Phosphorylation at the Cyclin-dependent Kinases Site (Thr85) of Parathyroid Hormone-related Protein Negatively Regulates Its Nuclear Localization*

Parathyroid hormone-related protein (PTHrP) is expressed by a wide variety of cells and is considered to act as a secreted factor; however, evidence is accumulating for it to act in an intracrine manner. We have determined that PTHrP localizes to the nucleus at the G1 phase of the cell cycle and is transported to the cytoplasm when cells divide. PTHrP contains a putative nuclear localization sequence (NLS) (residues 61–94) similar to that of SV40 T-antigen, which may be implicated in the nuclear import of the molecule. We identified that Thr85immediately prior to the NLS of PTHrP was phosphorylated by CDC2-CDK2 and phosphorylation was cell cycle-dependent. Mutation of Thr85 to Ala85 resulted in nuclear accumulation of PTHrP, while mutation to Glu85 to mimic a phosphorylated residue resulted in localization of PTHrP to the cytoplasm. Combined, the data demonstrate that the intracellular localization of PTHrP is phosphorylation- and cell cycle-dependent, and such control further supports a potential intracellular role (10, 34, 35) for PTHrP.

Parathyroid hormone-related protein (PTHrP) 1 is widely expressed (1-3) and acts as a paracrine, and possibly an autocrine and intracrine, factor. However, its intracellular roles have not been fully defined. Recently a nucleolar localization signal (NLS) was identified within PTHrP and deletion of this motif prevented PTHrP from entering the nucleolus, maintaining it as a cytoplasmic protein (4). Nucleolar localization of PTHrP delays apoptosis in chondrocytes (4) and increases smooth muscle cell proliferation (5). PTHrP has also been linked to the ras signaling pathway (6) and the hedgehog signaling pathways (7,8), indicating its importance in regulating growth and differentiation. PTHrP expression is cell cycle-dependent (9,10) and PTHrP mRNA expression responds to mitogenic factors only at the G 1 phase of the cell cycle (10). Furthermore, PTHrP localizes to the nucleolus at the G 1 phase of the cell cycle (10).
Cyclin-dependent kinases control the progression of the various phases of the cell cycle (11). CDKs are activated at different phases of the cell cycle by the formation of cyclin-CDK complex and deactivated when their cyclin partner is degraded. The prototype CDK is CDC2, which associates with cyclin B and regulates the transition between the G 2 and M phases of the cell cycle. Cyclin E-CDK2 and cyclin A-CDK2 complexes are involved in the G 1 to S transition, while CDK4 and CDK6 associating with the D-type cyclins are involved in the progression through G 1 (12). In addition to direct regulation of the cell cycle, cyclins and CDKs have functions in other biological processes such as transcriptional control (13), and protein phosphorylation by CDC2 and CDK2 results in increase of affinity for the cytoplasm of some molecules containing an NLS (14).
A stretch of basic residues, or a pair of basic residues separated by a 10 -12 amino acid spacer to form a bipartite NLS, characterizes NLSs. The archetypal protein used in nuclear localization studies is the SV40 T-antigen where nuclear/cytoplasmic localization is regulated by phosphorylation. CK2 and/or CDC2-CDK2 phosphorylations at sites near the NLS determine the rate and amount of localization within the nucleus in an NLS-dependent manner (15)(16)(17). The combination of a CK2 site, a CDC2-CDK2 site and an NLS constitute a CcN motif, which is involved in phosphorylation-regulated nuclear localization. The CcN motif is conserved in a number of nuclear localized proteins, including mammalian c-MYC, p53, and mouse c-ABL IV and A-MYB (14,18). PTHrP contains a putative CcN motif similar to SV40 T-antigen, consisting of a CK2 site, and a CDC2-CDK2 site immediately before the NLS (Fig.  2). Additionally, a putative nucleolar targeting sequence (NoS) is located immediately C-terminal to the NLS.
In the present work using overexpressing cells, PTHrP was found to be phosphorylated in vivo. Using in vitro assays cyclin E-CDK2, cyclin A-CDK2 and cyclin B-CDC2 phosphorylated PTHrP at Thr 85 , which is immediately prior to its NLS. With the use of wild type and mutant green fluorescent protein GFP-PTHrP fusion proteins overexpressed in HaCaT cells, it was found that phosphorylation of Thr 85 resulted in cytoplasmic retention/nuclear exclusion. The results of these studies indicate a cell cycle-dependent nuclear exclusion of PTHrP from the start of S phase to mitosis, providing support for a tightly regulated nuclear function for PTHrP at G 1 .

