Efficient secretion of human parathyroid hormone by Saccharomyces cerevisiae

Gene, 90 (1990) 255-262 Elsevier

255

GENE 03522

Efficient secretion of human parathyroid hormone by Saccharomyces cerevisiae (Recombinant DNA; yeast; mating factor ~ fusion; proteolytic processing; KEX2; osteoblast adenylate cyclase; enzyme

assay) Odd S. Gabrielsen', Sjur Reppe ~, Olav S~etherb, Ola R. Blingsmo b, Knut Sletten', Jan O. Gerdeladze b, Anders Hegset% Vigdis T. Gantvik b, Peter Alestmm b, Tordis B. Oyen" and Kke M. Gautvik b ° Department of Biochemistry, University of Oslo, Blindern, 0316 Osio 3 (Norway) and b Institute of Medical Biochemistry, University of Oslo. Biindern, 0316 Oslo 3 (Norway) Received by J.K.C. Knowles: 25 May 1989 Revised: 6 October 1989 Accepted: 11 November 1989

SUMMARY

A cDNA encoding mature human parathyroid hormone (hPTH) was expressed in Saccharomycescerevisiae, after fusion to the prepro region of yeast mating factor ~ (MFoQ. Radioimmunoassay showed high levels of hPTH immunoreactive material in the growth medium (up to 10/zg/ml). More than 95% of the immunoreactive material was found extracellularly as multiple forms ofhormone peptides. Three internal cleavage sites were identified in the hPTH molecule. The major cleavage site, after a pair ofbasic amino acids (aa) (Arg2SLys2e~Lys27),resembles that recognized by the KEX2 gene product on which the MF0e expression-secretion system depends. The use of a protease-deficient yeast strain and the addition of high concentrations of aa to the growth medium, however, not only changed the peptide pattern, but also resulted in a significant increase in the yield of intact hPTH (1-84) (more than 20% of the total amount of immunoreactive material). The secreted hPTH (1-84) migrates like a hPTH standard in two different gel-electrophoretic systems, co-elutes with standard hPTH on reverse-phase high-performance liquid chromatography, reacts with two hPTH antibodies raised against different parts of the peptide, has a correct N-terminal aa sequence, and has full biological activity in a hormone-sensitive osteoblast adenylate cyclase assay.

INTRODUCTION

Human parathyroid hormone (hPTH) is a key regulator of Ca2÷ homeostasis through its actions on kidney and bone (reviewed in Cohn and Elting, 1983; Morel, 1983; Norman et al., 1982). A net retainment of body Ca2÷ is Correspondence to: Dr. O.S. Gabrielsen, Department of Biochemistry, University of Oslo, Blindern, 0316 Oslo 3 (Norway) Tel. (47-2)-456173; Fax (47.2).454194.

Abbreviations: aa, amino acid(s); AC, adenyl~te cyclase; AR$, autonomously replicating sequence; bp, base pair(s); cDNA, DNA complementary to RNA; HPLC, high-performance liquid chromatoglaphy; hPTH, human parathyroid hormone; kb, kilobase(s) or 1000bp; KEX2, sene encoding yscF specific for cleaving on the C side of pairs of basic 0J?8-1119/90/SOJ.50© 1990Elsevier Scia:c¢ Publishers B.V. (Biomedical Division)

obtained via increased urinary Ca2 ÷ reabsorption and indirectly via vitamin D3-induced Ca2 + adsorption from the gut. The overall physiological action of hPTH is probably to generate a positive Ca 2 ÷ balance and enhance bone formation. The hormone is produced in the parathyroid gland as a l l5-aa prepro-peptide. Before secretion, the residues that contain arginine; MF0c,yeast mating factor 0c;nt, nucleotide(s); oligo, oligodeoxyribonucleotide; or/, origin of DNA replication; PAGE, polyacrylamide-gel electrophoresis; PVDF, polyvinylidene difluoride; e, resistance/resistant; $., Sacchavomyces;SDS, sodium dodecyi sulfate; STEI3, sene encoding ysclV involved in the N-terminai maturation of the a.factor pheromone; Tc, tetracycline; yscF, membranebound endopeptidase; ysclV, membrane-hound depeptidyl-aminopeptidase.

