INTRODUCTION

Parathyroid hormone (PTH)¹ represents a potent but complex effector of bone formation since it mediates catabolic as well as anabolic effects on bone in vivo, apparently related to its temporal pattern of administration. Continuous administration for short time periods decreases bone mass (1, 2, 3). However, intermittent administration over prolonged periods increases bone mass in normal and osteoporotic animals and humans (4, 5, 6, 7, 8). The precise mechanism of PTH in vitro appears controversial since PTH either stimulates or inhibits the growth of osteoblasts and chondrocytes and exerts differentiating as well as dedifferentiating effects on the osteoblastic phenotype (8, 9, 10, 11).

The PTH/PTHrP receptor has been cloned and characterized as a G protein-coupled receptor containing seven transmembrane domains (12, 13, 14). Activation of this receptor stimulates both the adenylate cyclase and the phospholipase C pathways. Receptor coupling to adenylate cyclase leads to the production of cyclic AMP and the activation of protein kinase A (PKA) (15, 16), while the phospholipase C-catalyzed hydrolysis of phosphatidylinositol 4,5-bisphosphate generates diacylglycerol and inositol 1,4,5-trisphosphate, resulting in transient elevations of the cytosolic calcium concentration and activation of protein kinase C (PKC), respectively (17, 18, 19).

Studies on structure-activity relations of PTH have revealed that some of the triggered cellular reactions are related to certain distinct domains on the PTH sequence. The hormone needs the first two N-terminal amino acids and its principal receptor-binding region (at positions 25-34) to stimulate adenylate cyclase (20, 21), but region 29-32 is sufficient to activate the inositol 1,4,5-trisphosphate/Ca²⁺/PKC pathway (22). Treatment with the mid-regional fragment PTH-(28-48) causes a cAMP-independent stimulation of DNA synthesis in chicken cartilage cells and rat osteoblasts in vivo and in vitro (23, 24). The responsible mitogenic core domain was localized to region 30-34 of the PTH molecule and uses the inositol 1,4,5-trisphosphate/Ca²⁺/PKC pathway. In addition, an increase in alkaline phosphatase activity was reported after treatment of different osteoblast-like osteosarcoma cells with the carboxyl-terminal fragment PTH-(53-84) (25). In summary, these observations led to the hypothesis that PTH is a polyhormone that is defined as a peptide containing several distinct regions, each possessing distinct biological activities (26).

The respective implications and contributions of the two above-mentioned signal transduction pathways (adenylate cyclase and phospholipase C) in mediating the alterations of gene expression observed in PTH-responsive cells are not well understood. An insight into the roles of PKA- and PKC-dependent signals in PTH action would require the selective activation of either one of the two messenger systems. Therefore, the purpose of this study was to examine whether or not the PTH fragments PTH-(1-34) (harboring the adenylate cyclase- and PKC-activating domains) and PTH-(28-48) (harboring the PKC-activating domain only) have distinct effects on gene expression in osteoblastic ROS17/2.8 cells.

We investigated the effects of the two PTH fragments on the expression of the two immediate-early genes c-fos and c-jun, which play important roles in regulating the expression of osteoblastic marker genes upon PTH application, and analyzed the response of osteoblastic marker genes upon PTH application. In addition, to identify further immediate-early genes involved in the signaling of the different PTH fragments, we have used subtractive cDNA cloning to characterize mRNAs expressed in ROS17/2.8 cells as a response to PTH induction.

