Mar 30, 1989 - Vivian Grill$, Jane Glatz$, Christine P. Rodda$, Jane M. Moseley$, William I. ..... Mundy, G. R., Ibbotson, K. J., D'Souza, S. M., Simpson, E. L.,.
THE JOURNAL OF BIOLOGICAL CHEMISTRY 0 1989 by The American Society for Biochemistry and Molecular Biology, Inc.
Vol. 264, No. 25, Issue of September 5 , pp. 14806-14811.1989 Printed in U.S.A.
Purification and Characterizationof Recombinant Human Parathyroid Hormone-related Protein* (Received for publication, March 30, 1989, and in revised form, May 8, 1989)
R. Glenn HammondsJr., Patrick McKay, Genine A.Winslow, Hannelore Diefenbach-Jagger$, Vivian Grill$, Jane Glatz$, Christine P. Rodda$, Jane M. Moseley$, William I. Wood, and T. John Martin$ From the Department of Developmental Biology, Genentech, Znc., South Sun Francisco, California 94080 and the $.Department of Medicine, University of Melbourne, Repatriation General Hospital and St. Vincent’s Institute of Medical Research,St. Vincent’s Hospital, Melbourne, Australia
Full-length human parathyroid hormone-related sufficient for hypercalcemia of malignancy in uiuo (at least protein (PTHrP-(1-141)) as well as a carboxyl-termi- for thetumors tested) is provided by rat studies in which the nal shortened form (PTHrP-(1-108))have been injection of PTHrP produces hypercalcemia (11)and in which expressed fromrecombinant DNA-derived clones. the injection of PTHrP antiserumcorrects hypercalcemia These proteins were expressed in Escherichia coli as induced by human tumors transplanted intonude mice (7). fusion proteins so that cyanogen bromide cleavage Although 8 of the first 13 amino acids of PTHrP are yields the desired product. Both proteins were purified identical to those of PTH, thereis no clear similarity between and then characterized by sodium dodecyl sulfate gel the rest of the two sequences ( 5 ) . Whereas full-length PTH electrophoresis, amino-terminal amino acid sequenccontains 84 amino acids, PTHrP continues with a basic ing, peptide mapping, andmass spectral analysis. Recombinant PTHrP-(1-141), PTHrP-(1-108), syn- sequence from amino acids 84 to 108 followed by a unique thetic PTHrP-(1-34), and naturally derived PTHrP peptide from amino acids 109 to 141. The hypercalcemic and are all equipotent in the stimulation of cyclic AMP PTH-like effects of PTHrP reside in thefirst 34 amino acids levels in the osteoblast-like cellline UMR 106-01. (7, 11, 12). We present here the use of recombinant DNA-derived However, PTHrP-(1-141) and -(l-108) aretwo to four times more active than PTHrP-( 1-34) in the stim- clones to express PTHrP-(1-141) as well as a shortened form ulation of plasminogen activator activity from this cell (PTHrP-(1-108)) in an Escherichia coli fusion protein system. line. PTHrP-(1-141) reacts equipotently with PTHrP- The desired proteins have been cleaved from the fusion pro(1-34) in a radioimmunoassay using an antiserum pre- teins, purified, and characterized by a variety of techniques. pared against PTHrP-(1-34). A comparison of the biological and immunological properties PTHrP-(1-141), -(l-108), and -(1-84) were used as of the E. coli-derived proteins is also presented. These wellPTHrP-specific mobility standards on sodium dodecyl characterized recombinantproteins are used as mobility sulfate gel electrophoresis to determine the approximate length of two forms of naturally derived PTHrP. standards to determine the approximate size of two naturally Thedata show that PTHrP purified from the lung occurring forms of PTHrP for which only small amounts of tumor cell line BEN contains a major form of about protein are available. 108 amino acids and another form of about141 amino EXPERIMENTALPROCEDURES acids.
Hypercalcemia is often associated with malignant tumors from a variety of sources (1). Humoral hypercalcemia of malignancy occurs most commonly in squamous cell tumors of the lung and kidney (1, 2). Recent work has identified PTHrP’ asa tumor-derived protein relatedto PTH(3-6) that is likely to be responsible for humoral hypercalcemia of malignancy (7). Based on its PTH-like biological activity, PTHrP has been purified from tumor cell lines as well as from human tumors (4,8-10). Determinationof the complete primary structure of PTHrP by amino acid sequencing and the subsequent isolation of cDNA clones show that the fulllength form of the mature protein contains 141 amino acids ( 5 , 6). Strong evidence that PTHrP is both necessary and * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. The abbreviations used are: PTHrP, parathyroid hormone-related protein; PTH, parathyroid hormone; SDS, sodium dodecyl sulfate; HPLC, high performance liquid chromatography.