256 prepro part is cleaved off, yielding the 84-aa mature hormone. Sufficient suppfies of authentic recombinant hPTH are of considerable interest for biological studies and medical applications. So far, degradation ofhPTH has been a major problem when expressed in heterologous organisms. Low expression levels were obtained when prepro-hPTH was synthesized in Escherichia coli and the major expression products were identified as prepro-hPTH, hPTH (3-84) and hPTH (8-84)(Born et al., 1987b). When prepro-hPTH was synthesized in yeast, very low expression levels were obtained with no detectable secretion (Born et al., 1987a). Rabbani et al. (1988) recently reported a more successful intracellular expression of the hormone in E. coli, where about 0.~ ;!g of total immunoreactive material per ml of medium ~ is produced and where hPTH (1-84) was one of three identttied forms. However, the hormone was unstable and ha0 a half-life of approx. 12 min, and required a relatively intricate purification procedure leaving maximally 10 ng/ml culture. One strategy to circumvent this problem was recently reported by Wingender et al. (1989). By fusing the hPTH-coding region after the cro-lacZ gene through a linker giving an Asp-Pro motif at the junction, it was possible to produce reasonable amounts of a hPTH polypeptide with an extra proline at the N terminus. This modified hormone was biologically active and behaved as an agonist. In this report we describe the use ofthe S. cerevisiae MF~ expression system (e.g., Brake et al., 1984; Vlasuk et al., 1986; Zsebo et al., 1986) to export hPTH from yeast in a correctly processed form. Yeast cells transformed with different plasmids containing the MF~ hPTH hybrid gene, efficiently secreted hPTH peptides to the growth medium. However, in standard selective minimal medium the secre. tion product was heterogeneous and the yield of authentic hPTH (1-84) very low due to internal proteolytic processing of the hormone and incomplete removal of the spacer peptide from the N-terminal end of hPTH. After removal of the sequence encoding the spacer peptide by in vitro mutagenesis and growth of a protease-deficient yeast transformant in the presence of high aa concentrations, a substantial amount of authentic hPTH (1-84) was obtained, although degraded forms were still present. The secreted hPTH had correct N-terminal aa sequence and was biologically active in a hormone-sensitive osteoblast adenylate cyclase assay.

MATERIALS AND METHODS

(a) Strains and culture conditions Plasmids were propagated in £. coli using standard methods (Maniatis et al., 1982). The $. cerevisiae strains

used were FL200 (~, leu2, ura3) and BJI991(~ trpl, ura3-52, leu2, prbl-l122, pep4-3). Yeast cells were transformed by standard methods (Carter et al., 1987). Transformants were selected by plating on omission medium. Yeast cells were grown at 30°C in YNB medium [0.67~o yeast nitrogen base/2% glucose supplemented with aa (50-75#g/ml)/adenine and uracil (10pg/ml)] or YNB medium supplemented with 1~o Casamino acids (Difco).

(b) In vitro mutagenesis To delete the STE13 recognition sequence located immediately N-terminal to hPTH by site-directed in vitro mutagenesis of the fusion gene, a 1495-bp XbaI fragment containing the MFo~ promoter, the MFo~ leader sequence and the hPTH-encoding gene including the stop codon, was isolated from p~LXPTH (see Fig. 1, legend) and subcloned into M13mpl9 (to give M13PTH-1). An oligo with the sequence 5'-GGATAAAAGATCTGTGAG-3' was annealed to single-stranded DNA prepared from the recombinant phage. The first 10 nt of the ofigo are complementary to the sequence of the MFo~ leader just preceding the Glu-Ala-Glu-Ala coding region, and the last 8 nt are complementary to the beginning of the hPTH sequence. After second-strand synthesis and ligation, closed-circular heteroduplex DNA was isolated after sedimentation through an alkaline sucrose gradient as described (Carter et al., 1985), and used to transform a BMH71-18 mutL strain of E. coil defective in mismatch repair (kindly provided by Dr. O. Winter). Positive clones with the looped-out sequence 3'-CTCCGACTTCGA-5' deleted were identified by colony hybridization, using the mutagenizing oligo as the probe, and by DNA sequencing. The plasmid in these clones was designated MI3PTH-2. The MF~ transcription terminator was then inserted into one of the positive Ml3 clones as a SalI-HindIII fragment isolated from pMF~ (Fig. 1, legend) to give M13PTH-3. The entire expression cassette between a BamHI and a f'dled-in EcoRI site was finally inserted between the BamHI and Pvull sites of YEp24 (Bothstein et al., 1979). This expression plasmid was designated p~UXPTH-2 (Fig. 1). (¢) Purification of hPTH from yeast culture medium Yeast culture media were adjusted to pH 3 and concentrated by passage through an S Sepharose Fast Flow column (Pharmacia AB), pre-equllibrated with 0.3 M glycine pH 3. The column was washed with 0.1 M acetic acid adjusted to pH 6.0, and hPTH peptides were eluted with 0.1 M Na2HPO4 pH 8.5. The concentrated medium was subjected to fhrther purification by reversed-phase HPLC using a Vydac protein peptide C 18 column (The Separation Group, Hesperia, CA). The column was eluted with a linear gradient of acetonitrile/0.1 ~o tri-fluoroacedc acid.