MATERIALS AND METHODS

Culture Conditions and PTH Treatment

Osteoblastic rat osteosarcoma ROS17/2.8 cells were maintained in Dulbecco's modified Eagle's medium supplemented with 5% fetal calf serum in a humidified atmosphere of 95% air and 5% CO₂. The medium was changed every other day, and the cells were subcultivated once a week. For cAMP assays, cells were plated in 6-well plates at 5200 cells/cm²; for thymidine incorporation assays, cells were seeded in microtiter plates at a density of 12,000 cells/cm²; and for RNA isolations, cells were seeded at 6000 cells/cm² in tissue culture dishes with an area of 64 or 165 cm². Upon reaching confluence, cells were grown in serum-free medium for 24 h. The cells were incubated, if not otherwise stated, for the indicated time periods with PTH-(1-34) (300 nM), PTH-(28-48) (300 nM), forskolin (100 µM), or TPA (1 µM), respectively. Cycloheximide treatment started 30 min before the addition of the inducers at 100 µg/ml. The synthetic human peptides PTH-(1-34) and PTH-(28-48) were obtained from the Peptide Synthesis Unit (Gesellschaft für Biotechnologische Forschung, Braunschweig, Germany). The purity of the peptides was verified by high pressure liquid chromatography (C₁₈) and NMR spectroscopy.

Adenylate Cyclase Stimulation and cAMP Determination

Cells were cultured as outlined above. The cells were washed twice with serum-free medium and preincubated for 20 min at 37 °C in serum-free medium containing 1 mg/ml bovine serum albumin and 1 mM isobutylmethylxanthine (Sigma), which inhibits phosphodiesterase activity. Stimulation was performed by an additional 15-min incubation with the corresponding PTH fragment. Incubation was stopped by aspirating the medium and washing the cells once with ice-cold phosphate-buffered saline. Incubations were conducted in triplicates. cAMP was extracted with 1 ml of acidified ethanol (1.75 ml of HCl, 100 ml of ethanol) overnight at 20 °C. The extract was removed, evaporated to dryness, and redissolved in 200 µl of phosphate-buffered saline. Duplicates were assayed for cAMP by a competitive binding method using a commercially available [³H]cAMP radioassay kit (Amersham Buchler). Experimental values are expressed as picomoles of cAMP/well.

Thymidine Incorporation Assay

The rate of DNA synthesis was assayed in monolayer culture by the incorporation of [³H]thymidine into perchloric acid-precipitable material. Cells were cultivated in 200 µl of medium for 30 h in 96-well microtiter plates as described above. Then, the cells were washed once with serum-free medium, followed by an additional 30-h incubation in the latter medium. The indicated effector molecules were added, and incubation was continued for another 16 h. Radioactive DNA labeling was achieved by the addition of 10 µl of [³H]thymidine during the final 2 h of incubation. Subsequently, the medium was discarded, and the cells were lysed by 0.1 N NaOH. The DNA was precipitated by the addition of 2% perchloric acid for 30 min at 4 °C. The precipitated material was transferred to Ready Filter X-tal (Beckman Instruments) with a Scatron Model AS semiautomatic cell harvester, washed with deionized water, and dried at 70 °C. Precipitated radioactivity was determined with a -counter using the X-tal program (Beckman Instruments). Incorporated radioactivity was compared with that of untreated control cells.

Statistics

Concerning the thymidine incorporation assay, each test consisted of 10 identically treated wells and was repeated at least three times, as outlined above. All other data were obtained in duplicates or triplicates in at least three independent experiments. Data are presented in a quantitative fashion as the means ± S.E.

RNA Analysis

ROS17/2.8 cells were cultivated as described above. Cells were harvested at the indicated time intervals, and total RNA was isolated by guanidinium/CsCl step gradients. Total cellular RNA (10 µg) was separated electrophoretically on a 2.2 M formaldehyde, 1.2% agarose gel and transferred to nitrocellulose. Hybridization was carried out with nick-translated ³²P-labeled gene-specific DNA probes. The level of induction is the mean of at least three replicate experiments in relation to -actin mRNA levels to correct for loading variations. Quantification was performed with an imaging system (WinCam 2.0, Cybertech, Berlin).