EXPERIMENTAL PROCEDURES
Cell Culture-The spontaneously immortalized human keratinocyte cell line, HaCaT, which expresses PTHrP, was a kind gift from Professor N. E. Fusenig and was grown in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% FBS at 37°C and equilibrated with 5% CO 2 , as described previously (19). The human T lymphoid cell line (CEM) was grown in RPMI 1640 medium supplemented with 10% fetal bovine serum, 50 units/ml of penicillin, and 50 g/ml streptomycin.
Fluorescence Immunocytochemistry-70% confluent HaCaT cells fixed and stained with a fluorescein-labeled monoclonal antibody against amino terminus PTHrP as described previously (10,20). The cells were counter-stained with the nuclei acid-specific dye, propidium iodide. Images were collected and generated on a confocal laser scanning microscope (CLSM; Bio-Rad MRC 1024, Bio-Rad Microscopy Division, Hemel Hempstead Herts, United Kingdom).
Expression and Purification of Cyclins and Cyclin-dependent Kinases-The histidine-tagged protein A-bovine cyclin A complimentary DNA (cDNA) in a pET16b vector and the human p34 CDK2 cDNA in pGEX was a generous gift from Dr. T. Hunt (ICRF Clare Hall Laboratories, UK). GST-p34 CDK2 and histidine-tagged protein A-cyclin A were overexpressed in Escherichia coli strain BL21(DE3) purified and activated as described in Poon et al. (21). Alternatively, cDNAs for CDK2, cyclin A (gifts from Dr. Tony Hunter, The Salk Institute, San Diego, CA) and cyclin E (a gift from Dr. Steven Reed, The Scripps Institute, La Jolla, CA) were amplified by polymerase chain reaction and individually cloned into the baculoviral transfer vector pVL1329 (PharMingen, San Diego, CA). Cyclins E and A were cloned with an N-terminal GST cassette, and the CDK2 was cloned with an N-terminal 6-histidine tag. The sequences of all final constructs were confirmed by dideoxy chain termination sequencing analysis. Viruses were generated according to the manufacturer's instructions (Life Technologies, Inc.) and SF9 insect cells were infected with high titer baculoviral cyclin and CDK stocks. Following infection for 3 days, the cells were harvested, lysed, and active cyclin-CDK complexes purified using glutathione affinity chromatography. Recombinant baculoviruses for human GST-cyclin B and CDC2 were a kind gift from Dr. Helen Piwnica-Worms (Washington University, St. Louis, MO) and were expressed and purified as described above.
Protein Kinase Assays-Protein kinase assays were performed in a buffer containing 50 mM MOPS, pH 7.0, 10 mM magnesium acetate, 0.25 mM [␥-32 P]ATP (500 cpm/pmol), 0.1% (v/v) Tween 80, and varying concentrations of substrate and diluted enzyme (5 g/ml final concentration) made up to 40 l total volume. The enzyme was diluted in a buffer containing 50 mM MOPS, pH 7.0, 0.1% (v/v) Tween 80, and 1 mM dithiothreitol. The reactions were incubated for 20 min at 30°C, then 20 l of the reaction mixture spotted onto Whatman P-81 paper, washed and dried, and then quantitated as described in Pearson et al. (22).
For comparison of PTHrP phosphorylation by various cyclin-CDK combinations (Table I) and kinetic studies (Table II) Phospho-peptide Analysis-Peptides and proteins were phosphorylated as described in the protein kinase assay protocol above. [␥-32 P]ATP (500 cpm/pmol) for phospho-peptide mapping experiments and unlabeled ATP were used for HPLC and mass spectrometry experiments. The reactions were carried out in a total volume of 320 l, incubated for 1 h at 30°C, and terminated by adding 500 l of 2% (v/v) trifluoroacetic acid. Removal of free [␥-32 P]ATP from phosphorylated peptides and proteins was performed by application to a AG1-X8 column (Bio-Rad Laboratories, Hercules, CA) that was equilibrated with 5 bed volumes of 0.1% (v/v) trifluoroacetic acid. Unbound PTHrP-(67-94) was eluted using 6 ml of 0.1% (v/v) trifluoroacetic acid, while PTHrP-(1-141) was eluted with 6 ml 30% (v/v) acetic acid in 0.1% (v/v) trifluoroacetic acid. Radioactive eluates were pooled and applied to a Sep-Pak C 18 Cartridge (Waters, Milford, MA) activated previously according to manufacturer's instructions. Phospho-proteins and phospho-peptides bound to the Sep-Pak cartridges were washed twice with 10 ml of 0.1% (v/v) trifluoroacetic acid, eluted with 6 ml of 60% (v/v) acetonitrile, then vacuum-dried.
