Me 2 SO to at least 1000 times the final concentration. The catalytic subunit of PKA was a gift from Dr. James Lundblad (Oregon Health Sciences University, Portland, OR). Recombinant His-tagged cAMP response element-binding protein (CREB) and His-tagged mutant CREB S133A were gifts from Dr. Richard A. Maurer (Oregon Health Sciences University, Portland, OR). The mutant regulatory subunit of mouse PKA (clone MtR(AB)) was a gift from Dr. G. Stanley McKnight (University of Washington, Seattle, WA). Polyclonal antibodies to C/EBP␦ were raised in chickens and purified as described previously (15). A monoclonal antibody to the flag epitope and Hoechst dye were obtained from Sigma. Other antibodies (to Akt, extracellular signal-regulated kinases 1 and 2, phospho-Akt, and phospho-extracellular signal-regulated kinase) were from New England Biolabs (Beverly, MA). Cy3-conjugated rabbit antichicken IgY was from Jackson ImmunoResearch Laboratories (West Grove, PA); fluorescein isothiocyanate-conjugated goat anti-mouse IgG and alkaline phosphatase-conjugated goat anti-rabbit IgG were from Southern Biotechnology Associates (Birmingham, AL); and horseradish peroxidase-coupled rabbit anti-chicken IgY was from Promega Corporation. All other reagents were purchased from commercial suppliers.
Cell Cultures and Transfections-Osteoblast-enriched cell cultures were prepared from isolated calvarial bones of 21-day-old Sprague-Dawley rat fetuses, as previously described (22,23). Cranial sutures were removed by dissection, and cells were dispersed by five sequential digestions with collagenase. The last three digestions, which are enriched in cells expressing the osteoblast phenotype, were pooled and plated at 8000 cells/cm 2 in minimum essential medium (Life Technologies, Inc.) containing 20 mM HEPES, 10% bovine serum (HyClone, Logan, UT), 100 units/ml penicillin, and 100 g/ml streptomycin (Life Technologies, Inc.). Cells were incubated at 37°C in humidified air containing 5% CO 2 . Human fetal osteoblast cell line hFOB 1.19 (24) was purchased from ATCC and propagated in a 1:1 mixture of Dulbecco's modified Eagle's medium and Ham's F12 supplemented with 10% fetal bovine serum and 0.3 mg/ml G418 (all from Life Technologies, Inc.). These cells express a temperature-sensitive mutant of SV40 large T antigen and proliferate at the permissive temperature (34°C) in humidified air containing 5% CO 2 . At confluence, T antigen expression is inhibited by shifting the cells to 39°C, and osteoblast markers are expressed (24). Both types of cells were transfected at ϳ50% confluent density with 2 g of total plasmid per 9.6 cm 2 using GenePORTER transfection reagent (Gene Therapy Systems, San Diego, CA) for rat osteoblasts and LipofectAMINE (Life Technologies, Inc.) for hFOB 1.19 cells, according to the manufacturers' instructions. Experiments were performed at 48 -72 h after transfection, when the cells had reached confluent density.
Immunocytochemistry-Confluent osteoblast cultures were pre-incubated in serum-free medium for 20 h, followed by addition of drugs or vehicle (ethanol or Me 2 SO or both diluted 1:1000) in serum-free medium for the times specified. Where indicated, after incubation with PGE 2 , cultures were rinsed twice with phosphate-buffered saline (PBS), and fresh medium was added for various times. Cycloheximide and all protein kinase inhibitors were added to cells 15 min prior to addition of PGE 2 . Following hormone and drug treatment, cultures were rinsed with PBS, fixed in 4% paraformaldehyde, permeabilized with a 1:1 mixture of acetone and methanol, and blocked with 10% BSA in PBS. After two washes with PBS, cells were incubated with primary antibodies, either polyclonal chicken anti-C/EBP␦ (1:500) or monoclonal anti-flag (1:440) in PBS plus 3% BSA for 2 h at 25°C. Cells were then washed three times in PBS and incubated with 100 ng/ml Hoechst dye and the appropriate labeled secondary antibodies, either Cy3-conjugated rabbit anti-chicken IgY (1:400) or fluorescein isothiocyanateconjugated goat anti-mouse IgG (1:1000), for 1-2 h in the dark. Cells then were washed in PBS and examined by fluorescence microscopy (Nikon Eclipse TE 300). Images were captured with an Optronics CCD camera using an Apple Macintosh G3 computer and Scion Image software, version 1.62. Images were saved in Photoshop 5.5 (Adobe Systems, San Jose, CA).