Construction of cDNA Libraries

Poly(A)⁺ RNA (mRNA) isolated from untreated (control) ROS17/2.8 cells or from cells after treatment with PTH-(1-34) or PTH-(28-48) was reverse-transcribed into double-stranded cDNA by a commercially available kit (Gibco-BRL). cDNAs were inserted directionally into the phagemid pTZ18R or pTZ19R (Pharmacia, Uppsala) following the method of Dorssers and Postmes (27). Four cDNA libraries were constructed in this respect: three cDNA libraries in pTZ18R from PTH-(1-34)-treated, PTH-(28-48)-treated, or control ROS17/2.8 cells (target libraries in (+)-orientation) and one control (driver) library in pTZ19R (in ()-orientation). The cDNAs of the target libraries are in inverse orientation in regard to the cDNAs of the driver library. Libraries were established in Escherichia coli F (SURE) by electroporation. The primary cDNA libraries consisted of 3 × 10⁵ individual cDNA clones in the case of the control (driver) cDNA library in ()-orientation. The target cDNA libraries consisted of 4.8 × 10⁵ individual cDNA colonies in the case of PTH-(28-48) treatment and 4.2 × 10⁵ in the case of PTH-(1-34) treatment. 8.2 × 10⁵ individual colonies were obtained from the control (target) cDNA library in (+)-orientation.

Subtractive and Differential cDNA Screening

Isolation of induced genes was performed by a combined method using subtraction followed by differential screening. The strategy is outlined in Fig. 1.

Basically, a modified method of Duguid et al. (28) was used. For the generation of single-stranded cDNA, libraries in E. coli SURE were infected with the interference-defective helper phage M13K07. The resulting single-stranded libraries were purified by QIAGEN column chromatography (Diagen, Düsseldorf, Germany). Then, the single-stranded driver library in ()- orientation was photobiotinylated in two rounds with a commercially available kit (Gibco-BRL) following the manufacturer's directions. 2.5 µg of single-stranded target cDNA from the two PTH-treated libraries in (+)-orientation were subjected to a two-round subtraction from 25 µg of the biotinylated single-stranded driver DNA library in ()-orientation. The same procedure was applied to 2.5 µg of single-stranded cDNA from the control library in (+)-orientation, which was subtracted twice from 25 µg of the biotinylated control (driver) library. The latter step was essential for generating the control probe for the differential screening procedure following the subtraction (see below). Separation was achieved by streptavidin/phenol treatment. Moloney murine leukemia virus reverse transcriptase and the ``reverse primer'' complementary to a part of the lacZ gene in pTZ18R were used to convert the resulting single-stranded subtractive libraries into the double-stranded form. These double-stranded subtractive libraries were transferred into E. coli SURE by electroporation. The subtractive control library (from untreated control cells) consisted of 4.9 × 10⁴ individual clones. The sizes of the subtractive libraries from PTH-(28-48)- and PTH-(1-34)-treated cells were 3.6 × 10⁵ and 2.3 × 10⁵ individual clones, respectively, indicating a reduction factor of 10²/round of subtraction.

The subtractive libraries originating from the PTH-induced cells were replica-plated onto nitrocellulose sheets. In parallel, the cDNA clones from all subtractive libraries were amplified, and the recombinant plasmids were isolated. The cDNA inserts were excised by digestion with the restriction enzymes EcoRI and HindIII. Inserts ranging from 0.5 to 3 kilobase pairs were eluted from the gel and used as hybridization probes in the differential screening step to follow. Replica-plated colonies from the PTH-(1-34) and PTH-(28-48) subtractive libraries were analyzed by colony hybridization with nick-translated ³²P-labeled inserts either from the subtractive control library (representing the background of the system) or from the PTH-(1-34) or PTH-(28-48) subtractive library. Colonies exhibiting positive hybridization signals with the latter two probes were considered for further (Northern) analyses. Accession numbers for PTH-regulated cDNA sequences in the EMBL Data Library are X95079-X95094[GenBank][GenBank][GenBank][GenBank][GenBank][GenBank][GenBank][GenBank][GenBank][GenBank][GenBank][GenBank][GenBank][GenBank][GenBank][GenBank].

