Identification of a hormonally regulated protein tyrosine phosphatase associated with bone and testicular differentiation.

Absence of the tyrosine kinase activity of c-src and c-fms results in impairment of bone remodeling. Such dysfunction underscores the importance of tyrosine phosphorylation, yet the role of protein tyrosine phosphatases in bone metabolism remains unexamined. We have isolated the cDNA for a novel receptor-like tyrosine phosphatase expressed in bone and testis named osteotesticular protein tyrosine phosphatase (OST-PTP). The deduced 1711-residue protein possesses an extracellular domain with 10 fibronectin type III repeats and a cytoplasmic region with two catalytic domains. In primary rat osteoblasts, the 5.8-kilobase OST-PTP transcript is up-regulated in differentiating cultures and down-regulated in late stage mineralizing cultures. In addition, a presumed alternate transcript of 4.8-5.0 kilobases, which may lack PTP domains, is present in proliferating osteoblasts, but not detectable at other stages. Parathyroid hormone, a modulator of bone function, as well as cyclic AMP analogues, increase OST-PTP mRNA 5-8-fold in UMR 106 cells. In situ hybridization of adult rat testis revealed stage-specific expression of OST-PTP. OST-PTP may function in signaling pathways during bone remodeling, as well as serve a broader role in cell interactions associated with differentiation in bone and testis.

mutation in the gene encoding its ligand, macrophage colonystimulating factor, resulting in the lack of biologically active macrophage colony-stimulating factor necessary for osteoclast maturation (5). The osteoblast, which secretes macrophage colony-stimulating factor, is also dependent on tyrosine phosphorylation of cellular proteins for bone matrix formation and coupled modulation of osteoclast function. Osteoblast proliferation and differentiation requires cell signaling through a number of receptor protein tyrosine kinases such as epidermal growth factor (6) and insulin-like growth factors (7).
To gain a better understanding of the specific roles of tyrosine phosphorylation in bone metabolism, the activity of the protein tyrosine phosphatases (PTPs),' as well as these protein tyrosine kinases, must be considered. The PTPs represent a diverse family of enzymes that serve as critical signal transduction proteins in cell division, proliferation and differentiation (8). These enzymes are multidomain proteins whose structural features separate them into two main groups: the transmembrane or receptor-like PTPs and the intracellular PTPs. Each protein possesses at least one 230-residue catalytic domain with the consensus motif, (W)HCXAGXXR(S/'I')G ( X = any amino acid), which bears no resemblance to that of the serinekhreonine or the alkaline or acid phosphatases (8). The receptor-like PTPs are characterized by their highly divergent extracellular domains which include glycosylated segments (human PTPa), tandem repeats of immunoglobulin or fibronectin type 111 domains similar to cell adhesion molecules (PTPp), or alternately spliced lengths of sequence containing Nand O-linked carbohydrates (CD45). The activity of these receptorlike PTPs may be regulated by ligand binding to these extracellular domains. The intracellular PTPs possess a diversity of sequences outside the catalytic domain which are thought to target these proteins to cellular membranes, to the cytoskeleton and the nucleus (9).
Our knowledge of the relevance of protein tyrosine phosphatases in bone metabolism is very limited. In osteoblast cultures, inhibition of PTP activity using orthovandate appears to enhance osteoblast proliferation and matrix formation (10). PTP activity has been previously described in osteoblasts and enzymes have been partially purified that exhibit activity toward phosphotyrosine substrates (11-13). However, neither the identity of these PTPs nor their function is known. In this report, we describe the isolation of a novel receptor-like PTP, designated as osteotesticular PTP (OST-PTP), whose expression appears to be restricted to bone and testis. Interestingly, 'The abbreviations used are: PTP, protein tyrosine phosphatase; OST, osteotesticular; PCR, polymerase chain reaction; R H , parathyroid hormone; GST, glutathione S-transferase; CAM, cell adhesion molecule; pNPP, p-nitrophenylphosphate; DPBS, Dulbecco's phosphatebuffered saline; FN, fibronectin; PKA, cyclic AMP-dependent protein kinase; bp, base paids); kb, kilobase paids); RACE, rapid amplification of cDNA ends.