Materials-Culture media and fetal calf sera were obtained from GIBCO. [3H]Adenine,sodium [‘251]iodide,and [32P]cAMPwere purchased from Amersham Corp. All other chemicals were of reagentgrade and obtained from standard suppliers. Human PTH-(1-34)was obtained from AUSPEP (Parkville, Australia). PTHrP-(1-34) was synthesized as previously described (12). Lysyl endopeptidase was from Wako Pure Chemical Industries (Osaka,Japan). Endopeptidase Asp-N was from Boehringer Mannheim. Cell Culture-The human lung cancer cell line BEN (13) was maintained in culture as previously described (4). The clonal osteogenic sarcoma cell line UMR 106-01 was used between passages 7 and 16 and maintained as described (14). PTH-related Protein-PTHrP was prepared from serum-free conditioned medium from BEN cells as previously described (4). Highly purified PTHrP with apparent molecular mass of 17 kDa on SDS gel electrophoresis (4) was used for the activity and antibody studies. Partially purified material was also harvested from earlier stages of purification on reversed phase high performance liquid chromatography. This material (peak A) was noted in our earlier work (4) and was found to have an apparent molecular mass of 21-24 kDa on SDS gel electrophoresis. PTHrP Expression Plasmids-Plasmids pTrpMLE.brf3.1, pTrpMLE.brf8.2, and pTrpMLE.brf7.2 were used to express PTHrP-(1141), PTHrP-(1-108), and PTHrP-(1-84) as fusion proteins. The PTHrP-(1-141) expression plasmid was assembled from 1)pMNCV, an E. coli expression vector containing the trp promoter followed by
14806
14807
Recombinant PTHrP the 101-amino acid carrier protein, "trp medium LE"'; 2) the PTHrP mature coding and 3'-untranslated region from pbrf.52 (5); and 3) a synthetic oligonucleotide to fuse the carrier protein to PTHrP with an in-frame methionine codon. The expression plasmid for PTHrP(1-108) and -(I-84) used the same carrier protein fused to theamino terminus and a synthetic oligonucleotide to place a stop codon after amino acid 108 or 84 of PTHrP. Expression was in E. coli strain "294 (15). flash cultures Purification of Recombinant PTHrP-(I-l4l)-Shake were grown overnight in M9 medium (16) containing 5 pg/ml tetracycline. Harvested bacteria were homogenized in 10 mlof 25 mM Tris-HC1, pH 7.5, 5 mM EDTA/g of paste, sonicated on ice, and centrifuged to collect the insoluble material; andthe pellet was washed in the same buffer. Rapid differential extraction of the PTHrP fusion protein was obtained with the same buffer containing 0.1% SDS. The extract was precipitated with acetone and dissolved in 7 M guanidine HCl, and the buffer was exchanged into 6 M urea, 20 mM sodium phosphate, pH 7.5, by Sephadex G-25 chromatography. This material was chromatographed on sulfopropyl-Sepharose and eluted with a 0-1 M NaCl gradient in the same buffer. Fractions containing the PTHrP fusion were acetone-precipitated. The precipitate was dissolved to 7.5 g/liter protein in 70% formic acid containing 20 g/ liter CNBr, cleaved overnight at room temperature, and evaporated to dryness. The dried product was dissolved in 1 volume of 7.5 M urea followed by the addition of 2 volumes of 0.1% trifluoroacetic acid. The material was filtered through a 0.45-pm filter and applied to a Vydac CIScolumn. Elution was with a 20-60% acetonitrile gradient in 0.1% trifluoroacetic acid. PTHrP-(1-141) containing fractions were pooledand lyophilized. From 20 liters of culture, 55 g of bacteria were harvested, and 144 mg ofpurified PTHrP-(I-141) were obtained. Purification of Recombinant PTHrP-(I-10B)"Shake flask cultures were grown as described above in LB medium (16). The harvested bacteria were homogenized in 10 ml of 50 mM sodium acetate, pH 5.0/g of paste, lysed with a Mantin-Gaulin press, and centrifuged. The insoluble material was suspended in 10 volumes of 70% formic acid; CNBr was added to 10 g/liter; and the cleavage was performed as described above. The dried product