3, Issue of January 25, pp. 1307-1313,1988. Printed in U. SA. ... World Health Organization standard of extracted na- tive hPTH-(1-84). ... osteosarcoma cells its activity was identical to that of ... at the Ninth Annual Meeting of the American Society for Bone and ...... of Endocrinology and Metabolism (Becker, K. L., ed) J. B.
Vol. 263, No. 3, Issue of January 25, pp. 1307-1313,1988 Printed in U.S A .
THEJOURNAL OF BIOLOGICAL CHEMISTRY Q 1988 by The American Society for Biochemistry and Molecular Biology, Inc.
Recombinant Human Parathyroid Hormone Synthesized in Escherichia coli PURIFICATION AND CHARACTERIZATION* (Received for publication, August 20, 1987)
Shafaat A. Rabbani3, Toshiyuki Yasuda, Hugh P. J. Bennett&Wing L. Sung, DianaM. Zahab, N. Cherk S. Tam, David Goltzmanv, and GeoffreyHendyll From the Departments of Medicine, McGiU University and Royal Victoria Hospital, Montreal, Quebec, Canada, H3A lA1, the Division of Biological Sciences,National Research Council of Canada, Ottawa, Ontario, Canada, K1A OR6, and the Departments of Pathology and Medicine, university of Toronto, Toronto General Hospital, Toronto, Ontario, Canada, M5G lL7
lated from human parathyroid tumors and its complete priRecombinant human parathyroid hormone (hPTH) was expressed inEscherichia coli harboring a plasmid mary structure determined by amino acid sequencing (4). In addition, using molecular cloning techniques the structures of containing a synthetic humanparathyroidhormone gene under the control of the E. coli lac promotor. the mRNA and gene for human prepro-PTH have been deThree major forms of the hormone were isolated by termined (5, 6). This has allowed the introduction of cDNA acid extraction and purified to homogeneity by high molecules encoding human prepro-PTH intoprokaryotic and performance liquid chromatography. By amino acid eukaryotic cells (7). When the human prepro-PTH cDNA analysis andNHs-terminalsequencing,these were was expressed in Escherichia coli and yeast, the human preidentified as hPTH-(1-84), formyl-methionyl-hPTH- pro-PTH synthesized was not processed to hPTH-(1-84) (1-84), and hPTH-(8-84). The recombinant hPTH-( 1-apparently because the mammalian signal sequence was not 84) was immunologicallyindistinguishable from a appropriately cleaved by the signal peptidase (8, 9). In this World Health Organization standard of extracted na- paper we describe the construction of a plasmid which directs tive hPTH-(1-84). Recombinant hPTH-(1-84) was the expression of hPTH-(1-84) from a synthetic gene under also bioactive in renal and skeletal adenylate cyclase the control of the E. coli lac promotor. The products of assays. In the skeletal bioassay performed inUMR 108 osteosarcoma cells its activity was identical to that of expression of the plasmid were isolated by acid extraction an hPTH-(1-84) standard. In this bioassay, formyl- followed byHPLC. They were characterized by region-specific methionyl-hPTH-(1-84) had 10% ofthe activity of immunoassay, bioassay, amino acid analysis, and NH,-terhPTH-(1-84) and hPTH-(8-84) was inactive. The re- minal sequencing. sults demonstrate the importance of isolating hPTHEXPERIMENTAL PROCEDURES (1-84) from other recombinant forms and metabolites to achieve full hormonal bioactivity and indicate that Materials-E. coli K12 strain JM103 (A(& pro), thi, str A, sup E, purified recombinant hPTH-( 1-84) can thereby be ob- end A , sbc B, hsd R, Ftra D36, pro AB, lac IQ, ZA M15) was used as tained which should be a useful source of hormone for the bacterial host. Synthetic oligonucleotides were prepared using an both basic andclinical studies. Applied Biospstems DNA Synthesizer, Model 380A. Enzymes and plasmid pUC8 were purchased from Bethesda Research Laboratories and Boehringer Mannheim. Construction of the Expression Plasmid-Fig. 1 outlines the construction scheme of the E. coli plasmid vector that directed the Parathyroid hormone (PTH)’ is the major regulator of expression of hPTH DNA. The construction of pPTH-84, which extracellular calcium concentrations (1-3). It has been iso- bears a synthetic hPTH gene, was detailed previously (10).pPTH-84 was digested with EcoRI and PstI to remove the 5’ portion of the hPTH gene. The large restriction fragment bounded by EcoRI and * This work wassupported by Grants MA 9315, MT 5775, and MT PstI sites was isolated and ligated with six synthetic overlapping 5917 from the Medical Research Council of Canada. This is NRCC oligonucleotides which reconstructed the 5’ end of the hPTH gene publication 28466. A preliminary report of this work was presented with an ATG starting codon at amino acid -1 (Fig. 2). A crossover at the NinthAnnual Meeting of the American Society for Bone and linker sequence was designed at theupstream end, which was homolMineral Research, Indianapolis, IL. The costs of publication of this ogous to thesequence encompassing the ribosomal binding site to the article were defrayed in part by the payment of page charges. This starting ATG of the @-galactosidasegene already present in the article must therefore be hereby marked “advertisement” in accord- opposite terminus of the plasmid intermediate. After transformation ance with 18 U.S.C. Section 1734 solely to