257 RESULTS AND DISCUSSION

URA3 marker and a 2/~m or/that results in a normal high

copy number. (a) Strategy for the expression of human parathyroid hormone in Sacckaromyces cerevisiae A cDNA clone encoding the human parathyroid hormone (Hegset etal., 1990) was used to make an in-frame gene fusion with the MFo~leader sequence including its spacer sequence. Three different yeast vectors were used for the hybrid gene. (1) pL5 with the LEU2d selective marker and a 2 pm or/ derived from pJDB207 (Beggs, 1981) giving a copy number in yeast cir + cells that is very high relative to that of other plasmids (Erhart and Hollenberg, 1983). (2) Vector YEp24 (Botstein et al., 1979) containing a

Bcll

(3) YCpS0 (Ma et al., 1987) containing a CEN4 sequence, a URA3 marker and a chromosomal A R S that yields a copy number near unity. The plasmids with inserted expression cassettes were designated p0cLXPTH, pgUXPTH-I and pgUCXPTH

(Fig. 1). Initial studies revealed incomplete removal of the GluAla-Glu-Ala spacer peptide leaving this N-terminal extension on the secreted hPTH, as observed previously for other heterologous proteins produced in yeast with the MFoc system (Brake et al., 1984; Loison et al., 1988; Bitter et al., 1984). To remove the Glu-Ala dipeptide, we deleted

Sol I ~Xbal

EcoRI Hlndlll Ap

2

~/Xbul

I

n~LXPTH

~"Oolll

EcoRI

paUXPTH- 1 URA3

Sail

EcoRI--..~m~.~

Sail Xbul

Htndlll EcoRI Sail, eam,-.if...~-IXbal

Hindlll EcoRI Hindl I I

IIII Ap

Xbal

IIPTH'~ MFm

kP

p~UCXPTH paUXPTH- 2

ARSI

EcoRI

CEN4

Hlndlll

URA3

Oicl~ '-~'''~" Dcll PVull

Oglll" ~ ~

II

XbOI

Fig. I. Recombinant plasmids for the expression of human parathyroid hormone in yeast. The expression plasmids shown were constructed as follows. On the basis of nt sequence data for the MFcd gene in $. cere~iae (Kurjan and Herskowitz, 1982), a synthetic oHgo with the sequence 5'-TGGCATTGGCTGCAACTAAAGC-Y was used to select from a yeast DNA library clones containing a 1.7-kb EcoRl fragment in pBR322 encoding the entire MFal sene. Of severai positive clones, one was selected and the plasmid designated pMF~l. A Bglll.£¢oRl fragment from pMFal was then inserted into the yeast E. call shuttle vector pLS. pL5 is a derivative of I~DB207 (Beggs, 1981) with the region between the (filled-in) Sall site in the Tca gene and the H/ndlll site in the 2/Am fragment, replaced with a synthetic oliso $'-GATCTGCAGGATGGATCCAAACsCF-Y. The resulting expression vector pozLX contains the a-fuctor promoter and transcription terminator as well as the entire MF~I gent. To insert the parathyroid hormone-encoding gene in frame, a Dpnl-$all fragment from a hPTH/pUCI9 subcione that encodes the mature APTH sequence with its stop codons (Hegset et ai., 1990), was cloned into the p0cLX vector between a filled-in Hbsdlll site and a Sa/l site to substitute the mature MFa-codins region giving the plasmid paLXPTH. To transfer the expression cassette to other vectors, the pozLXPTH plasmid was prepared from a dam- E. call strain and a Bglll.Bcll fragment containing the whole fusion gene with its promoter and terminator was excised and inserted into the BamHl site of the £. coli yeast shuttle vectors YEp24 (Botstein et ai., 1979) and YpC$0 (Ma et ai., 1987) to give the expression plasmids pocUXPTH-I and paUCXPTH, respectively. The expression plasmid pzUXPTH-2 was constructed as described in MATERIALS AND METHODS, section b.