RESULTS

The purpose of this study was to elucidate cAMP/ PKA- as well as calcium/PKC-mediated gene expression after PTH treatment of ROS17/2.8 cells. Therefore, we analyzed and optimized the response of rat osteoblast-like ROS17/2.8 cells upon treatment with two PTH fragments, PTH-(1-34) (harboring the adenylate cyclase- and PKC-activating domains) and PTH-(28-48) (harboring the PKC-activating domain). A proliferative influence has been reported for both PTH fragments, but only the amino-terminal hormone fragment PTH-(1-34) is able to elicit a cAMP response (23, 29). Regarding cAMP synthesis, under our experimental conditions, only PTH-(1-34) was able to invoke a response (Fig. 3). We observed a PTH-dependent stimulation of thymidine incorporation by both hormone fragments. However, the amino-terminal fragment always exhibited a reduced capacity for DNA synthesis in comparison with the cAMP-independent mid-regional PTH fragment at comparable concentrations (Figs. 2 and 3). Application of PTH-(28-48) at a concentration of 100 nM led to a 3-fold induction of DNA synthesis, which is not significantly elevated at higher fragment concentrations (data not shown). On the basis of these results, we treated ROS17/2.8 cells with PTH-(1-34) and PTH-(28-48) at concentrations of 300 nM, i.e. under conditions where the biological responses regarding cAMP-independent mitogenicity and cAMP synthesis are optimal in ROS17/2.8 cells.

Depending on the cell type and environment, the immediate-early genes c-fos and c-jun can be activated by second messenger pathways that stimulate PKA activity (30, 31), PKC activity (32, 33), or an increase in the cytosolic calcium concentration (32). PTH is able to activate all of these signal transduction pathways (see above). Both PTH fragments were applied in the presence or absence of cycloheximide. After 30 min, PTH-(1-34) as well as forskolin caused an increase of 2 orders of magnitude in c-fos transcripts. The response of c-fos to PTH-(28-48) treatment was considerably lower, while the protein kinase C activator TPA failed to elicit a significant c-fos response in these cells (Fig. 4A). In contrast, the c-jun transcripts (3.2 and 2.7 kilobases) were already constitutively expressed during the non-induced state in ROS17/2.8 cells. Treatment with PTH fragments led to a moderate increase in c-jun transcripts only (~3-fold) (Fig. 4B). As observed for c-fos, this induction was mimicked by treatment with the adenylate cyclase activator forskolin, but not with TPA, suggesting that the main pathway for c-fos and c-jun activation in these cells involves PKA activation. Induction was also observed in the presence of cycloheximide (CHX), indicating that this response is independent of de novo protein synthesis (Fig. 4, A and B). It has also been documented that ornithine decarboxylase, which is an essential enzyme in polyamine synthesis and plays an important role in proliferation and differentiation, readily responds to PTH-(1-34) treatment in osteoblasts of various sources (34, 35, 36). PTH-(1-34) was able to increase the levels of ornithine decarboxylase transcripts alone and in the presence of CHX, which is in contrast to PTH-(28-48) (Fig. 5A) and which substantiates that ornithine decarboxylase is predominantly activated by the PKA pathway.

ROS17/2.8 cells express a high level of osteoblastic marker genes such as collagen 1(I), osteonectin, osteopontin, alkaline phosphatase, and the PTH/PTHrP receptor. The osteocalcin transcript, which is highly specific for late stages of osteoblastic development, was detected, albeit at a rather reduced level (Fig. 5A). Short-term application (4 h) of PTH-(1-34) and PTH-(28-48) in the presence or absence of CHX influenced only marginally the transcript levels for collagen 1(I), osteonectin, osteopontin, and the PTH/PTHrP receptor. In contrast, long-term application of PTH-(1-34) caused a down-regulation of the PTH/PTHrP receptor and alkaline phosphatase mRNA levels, while in this case, the level of osteocalcin mRNA was stimulated considerably (~5-fold) (Fig. 5B).

ROS17/2.8 cells exhibited all traits of an osteoblastic cell line under our experimental conditions. We monitored fast as well as slow responses upon PTH treatment in this system. To further analyze the direct response of these osteoblastic cells upon PTH-(1-34) and PTH-(28-48) treatment, we performed subtractive cloning.