30659
~~~e~ Regulated PTP Specific to Bone and Testis the expression of OST-PTP is regulated during osteoblast differentiation and following parathyroid hormone stimulation. In addition, this PTP shows stage-specific expression during spermatogenesis in rat seminiferous tubules. Our studies suggest that OST-PTP is a unique member of the PTP family which may serve a critical function in cell signaling during bone remodeling, EXPERIMENTAL PROCEDURES Isolation of Putative PTP Clones from Bone-Three sets of degenerate oligonucleotide primers were designed to conserved regions within the catalytic PTP domain. These primers were used in reverse transcription PCR to amplify novel sequences from bone cells. Three different 5' primers corresponding to the conserved amino acids, DYINA and XbaI, respectively. The template for PCR reactions was first-strand cDNA synthesized using poly(A+) RNA isolated from the rat osteosarcoma cell line UMR 106 and fetal rat calvaria and random hexamer primers. The cDNA synthesis reactions were performed with the Invitrogen cDNA cycle kit, using recommended man~acturer's instructions. The PCR reactions included 0.2 to 0.1 volume of the original CDNAsynthesis reaction and 500 ng of each primer, along with the recommended reagent concentrations in the GeneAmp kit (Perkin-Elmer), The PCR conditions were: 94 "C, 1 min; 55 "C, 1 min; 72 "C, 1 min; for 35 cycles. For PCR from calvaria, reamplification of original PCR reactions was necessary to see products. Products oftheexpectedsize,usingtheDYINApairf"495bpjandthe~TQGPpair (-432 bp), were isolated by gel electrophoresis, digested with BamHI and &I and cloned into the pBluescript I1 vector (pBS II; Stratagene). Despite attempts to vary PCR conditions, no products were seen using the KC-DQYWP primer. Approximately 200 independent clones were sequenced by the dideoxy chain termination method, using the Sequenase kit (U.S. Biochemical Corp.). Isolation of cDNA C1one"Two rat UMR 106 cDNA libraries were screened to isolate the full-length clone of OST-FTP. One of these libraries was constructed from poly(A+) RNA of normal confluent UMR 106 cells using the A Zap 11 library const~ction kits (Stratagene). The second library was a custom A ZAP Stratagene library constructed with poly(A+) RNA from PTH-stimulated UMR 106 cells (100 OM PTH for 18 h), using both oligo-d(T) and random hexamer primers. The original PCR fragment, corresponding to OST-PTP (i.e. DYll), was labeled by random priming and used as a hybridization probe for library screening (14). Initial screenings of these libraries yielded incompIete, poly(A)tailed clones, representing approximately 3.2 kb of the predicted 5.8-kb RNA transcript. To obtain the 5' end of the clone, 5' RACE (15) was performed using the 5' RACE kit (Life Technologies, Inc.). Two rounds of RACE PCR were necessary to isolate the remainder of the full-length cDNA. To verify the authenticity of this sequence, an 870-bp probe (random-prime labeled) corresponding to one of the RACE products was used to re-screen the custom Stratagene library. Several clones, ranging in size from 1.5 to 2.0 kb, were isolated and found to span the collective sequence of the RACE products. The final sequence of the full-length was verified by sequencing of both strands using an automated fluorescent DNA sequencer (Applied Biosystems, Inc.).
RNA Isolation and Northern Analysis-Total RNA was isolated from cells and tissues by centrifugation through a cesium chloride cushion (16) or using the RNazol B reagent (Tel-Test). Polyadenylated RNA was isolated using the Poly(A)Ttract kit (Promega). Polyadenylated RNA from spleen, smooth muscle, pancreas, and retina was purchased from Clontech. Samples (2-10 pg) were resolved in a formaldehyde-agarose gel (14) and transferred to a Nytran membrane (Schleicher & Schuell). Prehybridization (at least 5 h) and hybridizations (overnight) were conducted at 42 "C in 5 x SSC, 5 x Denhardt's solution, 0.1% SDS, 50% formamide with 250 pg/ml of tRNA. Stringent washes were performed at 65 "C, 0.  prime labeled (Boehringer Mannheim biochemicals kit) DNA probes corresponding to these sequences were used, as indicated for each figure. Controls probes for RNA integrity and concentration were human @-actin and rat cyclophilin. Quantitative determination of mRNA was determined using a Betagen scanner or a PhosphorImager. Relative cpms for each band were calculated and normalized for RNA concentration using cyclophilin units.