258 the 12-bp region 5' to the hPTH coding sequence by sitedirected in vitro mutagenesis. The deletion positioned the KEX2 cleavage site (reviewed in Bussey, 1988; Fuller et al., 1988) precisely in front of the first aa of the mature hPTH. The mutagenized fusion gene was inserted into the YEp24 shuttle vector, to give the plasmid p~UXTPH-2 (Fig. 1). 0b) Human parathyroid hormone synthesis Parallel cultures of the yeast strain FL200 transformed with one of the three expression plasmids, p~UCXPTH, p=UXPTH-1 and p=LXPTH, were grown and the growth medium, a periplasmic fraction, and an intracellular soluble fraction assayed for hPTH immunoreactive peptides. The results are summarized in Fig. 2a. Highest production and secretion of hPTH immunoreactive material was obtained with the p=UXPTH-1 transformed cells. Fourfold less was secreted by cells transformed with the high-copy-number plasmid p=LXPTH and the single-copy-number plasmid p=UCXPTH. Apparently, there is no simple correlation between copy number and secretion level. A similar situation seems to exist for promoter strength (Green et al., 1986; Ernst, 1986). In all three cases, most ofthe produced immunoreactive material was secreted to the growth medium (p=UCXPTH and p=UXPTH-I transformed cells: more than 95% secreted). In the p~LXPTH transformanta significant proportion of the immunoreactive material (about 25 ~o) was detected in the intracellular and the periplasmic fraction. No immunoreactive hPTH was secreted from yeast cells transformed with only the vector p=LX, The secreted hPTH products dilute~,in parallel to hPTH (1-84) standard in the radioimmunoassay, showing

that the active material bore a close immunological resemblance to the human parathyroid hormone (results not shown). Because of high production and efficient secretion, cells transformed with p~UXPTH-I (and its derivative p~UXPTH-2) were chosen for further work. The yield of secreted hPTH was further increased by supplementing the standard selective minimal medium with 1% Casamino acids. This resulted in a significant increase in the amount of hPTH secreted. The secretion of hPTH immunoreactive material directed by the p=UXPTH-2 plasmid as a function of growth phase in the two growth media is shown in Fig. 2b. Highest production and secretion levels were observed in the later stages of growth where yields of hPTH immunoreactive material up to 10/~g/ml were obtained. (c) Analysis of secreted product Fig. 3 shows that the secreted products concentrated from the growth medium contained a complex pattern of protein bands. A total of more than ten distinct species with relative Mr below 15 000 could be observed at high loadings, none of which were present in the control medium. When analysed immunologically on a protein blot, all the bands seemed to react with antibodies specific for the H-terminal part or mid-region of the hormone (not shown). However, most of the peptides had an electrophoretic migration different from that of the 9418-Da hPTH (1-84) band (Fig. 3). Apparently, a significant degree of incorrect processing or degradation of this heterologous protein occurs during secretion in yeast. To attack this problem rationally, we first identified the

10o b

2000

. ~ ~ ~

M

•

o.oi°.o

• .........................

10e ~:o

~/

o.,, ~,

A m

=- 1000

10~ M

.,'." , .0;,,'" ..o"..."

4.0 4 /

~,o------o--

o

2.0

i... 4 ~ O"<~ 10e ._ .L_.~'..~. . . . . . . . . . . . . . . . . o.o 0 5 10 15 20 25 p~UCXFrH paUXPTH.! pctLXPTH Time (h) Fig.2. The yieldof hFrH secreted fromyeast.(a) The $. cere~lae strain FL200transformed withthe expressionplasmidsp=UXCPTH,paUXPTH.I or p'~LXPTHwas 8townin liquidmediumlackinguracilor leucine(dependingon the selectivemarker)to densitiesof 1.3 × 10s, 1.3 × I0 s, and 7.9 × 107 cells/ml,respectively.RadioimmunoassayofhPTHwascarriedout as described(Gautviket al., 1979)usingan antiserumspecificfor the 44-68 aa domain. Yeast culture media(M. shaded bars) were assayeddirectly. Periplasmicfractions (P, suppledbars) were obtained by incubatingwashed cells/'or I h at 28°C in a buffer containing 0.05% Triton X-100/0.1M NaH=PO4/0.5M NaCI. After esntrifugafion, the superuatants were assayed for hPTH immunoreactivity.Intra¢~Hularfi'actions(I, blackenedbars) werepreparedfrom the remainingcells by disruption in an Eaton press. Valuesshownare normalizedto the samecellofdensity.(b) The $. ¢eve~e strain FL200transformedwiththe expressionplasmidp=UXPTH-2was grownin liquidYNB medium lackingursed (opan symbols)or in the same mediumsupplementedwith I~o Casaminoacids (filled symbols).Cell densities (O, O) were measured in a hemocytom~or,and hPTH concentrations in the 8rowthmediummeasured by radioimmanoassay(O, 4k). I P