HPLC and Mass Spectrometry of PTHrP-(67-94)-PTHrP-(67-94) was phosphorylated as described above but using unlabeled ATP and chromatographed on a Vydac 208TP52, C8 reversed phase column using a SMART TM HPLC system (Amersham Pharmacia Biotech), with a linear gradient of acetonitrile in 0.1% aqueous trifluoroacetic acid at a gradient rate of 1% acetonitrile/min and a flow rate of 100 l/min (monitored at 214 and 280 nm). As control, an equivalent amount of unphosphorylated PTHrP-(67-94) was also applied to the column and subject to identical elution protocols. Peaks corresponding to phosphorylated and unphosphorylated PTHrP were collected and their identity confirmed by matrix-assisted laser desorption ionization time-of-flight (MALDI-TOF) mass spectrometry on a Voyager DE (PerSpective Biosystems, Farmingham, MA) mass spectrometer using ␣-cyano-4-hydroxycinnamic acid as the matrix.
Analysis of Tryptic Fragments of PTHrP-(67-94)-5 g of proteins or peptides phosphorylated with [␥-32 P]ATP were digested at 37°C for 24 h with 1 g of trypsin (Promega) in 100 l of trypsin digest buffer (50 mM NH 4 HCO 3 , pH 8.3, and 10% (v/v) acetonitrile). The peptide fragments were purified using HPLC through a Vydac 208TP52, C8 reversed phase column and confirmed with MALDI-TOF mass spectrometry as described above. Theoretical masses of each tryptic digest product were used to compare with the experimental masses.
Two-dimensional Phospho-peptide Mapping-Two-dimensional phospho-peptide mapping of trypsin-digested phosphorylated PTHrP was performed as described by Beemon and Hunter (24). The plates were exposed in a PhosphorImager cassette or an x-ray cassette for visualization. 32 P Labeling of Cells Transiently Transfected with GFP-PTHrP-10 -20 g of DNA (see below for plasmid constructs) was made up in 150 mM NaCl to a final volume of 50 l and then added to COS-1 cells that were grown to 80% confluence in a 76-cm 2 tissue culture plate, trypsinized, pelleted, and resuspended in 200 l of fresh DMEM 10% FBS. The cells were electroporated in 0.4-cm gap size cuvettes at 975 microfarads, 200 V with the time constant between 40 and 42 ms using a GenePulser II (Bio-Rad Laboratories, Hercules CA). The cells were then replaced into 10 ml of DMEM 10% FBS and allowed to recover for 4 -6 h. The medium was changed once the cells reattached. The cells were allowed to grow for 48 h before labeling experiments were performed. Transiently transfected COS-1 cells were washed three times in phosphate-free DMEM, then placed into phosphate-free DMEM containing 10% FBS for 2 h. 1 mCi of 32 P in the form of orthophosphoric acid in water (NEN Life Science Products) was added to the medium and left for 5 h. The cells were washed three times in phosphatebuffered saline, and harvested in 0.8 ml of lysis buffer (50 mM Tris⅐HCl, pH 8.0, 5 mM EDTA, 150 mM NaCl, 2 mM phenylmethylsulfonyl fluoride, 50 mM NaF) plus protease inhibitors (Complete TM , Roche Molecular Biochemicals GmbH, Mannheim, Germany). Lysates were diluted in 1 volume of the same buffer and precleared with Pansorbin (Calbiochem) for 1 h at 4°C. The precleared lysates were then centrifuged (13,000 rpm, 5 min, 4°C), and 3 l of rabbit polyclonal anti-GFP antibody (CLONTECH) was added to the supernatant and incubated overnight at 4°C. Immune complexes were then collected with 10 l of protein A-Sepharose CL-4B (Amersham Pharmacia Biotech), washed five times with lysis buffer, and resolved by SDS-polyacrylamide gel electrophoresis. Proteins were visualized following Coomassie Blue staining. The gel was then dried, and radioactivity was detected using autoradiography. Bands that corresponded in size to those of GFP-PTHrP-(1-141) (45 kDa) were excised from the gel, destained, digested with trypsin, and analyzed by two-dimensional phospho-peptide mapping.