indicate this fact. of E. coli JM103 the homologous termini recombined in oiuo (11) to $ Recipient of a fellowship from the Medical Research Council of yield plasmid pPTH-84c. The new plasmid pPTH-84c was present in Canada in support of this work. 4% of all transformants. The construction of the plasmid in the 5 Recipient of a fellowship from the Fonds de la recherche en santk region of the ribosomal binding site and hPTH coding region was du Quebec. confirmed by nucleotide sequencing. Expression in E. coli-One liter of Luria broth containingampicilll Recipient of a scientist award from the Medical Research Council of Canada in supportof this work. lin was inoculated with a 20-ml overnight growth of E. coli transform11 Recipient of a scholarship from the Medical Research Council of ants bearing pPTH-84c and incubated at 37 “C with shaking for 2 h Canada in support of this work. To whom reprint requests should be until an ASS6 of 0.1-0.2 was obtained. Then 1 mM IPTG was added addressed McGill University, Allan Research Bldg., Rm. 117, 1033 and incubation was continued for a further 7 h. A control incubation Pine Avenue West, Montreal, Quebec, Canada, H3A 1Al. to which no IPTG was added was set up in an otherwise identical ’ The abbreviations used are: PTH, parathyroid hormone; h, hu- manner. Cells were harvested by centrifugation (5000 rpm, 10 “C, 10 man; b, bovine, m e t , formyl-methionyl; HPLC, high performance min) and thecell pellets frozen at -20 “C. The medium was returned liquid chromatography; IPTG, isopropylthiogalactoside. to the culture flasks and incubation a t 37 ‘C continued for a further
1307
Characterization of Recombinant Human Parathyroid Hormone
1308
17 h. Cells were harvested as before and combined with the other cell pellets. Throughout the initial incubation period, 1-ml aliquots of the incubations were removed at hourly intervals, the cells were collected by centrifugation in a Microfuge for 20 s and lysed by the addition of
EcoRI
PstI
0
m
PIC PII Pm
pPTH-84
maq-J
Prnc Pm PE ligose
in vivo recornbinotion
FIG. 1. General scheme for the construction of E. coli plasmid vector pPTH-84c that directs the expression of hPTH DNA. The black bar represents the sequence from the ribosomal binding site to the startingcodon ATG of the E. coli &galactosidase gene. The stippled bars represent synthetic oligonucleotidesencoding the amino-terminal portion of hPTH and thehatched bar represents the NH, terminus of the &galactosidase gene. The dephosphorylated EcoRI-PstI restriction fragment from pPTH-84 was ligated with the phosphorylated oligonucleotides as described under "Experimental Procedures." In subsequent transformation of E. coli JM103, the homology searching sequence (black bar) of the P1c:PVIIIc duplex recombined intramolecularly with the lac promotor-operator to complete the circularization of the plasmid. Through this operation, the NH,-terminal @-galactosidasesequence was specifically deleted and the PTHgene wasdirectly linked to thelac operator to form plasmid pPTH-84~.
1acZpo
125 plof a solution of 4 M guanidinium thiocyanate, 25 mM trisodium citrate, 1mM EDTA, and 0.1 M j3-mercaptoethanol.Aliquots of these extracts were analyzed by radioimmunoassay. In some experiments cultures of E. coli transformants bearing pPTH-84c were grown as described above and 150 pg/ml chloramphenicol was added 5 h after induction with IPTG. Aliquots of the cultures were taken before and after addition of chloramphenicol and extracted as described above for radioimmunoassay. Extraction Procedure-Cells were homogenized at 0 "C in acetone (30 ml/g). The suspension was then filtered and the residue homogenized in n-hexane (30 ml/g). This new suspension was filtered and the residue homogenized in acetone (30 ml/g) and thenvacuum dried. The dried and defatted residue was extracted in a mixture (40 ml/g) of 1M HCI containing 5% (v/v) formic acid, 1%(w/v) NaCI, and 1% (v/v) trifluoroacetic acid (CF,COOH) (12). HPLC-Preparative scale chromatography was performed on a Waters Associate (Milford, MA) HPLC system consisting of a 1000 Prep-Pak Module and a 3000 system controller. Analytical scale chromatography was performed on a Waters HPLCsystem consisting of one 6000A pump, one "45 pump, and anM720 controller. Column eluates from analytical runs were monitored for UV absorbance at 280 or 210 nm using a variable wavelength flow-through spectrophotometer (model M480, Waters) and an M730 data module (Waters). The supernatantfrom the acid extraction of the dried and defatted E. coli cells was pumped directly onto the Prep-Pak C,, cartridge. Loading of the preparative column, housed within the Prep-Pak modulewas achieved with a Masterflex peristaltic pump (ColeParmer Instrument Co., Chicago, IL) a t a flow rate of 10 ml/min. Loading was not done through the HPLC pumps to avoid potential damage due to the very acid nature of the extraction medium. Once loaded, the Prep-Pakmodule wasconnected to theHPLC pumps and equilibrated with 0.1% CF3COOHfor 10 min at 100 ml/min. The C,, cartridge was then step-eluted with 80% acetonitrile containing 0.1% CF3COOH for 7 min at 100 ml/min. This initial Cls-silica extract was then subjected to reversed-phase HPLC by diluting it with 0.1% CF3COOH and pumping it at 75 ml/min onto the Prep-Pak module into which was fitted a custom-packed Vydac CIS (15-20 PM) cartridge. This was eluted with a gradient of acetonitrile in 0.1% CF3COOH at 50 ml/min. Reverse-phase HPLC was performed as described previously (12, 13) on CIS pBondapak columns (Waters). Samples for loading were diluted with 0.1%CF3COOH or 0.13%