Protein Extraction and Immunoblotting-Confluent osteoblast cultures were deprived of serum for 20 h and then treated with PGE 2 for varying intervals, as indicated above. For inhibitor studies, cells were pretreated with the inhibitors for 15 min followed by addition of R 3 IGF-I or PDGF-BB in the presence of inhibitors for 15 min. Cytoplasmic and nuclear protein extracts (13) or whole cell extracts (25) were prepared as described, and aliquots were stored at Ϫ80°C until use. Western immunoblotting was performed as described previously (13,25). Immunoreactive proteins were visualized by enhanced chemiluminescence, followed by exposure to x-ray film, or by enhanced chemifluorescence followed by detection and quantitation using Molecular Imager FX imaging system and Quantity One software (Bio-Rad).

Preparation of Recombinant and in Vitro Translated
Proteins-Preparation of recombinant S-tagged C/EBP␦ (S-C/EBP␦) and C/EBP␦ in Escherichia coli has been described (15). The 31-amino acid NH 2 -terminal S-tag includes a minimal consensus phosphorylation site for PKA (Arg-Gly-Ser). C/EBP␦ was translated in vitro using the pET29a-C/ EBP␦ plasmid (15) and TNT coupled reticulocyte lysate system (Promega Corporation), according to the manufacturer's instructions.
PKA Assay-The purified, recombinant catalytic subunit of PKA was diluted to 10 g/ml in 100 g/ml BSA. Recombinant CREB or C/EBP␦ proteins (1 g each) were mixed on ice with 0.5 Ci of [␥-32 P]ATP in assay buffer (100 M ATP, 10 mM MgCl 2 , 250 g/ml BSA, 12.5 mM Tris-Cl, pH 7.5). PKA (10 ng) was added, and the reactions were allowed to proceed for 2 min at 30°C. The reactions were stopped after being placed on ice by addition of EDTA to 80 mM final concentration. After boiling for 5 min in SDS sample buffer, the samples were separated by SDS-polyacrylamide gel electrophoresis. Gels were stained with Coomassie Brilliant Blue, dried, and exposed to x-ray film for 2 h at Ϫ80°C with intensifying screens.
Construction of Recombinant Plasmids-Flag-tagged rat C/EBP␦ in pcDNA3 (pcDNA3-flag-C/EBP␦) was generated by polymerase chain reaction-mediated mutagenesis. The flag epitope tag (codons underlined) was added to the 5Ј end of C/EBP␦ in pBluescript-C/EBP␦ (15) just 3Ј to a BamHI site and an ATG codon (bold) using the following oligonucleotides: 5Ј-GCGGATCCGCCACCATGGACTACAAGGACGA-CGATGACAAGAGCGCCGCTCTTTTCAGCCTA-3Ј (top strand) and 5Ј-CCAGTCGGGTTCGCGCTTCA-3Ј (bottom strand). The amplified fragment was digested with BamHI and PstI and inserted into BamHI-and PstI-digested pBluescript-C/EBP␦. After DNA sequencing to verify the intended changes, the entire C/EBP␦ coding region was excised by digestion with BamHI and EcoRI and inserted into the corresponding sites of pcDNA3 (Invitrogen, Carlsbad, CA) to produce pcDNA3-flag-C/EBP␦. The C/EBP␦ deletion plasmids diagramed in Fig. 8A were prepared using the QuikChange site-directed mutagenesis kit (Stratagene, La Jolla, CA). C/EBP␦⌬Zip was prepared by introducing a point mutation (C to T) after codon 216, which places a stop codon (bold) just beyond the COOH terminus of the basic region. A KpnI site (underlined) also was added following the stop codon to aid in identifying the mutant. The oligonucleotides used were as follows: 5Ј-CGCCGCAACT-AGGGTACCGAGATGCAGCAGA-3Ј (top strand) and 5Ј-TCTGCTGCA-TCTCGGTACCCTAGTTGCGG CG-3Ј (bottom strand). For C/EBP␦⌬B, the basic region (amino acids 193-216) was deleted in-frame and replaced with an EcoRV site (underlined) with oligonucleotides 5Ј-CGG-GGCAGCC CTGATATCCAGGAGATGCAG-3Ј (top strand) and 5Ј-CT-GCATCTCCTGGATATCAGGGCTGCCCCG-3Ј (bottom strand). For C/EBP␦⌬BZip, a stop codon (bold) followed by an XbaI site (underlined) was introduced after codon 192, immediately following the transactivation domain, with oligonucleotides 5Ј-CGGGGCAGCCCTTAGTCTA-GATACCGGCAGC-3Ј (top strand) and 5Ј-GCTGCCGGTATCTAGAC-TAAGGGCTGCCCCG-3Ј (bottom strand). For each plasmid generated, the coding region was verified by DNA sequencing.