was suspended in 7M guanidine HC1 and centrifuged, and the supernatant buffer was exchanged as above into 6 M urea, 50 mM sodium acetate, pH 5.0. The material was chromatographed on sulfopropyl-Sepharose as described above, and the pooled fractions were diluted with 2 volumes of 0.1% trifluoroacetic acid. Cia HPLC was performed as described above with a gradient of 20-40% acetonitrile. From 11.7 liters of culture, 49.4 g of bacteria were harvested, and 63 mg of PTHrP-(1-108) were purified. PTHrP Characterization-Amino-terminal sequencing was performed using a gas-phase sequenator (Applied Biosystems Inc., Model 470A) with HPLC quantitation of phenylthiohydantoin-derivatives of cleaved amino acid residues. Proteolytic digestion of PTHrP-(1-141) was performed using lysyl endopeptidase at a weight ratio of 1:lOO overnight at 37 "C in 0.1 M NH4HC03. The digest was dried by Speed-Vac; redissolved in 6 M guanidine HCI, injected directly onto a Vydac C, column (214TP54) equilibrated with 0.1% trifluoroacetic acid, 5% acetonitrile; and eluted with a gradient of 5 5 0 % acetonitrile in 0.1% trifluoroacetic acid over 60 min. Peaks were detected by absorption at 280 and 214 nm and collected individually for further analysis. Digestion of PTHrP-(1108) was performed with endopeptidase Asp-N at a weight ratio of 1:lOO for 3 h at 37 "C in 50 mM Tris-HC1, pH 8.0. The digest was applied directly to a Vydac Cr column and chromatographed as described above. Mass analysis was performed using a JEOL HX22OHF/HXlIOHF tandem double focusing mass spectrometer with fast atom bombardment (6 keV of xenon). Fractions collected by HPLC were reduced in volume by centrifugation under a vacuum and then redissolved in thioglycerol and applied directly to the mass spectrometer probe. Data were collected using an automated system, and the M, was determined by comparison with CsI standards. The accuracy of the M, was to within 0.5 in the range 1000-4500. Cyclic AMP Generation in UMR 106-01 Cells-Thecyclic AMP response in the PTH-responsive UMR 106-01 cells was measured by prelabeling cellular ATP pools by incubating with [3H]adenine,treatment with PTH or PTHrP, and assaying [3H]cAMP by sequential chromatography on Dowex and neutral alumina (14). Phminogen Activator-Plasminogen activator activity was assayed by subculturing UMR 106-01 cells onto 96-well tissue culture plates coated with 1*51-fibrinas previously described (17). Purified R. Derynck, unpublished data.
human plasminogen (0.008 mg/ml) was used to determine plasminogen-dependent fibrinolysis. Assays were carried out against purified plasminogen activator as standard (obtained from Dr. P. J. Gaffney, National Institute for Biological Standards, London, United Kingdom). Antibody Studies-Antisera against synthetic human PTHrP4134) were raised by immunizing New Zealand White rabbits with peptide emulsified in Montanide (ISA 50 (Tall-Bennett Industrial pty. Ltd., Melbourne, Australia)). Antisera were tested by binding to plates coated with PTHrP-(1-34),and all positive antisera were checked by radioimmunoassay using PTHrP labeled with I z 5 I by the chloramine-T method as previously described for calcitonin (18). Radioimmunoassays were carried out by incubation for 48 h at 4 "C in a total volume of 0.3 ml in 0.1 M phosphate, pH 7.4, with 1% bovine serum albumin. Bound was separated from free tracer using anti-rabbit globulin (SAC-cell, Wellcome Laboratories). In experiments in which the effects of antibody on biological activity were studied, samples were incubated overnight at 4 "C with antibody before assaying for cyclic AMP response in UMR 106-01cells (19). Immunoblotting-Immunostaining of proteins separated on 15% polyacrylamide slab gels (20) was carried out by electrophoretic transfer toa nitrocellulose paper in a Trans-Blot TM cell (Bio-Rad) for 60 min a t 60 V in 25mM Tris, 192 mM glycine, 0.1% SDS, and 20% (v/v) methanol. Immunobiochemical detection