-
c- 8-Gal-
PTH
29 28 1-27 ACAGGAAACAGCTATG ACC ATG A l l ACG AAT TC... TG-GTCGATACTGGTACTAATGC TTA AG.. RBS pPTH-84
............GAC CTG CAG........... GTC-
FIG. 2. Details of theconstruction of plasmid pPTH-84c from plasmid pPTH-84 by intramolecular recombination. lacZpo is the promotor-operator region of the lacZ gene and RES is the ribosomal binding site of the p-galactosidase gene (@-Gal).CZAP is calf intestinal alkaline phosphatase. PIC,PII, PIII, PVI, PVII, and PVIIIc aresynthetic oligonucleotides. PI1 and PI11 make up the coding strand andPVII and PVI make up the complementary strand of hPTH-(4-29). PIC and PVIIIc make up the coding strand andcomplementary strand, respectively, of the ribosomal binding site, the ATG and hPTH-(1-4). The homologous sequences for intramolecular recombination between the lac operator and theoligomers P1c:PVIIc in the plasmid intermediate are contained in the box.
PstI
1)EcoRI,PstInucleases 2 ) CIAP 3 ) P I C ,P I I , 1 igase
lacZpo
" " " " " " " C
AT^
B -Gal
-
PIII, P V I ,P V I I ,P V I I I C
~ACAGGAAACAGCT ACC ATG ATT ACG ~TG-GTCGA TAC: TGG TAC TM TGC TTA A 5 '
:
RBS
EcoRI
II
PTH
I *
:
PIC 1 2 5-27 3 4 ~ACAGG~AACAGCT ATGiTCGGTTTCT GA L>T~-X~Jjflfl~cj AGC CAA AGA CTCTAG.. PVI II C
28 29 .................... .CTG CAG...............GACGTC
-1 PstI
recombi n a t i on
1 acZpo
start I 2 329 4 28 5-27 ACAGGAAACAGCTATGTCG GTT TCT GA T G W G T C G A TACAGC CAA AGA CTC TAG..
PTH
................... ..CTG CAG............... GAC GTC-
AS
PPTH-84~
PstI
Characterization of Recombinant Human Parathyroid Hormone I PTG 12r
1.0.
-
.+ a -
IPTG IPTG
0.8 '
I 0.6. I
4
04.
0.2 .
Time ( h l
FIG. 3. Effect ofaddition of IPTG to cultures of E. coli cells harboring plasmid pPTH-84c. The cultures were grown, aliquots were taken at the times .indicated, cell extracts made, and radioimmunoassay (C assay) carried out, as described under "Experimental Procedures." D is the detection limit of the radioimmunoassay. Cell growth was monitored by measurement of A596. heptafluorobutyric acid (C,F,COOH) and columns weredeveloped over 60-100 min a t a flow rate of 1.5 ml/min with linear gradients of acetonitrile containing 0.1% CF3COOH or 0.13%C3F7COOH as a counter ion. Gel filtration HPLC (14) was carried out on an 1-125 gelpermeation HPLC column (Waters) connected in series with a Protein Pak 300SW gel permeation column (Waters), eluting at a flow rate of 1 ml/ min with 40% aqueous acetonitrile containing0.1% CF3COOH.Loading onto the column was achieved by trace-enrichment ontoa GuardPak C,, column (Waters) followed by step elution onto the gel filtration column using the 40% acetonitrile, 0.1% CF3COOHsolvent system. This procedure was repeated several times to load the entire sample.*Calibration was achieved by injecting a mixture of 0.5 pg of each of the following in 25 r l of 40% acetonitrile containing 0.1% CF,COOH, chymotrypsinogen, ribonuclease, ubiquitin, adrenocorticotropic hormone, insulin &chain, and a-melanocyte-stimulating hormone. Aliquots of isolated peptides were injected in 25 p1 of 0.1% CF,COOH. Zmmumsays-Radioimmunoassays for parathyroid hormone were carried out aspreviously described (15) using either guinea pig antiserum GP 26 raised against bPTH-(1-84) and ['261-Tyr6Z]hPTH(52-84) as tracer (C assay) or rabbit antiserum 3149 raised against bPTH-(1-34) and "I-bPTH-(l-84) as tracer (N assay). In most assays, the standardused was synthetic hPTH-(1-84)purchased from Bachem Corp., Torrance, CA. In additional assays with highly purified recombinant hPTH peptides, a World Health Organization (WHO) standard (16) was used. Bioassay-Zn vitro adenylate cyclase assays were performed in cloned rat osteosarcoma cells (UMR 108) as described previously (13). An adenylate cyclase assay using purified rat renal cortical membranes was carried out as described previously (17). In both bioassays the standard used was synthetic hPTH-(1-84) (Bachem). This peptide was equipotent with a Medical Research Council (MRC) standard of bPTH (National Institutes of Biological Standards and Control reagent 77/651) which had a biological activity of 2500 MRC units/mg (approximately equivalent to 2500 USP units/mg). Amino Acid Analysis-Approximately 0.1-0.2 pmol of each of the purified peptides was subjected to HCl vapor hydrolysis at 110 'C using a Pico-Tag Work Station (Waters). Amino acid analyses were performed on a Beckman high performance system 6300 analyzer using a step elution program. Amino Acid Sequencing-Sequencing of purified peptides was performed on 0.5 nmol using an Applied Biosystems (gas-phase) sequencer. Phenylthiohydantoin-derivativeswere identified by an Applied Biosystems on-line HPLC. RESULTS
Characteristics of the Expression of Immunoreactive Recombinant PTH-A substantial increase in immunoreactive S. James and H. P. J. Bennett, manuscript in preparation.