Centrifugal Elutriation-Elutriation of CEM cells was performed in the Beckman J6 ME centrifuge with a JE-5 elutriation rotor, at 1500 rpm, 23°C. The elutriation was carried out in RPMI medium with 5% fetal bovine serum, 2 mM dithiothreitol, and 5 mM EDTA. The G 1 cell population was plated in RPMI medium with 10% fetal bovine serum, and fractions were collected at required time points. Flow Cytometry-Cell cycle distributions of fractions of elutriated cells were determined by staining with propidium iodide and analyses performed using the CellFIT program resident in a FACScan flow cytometer (Becton Dickinson, Lincoln Park, NJ).
Generation and Transfection of GFP Fusion Peptides-Oligonucleotide-directed mutagenesis was used to generate peptide mutants mimicking phosphorylated and dephosphorylated forms of PTHrP. These 25) as template, and products were cloned in frame C-terminal to the GFP reporter in the vector pEGFP-C1 (CLONTECH) using BglII and HindIII. The fidelity of all cloned DNA was confirmed by singlestranded DNA sequencing using a T7 DNA sequencing kit (Amersham Pharmacia Biotech). Large scale plasmid DNA was purified (Qiaex Maxi Prep Kit, Qiagen Inc., Chatsworth, CA) and transfected into HaCaT cells using the calcium phosphate method as described in Sambrook et al. (26). Cells were counter-stained with the DNA-specific dye, propidium iodide (Sigma) or with the actin specific dye, Texas Red-X phalloidin (Molecular Probes, Eugene, OR). Images were acquired by CLSM (Bio-Rad MRC1024).

RESULTS
Fluorescence immunocytochemistry and CLSM of PTHrP expression in a asynchronous proliferating culture of HaCaT cells showed that PTHrP was localized to the nucleolus and the cytoplasm in cells that show a G 1 phase phenotype (Fig. 1, A and B: i). In cells that show a G 2 and M phenotype, which have a more intense nuclear stain, PTHrP was not detected in the nucleus but was expressed at higher levels in the cytoplasm (Fig. 1, A and B: ii and iii). PTHrP expression was greatest in actively dividing cells (Fig. 1, A and B, iii). Because of this interesting observation, we looked at the sequence of the mature PTHrP protein compared with that of various nuclear proteins. The amino acid sequence of PTHrP (Fig. 2) shared similarities with the CcN motifs of SV40 T-antigen, p53, c-MYC, and A-MYB (13,17). The CcN motif included a putative CDK site at Thr 85 . C-terminal to the CcN is a putative NoS with characteristic strings of basic amino acids flanking an arginine hinge, which is commonly found in viral proteins such as HIV-1 Tat (27), HIV-1 Rev (28), and HTLV-1 Rex (29). Interestingly, the combination of parts of the NLS (Lys 88 -Lys-Lys-Lys 91 ) and the NoS (Lys 102 -Lys-Lys-Arg-Arg-Thr-Arg 108 ) could constitute a bipartite NLS (Fig. 2) where two blocks of basic residues are separated by a 10 -12-amino acid spacer (30). Because of these similarities it was thought important to determine whether PTHrP could be phosphorylated by CDC2-CDK2. PTHrP peptides and recombinant PTHrP proteins were incubated with recombinant cyclin A-CDK2. All peptides and proteins were adjusted to a concentration of 6 M and phosphorylated using 5 g/ml cyclin A-CDK2. PTHrP-(1-108), PTHrP-(1-141), PTHrP-(67-94), and the hCK2 ␤ peptide were phosphorylated by cyclin A-CDK2, while the PTHrP peptides PTHrP-(1-34), PTHrP-(1-84), PTHrP-(50 -69), and PTHrP-(67-84) were not phosphorylated (Fig. 3). Cyclin A-CDK2 phosphorylated only PTHrP proteins and peptides containing the putative phosphorylation motif Lys 84 -Thr-Pro-Gly-Lys 88 . These results indicate that the phosphorylation site of PTHrP lies within amino acids 85 and 94, thereby implicating Thr 85 , the only phosphorylated amino acid present.