The EGFP⅐C/EBP␦ fusion proteins are diagramed in Fig. 8A. To generate EGFP containing the basic and leucine zipper regions of C/EBP␦ (EGFPϩBZip), the BZip DNA region of C/EBP␦ was excised from pcDNA3-flag-C/EBP␦⌬BZip by digestion with XbaI, followed by filling in the overhang with the Klenow fragment of DNA polymerase I, and digestion with BamHI. This DNA fragment was then inserted in-frame into EcoRI (blunted with Klenow) and BamHI-digested pEGFP-C1 (CLONTECH Laboratories, Palo Alto, CA). To produce EGFP with the leucine zipper region of C/EBP␦ (EGFPϩZip), the EcoRV-EcoRI fragment was excised from pcDNA3-C/EBP␦⌬B and inserted into pEGFP-C1 that had been digested with HindIII (blunted with Klenow) and EcoRI. EGFP containing just the basic segment of C/EBP␦ (EGFPϩB) was prepared as follows. CEBP␦⌬Zip was excised from pcDNA3-CEBP␦⌬Zip by digestion with BamHI and EcoRI and ligated into corresponding sites in the polylinker of pEGFP-C1. In this construct there is a stop codon after the basic region of C/EBP␦. The portion of C/EBP␦ 5Ј to the basic region was then eliminated by mutagenesis using oligonucleotides that overlapped the 3Ј end of the EGFP coding region (5Ј-TCCGGACTTGTACAGCTCGTCCATGCCGAGTG-3Ј (top strand)) and the 5Ј end of the C/EBP␦ basic domain (5Ј-GAGTAC-CGGCAGCGACGCGAGCGCAACAACATC-3Ј (bottom strand)). The amplified region was verified by sequencing.
Statistical Analysis-Data are presented as the means Ϯ S.E. Statistical significance was determined using the Student's t test for paired samples. Results were considered statistically different when p Ͻ 0.05.

PGE 2 Stimulates Nuclear Translocation of C/EBP␦ in Rat
Osteoblasts-We previously identified C/EBP␦ as the key transcription factor mediating cAMP-activated IGF-I gene transcription in primary rat osteoblasts (14,15). We showed that stimulation of DNA binding of C/EBP␦ to its critical recognition site in the major IGF-I gene promoter and the subsequent induction of IGF-I gene expression were independent of the new protein synthesis (13). These results indicated that C/EBP␦ was activated by PGE 2 through post-translational mechanisms in osteoblasts. The current experiments were designed to determine how C/EBP␦ was regulated in these cells. Fig. 1 shows that incubation of osteoblasts with PGE 2 stimulated the nuclear accumulation of C/EBP␦. As seen by immunocytochemistry in Fig. 1A, under control conditions C/EBP␦ was diffusely distributed within the cell, but after 4 h of incubation with 1 M PGE 2 the protein was predominantly nuclear. Pre-incubation with cycloheximide at a concentration (2 M) found previously to block Ͼ90% of ongoing protein synthesis in osteoblasts (13) did not prevent accumulation of C/EBP␦ in nuclei, indicating that pre-existing C/EBP␦ was translocated to the nucleus in a protein synthesis-independent manner. This interpretation was validated by Western immunoblotting of osteoblast protein extracts (Fig. 1B). In control cells, C/EBP␦ was detected in cytoplasmic but not nuclear extracts. However, within 2 h of hormone treatment, it was found predominantly among soluble nuclear proteins and was depleted from the cytoplasm. Thus, PGE 2 induces nuclear translocation of C/EBP␦ in primary cultures of rat osteoblasts.