1309
hPTH synthesis occurred when transcription of the hPTH gene wasinduced with IPTG (Fig. 3). The peak production of hPTH, whichwas approximately 200 pg/liter of medium, occurred 7 h after additionof IPTG. Thisrepresented greater than 2% ofthe total acid-extractable E. coliprotein. Although some synthesis of hPTH occurred in non-induced E. coli cells containing pPTH-84c, it was at a reduced level and was delayed by 2 h relative to that observed in the induced cells. No immunoreactive PTH was detected in extracts of E. coli cells containing the control plasmid pUC8. When protein synthesiswas inhibited by addition of chloramphenicol 5 hafter induction with IPTG,the levels of immunoreactive hPTH fell rapidly with a tlhof approximately 12 min (Fig. 4). Equivalent amounts of N- and C-immunoreactivity were found in crude E. coli extracts (Fig. 5). In both assays the E. coli extract immunoreactivity diluted in parallel to the hPTH(1-84) standard, suggesting that it bore a very close immunological similarity to human PTH.
I
n
1160
FIG. 4. Study of the stability of immunoreactive PTH in E. coli cells containing pPTH-84c. Transcription of the hPTHgene was induced by addition of IPTG at time 0. Chloramphenicol (Cm) was added after 5 h to inhibit protein synthesis. The clearance of PTH was monitored by radioimmunoassay (N assay) of cell extracts and cell growth was monitored by measurement of A, (0).Immunoreactive PTH (iPTH) (A) was rapidly cleared with t,+ of approximately 12 min.
\.
loo"-.
c assay
N assay
so.
\\.
60 B/ x 100 00
40 20 -
hPTH
- ( I - 8 4 ) , moles I
01
I
10
Recomblnant
01
hPTH
x
.
IO"'
..
,..I.
I
I
I
. . ....J10
I pl I
FIG. 5. Effect of increasing amounts of extracts of E. coli cells containing the hPTH expression plasmid in radioimmunoassays specific for both the amino-terminal ( N assay) or carboxyl-terminal (Ccrssay) regions of the molecule. Aliquota of acid-extracted E. coli cells (A) and of standard hPTH-(1-84) (0)were radioimmunoassayed as described under "Experimental Procedures."
1310
Characterization of Recombinant Human ParathyroidHormone Acid e x t r a c t i o n ( 1 0 0 )
(67
RF-TFA
FIG. 6. Procedure used to isolate hPTH peptides from extracts of E. coli cells harboring plasmid pPTH84c. Numbers in parentheses represent the percentage of immunoreactive hPTH relative to theamount (100) present in the initial acid extract. Numbers in square brackets are the total relative amounts of immunoreactive hPTH (C assay) as represented in either the a and b forms, or a, b,, and b, forms of recombinant hPTH. RF-TFA indicates the use of reversed-phase chromatography with 0.1% CFsCOOH as counter ion and RFHFBA indicates the use of reversedphase chromatography with 0.13% C3F7COOHas counter ion.
on
RF-HFBA
(59)
.
1
1
RF-TFA
RF-TFA
Gel f i l t r a t i o n
Gel f i l t r a t i o n
Gel f i l t r a t i o n
/
FIG. 7. Selected elution profiles of HPLC purification of recombinant hPTHpeptides. HPLC and radioimmunoassay were carried out as described under “Experimental Procedures.” D is the detection limit of the radioimmunoassay. P a w l A , elution profile of immunoreactive PTH (iPTH) on a Vydac C,, column eluted with a gradient of acetonitrile, 0.1% CF3COOH (CH3CN/TFA).Panel B, elution profile of immunoreactive PTH purified bygel filtration chromatography using an acetonitrile, 0.1% CF3COOH solvent system. The arrow indicates the elution position of standard hPTH-(184). The bar indicates the fractions taken and processed further. Panel C, separation of recombinant hPTH into two forms, a and h, on reversed-phase HPLC eluted with a gradient of acetonitrile, 0.13% C3F,COOH (CH,CN/HFBA). Panel D,separation of hPTH peptide b into two forms, b, and bz, on reversed-phase HPLC eluted with an acetonitrile, 0.1% CF3COOH(CH3CN/TFA)gradient.