To ascertain the stoichiometry of phosphorylation, PTHrP-(67-94) phosphorylated with recombinant cyclin A-CDK2 was chromatographed on a Vydac C-8 reversed phase column and the eluted peaks analyzed by mass spectrometry. The masses of the HPLC peaks were in agreement with the theoretical masses of phosphorylated PTHrP-(67-94), which is 3341.80 mass units, and unphosphorylated PTHrP-(67-94), which is 3261.80 mass units. Thr 85 was confirmed to be the CDC2-CDK2 phosphorylation site of PTHrP by MALDI-TOF mass spectrometry analysis of tryptic fragments of cyclin A-CDK2phosphorylated PTHrP-(67-94). Two-dimensional tryptic peptide mapping showed that both PTHrP-(67-94) and PTHrP-(1-141) shared the same CDC2-CDK2 phosphorylation site.
Kinetic studies comparing the efficiency of PTHrP-(67-94) as a substrate to cyclin E-CDK2, cyclin A-CDK2, and cyclin B-CDC2 were performed as described under "Experimental Procedures" (Table II). The K m for cyclin E-CDK2 was 80.71 Ϯ 7.  (Table I). Histone H1 was used as a positive control for kinase activity and also to demonstrate the efficiency of PTHrP as a substrate compared with a known CDC2-CDK2 substrate. PTHrP-(50 -69) was used as a negative control and was not phosphorylated by any of the cyclin-CDK combinations as expected. Comparisons could not be made between each of the cyclin-CDK assays as the purified cyclin-CDKs were of different concentrations and also different activation states.
To determine whether phosphorylation of PTHrP-(67-94) was cell cycle stage-specific, whole cell lysates of CEM cells were collected at various points of the cell cycle following G 1 selection by centrifugal elutriation and used to phosphorylate PTHrP-(67-94). Consistent with the kinase activity in vitro (Table I), this in vitro enzyme assay for kinase activity using PTHrP-(67-94) as substrate shows that kinase activity was low at G 1 (0 h), increased 2-fold (Fig. 4A) as the cells moved on to S (4 -6 h) (Fig. 4B), and increased to 4-fold (Fig. 4A) at G 2 and M (8 -10 h) (Fig. 4B). In contrast, the activity of CK2 with hCK2␤ peptide (Arg-Arg-Arg-Asp-Asp-Asp-Ser-Asp-Asp-Asp-NH 2 ) as a substrate was constant throughout the cell cycle (Fig. 4, A  and B).
To confirm that PTHrP was phosphorylated in vivo, COS-1 cells transfected with GFP-PTHrP-(1-141) were labeled with   -(67-94) is a substrate for cyclin E-CDK2, cyclin A-CDK2, and cyclin B-CDC2 1 g of PTHrP-(67-94) was used as in vitro substrate for recombinant cyclin E-CDK2, cyclin A-CDK2, and cyclin B-CDC2, and 1 g of PTHrP-(50 -69) was used as a negative control, while 3 g of histone H1 was used as positive control and also to demonstrate the relative efficiency of PTHrP as a substrate for these kinases. The G 1 -type cyclin-kinases combinations of cyclin D 1 -cdk4, cyclin D 1 -cdk4, and cyclin D 3 -cdk6 did not phosphorylate PTHrP-(67-94) (results not shown).

TABLE II Kinetic studies of PTHrP phosphorylation
Kinetic studies were performed as described under "Experimental Procedures." The substrate used was PTHrP-(67-94), and assays were performed using recombinant cyclin E-CDK2 (n ϭ 3), cyclin A-CDK2 (n ϭ 3), and cyclin B-CDC2 (n ϭ 2). a Only two assays for cyclin B-CDC2 were performed, since there was insufficient enzyme from the same batch for further replicates. 32 P. Immunoprecipitation from the cell lysates with an antibody against GFP resulted in a radiolabeled protein band at 45 kDa, corresponding to GFP-PTHrP-(1-141) (Fig. 5A). When this band was excised and analyzed by two-dimensional phospho-peptide mapping, the resultant map (Fig. 5B, ii) corresponded to that of cyclin A-CDK2-phosphorylated recombinant PTHrP-(1-141) (Fig. 5B, i), implicating Thr 85 to be the site of phosphorylation. The heavy loading in Fig. 5B, ii, was necessary to enable visualization of tryptic phosphopeptides, but this resulted in a slight retardation of the phosphopeptide in the second dimension.