We next looked at the kinetics of nuclear accumulation of C/EBP␦. Fig. 2 shows results of time course studies. Again, under basal conditions C/EBP␦ was concentrated in Ͻ1% of osteoblast nuclei. Within 15 min of PGE 2 treatment, C/EBP␦ expression was primarily nuclear in 25.5 Ϯ 2.0% of cells. The proportion of cells with predominantly nuclear C/EBP␦ increased to 64% by 30 min and to ϳ94% at 1 and 2 h (t1 ⁄2 of nuclear accumulation ϭ 27.1 min). Upon removal of hormone, C/EBP␦ rapidly disappeared from nuclei and reaccumulated in the cytoplasm (t1 ⁄2 ϭ 11.6 min). Only 20.0 Ϯ 2.2% of cells retained prominent nuclear expression by 15 min (the earliest time point examined), and Ͻ1% retained prominent expression by 1 h. Thus, in response to PGE 2 , C/EBP␦ was rapidly redistributed from cytoplasm to nucleus and returned to the cytoplasm quickly after termination of the hormonal stimulus.
Nuclear Translocation of C/EBP␦ Is Dependent on PKA-We next looked at the signaling mechanisms involved in hormonal regulation of the subcellular distribution of C/EBP␦. Primary osteoblasts were treated with PGE 2 after pre-incubation with specific protein kinase inhibitors (Fig. 3A). The drugs LY294002 (phosphatidylinositol 3-kinase) or UO126 (MEK1 and 2) did not prevent PGE 2 -induced nuclear translocation of C/EBP␦ and did not alter the primarily cytoplasmic distribution of C/EBP␦ under control conditions. In contrast, the PKA inhibitors H-89 and KT5720 each prevented the appearance of C/EBP␦ in nuclei after PGE 2 treatment. To demonstrate that LY294002 and UO126 were effective in osteoblasts, cells were treated with either IGF-I or PDGF in either the absence or the presence of the two inhibitors. As shown in Fig. 3B, pre-incubation with LY294002 inhibited IGF-mediated phosphorylation of the phosphatidylinositol 3-kinase target, Akt, and UO126 prevented phosphorylation of the MEK targets, extracellular signal-regulated kinases 1 and 2, by PDGF.