b
20
30
40
50
20 40 30
Time ( min)
50
Characterization of Recombinant Human Isolation and Purification of PTH-related Moieties-Fig. 6 outlines the extraction and isolationprocedures used. Numbers in parentheses a t each step refer to immunoreactive PTH (C assay) relative to the amountof immunoreactive PTH in the acid extract (100%). At all steps the amounts of N assayimmunoreactive P T H were very similar to C assay-immunoreactive PTH (data not shown). Approximately 20% of the immunoreactive P T H present in the initial acid extract was recovered at the final purification stage in three major forms designated a, b,, and b,. These were present in approximately equal amounts. HPLC elutionprofiles of acid extracts are shown in Fig. 7. Two incompletely resolving peaks of immunoreactivity were observed on chromatography over a Vydac C,, column using a gradient of acetonitrile, 0.1% in CF3COOH (panel A ) . Bot,h peaks of immunoreactivity co-eluted with standard hPTH(1-84) upon gel filtration HPLC (panel B ) . When this material was chromatographedon areversed-phasecolumn whichwas eluted with a gradient of acetonitrile in 0.13% C,F,COOH, two peaks (a and b) of immunoreactivity were
l o5o /0
Parathyroid Hormone
1311
resolved (panel C). When peak b was rechromatographed on a reversed-phase column, eluting with a shallow gradient of acetonitrile in 0.1% CF&OOH, two discrete peaks, bl andb,, were seen (panel D). Each of the three major forms a, bl, andb, showed homogeneous UV profiles when examined by gel filtration HPLC and each had an apparent molecular weightequivalent to that of standard human PTH-(1-84) (Fig. 8). Chemical and Biological Characterization of PTH-related Moieties-When each peptide was subjected to amino acid analysis, the composition of bl was in keeping with that of hPTH-(1-84). The composition of b, was similar except for the presence of 1 extra methionine residue per mol. Amino that acid analysis of a, also revealed a composition resembling of hPTH-(1-84), but with lower amounts of serine, valine, glutamic acid, isoleucine, and leucine (Table I). Each peptide was then subjected to amino acid sequence analysis(Table 11). Peptide a had the sequence of hPTH beginning with methionine at residue 8. Peptide b, had the sequence of hPTH beginning with serine at position 1.Peptide b, could not be sequenced, consistent with the presence of a blocked NH, terminus. Taken together, the retention timesof the different forms on reversed-phase HPLC, the amino acid compositions, and NH,-terminal sequence data identified peptide a as hPTH(8-84), peptide b, as hPTH-(1-84), and peptide b, a s M e t hPTH-(1-84). Ininitialstudiesthe bioactivity of thethreeforms of TABLE I1 Comparison of the NH2-terminalsequences of recombinant hPTH peptides and native hPTH 500 pmol each of a, b,, and bz were subjected to 25 cycles of Nh2terminal sequencing as described under “Experimental Procedures.” Only the results for the first 10 cycles are shown; no sequence was obtained for fraction b,. Native human PTH
Time (mln 1
8
1
NHZ-Ser-Val-Ser-Glu-Ile-Gln-Leu-Met-His-Asn-a
FIG. 8. Gel filtration HPLC of the three major forms of hPTH (a, b,, and b,) after purification. HPLC was carried out as described under “Experimental Procedures” using gel filtration columns calibrated with a variety of molecular weight peptide markers (0).All three recombinant hPTH forms (a, b,, and bz), as well as standard hPTH-(1-84) (O),had molecular weights of approximately 10,000.
Recombinant human PTH a NH2-Met-His-Asn-Leu-Gly-Lys-His-Leu-Asn-Ser-
b, NH,-Ser-Val-Ser-Glu-Ile-Gln-Leu-Met-His-Asn-
b, a
-
-
.
-
-
-
-
-
-
-
From Ref. 4.
TABLEI Amino acid comwosition of HPLC-wurified recombinant human PTH weaks ~~
~~
Mol fraction
Residues/ mol
1.55 0.56 0.86 0.57 0.22 0.81 1.38 0.65 0.69 1.27 1.11 0.36 0.05 1.50
9.6 3.5 5.3 9.8 1.3 5.0 8.6
LYs His Arg Aspb Thr Ser Glub Pro GlY Ala Val Met Ile Leu TYr Phe
0.04 0.16 a
4.0 4.3 7.9 6.9 2.2 0.3 9.3 0.2
1.o
Human pTH-(8-84)”
9 4 5
10 1 5 9 3 4 7 7 2
0 9 0
1
Mol fraction
1.70 0.77 1.oo 1.90 0.16 1.29 2.00 0.56 0.79 1.30 1.50 0.36 0.20 1.90 0.03 0.22
Predicted amino acid compositions. Amide nitrogen and tryptophan not determined.