Expression and Characterization of Recombinant OST-PTP Protein-The recombinant OST-PTP protein was produced as a glutathione Stransferase (GST) fusion protein using the pGEX-KG Escherichia coli expression vector (17). A blunted NotI-XbaI fragment corresponding to the entire cytoplasmic domain of OST-PTP was subcloned into SmaI-XbaI digested pGEX-KG vector. For protein expression, this construct, pKG-OST, was freshly transformed into the BL21 strain of E. coli and the GST-OST PTP fusion protein produced and purified as described previously (17). The recombinant fusion protein was approximately 70% pure as determined by SDS-polyacrylamide gel electrophoresis and Coomassie Blue staining.
Basic kinetic analysis of the recombinant protein was performed using the hydrolysis of the artificial substrate, p-nitrophenylphosphate (pNPPf, as the measure of phosphatase activity. Assay for hydrolyzed p W P were conducted as described previously 118). The extinction coefficient of 1.8 x lo4 M-' cm" was used to determine the molar concentration of hydrolyzed pNPP. Subsequent Michaelis-Menton analysis of data was performed using KinetAsyst software (IntelliKinetics, State College, PA).
Substrate specificity was determined using tyrosine-phosphorylated Raytide (Oncogene Sciences) and serine-phosphorylated Kemptide (Sigma). Raytide was phosphorylated on tyrosine by incubating 10  of 120 pl of 10% phosphoric acid and then applied to P81 filters. Free ATP was removed by 3 x 100-ml washes with 0.5% phosphoric acid, and the peptide was eluted with 2 x 1-ml aliquots of 0.5 M ammonium bicarbonate. Aliquots were lyophilized and resuspended in distilled water, and radioactivity was determined. Phosphatase activity was determined in a 50-pl reaction containing assay buffer (0.1 M imidazole (pH 5.6-7.01, 5 m M EDTA, 0.2% ~-mercap~ethanolj, 100,000 cpm of phosphorylated substrate, and 100 ng to 4 pg of enzyme, incubated at 30 "C for 10 min. The reaction was terminated by addition of 750 pl of a charcoal mix (0.9 M HCl, 90 m M sodium pyrophosphate, 2 mM sodium phosphate dibasic and 4% (dv) Norit). The resulting supernatant after centrifugation was counted to determine the amount of 32P (counts/min) released.
Culturi~g of Primary Rat Ost~bZasts-Primary rat osteoblasts were isotated using a modification of the procedure of Aronow et al. 119).
Cells were plated at 1.2 x lo6 celld75 cmz flask in essential modified Eagle's medium containing 10% fetal bovine sera and 1 x penciuin, streptomycin, and amphotericin and maintained in a water-jacketed, humidified CO, (5-7%) incubator. At 5 days post-plating, the medium was replaced with BGJb medium containing 10% fetal bovine serum with the additives, ascorbic acid (50 pg/ml) and 0-glycerol phosphate ( protein tyrosine phosphatases amplified from bone. A, "' P random-prime labeled PCR fragments corresponding to the DY 5, DY 6, and OST-PTP clones were used as probes to hybridize a rat multitissue Northern (left panel, 2.5 pg poly(A+) RNMane; Clontech) and a blot of polv(A') RNA from UMR 106 osteosarcoma cells (right panel, 5 pg). RNA transcripts of the DY5 and DY6 clones were expression of markers did not vary more than 48-72 h.
Culturing of Osteosarcoma Cells-The rat osteosarcoma cells (UMR 106) were obtained from American 'Qpe Culture Collection. Cells were maintained in Dulbecco's modified Eagle's medium (high glucose) containing 10% fetal bovine sera at 37 "C in a water-jacketed, humidified C02 (5-7%) incubator. Cells were subcultured at confluence (approximately every 3 days) into multiple 75-cm2 tissue culture flasks a t a density of 5 x lo5 cells/flask. At least 24 h prior to hormonal treatment, cells a t 60-70% confluency were rinsed and treatment medium (Dulbecco's modified Eagle's medium, 0.3% fetal bovine seum), without hormone/drug, was added to the flasks. Treatments with rat PTH-1-34 (Peninsula Labs) or chlorophenylthiol-CAMP (Boehringer Mannheim) were conducted at times and concentrations indicated. Total RNA was isolated by centrifugation through a cesium chloride cushion (16) and polyadenylated mRNA using the PolyATtract system (Promega). The cDNA probe encoding rat collagenase, kindly provided by Dr. Nicola Partridge, was used to verify biological activity of substances and consistency of culturing methods.