was performed with the following buffer: 50 mM Tris, pH 8, 2 mMCaC12, 80 mM NaCI, 5% (w/v) non-fat dry milk, 0.2% (v/v) Nonidet P-40, and 0.02% sodium azide. The nitrocellulose paper was immersed in buffer for 60 min at 25 "C with gentle shaking. Incubation with first antibody was performed in buffer as described above with polyclonal rabbit anti-PTHrP-(1-34) at a dilution of1:300. Following threeshort washes in buffer, incubation with second antibody was performed in buffer as described above with goat anti-rabbit "'I-Fab (Du PontNew England Nuclear) at a final concentration of IO6cpm/ml. Three washes in buffer followed by two rinses with 50 mM Tris-HCI, pH 8, were performed. Autoradiograph was at -70 "C for 24 h. RESULTS
Whereas activePTHrP is readily detected in culturemedia after the transfection of recombinant DNA-derived clones into mammalian cells (5), we have elected to purify and characterize larger amounts of the protein produced by an E. coli expression system. Fig. 1 shows the plasmid used to produce PTHrP asa fusion protein. One-hundred-one amino acids of the TrpMLE protein followed by a methionineresidue precede the 141 amino acids of mature PTHrP. Since PTHrP contains no methionine (or cysteine residues), cyanogen bromide cleavage of the fusion protein will liberate intact, fulllength PTHrP. Thissystem was selected both because of the high level of protein expression and because these fusion proteinsareoften insoluble and form refractile inclusion bodies within the bacteria. We expected that this should minimize any possible proteolysis of PTHrP which contains a stretch of basic residues (positions 84-108) as well as the sequence RRH at thecarboxyl terminus. Microscopic examination of bacterial cultures expressing the PTHrP fusion protein confirmed that refractile bodies were present (data not shown). Purification of PTHrP-(1141) was achieved by differential extraction of the insoluble protein, ion exchange chromatography on sulfopropyl-Sepharose, cleavage with cyanogen bromide, and hydrophobic interaction HPLC. SDSgel analysis of various steps duringthe purification is shown in Fig. 2A. The final material contains one major band and a number of smaller bands. Both the major band and thesmaller bands react with amino-terminal PTHrP antiserum on immunoblotting (data not shown), indicating that the smaller bands are probably the result of a low level of proteolysis near the carboxyl terminus. Similar fusion protein vectors were also constructed to express shortened forms of PTHrP, containing amino acids 1-108 and 1-84. These shortened forms contain the complete amino-terminal region of PTHrP required for its known
Recombinant PTHrP
14808
TrpULE
FIG.1. A, homology of PTH and PTHrP. Diagram is shown of the Trp medium LE fusion protein to PTHrP. Locations are shown of PTHrP-(1-141), -(1-108), and -(l-84). B, PTHrP bacterial expression plasmid.
PlHrP (1-141)
I
PlHIP (1.108)
I I
mrP(1.8.1)
I
"
B PTHrP
Tet I 1
A. top -
1
2
3
4
I
I
10
20
I
I
I
30
40
A
't N
5 6
a
0.1
97k 66k -
0
Time (min)
50
43k c
30k . . "
I I 21k- .
'1-
r~
Fusion Protein 0.2
11
I
I
1
1
4 4161 4162
PTHrP (1-141)
:-ii
14k-
B. top 97k 66k 43k 30k 21k 14k -
ui; -
" "
0' 0
I
10
I
20
30
Time (min) 1 2 3 4 5 6
FIG.3. Preparative HPLC of protease-digested PTHrP. A, PTHrP-(1-141) digested with lysyl endopeptidase; B, PTHrP-(1108) digested with endopeptidase Asp-N. The peaks are numbered, and thelocations of the peptides are shown. c
-
Fusion Protein PTHrP (1-108)
FIG.2. Nonreduced SDS gel summary of purification of recombinant PTHrP. A, PTHrP-(1-141);lune I, standards; lane 2, total cell extract; lane 3, pellet following sonication; lane 4, sulfopropyl-Sepharose pool; lane 5,cyanogen bromide cleavage; lane 6, HPLC pool. B, PTHrP-(1-108): lane I, standards; lane 2, total cell extract; lune 3, pellet following cell lysis; lune 4, cyanogen bromide cleavage; lane 5,sulfopropyl-Sepharose pool; lane 6, HPLC pool.