’
Peak bt
Peak b,
Peak a Amino acid
Residues/ mol
Human pTH-(1-84)”
9.2
9
4.0
4
5.0 10.0 0.8 6.7 10.5 2.9
5 10 1 7 11
4.1 6.8 7.8 1.9 1.0 10.3 0.1
1.1
3 4 7 8 2 1
10 0 1
Human PTH fMet-(1-84)’
Mol fraction
Residues/ mol
0.50 0.15 0.25 0.52 0.04 0.34 0.59 0.16 0.20 0.36
10.0
9
3.1
4
5.0 10.6 0.8
10
0.40 0.16 0.06 0.52 0.01 0.06
6.8 11.8 3.2 4.0 7.0
8.0 3.3
1.3 10.4 0.2 1.3
5 1 7 11 3
4 7 8 3 1 10
0 1
1312
Characterization of Recombinant Human Parathyroid Hormone
DISCUSSION recombinant hPTH were tested in a renal adenylate cyclase assay. Peptide b, (hPTH-(1-84)) was active in the two conWe describe here the successful construction of an exprescentrations examined. Peptide bp (fMet-hPTH-(l-84)) and sion plasmid which efficiently directs the synthesis of hPTHpeptide a(hPTH-(8-84)) were without significant activity (1-84) in bacteria. This was achieved using a synthetic gene (Table 111). encoding hPTH-( 1-84) preceded by an ATG. This construct In more extensive studies, increasing concentrations of each was under the control of the E. coli lac promotor and, as we peptide were bioassayed in an adenylate cyclase assay per- show here, expression of the gene was enhanced severalfold formed in osteoblast-derived osteosarcoma cells (Fig. 9). Pep- by induction with IPTG. Thelevel of expression achieved was tidea(hPTH-(8-84)) was inactive inthis bioassay. The approximately 200 pglliter which is comparable to that obpotency of peptide b, (hPTH-(1-84)) was equivalent to that tained for other proteins expressed in E. coli under the control of standard hPTH-(1-84) and therefore had a biological ac- of the lac promotor (18).Increased levels of expression would tivity of 2500 MRC units/mg. However, peptide b2 displayed be anticipated using an even stronger promotor, for example, only about 10% of the potency of the standard on a molar the tac promotor and we are presently constructing such a basis. As peptide bl (hPTH-(1-84)) was the only immuno- plasmid. reactive moiety with full bioactivity the overall yield of bioacIt has been reported previously that plasmid constructs tivity was 5% of the total immunoreactivity present in the encoding an ATG followed bya genomic restriction fragment initial acid extract (see Fig. 6). encoding hPTH-(1-84), which was placed under the control When tested in a radioimmunoassay (C assay), the three of a variety of promotors, were successful in directing the highly purified recombinant hPTH peptides, peptide a synthesis, in E. coli, of hPTH which was immunologicallyand (hPTH-(8-84)), peptide b, (hPTH-(1-84)), and peptide bp biologically active (19). However, nopurification of the crude (fMet-hPTH-(1-84)) diluted in parallel with, and were equi- extract material described in that study was attempted. The potent with, a World Health Organization standard of ex- present work emphasizes the importance of the use of approtracted native hPTH-(1-84) (data not shown). priate purification methods to isolate authentic hPTH-(1-84) not only from E. coli proteins, but from other recombinant forms of hPTH. TheE. coli cells harboring the plasmid were TABLE 111 extracted using an acid-extraction technique followed by isoEffect of standard hPTH-(1-84) and of recombinant PTH fractions lationand purification on HPLC. This method has been on rat renal membrane adenylate cyclase activity Adenylate cyclase assay was performed as outlined under "Exper- shown previously to be suitable for a large number of native imental Procedures." Recombinant PTH fractions a, b,, and b2 were basic peptides, including parathyroid hormone (12). Three purified as shown in Fig. 6. major forms having PTH immunoreactivity were identified and were purified to homogeneity. They were characterized Adenylate Peptide cvclase activitv" as Met-hPTH-(l-84), hPTH-(1-84), and hPTH-(8-84). In E. coli, peptides are synthesized with an amino-terminal forngl0.l ml pmol cAMP/mg protein/30 min myl-methionine which may be removed enzymatically by a None (basal) 38.6 f 2.5 61.1b f 3.6 hPTH-(1-84) 180 methionyl-aminopeptidase.The relative efficiency with which 51.8* f 3.3 36 this occurs is determined in part by the nature of the neigh36.8 f 2.4 a 200 boring amino acid (20). It has been suggested previously that 40 38.5 f 0.4 the presence of serine next to the initiator methionine, as is 60.7bf 2.6 b, 200 the case in our hPTH expression plasmid, leads to efficient 52.gb f 0.7 40 cleavage bythe methionyl-aminopeptidaseenzyme. Certainly, 42.9 f 0.7 200 b, 38.1 f 0.5 40 in the study described here, a large proportion of the