PTH-(1-34) and PTH-(28-48) were applied in the presence of CHX for 4 h to subconfluent ROS17/2.8 cells. The presence of CHX during the PTH treatment should enable the isolation of mRNAs synthesized or stimulated by an immediate response in the absence of de novo protein synthesis. Furthermore, CHX exerts a stabilizing effect upon many unstable RNA species. The subtractive cDNA cloning of PTH-(1-34)- and PTH-(28-48)-inducible genes was performed with a modified phagemid system (28). This strategy was combined with a differential screening procedure as detailed under ``Materials and Methods'' and outlined in Fig. 1. A total of 288 clones were scored as positive after the differential screening step. 75 of these were chosen at random to be further analyzed by Northern analyses (30 from the PTH-(28-48)- and 45 from the PTH-(1-34)-induced subtractive libraries). 29 clones were positive in detecting mRNAs stimulated by CHX, by PTH, or by combined PTH/CHX treatment. 18 of these were induced by CHX treatment only, while nine were induced either by PTH alone or by combined PTH/CHX treatment. These nine clones were characterized further by sequencing (Fig. 6 and Table I).

Table I. Features of isolated cDNAs of enriched mRNAs after PTH treatment The position numbers of the clones refer to the subtractive library from which they were isolated. The level of induction is given as the level of induction/control. Values represent the means of three replicate experiments in relation to -actin mRNA levels to correct for loading variations. cDNA clone mRNA Induced by Level of induction after treatment with inducers PTH-(28-48) PTH-(28-48) + CHX PTH-(1-34) PTH-(1-34) + CHX CHX PTH-(1-34) PTH-(28-48) kba 1-34I-11 (phosphodiesterase) 4.5   +   +   3.8 1.6 1-34II-1 4.2   +   +   9.5 5.8 1-34II-5 (methyltransferase) 5.2 + + + + + 2.0 1.3 1-34II-10 4.3       +   8.8 0.9 28-48II-18 1.5   +   +   17 9.1 28-48II-22 3.8       +   5.4 0.9 28-48II-28 (PKC substrate) 4.4 + + + + + 5.1 2.6 28-48II-36 (GC-binding protein) 4.4 + + + + + 14.4 4.1 28-48II-43 (junB) 2   +   + + 3.1 2.5 a  kb, kilobase; +, induced; , non-induced.

Five cDNAs were identified by sequence homology. Clone 1-34I-11 corresponds to the gene encoding the cyclic phosphodiesterase (37), and clone 1-34II-5 corresponds to the gene coding for the (cytosine 5)-methyltransferase (38). Clone 28-48II-28 exhibits homology to a gene encoding an 80-kDa protein kinase C substrate (39). These cDNAs have been identified by homology to human cDNA sequences. The cDNA clone 28-48II-36 represents the rat version of a cDNA encoding a novel GC-binding protein isolated from a murine cDNA library.² Another clone (28-48II-43) harbors part of the rat immediate-early response gene junB. Four cDNAs for PTH-activated mRNAs are not functionally characterized at present (1-34II-1, 1-34II-10, 28-48II-18, and 28-48II-22; EMBL Release 42.0) (Fig. 4 and Table I).

The isolated cDNAs harbor only a relatively small part of the coding sequence (300-700 base pairs). Prior to subtractive cloning, a size selection was performed (800 base pairs); therefore, it seems that the single-stranded DNA intermediate is relatively sensitive to the various steps of DNA manipulations performed during the subtractive cloning routine. Based on the cDNA isolates described here, strong to moderate responses may be monitored upon PTH treatment in ROS17/2.8 cells (Fig. 4). Strikingly, in all cases reported here, the ability of PTH-(28-48) to induce genes was limited in comparison with PTH-(1-34). This feature is independent of the subtractive library from which the cDNA was originally isolated (the position numbers in the clones refer to the original subtractive library). Moreover, we were unable to identify any cDNA that was activated by PTH-(28-48) alone. In Northern analyses, several mRNAs were detected only after combined PTH/CHX treatment of ROS17/2.8 cells, indicating the involvement of unstable regulatory factors for gene induction or RNA stability. Alternatively, a cross-talk between the various signaling pathways for PTH-mediated activation and the CHX-activated stress-activated protein kinase pathway (41) could be imagined, leading to higher transcript levels for some PTH-activated genes.