In Situ Hybridizations of Adult TestisSprague-Dawley male rats (200-250 g) were anesthetized with sodium pentobarbital and unperfused testes were collected, frozen in Tissue-Tek embedding medium (Miles) on dry ice and stored at -80 "C. Tissue sections (10-15 pm) were cut at -15 "C and thaw-mounted on poly-L-lysine slides. Sections were stored a t -80 "C until use. Double-labeled "S-UTP and -CTP) riboprobes were prepared using the Maxiscript (T3/"7/SP6) in vitro transcription kit (Ambion) with minor modifications to the manufacturer's protocol. The template for the antisense and sense riboprobes were BarnHI linearized and XbaI linearized pBS I1 containing the original 470 bp PCR fragment. Prior to hybridization, sections were warmed to room temperature, fixed for 1 h in 4% paraformaldehyde, and washed twice with 2 x SSC. RNase control slides were treated for 1 h a t 37 "C with 200 pg/ml RNase A dissolved in 10 mM " i s (pH 8.0) and 0.5 M NaCI. All sections were acetylated to reduce background by incubation a t room temperature for 10 min. in 0.1 M triethanolamine (pH 8.0) and 0.25% acetic anhydride. Following washes in 2 x SSC, sections were dehydrated in a series of ethanol washes and air dried. Hybridizations were conducted at 55 "C overnight in hybridization buffer (50% formamide, 10% dextran sulfate, 3 x SSC, 50 mM sodium phosphate (pH 7.4), 1 x Denhardt's solution, 0.1 mg/ml yeast tRNA, 10 mM dithiothreitol) containing 1-2 x lofi cpm of riboprobe/slide. Sections were washed in 2 x SSC, RNase treated as above and final stringent washes conducted at 65 "C in 0.1 x SSC, 0.1% SDS for 1 h. Following this incubation, slides were rinsed in 0.1 x SSC, dehydrated with serial ethanol washes and air dried. Slides were exposed to Hypefilm (Amersham) for 5-7 days to determine quality. Appropriate slides were dipped in Kodak photographic emulsion (NBT-2) and exposed for 14 days. Sections were developed and counterstained with hematoxylin-eosin (Richard Allan, Inch Slides were viewed and photographed on a Zeiss Axioskop with a Microvideo darkfield illumination system.  (24)). Three of the 13 PCR fragments encoded novel PTPs, referred to as DY 5 , 6, and 11. Northern analysis, using these PCR fragments as probes, revealed that DY 5 and 6 were expressed in numerous tissues (Fig. L4). The DY 5 transcript (8.1 kb) was found in highest concentration in bone (osteosarcoma) and brain as well as in lung, liver and kidney. The DY 6 transcript (6.3 kb) was expressed highest in bone, brain and heart with moderate to low expression in lung, liver and skeletal muscle. In contrast, the DY 11 transcript was found exclusively in bone and testis (Fig. JA). This 5.8-kb transcript is in low abundance in comparison to the other novel clones, requiring at least 2-3 pg of polyadenylated RNA and 2-3-fold longer exposure times (autoradiogram or scanner) to visualize the transcript. Additional Northern analysis using approximately 10 pg of polyadenylated RNA and high specificity riboprobes further verified this tissue expression (Fig. U?). The isolation of the cDNA clone for DY 11, now called osteotesticular protein tyrosine phosphatase (OST-PTP), was pursued in light of this restricted tissue distribution.
Structural Features of OST-PTP-The full-length OST-PTP cDNA is 5455 bp in length, with a single open reading frame of 1711 amino acid residues starting from the ATG at bp 205. The invariant adenine at the -3 position of the Kozak consensus sequence (25) is present upstream of this codon, and an adenine exists a t position +4. The sequence following this putative initiating methionine encodes a typical signal peptide (261, containing the positively charged n-region, the hydrophobic hregion and the expected peptidase cleavage site (Fig. 2, underlined). An alternative ATG does exists 74 bp upstream of this methionine but the predicted translation from this codon fails to encode a putative signal peptide and a terminator codon exists just prior to bp 205. There is a 204-bp 5"untranslated sequence and a 115-bp 3"untranslated sequence containing the consensus polyadenylation signal and a poly(A) tail. The nucleotide sequence of OST-PTP has been deposited to Gen-BankTM (accession no. L36884).