biological activities (11, 12) but lack the uniquecarboxylterminal peptide (PTHrP-(1-108)) or the unique peptide as well as the preceding basic region (PTHrP-(1-84)). Fig. 2B shows SDS analysis of the purification of PTHrP-(1-108). In contrast to the purification of PTHrP-(1-141), the PTHrP(1-108) fusion protein was cleaved with cyanogen bromide immediately after extraction rather thanafter purification by sulfopropyl-Sepharose chromatography. This was intended to further reduce any proteolysis of the product during its isolation. In this case, only a single band was observed for the final material (Fig. 2B). Amino-terminalamino acid sequencingof purified PTHrP(1-141) and -(l-108) gave a single sequence for the first 15 residues identical to that expected for the mature protein. Characterization of the carboxyl termini was achieved by proteolysis of the proteins and analysis of the isolated peptides by molecular ion mass spectrometry. Fig. 3A shows the reverse phase separation of peptides generated by the lysyl endopeptidase cleavage of PTHrP-(1-141). The four expected major
Recombinant PTHrP TABLEI Summary of mass spectral data Mass spectral molecular mass
Amino acid
HPLC oeak
Calculated
Observed
Lysyl endopeptidase digest of PTHrP-(1-141) 1-11 14-47 14-53 54-74 105-141
1277.41 3974.51 4599.20 2464.61 4058.32
1277.3 3975.3 4599.8 2464.4 4058.9
1 4 4 2 3
Endopeptidase Asp-N digest of PTHrP-(1-108) 1-9 17-29 41-61 41-62 62-79 63-75 63-79 so- t os
1034.16 1695.02 2280.50 2395.59 2190.32 1553.66 2075.23 3475.17
1034.6 1695.3 2280.4 2395.6 2190.5 1554.0 2075.4 3474.4
2 12 4 4 9 5 7 2
~~
100
'
I
4058.9
I
Peak 3 K a U
plasminogen activator activity, a consistent result found in several experiments. Recombinant PTHrP-(1-141) gives the same response (on a molar basis) as chemically synthesized PTHrP-(1-34) in a radioimmunoassay using an antiserumpreparedagainst PTHrP-(1-34) (Fig. 6). Furthermore, preincubation of PTHrP-(1-141) with this antiserum abolished the stimulation of cyclic AMP levels in theUMR 106-01biological assay (Fig. 7). This antiserum is specific for PTHrP and hasno effect on the response to PTH. Conversely, antiserum prepared with PTH-(1-34) inhibits its action, but does not affect PTHrP(1-141) or -(1-34) (Fig. 7). In the course of purifying PTHrP from BEN cell culture media (4), two major peaks of active material were obtained at an intermediate stage of purification (Fig. 8). Peak B was further purified to yield a 17-kDa band on SDS gel electrophoresis which gave the amino acid sequence data used to isolate clones of PTHrP (7), but peak A has not been fully purified. Fig. 9 shows an immunoblot in which we have compared the electrophoretic mobility of recombinant PTHrP-(1-141) and -(l-108) (as well as less characterized PTHrP-(1-84)) to BEN cell-derived PTHrP peaks A and B. These data show that peak Ahas the same mobility as PTHrP-(1-141), whereas peak B hasthat of PTHrP-(1-108).
-
75 -
14809
A.
c
3
-
50
-
.-9 a U
4042.8 4014.5
25 -
0 3950
3999.3
-
4027.6
I
I
I
4000
4050
4100
5
1
10
2030
peptlde lnM1
4150
E.
MI2 FIG.4. Molecular ion mass spectrometry of peak 3 from lysyl endopeptidase digest of PTHrP-(l-141).
peaks were identified by mass spectrometry (and in some cases by direct amino acid sequencing). Table I shows that good agreement was obtained between the predicted and actual molecular masses. In particular, the mass spectrum of peak 3 (Fig. 4) shows that it is composed of the expected carboxyl-terminal peptide, demonstrating that the majority of purified PTHrP-(1-141) is in factfull-length. No indication of any modification such as formylation of the amino terminus was found in the mass spectrum of peak 1. PTHrP-(1-108) was digested with endopeptidase Asp-N, and the peptides were separated (Fig. 3B). Again, molecular ion mass spectrometry gave the expected molecular masses (Table I), demonstrating that thematerial does in fact contain amino acids 1108. The biological activity of the two purified forms of PTHrP was determined with the ratosteogenic sarcoma cell line UMR 106-01. Both the initial rapid production of cyclic AMP and the longer term production of tissue plasminogen activator were assayed. Previous work (4, 12) hadshown that PTHrP(1-34) or PTHrP isolated from BEN cell culture medium is several times more active in both these assaysthan bovine or human PTH-(1-34). Fig. 5A shows that PTHrP-(1-141) and -(l-108) are about equipotent on a molar basis with chemically synthesized PTHrP-(1-34) in the stimulation of cyclic AMP production. In contrast to the equipotent cyclic AMP stimulation, Fig. 5B shows that recombinant PTHrP-(1-141) and -(l-108) are 2-4-fold more active in the stimulation of
001
0
01
1
10
25
peptlde InMI
FIG.5. UMR 106-01 response to PTHrP. A, cyclic AMP; B, plasminogen activator. 0, PTHrP-(1-34); A, PTHrP-(1-108); 0, PTHrP-(1-141).