DISCUSSION

PTH is able to activate at least two different intracellular transduction pathways in osteoblasts (42). These are the cAMP-dependent PKA (cAMP/PKA) (43, 44) and the calcium/protein kinase C (Ca²⁺/PKC) (19, 45) pathways. The cAMP/PKA-activating domain is located in the N-terminal domain of PTH (20, 21), while a PKC-activating domain is located mid-regionally (22). It was our goal to investigate the immediate response of the osteoblastic cell upon PTH treatment. Osteoblastic target genes that are activated by PTH treatment in a direct manner have to be evaluated concerning a potential contribution to the contrasting action of PTH and the signaling pathway by which they are activated. We addressed this problem by use of two PTH fragments activating different signaling cascades and the isolation of target genes that may respond in a direct manner to these two PTH domains. In addition, we examined the response of the immediate-early genes c-fos and c-jun as well as of osteoblastic marker genes upon application of either PTH-(1-34) (bearing the cAMP/PKA- and Ca²⁺/PKC-activating domains) or PTH-(28-48) (harboring the Ca²⁺/PKC-activating domain only).

PTH-(1-34) treatment of ROS17/2.8 cells activates the immediate-early response genes c-fos and c-jun. PTH-(28-48) is also able to activate these genes, albeit to a far lower extent. PTH-mediated induction in the presence of CHX indicates that de novo protein synthesis is not required to activate these genes (Fig. 4). To delineate the signaling cascades involved, the adenylate cyclase activator forskolin and the PKC activator TPA were used. In the osteoblastic ROS17/2.8 cells, only treatment with forskolin resulted in transcript levels of c-fos and c-jun comparable to those induced by PTH-(1-34) treatment, indicating that mainly the cAMP/PKA cascade is responsible for the PTH-mediated response. This response seems to be independent of the TPA-inducible PKC (Fig. 4). The c-fos promoter contains several elements mediating cAMP responsiveness including one consensus cAMP response element (46, 47). The c-jun promoter also contains sequence elements conferring sensitivity to cAMP (30). In conclusion, these results show that PTH increases c-fos and c-jun transcript levels in ROS17/2.8 cells by a mechanism that is mimicked by forskolin, but not by TPA. The cAMP-independent fragment PTH-(28-48) is only able to elicit a weak c-fos and c-jun transcript level response. This could be due to a PTH-(28-48)-induced increase in intracellular Ca²⁺ levels. These results are in line with an investigation of Clohisy et al. (48), who also showed that PTH-mediated induction of c-fos and c-jun in another osteoblastic cell line (UMR106-01) is mainly a cAMP-dependent event. It has also been documented that ornithine decarboxylase, the rate-limiting enzyme in polyamine synthesis, is regulated by PTH-(1-34) in osteoblastic cells (34, 35, 36). Only PTH-(1-34) was able to significantly enhance the level of ornithine decarboxylase transcripts in the presence or absence of CHX (Fig. 5A). The osteoblastic marker genes investigated (Fig. 5A) do not significantly respond to short-term treatment with either PTH fragment in the presence or absence of CHX. However, treatment of ROS17/2.8 cell with PTH-(1-34) for longer time periods influences the transcript levels of several osteoblastic marker genes. An effect for PTH-(28-48) could not be monitored (data not shown). While the alkaline phosphatase and PTH/PTHrP receptor mRNAs are down-regulated after 2 and 3 days of PTH-(1-34) treatment, respectively, the level of osteocalcin transcripts is enhanced considerably. For alkaline phosphatase, this is in the range of the reduction of enzymatic activities (30-40%) (49, 50). For the PTH/PTHrP receptor, however, a PTH-mediated down-regulation of up to 95% is mainly a post-translational event and cannot be explained by the reduction of transcript levels as observed here (51). The considerable enhancement of osteocalcin levels in osteoblastic cells (~5-fold) has been attributed to PTH-mediated cAMP-dependent post-transcriptional mechanisms (Fig. 5B) (52). These results are in agreement with other studies that show that most osteoblastic marker genes respond to long-term application of hormones and factors like retinoic acid, vitamin D₃, and fibroblast growth factor.