0

i

• ii ii

259

kDa ~!:q ,.--, 94.0

.........

43.o

....

~,.~: i , , ~ ¸i ~

30.0

~,~ml) 20.1 i "~'~ t

1

~~

2

3

14.4

M

Fig. 3. SDS-PAGE of concentrated yeastgrowth medium. Aliquotsof concentrated growth medium (see MATERIALS AND METHODS, section e) from yeast strain FL200 transformed with the expression plasmids p~tUXPTH-I(lane 1) or paUXPTH-2(lane 2) was analysedby 15% PAGEin the presenceof 0.! ~ SDS (Laemmli, 1970).Proteinwas visuaiised by silver staining (Btlrk et ai., 1983). Concentrated growth medium from an untransformedstrainwas used as control (lane3). The arrowshowsthe migrationofa hPTH standard. A molecularsizemarker is shown in lane M.

nature of the processings that had occurred by microsequencing of the different hormone forms. The products from the p~UXPTH-I transformed cells were found to contain N-terminal peptides with the Glu-Ala dipeptide extension suggesting an incomplete removal of the STEI3 recognition signal. However, after removal of the spacer peptide encoding region, correct N-terminal aa sequence

M Fm

~

was founu m small N-terminal peptides from p~UXPTH-2 transformed cells. The microsequencing of the other bands revealed three internal processing sites as illustrated in Fig. 4. A major cleavage site was found after a pair of basic aa in position 25 and 26 (Arg~, Lys26). Most probably this site is recognized by the KEX2 endopeptidase (reviewed in Bussey, 1988; Fuller et al., 1988). We found, however, no evidence for cleavage of a second potential KEX2 site in the hPTH sequence at positions 52-54 (Arg-Lys-Lys) even the latter is located in an extremely hydrophilic region of the molecule. It is noteworthy that hPTH is cleaved internally when it is synthesized in yeast fused to a heterologous leader sequence but not when it is synthesized in the parathyroid gland fused to its homologous presequence. The KEX2 endopeptidase is very similar to mammalian prohormone processing enzymes (Fisher and Scheller, 1988; Fuller et al., 1988), and when artificially introduced into mammalian cells, the KEX2 endopeptidase seems to process prohormones correctly (Thomas etal., 1988). Either there are subtle differences in substrate specificities between the yeast and the mammalian enzymes, or the mature hPTH sequence is protected against such cleavage by its homologous prosequence. Weaker bands on the blots had an N-terminal sequence that started with aa 35 and 45 in the hPTH sequence. This corresponds with a cleavage after a Phe ~4 and Arg~ residue, respectively, indicating a certain level also of a chymotrypsin-like and a trypsin-like cleavage. A similar type of cleavage has previously been reported for a human p-endorphin analogue when secreteded from yeast (Bitter et al., 1984). Partial cleavage of the fusion protein at so many sites resulted in a very low yield ofintact hPTH (1-84). Attempts were therefore made to reduce the extent of degradation.

(d) Approaches to increase the yield of intact hPTH (1-84) To minimize proteolytic degradation, a protease-deficient yeast strain (BJI991) was first tried as a host. This strain secreted less hPTH-related peptides than did the

hPTH

1

I,..

i

,..

,,r,

., t, t.