20 % baund
10 -
01 0
10
102
IPeptldel
lo3
10'
ifmol/tubel
FIG.6. Radioimmunoassay of PTHrP. 0, PTHrP-(1-141); 0, PTHrP-(1-34).
14810
Recombinant PTHrP
gel electrophoresis, amino acid sequencing, protease mapping, and mass spectral analysis. These data show that the two forms of PTHrP arehighly purified, especially in the case of PTHrP-(1-108). The minor contaminating bands visible on SDS gel analysis of PTHrP-(1-141) appear to be a series of proteolytic degradation products cleaved at thecarboxyl terminus. Proteolytic cleavage or degradation also appears to be one mechanism for the generation of various forms of PTHrP isolated from natural sources (see below). Mass spectral analysis of lysyl endopeptidase- and endopeptidase Asp-N-derived peptides demonstrates that bothrecombinant forms of bPTH(1-341 PTHrP ( 1 - 3 4 1 PTHrPI1.141) PTHrP areprimarily the expected lengths, containing 141or lOng/ml 2 ng/ml 20 nq/ml 108 amino acids. FIG. 7. Specific antiseruminhibition of cyclic AMP reThe PTH-like biological activity of the two recombinant sponse of U M R 106-01 cells to PTHrP and bovine PTH (bPTH). Open burs, nonimmune rabbit serum; stippled burs, PTH- forms of PTHrP was assessed in two assay systems. One (1-34) antiserum; solid burs, PTHrP-(1-34) antiserum. measures an early signal response, stimulation of cyclic AMP levels, whereas the otherdetects a later cellular event, increase B in plasminogen activator activity. In previous work (17), we h have shown that the stimulation of plasminogen activator activity is a cyclic AMP-mediated response to PTH. Studies with PTH antagonists have suggested that both the PTH-(134) and PTHrP-(1-34) responses in this system are mediated by the PTH receptor (4). Inthis work, the cyclic AMP response to recombinant PTHrP-(1-141) is the same on a molar basis as that toPTHrP-(1-34). Thus, in this respect, it is concluded that the PTH-like action of PTHrP is contained completely within the first 34 residues. Similar experfraction no. iments have never been carried out with PTH itself because FIG. 8. Vydac CIS HPLC separation of two forms of BEN of the uncertain purity of full-length PTH, although it is cell-derived PTHrP (4). -, absorbance at 210 nm; O ” 0 , generally believed that all the cyclic AMP-related effects of cyclic AMP response of UMR 106-01 cells to each fraction; - - -, PTH are contained in the amino-terminal 34 residues. Our acetronitrile gradient. The two main peaks of activity eluted at 32% consistent finding of slightly greater potencies of PTHrP-( 1(peak A ) and 37% (peak B ) acetonitrile. 141) and -( 1-108) as compared to PTHrP-(1-34) in promoting plasminogen activator activity may with further study throw light on a cellular response distinguishable from those of PTH. Thiswork provides no clear evidence that PTHrPbinds to a different receptor than PTHrP-(1-34) (or PTH-(1-34)) to produce the PTH-like effects. This does not exclude the -30 K possibility, however, that PTHrPmay also exert other effects -20 K through its own specific receptor. Such a possibility is suggested by our studies (19) which show that PTHrP, but not -14.4K PTH, stimulates calcium transport across sheep placenta. A comparison of antibody reactivity as well as theneutralization of biological activity was made between PTHrP-(134) and -( 1-141) using an antiserum prepared against PTHrP(1-34). Essentially identical results have also been obtained 1 2 3 4 5 6 with antiserum to PTHrP-(1-16) (data notshown). The antiFIG. 9. Size analysis of various PTHrP samples. Gel and serum recognized PTHrP-(1-34) and full-length PTHrP-(1immunoblot analyses were performed as described under “Experi141) equally on a molar basis under radioimmunoassay conmental Procedures.” Lune 1, PTHrP-(1-141); lane 2, PTHrP-(1-108); lune 3, PTHrP-(1-84); lune 4, BEN cell-derived peak A material; lane ditions. Furthermore, incubation at high antiserum concentrations completely neutralized the biological activity of re5, BEN cell-derived peak B material (with additional purification combinant PTHrP. It is important to note that even under (4)); lane 6, size standards. these conditions of high antibody concentration,there was no