As a further test that PKA was required to stimulate the nuclear translocation of C/EBP␦ in osteoblasts, cells were co-transfected with expression plasmids for the marker protein, EGFP, and for a modified regulatory subunit of PKA that cannot bind cAMP. This latter protein thus acts to block endogenous enzyme activity (26). Following incubation with forskolin (10 M for 2 h) to activate adenylate cyclase, the subcellular location of C/EBP␦ was assessed by immunocytochemistry (Fig. 4). Treatment with forskolin stimulated the nuclear accumulation of C/EBP␦ in nontransfected cells but did not alter the predominantly cytoplasmic distribution of C/EBP␦ in osteoblasts expressing the dominant-interfering PKA regulatory subunit (Fig. 4A, left panels). As shown in Fig. 4B, forskolin treatment induced nuclear translocation of C/EBP␦ in 91.7 Ϯ 3.5% of cells transfected with the empty expression plasmid but only in 11.2 Ϯ 1.7% of cells transfected with the dominantinterfering regulatory subunit of PKA (p ϭ 0.0013). Thus, PKA activity is required for hormone-regulated nuclear translocation of C/EBP␦. C/EBP␦ Is Not a Direct Substrate for PKA in Vitro-The next series of experiments was designed to determine whether C/EBP␦ was phosphorylated by PKA. Other studies have shown that the related transcription factor, C/EBP␤, is a substrate for PKA (27,28), and inspection of the protein sequence of C/EBP␦ revealed several potential PKA phosphorylation sites. To test the hypothesis that C/EBP␦ is a substrate for PKA, recombinant C/EBP␦ was generated in E. coli, purified, and used in in vitro kinase assays with the purified, recombinant catalytic subunit of PKA. As shown in Fig. 5, under the conditions described under "Experimental Procedures," the cAMP-regulated transcription factor, CREB, was readily phosphorylated by PKA, whereas a mutant CREB lacking the PKA phosphorylation site at serine residue 133 was not labeled (29). These results demonstrate the specificity of the in vitro kinase assay. C/EBP␦ also was not phosphorylated by PKA, but S-C/ EBP␦ was labeled. S-C/EBP␦ contains a consensus PKA site in the NH 2 -terminal S-tag. These in vitro experiments show that C/EBP␦ does not appear to be a high affinity substrate for PKA. Figs. 1-4 did not allow us to determine whether C/EBP␦ was constitutively cytoplasmic under basal conditions or whether it rapidly shuttled between cytoplasmic and nuclear compartments. To distinguish be- . Under control conditions, fewer than 1% of cells had predominantly nuclear C/EBP␦. The process of nuclear accumulation fitted a linear progression curve with t1 ⁄2 ϭ 27.1 min; nuclear disappearance was found to be an exponential function with t1 ⁄2 ϭ 11.6 min.

FIG. 3. Prevention of nuclear translocation of C/EBP␦ in rat osteoblasts by inhibitors of PKA but not of phosphatidylinositol 3-kinase or MEK.
A, immunocytochemistry for C/EBP␦ of primary rat osteoblasts after incubation with vehicle (con) or 1 M PGE 2 for 4 h in the absence or presence of the inhibitors listed below the fluorescence micrographs. B, inhibition of target kinases by LY294002 and UO126. The graph shows the percent inhibition of IGF-I-stimulated phosphorylation of Akt by LY294002 (LY) and PDGF-stimulated phosphorylation of extracellular signal-regulated kinases 1 and 2 by UO126 (UO). The IGF-I analogue, R 3 IGF-I (2 nM) was used to activate the phosphatidylinositol 3-kinase-Akt pathway, and PDGF-BB (0.  tween these possibilities, we employed the antibiotic leptomycin B (30,31), which specifically inhibits chromosome region maintenance 1 (CRM1), the receptor that functions to export proteins from the nucleus (32,33). Fig. 6 shows that leptomycin B did not alter the subcellular distribution of C/EBP␦ under basal conditions and did not influence the ability of PGE 2 to stimulate its nuclear translocation or of H-89 to prevent it. To demonstrate that leptomycin B was effective in osteoblasts, cells were pre-treated with PGE 2 for 2 h, washed with PBS, and then incubated with leptomycin B or with vehicle. Under these experimental conditions, leptomycin B inhibited the exit of C/EBP␦ from nuclei (Fig. 7). In vehicle-treated cells, export of C/EBP␦ to the cytoplasm was rapid, being nearly complete within 30 min. By contrast, in osteoblasts incubated with leptomycin B, C/EBP␦ remained predominantly nuclear for up to 4 h. Based on these results, we conclude that in the absence of hormonal stimulation, C/EBP␦ is primarily cytoplasmic and that PKA induces transport of C/EBP␦ into the nucleus.