This cDNA is predicted to encode a novel transmembrane  PTP possessing an extracellular domain (1068 amino acids in length), a hydrophobic membrane spanning region (33 amino acids) and a cytoplasmic region (610 amino acids) containing two PTP domains (Figs. 2 and 3 A ) . The extracellular domain contains a signal peptide (17 amino acids) and 10 fibronectin type 111-like (FN-111) domains which share 21% sequence identity with fibronectin (27) and human PTPP (20) (Fig. 2). These FN-I11 domains possess the conserved residues characteristic of FN-I11 repeats found in numerous cell adhesion molecules (CAMS) and extracellular matrix proteins (Le. tenascin) as well as other receptor-like PTPs (8,28). Homophilic binding of such structural domains in PTPp and PTPK can mediate cell-cell aggregation (29,30). In addition to the FN-I11 domains, there are 16 potential N-glycosylation sites scattered throughout the extracellular portion of OST-PTP, as indicated in Fig. 2. The cytoplasmic region of OST-PTP has two tandemly repeated phosphatase domains designated domain I and I1 (Fig.  2). Domain I is a typical 250 residue catalytic domain, containing the characteristic 11 conserved residues at the active site of the enzyme, with 45% sequence identity to human PTPP (Fig.  2). In contrast, domain I1 has a highly divergent active site but possesses many of the conserved motifs surrounding this site, as shown in the sequence alignment in Fig. 3B. The active site of domain I1 lacks the invariant cysteine necessary for catalysis, retaining only the histidine and glycine residues and sharing 42% identity with domain I1 of CD45 (31) (Figs. 2 and 3B). Such divergent PTP domains are presumed to be inactive and are found in a limited number of transmembrane PTPs (8).

T T W P D H S V P E A P S S L L A F V B L V Q E Q V Q A T Q G K G P I L~~~~~~~~~~~T F V A L L R L L R Q L E B E K V
A schematic diagram showing the overall structural features of the predicted OST-PTP protein is shown in Fig. 3A. Overall, OST-PTP has the highest degree of similarity to the human PTPP, sharing 45% sequence identity within the entire coding region. This low homology to other PTPs as well as the presence of multiple FN-111 repeats in the extracellular domain of this phosphatase indicate that OST-PTP is a new member of the type I11 class of receptor tyrosine phosphatases Kinetic Analysis of Recombinant OST-PTP-To verify the phosphotyrosine specificity of OST-PTP and determine kinetic parameters, the cytoplasmic region of the molecule (amino acids 1112-1711) was expressed in E. coli as a GST fusion protein (GST-OST). Initial kinetic analysis of phosphatase activity was performed with the artificial substrate, pNPP, The pH optimum of GST-OST for pNPP hydrolysis was approximately 5.6 (Fig. 4A). This value is very similar to pH optimums for other receptor-like PTPs. The neuronal specific PTP NE-3 is also reported to exhibit optimal activity at 5.6 (32) and recombinant rat LAR has a pH optimum of 5.0 (18). Michaelis-Menton analysis of GST-OST activity toward pNPP showed a K , of 0.52 mM.
Using an estimate of 70% purity for the fusion protein (based on SDS-polyacrlyamide gel electrophoresis and Coomassie Blue staining), a keat value of 41 s-l was obtained. These K,,, and kc,, values are similar those reported for other PTPs, using pNPP as a substrate (33). Substrate specificity was determined using tyrosinephosphorylated Raytide and serine phosphorylated Kemptide as substrates and monitoring dephosphorylation as a function of 32P released. The GST-OST protein readily dephosphorylated Raytide at concentrations as low as 50 ng, yet could not dephosphorylate Kemptide at any protein concentration (Fig.  4B). This dephosphorylation of Raytide was virtually elimi- nated by inclusion of 5 mM sodium vandate in the assay reaction (Fig. 4 0 . Expression of OST-PTP mRNA during Osteoblast Differentiation-Bone remodeling involves complex cell-cell and cellmatrix interactions which dictate the differentiation and function of the osteoblast and osteoclast. Because of the structural similarity of OST-PTP to the CAMs, we hypothesized that this protein might function in signaling pathways important in osteoblast differentiation. Established methods for culturing fetal osteoblasts provide a n excellent experimental paradigm to study the differentiation of the osteoblast phenotype and the associated extracellular matrix (34). In these primary osteoblasts, the regulated expression of genes important for cell growth, osteoblast function, and matrix formation can be monitored during progression through the stages of proliferation, differentiation and mineralization of the bone matrix. This system was employed to examine the expression of OST-PTP mRNA in relation to other regulated genes. Primary osteoblasts from the calvaria of 21-day fetal rats were isolated and maintained in culture.