immunoreactive PTH was present in the form of mature hPTHMean f S.E. of triplicate determinations. (1-84). In future studies, it may be possible to generate an b p< 0.05 compared to basal (Student's t test). even greater proportion as mature hPTH-(1-84)by introducing the hPTH expression plasmid into E. coli cells already containing multiple copies of a plasmid expressing the methionyl-aminopeptidase gene (20). Some proteolytic digestion of the 1-84 product was also evident by the identification of hPTH-(8-84) as amajor form in the E. coli extracts. In a previous study, an expression plasmid encoding the precursor, prepro-PTH, was constructed from human prepro-PTH cDNA (8).Microsequence analysis of extracts of E. coli maxicells containing this expression plasmid revealed prepro-hPTH, hPTH-(3-84), and hPTH(8-84) as products. In thisstudy, it was hypothesized that the latter two molecular forms had arisen by aberrant initiation at nonauthentic start sites in the mRNA encoding the preprohPTH molecule (8).However, the hPTH-(8-84) characterized in our study clearly did not have a blocked amino terminus, * B-I j"r-I-I-r--I which would have indicated the presence of a formyl-methiI lo-~o 10-9 ,O-e 10-7 onine. This product is therefore likely to have arisen by P T H Concentration ( M ) proteolytic digestion. In future studies, it will be important to FIG. 9. Effect of increasing quantitiesof a (O), bl (A), and either use E. coli mutants which are deficient in proteolytic bn (A) and standard hPTH-(1-84) (0)on adenylate cyclase activity or to test different enzyme inhibitors in order to activity in osteosarcoma cells (UMR 108). The osteosarcoma cell bioassay was carried out as described under "Experimental Pro- determine if any of these can retarddegradation of the mature cedures." B (x), represents the basal activity of the adenylate cyclase hPTH-(1-84) and thus improve yield. The recombinant hPTH-(1-84) had 100% of the biological assay. Each point is the mean f S.E. of triplicate determinations. c
0
Y
,
Characterizationof RecombinanLt Human Parathyroid Hormone and immunological activity of standard hPTH-(1-84). The Met-hPTH-( 1-84) whilemaintaining immunologicalreactivity, demonstrated only 10% of the bioactivity of hPTH-(184) in the osteosarcoma bioassay. This is consistent with previous structure-function studies which indicated that extension at the amino terminus markedly reduced biological +34) has apactivity. For example, bovine pro-PTH-(-6 proximately 5% of the activity of bPTH-(1-34) (21). Thus, while other recombinant peptides which retain an aminoterminal formyl-methionine, such as recombinant human growth hormone, are fully biologically active (22), parathyroid hormone differs markedly because of the profound decrease in hormonal activity which accompanies modification of its amino terminus, whether substitution, extension, or deletion (23). Our studies therefore emphasize the importance of careful purification of recombinant PTH which may be composed of a mixture of peptides altered at theNH, terminus, in order to achieve yields of hormone of optimal bioactivity. Isolation of recombinant hPTH-(8-84) allowed us to examine its potential bioactivity. In both skeletal and renal bioassays it was without agonist activity; previous studies correlating PTH function with structure have indicated that inert synthetic analogues of PTH-(1-34) truncated at the NH, terminus may act as hormonal antagonists (1, 24, 25). Although these have been of great benefit for in vitro studies, the in vivo efficacy of such truncated analogues based on the NH,-terminal 34 residues continues to be evaluated, and highly potent forms remain to be developed. We and others (17, 26) have demonstrated the presence of binding sites in target tissues within the carboxyl region of PTH-( 1-84) whose role in hormonal action remains to be delineated. In future studies, it will be important to prepare a series of analogues with deletions at the amino terminus but based upon the complete 1-84 sequence, thus including the carboxyl region, to determine if these will provide more potent in vitro or in vivo antagonists. The use in the present studies of synthetic oligonucleotides containing crossover linker sequences which we have employed to construct the hPTH-(1-84) gene should facilitate the construction of analogues of the intacthormone which can be tested for modified biological activity. In recent years an anabolic action of PTH within the skeleton has been recognized (27). Consequently, clinical studiesin osteopenic states have been initiated with the synthetic amino-terminal fragment of PTH, and these have yielded promising results (28, 29). Large-scale growth of the expression plasmid described here, possibly in conjunction with high density fermentation, should yield substantial quantities of recombinant hPTH-(1-84). Thiswould make possible clinical trials employing intact hPTH-(1-84) for the first time and allow an exploration of the potential of hPTH-(1-84) as an anabolic agent in disorders such as osteoporosis. --$