The Basic and Leucine Zipper Domains of C/EBP␦ Are Required for Nuclear Targeting in Osteoblasts-We next sought to identify the domains of C/EBP␦ that were required for its nuclear localization. The full-length protein contains three major functional segments: an NH 2 -terminal transcriptional activation domain, a basic region that mediates DNA binding, and a leucine zipper segment that is responsible for dimerization (16). We generated expression plasmids for full-length and truncated rat C/EBP␦, each containing an NH 2 -terminal flag epitope tag to distinguish them from endogenous C/EBP␦ (Fig.  8A). Upon transient transfection into primary rat osteoblasts (data not shown) or into the human osteoblast cell line hFOB 1.19, full-length C/EBP␦ was found in the nucleus even in the absence of hormone treatment (Fig. 8B, left top panel). This precluded us from investigating the regulation of transfected C/EBP␦ by PKA. We instead used the various truncation mutants of C/EBP␦ to establish the structural requirements for its nuclear localization. A truncation mutant lacking the leucine zipper (C/EBP␦⌬Zip) was primarily nuclear when expressed in hFOB 1.19 cells. In contrast, mutant proteins lacking the basic domain (C/EBP␦⌬B) or both basic and leucine zipper regions (C/EBP␦⌬BZip) were predominantly cytoplasmic (Fig. 8B, top  panels). These observations are consistent with results obtained with the related transcription factor, C/EBP␤, which show that the highly conserved basic region contains the nuclear localization sequence (34). In C/EBP␦, however, removal of the leucine zipper led to partial expression in the cytoplasm (Fig. 8B, top, compare the two left panels), indicating that this latter region also contributes to nuclear localization.
These results were confirmed after transfection of EGFP fusion constructs containing different segments of C/EBP␦ at their COOH termini (diagrammed in Fig. 8A). EGFP fused to the basic and leucine zipper regions (EGFPϩBZip) was exclusively nuclear when expressed in hFOB 1.19 cells. EGFP plus the basic domain (EGFPϩB) was predominantly nuclear, and EGFP plus the leucine zipper (EGFPϩZip) was diffusely distributed, as was EGFP (Fig. 8B, lower panels). We interpret these experiments to indicate that a nuclear localization sequence (NLS) resides within the basic region of C/EBP␦ but that the leucine zipper contains additional determinants that facilitate full expression within the nucleus.
One NLS Is Sufficient to Translocate a C/EBP␦ Dimer into the Nucleus-The hFOB 1.19 cell line does not produce C/EBP␦ (assessed by immunocytochemistry and immunoblotting; data not shown). In these cells, transfected C/EBP␦⌬B was found exclusively in the cytoplasm (Figs. 9, top left panel, and 8B). However, when expressed in primary rat osteoblasts, C/EBP␦⌬B was concentrated in the nucleus (Fig. 9, top row, third panel from left). These results suggested the possibility that the transfected protein dimerized with endogenous C/EBP␦ and used its NLS for translocation into the nucleus. Consistent with this hypothesis, blocking basal PKA activity with H-89 resulted in retention of C/EBP␦⌬B in the cytoplasm (Fig. 9, top right panel). In confirmation of these results, transfected C/EBP␦⌬BZip was located in the cytoplasm of both hFOB 1.19 cells and primary rat osteoblasts (Fig. 9, lower panels). C/EBP␦⌬BZip lacks both the basic and leucine zipper domains and can neither dimerize nor translocate to the nucleus by itself. We interpret these observations to indicate that C/EBP␦ forms dimers in the cytoplasm and that one NLS is sufficient for nuclear localization of the dimer when at least basal PKA activity is present in the cells.