"

Fig. 4. Schematicrepresentationof cleavagesites in the MF~t-hPTHfusion protein as detected by aa sequencingof immunoreactivepeptides secreted from yeast. To sequencethe hPTHorelatedpeptides, concentratedgrowth medium from yeast strain FL200 transformedwith the expressionplasmids p0tUXPTH-i (upper)or p~tUXFrH-2(lower)were separated either by acetic acid PAGE and electroblotteddirectlyonto glassfibersbe,ts (Abersold et ai., 1986),or by SDS-PAGE followedby blottingonto PVDF membranes(Matsudaira, 1987).Automated Edmandegradationwas performedon a 477A Protein Sequencerwith an on-fine 120Aphenylthiohydantoinaa anaiyserfromAppliedBiosystems(Foster City,CA).All reagentswereobtained from Applied Biosystems.Thick arrows indicate major cleavagesites found, thin arrows minorones. Numbers refer to the aa positions in the mature hPTH sequence(Kentmannet ai., 1978).

260

FL200 strain (up to 2 gg/ml immunoreactive material), but the pattern ofpepfides was still complex (Fig. 5, lane 1) and no major band migrated like the hPTH standard. Our second approach was to change the growth conditions by substituting the standard selective minimal medium by a Casamino acid-supplemented minimal medium. This is possible since we employ the URA3 marker. As shown in Fig. 5, this did not only increase the levels of hPTH peptides found in the medium, but also changed the band pattern (compare lanes I with 2, and 4 and 3). Most significandy, a strong band appeared in the lane with growth medium from the BJ 1991 transformant, which had the same migration on SDS-PAGE as a commercial hPTH standard 0ane 2). Only a faint band of similar migration was seen in the concentrated medium from the FL200 transformant (lane 3). The hPTH (1-84) candidate from the protease deficient strain was analysed further by electrophoresis and protein blotting (Fig. 6). The hPTH (1-84) pepfide (band 2 in Fig. 6a) reacted with two antibodies, one specific for the N-terminal part and one for the mid-region ofhPTH. It also showed the same migration as the hPTH standard when analysed by electrophoresis in two dimensions followed by protein blotting (acetic acid/urea + SDS-PAGE), conf'wming the correct size of the peptide (not shown). The identity ofthe hPTH(I-84) candidate was demonstrated by aa sequence analysis of the peptide after blotting onto an PVDF membrane filter. The first 19 aa were analysed and all were identical to that of human PTH. Scanning of a

:~

18.4

ii~

14.3

6.2

3.4

1

2

3

4

M

Fig. $. Effect of'strain and growth medium composition on the production of intact hFFH (I-84) in yeast. S. cerevivlae strain FL200 (lanes 3 and 4), or KlgPl (lanes I and 2) transformed with the plasmid paUXPTH-2 were caltured in YNB liquid medium lacking uracil (lanes I and 4) or in the same medium supplemented with 1% Cmamino acids

(lanes 2 and 3). Secretedproductsfrom~rowthmediumcorresponding to S x 10s cellswere analyxd by IS~o PAGEin the pru~ce of O.l~o SDS, as describedin the lesendto Fill.3. A silver-stainedgel is shown. The arrowshowsthe migrationof a hPTH standard.

kOa 29.0

a

b

c

...........

~e.4 a m

14.3

~

3.4

M

1

S

S

1

1

S

Fig. 6. Immunc,:e~¢V.'~,:hPTH peptides in the yeast growth medium. S. cere~siae strain BJI991 transformed with the plasmid p~UXPTH-2 was cultured in a Casamino acid-supplemented minimal liquid medium lacking uracil at 30°C. At a cell density of 5 × 107/ml, secreted products were concentrated from the growth medium and analysed by 15% PAGE in the presence of0.1% SDS, as described in the legend to Fig. 3. (Panel a) Silver-stained gel with molecular size marker (lane M), hPTH standard (Peninsula Laboratories, USA, lane S), and concentrated yeast growth medium (lane 1). (Panels b and ¢)The corresponding protein blots. Immobilion PVDF Transfer Membranes (Millipore) and the butters ofTowbin et ai. (1979) were used. Blots were visualized with cock anti-hPTH antiserum (Oautvik et al,, 19"/9) as the primary antibody, rabbit anti-cock-lgO (Gautvik et al., 1979) as secondary antibody and donkey [Z2Sl]antirabbit.18G (Amersham) as tertiary antibody. Auto. radiography was performed overnisht at -70°C with Kodak X.omat AR$ film and intensiPying screen. Two hPTH-speeific antibodies were used, one reactive ~ainst the N-terminal part of the hormone (panel b), another reactive against the middle region of the hormone (panel c). Lanes in panels b and c are numbered as in panel a.