The five cDNAs that were identified by their sequence homology to known sequences seem to be involved in the PTH-mediated signaling pathways, so the clone coding for the 61-kDa calmodulin-stimulated cyclic nucleotide phosphodiesterase (clone 1-34I-11) was assigned to this gene due to its homology in the 3-untranslated region. This phosphodiesterase was originally isolated from a bovine brain DNA library (53) and belongs to a family of phosphodiesterases that are activated by calcium and calmodulin and antagonize a variety of physiological processes regulated by cAMP or cGMP. Stimulation of the PTH/PTHrP receptor also increases phosphoinositide turnover and intracellular calcium levels (see above), so it seems that PTH-mediated enhanced levels of calmodulin-stimulated phosphodiesterase may be part of the machinery that desensitizes the cell to further PTH treatment via negative feedback by decreasing the cAMP release. In addition, activation of calmodulin-stimulated phosphodiesterase may augment biological responses mediated by the Ca²⁺/PKC pathway by decreasing intracellular cAMP as suggested in other investigations (54). PTH also exerts proliferative action upon osteoblastic cells. The up-regulation of a methyltransferase could be a necessary component of this feature. Clone 1-34II-5 corresponds to the 3-untranslated region of the human DNA (cytosine 5)-methyltransferase. This enzyme modifies cytosine in the 5-C position (38), and important functions regarding gene expression and cell proliferation have been suggested (40, 55). Another functionally uncharacterized protein is encoded by clone 28-48II-28, corresponding to a PKC substrate that was originally cloned from a human cDNA library (39), and in this respect, this gene may also be activated to play a role in the PTH-mediated Ca²⁺/PKC signaling pathway. That PTH activates directly several transcription factors has already been discussed above. We isolated two cDNAs for two PTH-up-regulated mRNAs encoding transcription factors. junB (clone 28-48II-43) is another potential member for AP-1. The other transcription factor has recently be characterized in an embryonic carcinoma line (clone 28-48II-36). This factor reconstitutes a novel GC-rich binding protein related to SP1.² It is activated in direct response to PTH. In addition, this factor is strongly up-regulated by the presence of CHX alone (data not shown). The characterization of these two factors as PTH-responding genes is, as discussed above, another implication for AP-1 as an important factor in PTH-dependent gene expression and for more transcription factors being involved in the manifestation of the complex PTH-dependent effects in osteoblasts. The other cDNAs encoding genes that respond to PTH in the presence of CHX in a direct manner are currently under investigation.

In conclusion, in osteoblast-like ROS17/2.8 cells, PTH-(1-34) (bearing the cAMP/PKA- and Ca²⁺/PKC-activating domains) is the dominant biologically active compound, and its effects can be demonstrated after long- and short-term applications. PTH-(28-48) exerts only minor effects. This finding was substantiated by the isolation of various cDNAs that detect mRNAs responding to PTH treatment. So far, the N-terminal fragment harboring the cAMP/PKA domain is superior for gene induction in comparison with the Ca²⁺/PKC domain located on PTH-(28-48). We also did not identify a gene that is induced by the latter fragment alone, indicating that the major pathway for PTH-mediated gene activation involves the cAMP/PKA pathway. In this respect, it should be pointed out that, in spite of many studies that show alternative signaling pathways for PTH in bone and kidney cells, up to now, no significant biological effect of PTH could exclusively be attributed to the Ca²⁺/PKC pathway. For short-term PTH application, the latter may therefore constitute a minor pathway able to modify, but not significantly change, the biological action mediated by the dominant cAMP/PKA cascade.