DISCUSSION
Our current studies define a mechanism for activation of the transcription factor C/EBP␦ in primary rat osteoblasts through its regulated nuclear import by a PKA-mediated pathway. In previous studies, we identified C/EBP␦ as the critical transcription factor for induction of IGF-I gene expression in response to PGE 2 (14,15). We showed that PGE 2 stimulated PKA in osteoblasts (12) and that PKA induced C/EBP␦ to bind to a site termed HS3D located within the major IGF-I gene promoter, leading to activation of IGF-I gene transcription (10,13,14). This pathway of hormonal stimulation of gene expression also has been shown to be independent of new protein synthesis (13). Using a combination of experimental approaches we now demonstrate that activation of PKA by forskolin or PGE 2 leads to the rapid accumulation of C/EBP␦ in osteoblast nuclei. Nuclear translocation was induced within 15 min of hormone treatment (the earliest time point examined), was detected in the majority of osteoblasts by 60 min, and persisted for at least 4 h when cells were continually incubated with PGE 2 . Translocation of C/EBP␦ into osteoblast nuclei occurred in the presence of the protein synthesis inhibitor cycloheximide, indicating that it was mediated by post-translational mechanisms. Upon removal of hormone, C/EBP␦ exited the nucleus rapidly (t1 ⁄2 Ͻ 12 min), suggesting that a continuous stimulus was needed to maintain its nuclear localization. Regulated nuclear translocation of C/EBP␦ was blocked by the specific PKA inhibitors H-89 and KT5720 and by forced expression of a dominant-interfering regulatory subunit of PKA, indicating that nuclear translocation was stimulated by PKA. Surprisingly, C/EBP␦ did not appear to be a direct substrate for PKA, because the purified enzyme failed to phosphorylate C/EBP␦ in vitro to a measurable extent. Therefore, activation of C/EBP␦ by PKA occurs through an indirect mechanism, perhaps by a PKA-initiated signaling cascade. However, attempts to block this postulated pathway with LY294002 or UO126 were unsuccessful. These latter results may be interpreted to indicate that neither phosphatidylinositol 3-kinase-Akt nor MEK-extracellular signal-regulated kinase pathways control the activity of C/EBP␦ in osteoblasts.
This report presents the first example of PKA-dependent nuclear import of C/EBP␦ in any cell type, although in a cultured hepatocyte cell line, treatment with tumor necrosis factor-␣ induced its nuclear accumulation (35). The regulated nuclear import of C/EBP␦ defined here appears to resemble the pathway of nuclear translocation of the related transcription factor, C/EBP␤. As shown by several investigators, C/EBP␤ resided in the cytoplasm in unstimulated cells and accumulated in the nucleus after treatment with agents that activated PKA with kinetics similar to those observed here for C/EBP␦ (28,36). However, C/EBP␤ is a substrate for PKA, and its phosphorylation is required for its nuclear translocation (28). In addition, C/EBP␤ can be phosphorylated in vitro by PKA on serines 277 and 299 (27) and in cells on serine 299 (28). An alanine substitution at residue 299 blocked regulated nuclear translocation of C/EBP␤ in DKO-1 colon carcinoma cells (28), providing clear evidence for control of nuclear localization by direct phosphorylation. In contrast, we were unable to demonstrate that C/EBP␦ was a substrate for PKA in vitro, confirming results of Kageyama et al. (37). C/EBP␦ has been shown to become phosphorylated after treatment of HepG2 cells with interleukin-1 (38), and changes in phosphorylation induced by other cytokines have been implicated in its transcriptional activation (38,39). DNA binding of C/EBP␦ to a consensus site also has been shown to be increased ϳ3-fold after phosphorylation in vitro by casein kinase II (40). No information is available on a role for casein kinase II in modulating the function of C/EBP␦ in cells. Other members of the C/EBP family, including C/EBP␤ and CHOP, undergo regulated phosphorylation by several different protein kinases. These modifications result in either altered DNA binding (27,41) or transcriptional activity (28,(42)(43)(44)(45)(46)(47). However, in preliminary experiments we were unable to detect an increase in phosphorylated C/EBP␦ in primary rat osteoblasts after incubation with PGE 2 (data not shown), indicating that this modification may not be part of the mechanism of nuclear translocation induced by PKA in these cells.
Very little is known about the cellular steps controlling nuclear import of members of the C/EBP family of transcription factors. Morever, no information is available regarding which nuclear import receptors interact with C/EBP␦ or which importins are expressed in osteoblasts. We find that C/EBP␦ is FIG. 6. Leptomycin B does not interfere with PKA-mediated nuclear translocation of C/EBP␦ in rat osteoblasts. Immunocytochemistry for C/EBP␦ of primary rat osteoblasts after incubation with vehicle (con), 1 M PGE 2 , or 10 M H-89 in the absence or presence of 10 ng/ml of leptomycin B (lept B). Nuclei stained with Hoechst dye are blue.