Polyadenylated RNA was harvested from proliferating (day 41, differentiated (day 181, and mineralizing (day 35) cultures. Northern analysis was conducted using riboprobes corresponding to the PTP domain I ( D Y l l ), the entire cytoplasmic region or a portion of the extracellular domain (OST-PTP,,). A representative experiment is pictured in Fig. 5 . The transition from proliferation ( P ) to differentiation ( D ) is marked by a 3.7-fold increase in the 5.8-kb OST-PTP transcript at day 18. In independent experiments, this increase in OST-PTP mRNA varied between 4-6-fold in differentiating cultures at days 15-18. Levels of alkaline phosphatase mRNA, a marker of the mature osteoblast, paralleled the changes in OST-PTP mRNA, increasing 8-fold above the expression of proliferating cultures. The further progression from differentiation to matrix mineralization ( M ) was also marked by a dramatic change in OST-PTP expression (Fig. 5 ) . Mineralizing cultures a t day 35 showed no detectable expression of the 5.8-kb OST-PTP transcript. A 50% decrease in alkaline phosphatase mRNA to a level only 4-fold of the proliferating cultures was seen.
This increase and decline of alkaline phosphatase is the expected pattern of expression in cultures spanning the stages of proliferation to late, heavy mineralization (34). If cultures are harvested earlier in the mineralization stage, the expression of both OST-PTP and alkaline phosphatase mRNA remains elevated, similar to that observed in differentiating cultures. Therefore, OST-PTP and alkaline phosphatase consistently exhibited parallel trends in expression during the development of the osteoblast phenotype. In addition, the other markers of osteoblast function (histone H4, type I collagen, and osteocalcin) exhibited the expected fluctuations during each stage (data not shown). These results indicate that the expression of OST-PTP in primary osteoblasts is tightly regulated in a manner similar to genes known to be critical in osteoblast differentiation.
In addition to the increased expression of the 5.8-kb transcript during differentiation, our analysis revealed the existence of an additional RNA transcript which appears to be down-regulated during the developmental progression (Fig. 5). Hybridization with the OST-PTP,, probe reveals two transcripts in primary rat osteoblasts, the 5.8-kb transcript as well as a smaller 4.8-5.0-kb transcript. This smaller transcript (lower arrow in Fig. 5 ) is abundant in proliferating cells, yet is essentially nondetectable in differentiating and mineralizing cultures. Probes to a portion ( D Y l l , upper panel in Fig. 5 ) or the entire cytoplasmic domain (data not shown) hybridize only to the larger transcript. In addition, the smaller transcript does not appear to be expressed in UMR 106 osteosarcoma cells or in whole adult testis (data not shown) regardless of the hybridization probe. These results suggest that the smaller transcript may encode an isoform which lacks one or both of the PTPase domains. Evidence of such PTP isoforms has been reported for the neuronal specific, receptor-like RPTPUP which appears to have an extracellular splice variant encoding a soluble protein lacking the PTP domains (35,36). This protein, named phosphacan, is thought to be a brain proteoglycan which interacts with neural CAMs and modulates neurite outgrowth. Definitive proof that this smaller RNA is an alternate transcript encoding an non-signaling isoform of OST-PTP awaits the isolation and characterization of this RNA species. This work is currently in progress.