Acknowledgments-We thank Suzanne Bernier, Isabel Bolivar, Miren Gratton, Susan James, and Jane Mitchell for assistance with aspects of these studies, Drs. James Schilling and Michael West for protein sequencing, and Vicki Armstrong and Diane Allen for preparation of the manuscript. REFERENCES 1. Potts, J. T., Jr., Kronenberg, H. M., and Rosenblatt, M. (1982)
1313
Ado. Protein Chem. 36,323-395 2. Kemper, B. (1986) CRC Crit. Rev. Biochem. 1 9 , 353-379 3. Goltzman, D., and Hendy, G. N. (1987) in Principles and Practice of Endocrinology and Metabolism (Becker, K. L., ed) J. B. Lippincott Co., Philadelphia, in press 4. Keutmann, H. T., Sauer, M. M., Hendy, G. N., O'Riordan, J. L. H., and Potts, J. T., Jr. (1978) Biochemistry 17,5723-5729 5. Hendy, G . N., Kronenberg, H. M., Potts, J. T., Jr., and Rich, A. (1981) Proc. Natl. Acad. Sci. U. S. A . 7 8 , 7365-7369 6. Vasicek, T. J., McDevitt, B. E., Freeman, M. W., Fennick, B. J., Hendy, G. N., Potts, J. T., Jr., Rich, A., and Kronenberg, H. M. (1983) Proc. Natl. Acad. Sci. U. S. A . 80, 2127-2131 7. Kronenberg, H. M., Igarashi, T., Freeman, M. W., Okazaki, T., Brand, S. J., Wiren, K. M., and Potts, J . T., Jr. (1986) Recent Prog. Horm. Res. 42,641-663 8. Born, W., Freeman, M., Hendy, G. N., Rapoport, A., Rich, A., Potts, J. T., Jr., and Kronenberg, H. M. (1987) Mol. Endocr. 1, 5-14 9. Born, W., Freeman, M., Bornstein, W., Rapoport, A., Klein, R. D., Hendy, G . N., Khorana, H. G., Rich, A., Potts, J. T., Jr., and Kronenberg, H. M. (1987) J. Bone Min. Res. 2,353-360 10. Sung, W.L., Zahab, D.M.,Yao, F.-L., and Tam, C. S. (1986) Biochem. Cell Bioi. 64,133-138 11. Sung, W. L., Zahab, D.M., MacDonald, C.A., and Tam, C. S. (1986) Gene (Amst.)4 7 , 261-267 12. Bennett, H. P. J., Solomon, S., and Goltzman, D. (1981) Biochem. J. 197,391-400 13. Rabbani, S. A., Mitchell, J., Roy, D. E., Kremer, R., Bennett, H. P. J., and Goltzman, D. (1986) Endocrinology 1 1 8 , 1200-1210 14. Bennett, H. P. J., Browne, C.A., and Solomon, S. (1983) Anal. Biochem. 128,121-124 15. Goltzman, D., Gomolin, H., DeLean, A., Wexler, M., and Meakins, J. L. (1984) Am. J. Physiol. 68, 70-75 16. Zanelli, J. M., and Gaines Das, R. E. (1983) J. Clin. Endocrinol. Metab. 67,462-469 17. Demay, M., Mitchell, J., and Goltzman, D. (1985) Am. J. Physiol. 2 4 9 , E437-E446 18. Harris, T. J. R. (1983) in Genetic Engineering 4 (Williamson, R., ed) pp. 127-185, Academic Press, London 19. Breyel, E., Morelle, G., AufmKolk, B., Frank, R., Blocker, H., and Mayer, H. (1984) in 3rd European Congress on Biotechnology, (Dechema, ed) Vol. 3, pp. 363-369, Verlag Chemie, Weinheim, West Germany 20. Ben-Bassat, A., Bauer, K., Chang, S.-Y., Myambo, K., Boosman, A., and Chang, S. (1987) J. Bacteriol. 1 6 9 , 751-757 21. Peytremann, A., Goltzman, D., Callahan, E. N., Tregear, G. W., and Potts,J. T., Jr. (1975) Endocrinology 9 7 , 1260-1280 22. Hintz, R. L., Wilson, D. M., Finno, J., Rosenfeld, R. G., Bennett, A., McClellan, B., and Swift, R. (1982) Lancet 8284 1276-1279 23. Parsons, J. A., Rafferty, B., Gray, D., Reit, B., Zanelli, J. M., Keutmann, H. T., Tregear, G . W., Callahan, E. N., and Potts, J. R., Jr. (1974) in Calcium-regulating Hormones (Talmage, R. V., ed) pp. 33-39, Excerpta Medica, Amsterdam 24. Goltzman, D., Peytremann, A., Callahan, E., Tregear, G. W., and Potts, J. T., Jr. (1975) J. Bwl. Chem. 2 6 0 , 3199-3203 25. Horiuchi, N., Holick, M. F., Potts, J. T., Jr., and Rosenblatt, M. (1983) Science 2 2 0 , 1053-1055 26. McKee, M. D., and Murray, T. M. (1985) Endocrinology 1 1 7 , 1930-1939 27. Tam, C. S., Heersche, J. N. M., Murray, T. M., and Parsons, J. A. (1982) Endocrinology 110,506-512 28. Reeve, J., Meunier, P. J., Parsons, J. A., Bernaut, M., Bijvoet, 0. L. M., Courpon, P., Edouard, C., Klenerman, L., Neer, R. M., Renier, J. C., Slovik, D., Vismans, F. J. F. E., and Potts, J. T., Jr. (1980) Br. Med. J. 2 8 0 , 1340-1344 29. Slovik, D. M., Rosenthal, D. I., Doppelt, S . H., Potts, J. T., Jr., Daly, M. A., Campbell, J. A., and Neer, R. M. (1986) J. Bone Min. Res. 1 , 377-382