FIG. 7. Treatment with leptomycin B prevents the exit of C/EBP␦ from the nucleus of rat osteoblasts. Immunocytochemistry for C/EBP␦ of primary rat osteoblasts after incubation with 1 M PGE 2 for 2 h followed by washes with PBS and addition of vehicle or 10 ng/ml leptomycin B (Ϫ lept B or ϩ lept B, respectively) for the times indicated. In the absence of leptomycin B, the t1 ⁄2 for the disappearance of C/EBP␦ from the nucleus was 11.8 min, and in the presence of leptomycin B, t1 ⁄2 was Ͼ4 h. exclusively cytoplasmic when PKA activity is suppressed by H-89 or KT5720, providing more evidence for a key role for PKA in promoting nuclear translocation, but not further identifying cellular mechanisms. Our transfection experiments show that the basic region of C/EBP␦ is essential for nuclear import, as has been described for other C/EBPs (34) and other members of the basic-leucine zipper transcription factor family (48 -50). Our results additionally suggest a secondary function for the leucine zipper domain of C/EBP␦ in maintaining full nuclear expression, because proteins lacking this domain were partially distributed in the cytoplasm. The leucine zipper also may play a role in regulated nuclear expression of C/EBP␦, because we find that a modified C/EBP␦ lacking the basic segment is located in the nucleus in rat osteoblasts under basal conditions but in the cytoplasm after treatment of cells with H-89. The same protein is exclusively cytoplasmic in hFOB 1.19 cells that do not express C/EBP␦ endogenously. These observations in aggregrate indicate that dimerization may occur between transfected and endogenous C/EBP␦, potentially through the leucine zipper regions, and that a single NLS within the dimer was sufficient for nuclear localization.
The mechanisms responsible for maintaining C/EBP␦ in the nucleus after activation by PKA or for inducing its nuclear export are unknown. Our results show that C/EBP␦ did not undergo continual shuttling between subcellular compartments under basal conditions or after hormone treatment. Rather, continuous activity of PKA was required to retain C/EBP␦ in the nucleus of primary rat osteoblasts, because removal of hormonal stimulus led to rapid redistribution into the cytoplasm (t1 ⁄2 Ͻ 12 min). The pathways of nuclear export of C/EBP␦ involved CRM1, because inhibition of this receptor with leptomycin B (30, 31) caused prolonged retention of C/EBP␦ in nuclei after removal of hormone. The segment of C/EBP␦ that interacts with CRM1 is not known. The currently recognized consensus sequences for binding to the nuclear export receptor include closely spaced short stretches of leucine residues (51,52). No typical consensus sequence is found in C/EBP␦. To date, however, no functional studies have been performed to demonstrate direct interactions of C/EBP␦ with CRM1.
An intriguing question arising from our current and previous results is whether nuclear localization alone is sufficient for full transcriptional activity of C/EBP␦ or whether other modifications of the protein are required. As shown in this report, forced expression of C/EBP␦ in osteoblasts resulted in its accumulation in the nucleus and is sufficient to transactivate an IGF-I promoter-reporter gene, although treatment of cells with PGE 2 further increases the level of IGF-I promoter function (14,15). Thus, potentially more than one mechanism controls the transcriptional response of the IGF-I gene to PKA.
In summary, we have shown that stimulation of PKA in primary rat osteoblasts led to the rapid activation of the transcription factor C/EBP␦ through its regulated nuclear import. Nuclear targeting required the basic region of C/EBP␦ and was enhanced by the presence of the leucine zipper motif. Our results provide a framework for defining the specific cellular FIG. 9. One NLS is sufficient to translocate a C/EBP␦ dimer into the nucleus in osteoblasts. Immunocytochemistry for the flag epitope tag of hFOB 1.19 cells (no endogenous C/EBP␦ expression) and primary rat osteoblasts (rOB) transfected with the indicated C/EBP␦ deletion and truncation mutants and treated with vehicle (Ϫ) or 10 M H-89 for 2 h. machinery and molecular mechanisms by which hormones that activate cAMP induce the nuclear translocation of a critical transcription factor that regulates expression of IGF-I, a key gene product for bone growth and maturation.