Phosphorylation at the CDC2-CDK2 site adjacent to the SV40 T-antigen NLS decreases the rate and amount of nuclear localization of this protein (16). In order to determine whether this is the case also with PTHrP, GFP-PTHrP fusion products containing wild type and point mutations of Thr 85 were constructed. Thr 85 was mutated to Ala 85 to preclude phosphorylation at this site or to Glu 85 to mimic the negative charge of a phosphorylated threonine. Mutant PTHrP-GFP fusion proteins encompassed the NLS (amino acids 86 -91) and the putative NoS (amino acids 94 -106) (Fig. 6A, c-e). The vector pEGFP-C1-containing GFP without any fusion sequences was used as a vector nonfusion control (Fig. 6A, a) fused at the N terminus to GFP was used as a positive control (Fig. 6A, b).
A total of 30 cells from three independent experiments were scored, with the cells graded for GFP-PTHrP localization to the nucleus or cytoplasm (Fig. 6C).

DISCUSSION
This is the first report of an endocrine/paracrine factor that possesses a phosphorylation-regulated NLS and displays differential cellular localization (nuclear/nucleolar versus cytoplasm), inferring that it acts intracellularly. To date, proteins shown to possess such motifs have intracellular functions only, such as viral proteins (e.g. SV40 T-antigen), oncogenes (e.g. c-MYC) and tumor suppressors (e.g. p53) (14,33).
Within PTHrP we identified a putative CcN motif similar to those of SV40 T-antigen, c-MYC, p53, c-ABL, and A-MYB. Like these "CcN motifs," PTHrP contains a consensus CDC2-CDK2 phosphorylation site near the putative NLS. We showed that PTHrP was phosphorylated by CDC2-CDK2 specifically at Thr 85 in the CcN motif and that this phosphorylation occurred in cells overexpressing PTHrP.
Phosphorylation of residues in the vicinity of the NLS in other CcN-containing proteins, such as SV40 T-antigen, alters their cellular distribution (14,16). Phosphorylation at the CDK site N-terminal to the NLS redistributes these proteins from a nuclear/nucleolar location and confines them to the cytoplasm (14,16). To address whether phosphorylation of PTHrP similarly results in a cellular redistribution of PTHrP, we employed GFP-PTHrP expression constructs in which Thr 85 was mutated to Ala 85 or Glu 85 . These residues were chosen to preclude phosphorylation or to mimic a phosphorylated residue, respectively. These studies indicate that PTHrP in a dephosphorylated state is maintained in the nucleus, while in the phosphorylated state, PTHrP is excluded from the nucleus and is essentially cytoplasmic. In common with proteins possessing a CDK phosphorylation site immediately adjacent to a NLS (14,16), phosphorylation at this site inhibited NLS function (14), and this also appeared to be true for PTHrP.
CDC2-CDK2 phosphorylation activity was found to begin at S phase and peaked at the G 2 /M phase of the cell cycle, coincident with the localization of cyclin A and B 1 to the nucleus (32), suggesting that PTHrP might be phosphorylated and thus excluded from the nucleus at these phases. In agreement, we observed that PTHrP was localized to the nucleus during G 1 , further supporting the notion that cell cycle-related nuclear localization of PTHrP may be regulated by CDK phosphorylation.
To date, evidence has indicated that PTHrP acts as a hormone in certain patients with cancer and possibly also in lactating women, but it acts as a paracrine factor in several normal tissues (2,3). The demonstration that PTHrP is localized to the nucleus and nucleolus indicates that PTHrP may have intracellular roles (10,34,35). A nuclear/nucleolar location for PTHrP may arise as a result of two distinct mechanisms: (i) as a normal consequence of PTHrP production within a cell expressing PTHrP or (ii) in a target cell as a result of binding to its cognate receptor and internalization of the complex with consequent nuclear localization. The control mechanisms for determining intracellular trafficking of PTHrP, whether via the trans-Golgi network and secretory vesicles versus a nuclear localization is not known.
Since PTHrP demonstrates phosphorylation and cell cycledependent nuclear localization, it is conceivable that PTHrP may have a defined nuclear role, possibly to regulate growth and differentiation. Support for this proposition results from the demonstration that intracellular targeted PTHrP increases proliferation of vascular smooth cells (5). It is likely that other secreted growth factors, such as acidic and basic fibroblast growth factors, platelet derived growth factor, and angiogenin (36,37), which possess basic amino acid sequences analogous to those of a NLS, may also display regulated intracellular localization similar to that of PTHrP.