The analysis of the purified proteins by SDS gel electropho- neutralization of the activity of PTH, and the anti-PTH antiserum did not affect the PTHrP activity. Thus, despite resis (Fig. 2) shows that PTHrP-(1-141) has an apparent molecular mass of 23 kDa and that PTHrP-(1-108) has an the close structural similarities of PTHrP and PTH, their apparent molecular mass of 16 kDa. Based on the calculated immunological differences are striking. We have made use of molecular mass for PTHrP-(1-141) of 16 kDa, we had sug- this specificity to localize PTHrP immunohistologically in gested previously (5) that peak B would correspond to full- tumors and normaltissues (21, 22) and to neutralize the length PTHrP-(1-141). Fig. 9 shows that in fact peakB, with biological actions of PTHrP in uiuo (7). an apparentmolecular mass of 17 kDa, corresponds approxiWe had proposed previously (4, 5) based on a calculated mately to PTHrP-(1-108). molecular mass of 16 kDa that full-length PTHrP-(1-141) corresponds to material purified from BEN cell culture meDISCUSSION dium which has an apparentmolecular mass of 17 kDa (peak Two forms of PTHrP produced by recombinant DNA tech- B material). The finding that recombinant PTHrP-(1-141) niques in E. coli have been purified and characterized by SDS hasanapparent molecular mass of 23 kDa on SDS gel
Recombinant PTHrP electrophoresis shows that in fact thisproposal was incorrect. We have now used the full-length PTHrP-(1-141) and two shortened forms, PTHrP-(1-108) and -(1-84), as specific size 6. standards to establish the actual length of BEN cell-derived PTHrP (peak B) as well as a less purified BEN cell fraction (peak A) (4). Based on the SDSgel mobility of these proteins, 7. the peak A material is approximately full-length, 141 amino acids, whereas the peak B material is about 108 amino acids long. It should also be noted that alternative splicing has been 8. reported at thecarboxyl terminus of PTHrP which produces a form of the protein shortened by two amino acids (23) (as well as some other forms (24)). Immunoblottingas performed 9. in this work would not be expected to resolve these forms of PTHrP which may be present in the peak A material. The 10. BEN cell line has consistently yielded a substantial amount of a biologically active species (peak B) with actual molecular 11. mass of about 13 kDa (4). In contrast, PTHrPpurified from a kidney cancer cell line has an apparent molecular mass of 6-9 kDa on SDS gel electrophoresis (10). It will beinteresting to determinewhether processing of the carboxyl-terminal 12. portion of PTHrP is in any way specific either tonormal cells producing the hormone or to tumors. Information on this point will provide essential background for the development 13. of two site assays for PTHrP in plasma samples. The discovery of PTHrP has implications for our understanding of normal calcium metabolism as well as its abnor- 14. malities which commonly occur in cancer. It will beimportant 15. to define the action of PTHrP more fully, as well as the nature 16. of its secreted forms. Recombinant expression of altered forms of the molecule provides a valuable approach to thiswork. While this article was being prepared for publication, a 17. reporton the expression and purification of recombinant 18. PTHrP-(1-141) was published (25). Acknowkdgments-We thank J. Stults and J. Bourell for mass spectrometry, M. Struble for preparative HPLC, B. Kohr for amino acid sequencing, and E. Allan for plasminogen activator assays.
1. 2. 3. 4.
5.
REFERENCES Mundy, G. R., Ibbotson, K. J., D’Souza, S. M., Simpson, E. L., Jacobs, J. W., and Martin, T. J. (1984) N . Engl. J. Med. 3 1 0 , 1718-1727 Broadus, A. E., Mangin, M., Ikeda, K., Insogna, K. L., Weir, E. C., Burtis, W. J., and Stewart, A. F. (1988) N . Engl. J. Med. 319,556-563 Martin, T. J. (1988) Bone Miner. 4, 83-89 Moseley, J. M., Kubota, M., Diefenbach-Jagger, H., Wettenhall, R. E. H., Kemp, B. E., Suva, L. J., Rodda, C. P., Ebeling, P. R., Hudson, P. J., Zajac, J. D., and Martin, T. J. (1987) Proc. Natl. Acad. Sci. U. S. A. 8 4 , 5048-5052 Suva, L. J., Winslow, G. A., Wettenhall, R. E. H., Hammonds,
19. 20. 21. 22. 23. 24. 25.