silver-stained gel indicates that more than 20~o of the secreted pepfides are intact hPTH(I-84). The as sequence analysis of the other bands in Fig. 6a showed that the major product (band 4) had an N-terminal starting Lys27 in the hPTH sequence. The same was found for the minor band 3. Combined with a size estimate, this suggests that band 4 corresponds to the hPTH(27-84) fragment. Consequently, the altered conditions of strain and growth medium had not eliminated the cleavage at the potential K£X2 site in the hPTH sequence. Band 5 revealed a correct N-terminal sequence like that found for band 2. Thus, band 5 is probably the small hPTH(l-26) fragment. No bands corresponding to cleavage at positions 35 and 45 in hPTH were found when Casaminc~-supplemented medium was analyzed, suggesting that it is primarily these

261 cleavage events that are affected by the change in strain and growth conditions. Band 1 also had a correct N-terminal sequence. It was not further investigated. Since the hPTH gene contains two stop codons, a peptide longer than 84 aa is highly improbable. Some kind of posttranslational modification, possibly O-linked glycosylation, could explain its slower migration. The same phenomenon probably is the reason for the appearance of band 3 in addition to band 4 both having the same N-terminal sequence, and for the band appearing between bands 2 and 3 in a growth medium without Casamino acids (Fig. 5, lanes 1 and 4) which also had an N-terminal aa sequence starting with Lys27. The latter observation suggests that the extent of the assumed modification also varies with the conditions of strain and growth medium. The mechanism for the protective effect of aa on aberrant hPTH processing was not further investigated. High aa concentrations could have diverse effects on the physiology of the yeast cell as well as on the fermentation process. The important point is that conditions have been found where yeast cells efficiently express authentic recombinant hPTH. (e) Biological activity The biological activity of the secreted hPTH(1-84) was tested in a hormone-sensitive osteoblast adenylate cyclase assay (Oautvik et al., 1983; 1984). Concentrated growth medium of p0cUXPTH-2 transformed BJI991 cells was purified by HPLC. The peak that comigrated with a hPTH standard was analysed for its ability to stimulate the adenylate cyclase activity of OMRI06 osteosarcoma cell membranes'" above the basal level. Commercial hPTH(I-84) was used as a reference. As shown in Fig. 7, hPTH produced from p~UXPTH-2 transformed cells had a clear stimulatory effect comparable to that of the hPTH reference. if) Conclusions We have demonstrated that biologically active human parathyroid hormone can be synthesized and secreted in S. cerevisiae. A maximum of l0 pg/ml medium of immunoreactive material was obtained. (2) The main problem with the expression and secretion of human parathyroid hormone in yeast, was not the expression level, but rather a phenomenon of aberrant processing of the heterologous protein that reduces the yield of the intact hPTH(l-84). A very simple remedial action was found effective in increasing the yield of unproteolyzed hPTH(I-84). The combination of a growth medium rich in aa and the use of a protease-deficient yeast strain was sufficient to obtain intact hPTH as one of the major products (about 20% of the hPTH peptides secreted). (3) The secreted hPTH(1-84) behaves like a hPTH standard with respect to electrophoretic migration, HPLC re-

7.0

.""

6.0 5.0 4.0

~ 3.0

~ 2.o! t.O 0.0

.t

.

.

.

.

.

.

.

.

.

t

'

'

.

.

.

.

.

.

.

10

Concentration (nglml)

Fig. 7. Biological activity. Recombinant hPTH (&) was assayed for biological activity in a hormone-sensitive osteoblast adenylate cyclase (AC) assay, hPTH was purified on HPLC, freeze-dried, dissolved in distilled water and diluted to yield doses capable of stimulating the adenylate cyclase of UMRI06 rat osteosarcoma cell membranes. Membranes were prepared and the assay carried out as previously described (Gautvik et al., 1983; 1984). The experiments were carried out in tripli. care determinations which differed by less than 17%. hPTH (I-84) from Sigma was used as reference (O).

tention, reaction with hPTH specific antibodies, N-terminal aa sequence, and biological activity in a hormone-sensitive osteoblast adenylate cyclase assay. (4) When available in sufficient quantities, recombinant hPFH has many potential uses, for example in diagnostics and as a drug in veterinary medicine. A fragment of hPTH together with 1,25(OH)2 vitamin D3 have also been reported to induce bone formation in humans (Reeve et al., 1980; Slovik et al., 1986), and one of the major areas of potential use of a recombinant hPTH is therefore in the treatment of osteoporosis. Sufficient supplies of authentic recombinant hPTH are of considerable interest to evaluate such applications.