Modulation of OST-F'TP Expression by Parathyroid Hormom-Because OST-PTP is regulated during osteoblast differentiation, its expression could also be modulated by factors known to influence osteoblast k c t i o n . PTH, a major calciotrophic hormone, is a potent modulator of bone remodeling that exhibits both catabolic effects to enhance bone resorption (37) and anabolic effects to augment bone formation (38). PTH modifies the gene expression of critical enzymes such as alkaline phosphatase (39) and collagenase (40), along with matrix and matrix-associated proteins such as osteocalcin (41). Experiments with the osteoblast-like cell  (Fig. 6A). In contrast, PTH treatment (100 n~ rat PTH-1-34, 18 h) resulted in an increase in the OST-PTP transcript, with no significant effect on the expression of the other putative PTP clones (Fig. 6A). Within 4 h of PTH treatment, a 5-fold increase in OST-PTP mRNA was seen, with further increases at later time points (Fig. 6B). PTH is known to increase collagenase mRNAin UMR 106 cells, with maximal abundance at 4 h after treatment (40). In our experiments, PTH induced a maximal 28-fold increase in rat collagenase mRNA at 4 h (Fig.   6B). PTH concentrations as low as 1 IIM administered for 18 h were effective in increasing OST-PTP transcripts, but exposure for less than 4 h, regardless of concentration (10 PM to 100 n~), had no significant effect (data not shown).
G protein-coupled PTH receptors activate a cyclic AMP-dependent protein kinase (PKA) signaling pathway (42). Stimulation of the PKA pathway in the osteosarcoma cells using the cyclic AMP analogue, chlorophenylthio-CAMP (CPT-CAMP) for 18 h (10 J~M ) resulted in an increase in OST-PTP mRNA levels (5-fold) comparable to that seen with PTH (Fig. 6B). Further increases in mRNA levels were noted with 100 PM and 1 mM CPT-CAMP treatment. The expression of the collagenase gene is also CAMP-dependent (40) and showed the expected dosedependent increase in transcript levels following CPT-CAMP treatment (Fig. 6B). These results are consistent with the hypothesis that OST-PTP is a PTP whose expression is modulated by PTH through stimulation of the PKA pathway, and such regulation is further support of its role in osteoblast differentiation.
Expression of OST-PTP during Spermatogenesis-Although our tissue distribution analysis revealed a seemingly incongru-ous restriction in expression to bone and testis, both these tissues possess highly regulated temporal and spatial organization which is necessary for the continuous differentiation and function of specific cell populations. Within the seminiferous tubule, coupled paracrine and cell surface interactions among the Sertoli cell, germ cell, and the extracellular matrix (basement membrane) are essential for spermatogenesis (43). In addition to its role in bone, OST-PTP might also function in the regulation of germ cell differentiation in the testis. Numerous proteins important in signaling pathways show differential expression during germ cell differentiation including protein kinases such as c-Kit (441, protein phosphatases such as the calmodulin-dependent phosphatase (45) and nuclear transcription factors such as AMP-responsive element modulator (46). I n situ hybridizations were performed with adult rat testis to determine the pattern of expression of OST-PTP during spermatogenesis. OST-PTP transcripts were found to be spatially restricted to the basal portion of the seminiferous tubule (Fig.  7). The specific clustering of silver grains appeared to be over cells possessing lightly stained, irregularly shaped nuclei suggesting localization to the Sertoli cell and/or primary spermatogonia (Fig. 7, C and D). Interestingly, this expression is stagespecific (Fig. 7, A and B). Abundance of OST-PTP transcripts appears greatest between stages I and VI1 when maturing spermatids remain buried within the Sertoli epithelium. Those tubules with low or nondetectable signal possess mature spermatids at the luminal surface of the Sertoli epithelium (stage VIII-IX), or immature spermatids with heads lacking the densely staining chromatin or the strong "bent rod" appearance (stage X-XI) (Fig. 7, C  We have reported the isolation and characterization of a novel type I11 receptor-like PTP, named OST-PTP, whose expression is restricted to bone and testis. The cellular studies presented here highlight numerous properties of this new molecule which make it a very unique member of the PTP family. First, OST-PTP is one of the few tyrosine phosphatases whose expression has been shown to be tightly regulated during the differentiation of a specific cell type. The OST-PTP mRNA is up-regulated following differentiation and matrix formation of primary osteoblasts and subsequently down-regulated in the osteoblasts which are actively mineralizing bone in culture. The expression of this phosphatase also appears to be tightly regulated during the differentiation of the germ cell to mature sperm during the process of spermatogenesis. Of course, i t remains to be proven that these