14811
R. G., Moseley, J . M., Diefenbach-Jagger, H., Rodda, C. P., Kemp, B. E., Rodriguez, H., Chen, E. Y., Hudson, P. J., Martin, T. J., and Wood, W. I. (1987) Science 237,893-896 Mangin, M., Webb, A. C., Dreyer, B. E., Posillico, J. T., Ikeda, K., Weir, E. C., Stewart, A. F., Bander, N. H., Milstone, L., Barton, D.E., Francke, U., and Broadus, A. E. (1988) Proc. Natl. Acad. Sci. U. S. A. 8 5 , 597-601 Kukreja, S. C., Shevrin, D. H., Wimbiscus, S. A,, Ebeling, P. R., Danks, J. A., Rodda, C. P., Wood, W. I., and Martin, T. J. (1988) J. Clin. Znuest. 8 2 , 1798-1802 Burtis, W. J., Wu, T., Bunch, C., Wysolmerski, J. J., Insogna, K. L., Weir, E. C., Broadus, A. E., and Stewart, A. F. (1987) J. Biol. Chem. 262,7151-7156 Stewart, A. F., Wu, T., Goumas, D., Burtis, W. J., and Broadus, A. E. (1987) Biochem. Biophys. Res. Commun. 146,672-678 Strewler, G. J., Stern, P.H., Jacobs, J. W., Eveloff, J., Klein, R. F., Leung, S. C., Rosenblatt, M., and Nissenson, R. A. (1987) J. Clin. Inuest. 8 0 , 1803-1807 Horiuchi, N., Caulfield, M. P., Fisher, J. E., Godman, M. E., McKee, R. L., Reagan, J. E., Levy, J . J., Nutt, R. F., Rodan, S. B., Schofield, T. L., Clemens, T. L., and Rosenblatt, M. (1987) Science 238,1566-1568 Kemp, B. E., Moseley, J. M., Rodda, C. P., Ebeling, P.R., Wettenhall, R. E. H., Stapleton, D., Diefenbach-Jagger, H., Ure, F., Michelangeli, V. P., Simmons, H. A., Raisz, L. G., and Martin, T. J. (1987) Science 238, 1568-1570 Ellison, M., Woodhouse, D., Hillyard, C. J.,Dowsett, M., Coombes, R. C., Gilby, E. D., Greenberg, P. B., and Neville, A. M. (1975) Brit. J. Cancer 3 2 , 373-379 Martin, T. J., Ng, K. W., Partridge, N.C., and Livesey, S. A. (1987) Methods Enzymol. 1 4 5 , 324-336 Miller, H. (1987) Methods Enzymol. 152,145-170 Maniatis, T., Fritsch, E. F., and Sambrook, J. (1982) Molecular Cloning: ALaboratoryManual, p. 440, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY Hamilton, J. A., Lingelback, S. R., Partridge, N. C., and Martin, T. J. (1985) Endocrinology 116,2186-2191 Findlay, D.M., DeLuise, M., Michelangeli, V. P., Ellison, M., and Martin, T. J. (1980) Cancer Res. 4 0 , 1311-1317 Rodda, C. P., Kubota, M., Heath, J. A., Ebeling, P. R., Moseley, J. M., Care, A. D., Caple, I. W., and Martin, T. J . (1988) J. Endocrinol. 1 1 7 , 261-271 Laemmli, U. K. (1970) Nature 227, 680-685 Danks, J . A., Ebeling, P. R., Hayman, J., Chou, S. T., Moseley, J. M., Dunlop, J., Kemp, B. E., and Martin, T. J. (1989) J. Bone Miner. Res., 4, 273-278 Hayman, J., Danks, J. A., Ebeling, P. R., Moseley, J. M., Kemp, B. E., and Martin,T. J. (1989) J. Pathol., in press Thiede, M. A., Strewler, G. J., Nissenson, R. A,, Rosenblatt, M., and Rodan, G. A. (1988) Proc. Natl. Acad. Sci. U. S. A . 85, 4605-4609 Mangin, M., Ikeda, K., Dreyer, B. E., Milstone, L., and Broadus, A. E. (1988) Mol. Endocrinol. 2 , 1049-1055 Thorikay, M., Kramer, S., Reynolds, F. H., Sorvillo, J. M., Doescher, L., Wu, T., Morris, C. A., Burtis, W. J., Insogna, K. L., Valenzuela, D. M., and Stewart, A. F. (1989) Endocrinology 124,111-118