fluctuations in mRNA are associated with increases in protein and phosphatase activity and that OST-PTP is necessary for these processes. Yet, these results strongly suggest that this PTP could be a relevant signaling molecule in the differentiation of these cells. Two of the PTPs that have been proven to be essential for cell differentiation are CD45, expressed in T cells, and PTPlC, expressed in cells of hematopoietic origin. Gene knockouts and mutation linkage studies have shown that CD45 is necessary for proper T cell maturation (47) and PTPlC is required for hemopoietic cell proliferation and differentiation (48,49) The existence of a second, regulated transcript is another intriguing property of OST-PTP. It is possible that this phosphatase may be expressed as a transmembrane protein possessing phosphatase activity or as an inactive transmembrane1 soluble protein lacking phosphatase domains. Such isoforms in the PTP family may prove to be an interesting means of modulating phosphatase activity and subsequent intracellular signaling events. Finally, OST-PTP is one of the first tissuespecific PTPs whose expression is modulated through a G protein-coupled receptor via the stimulation of the protein kinase A pathway. Stimulation of the G protein-coupled somatostatin crease PTP activity, but the identity of these activated PTP(s) involved in the anti-proliferative effects of these hormones is unknown. Parathyroid hormone is known to have both proliferative and anti-proliferative effects on bone, activating both the protein kinase A as well as Ca"1protein kinase C pathways (52). Studies on the role of OST-PTP in the biological effects of PTH could provide an interesting model for investigating the involvement of PTPs in G protein-activated signaling pathways.
At present, we can only speculate as to the role of this phosphatase in osteoblast differentiation or spermatogenesis. The changes in expression observed during osteoblast differentiation and following PTH stimulation suggest that OST-PTP may function as a critical modulator of tyrosine phosphorylation status during bone metabolism, as has been shown for the protein tyrosine kinases, c-src and c-fms (2,3). Early studies on phosphotyrosine phosphatase activity in calvarial osteoblast cultures revealed PTP activity, which was correlated with alkaline phosphatase activity, was inhibited by vanadate and was enhanced by treatment with PTH and 1,25-(OH),D, (vitamin D,) (14,15). It was proposed that this PTP activity could be attributed to alkaline phosphatase itself or to a n unidentified osteoblastic PTP. In light of our studies, OST-PTP could potentially be one of the PTPs responsible for this activity. As a receptor-like molecule similar to CAMS, this PTP could serve to transduce osteoblast-osteoblast and osteoblast-matrix interactions into intracellular signals, modifying osteoblast function and thereby influencing bone remodeling. Extrapolating from osteoblasts in culture to bone tissue, for example, the preosteoblast or immature osteblast situated away from the bone surface may express little active OST-PTP. Instead, this cell may express a n inactive OST-PTP isoform, a soluble or cell surface protein, which may mediate cell proliferation or cell-cell interactions. When such a cell migrates to the surface of the bone, this differentiating osteoblast may require the cell surface expression of OST-PTP to mediate interactions with matrix pro-(50) and dopamine D2 (51) receptors has been shown to in-teins and neighboring osteoblasts, all necessary for maturation of the osteoblast and optimal bone formation. Following new matrix formation and mineralization, the detached or boneencased osteoblast ceases to require OST-PTP. The regulated expression of this phosphatase in the testis suggests that it may also be relevant in the establishment and disruption of cell basement and Sertoli cell-germ cell interactions that occur during spermatogenesis. A conspicuous absence of OST-PTP RNA transcripts was observed between stages VI11 and X in the basal portion of the seminiferous tubule. During these stages, there is a movement of differentiating spermatocytes from the basal to the adluminal compartment entailing the dissolution and subsequent formation of junctions between a Sertoli cell and a neighboring Sertoli cells, a neighboring germ cell or the basal lamina (53). Such reconstruction of cell junctions is necessary to maintain the bloodtestis barrier. One could speculate that OST-PTP may be important in the maintenance of such interactions and expression during these stages might be undesirable.
Future research is necessary to address these speculative roles of OST-PTP in the signaling pathways important during osteoblast differentiation and spermatogenesis. Considering our limited understanding of tyrosine phosphorylation in bone, such studies should provide insights into the relevance of this phosphatase and tyrosine phosphorylation during normal bone metabolism and, ultimately, in the etiology and treatment of common bone diseases.