ACKNOWLEDGEMENTS

This project has received economical support from the company Selmer-Sande A/S and from The Royal Norwegian Council for Scientific and Industrial Research (Grant PT 15.18471, and PT 15.15449). The project was also in part funded by a grant from the Nordic Yeast Research Program to OSG.

REFERENCES Abersold, R.H., Teplow, D.B., Hood, L.E. and Kent, S.B.H.: Electroblotting onto activated glass. High efficiency preparation of proteins from analytical sodium dodecyl sulfate.polyacrylam/de gels for direct sequence analysis. J. Biol. Chem. 261 (1986) 4229--4238. Beggs, J.D.: Multiple-copy yeast plasmid vectors. In Von Wettstein, D., Friis, J., Kielland-Brandt, M. and Stenderup, A. (Eds.), Molecular

262 Genetics in Yeast. Alfred Benzon Symposium,Vol. 16. Munksgaarde, Copenhagen, 1981, pp. 383-390. Bitter, G.A., Chen, K.K., Banks, A.R. and Lai, P.-H.: Secretion of foreign proteins from Saccharomyces cere~iae. Proc. Natl. Acad. Sci. USA. 81 (1984) 5330-5334. Born, W., Freeman, M., Bornstein, W., Rapoport, A., Klein, R.D., Hendy, G.N., Khorana, H.G., Rich, A., Potts Jr., J.T. and Kronenherg, H.M.: Signal sequence of human preproparathyroid hormone is inactive in yeast. J. Bone Min. Res. 2 (1987a) 353-360. Born, W., Freeman, M., Hendy, G.N., Rapoport, A., Rich, A., Potts Jr., J.T. and Kronenherg, H.M.: Human preproparathyroid hormone synthesized in Eseh.~c~da coli is transported to the surface of the bacterial inner membrane, but not processed to the mature hormone. Moi. Endocrinol. 1 (1987b) 5-~4. Botstein, D.° Falco, S.C.° Stewa~% S.E., Brennan, M., Scherer, S., Stinchcomb, D.T., Stn~, K. and Davis, R.W.: Sterile host yeast (SHY): a eukaryntic system of biological containment for recombinant DNA experiments. Gene 8 (1979) 17-24. Brake, A.J., Merryweather, J.P., Coit, D.G., Heberlein, U.A., Masiarz, F.R., Mullenbach, G.T., Urdea, M.S., Valenzuela, P. and Ban', J.P.: a-Factor-directed synthesis and secretion of mature foreign proteins in Saccharomyces cere~siae. Proc. Natl. Acad. Sci. USA. 81 (1984) 4642-4646. B/Irk, R.R., Eschenbruch, M., Leuthard, P. and Steck, G.: Sensitive detection of proteins and peptides in polyacrylamide gels after formaldehyde fixation. Methods Enzymol. 91 (1983) 247-254. Bussey, H.: Proteases and the processing of precursors to secreted proteins in yeast. Yeast 4 (1988) 17-26. Carter, P., Bedouelle. ! 1.,Waye, M.M.Y. and Winter, G.: Oligunucleotide Site-Directed Mut,tgenesis in MI3. An Experiment Manual. MRC Laboratory of Molecular Biolo~, Cambridge (U.K.) 1985. Carter, B.L.A., Irani, M., Mackay, V.L., Seale, R.L., Sledziewski, A.V. and Smith, R.A.: Expression and secretion of foreign genes in yeast. In Glover, D.M, (Ed.), DNA Cloning, Vol. II1, A Practical Approach, IRL Press, Oxford, 1987, pp. 141-161. Cohn, D.V. and Elting, J.: Biosynthesis, processing, and secretion of parathormone and secretory protein.l. In Greep, R.O. (Ed.), Recent Progress in Hormone Research, Vol. 39. Academic Press, New York, 1983, pp. 181-209. Erhart, E. and Hollenberg, C,P.: The presence of a defective LEU2 gene on 2 ;, DNA recombinant plasmids of Saccharomyces cere~lae is responsible for curing and high copy number. J. Bacteriol. 156 (1983) 625-635. Ernst, J.F.: Improved secretion of heterologuus proteins by $ac~aro. myces cerev~iae: effects of promoter substitution in alpha-factor fusions. DNA 5 (1986) 483-491. Fisher, J.M. and Scheller, R.H.: Prohormone processing and the secretory pathway. J. Biol. Chem. 263 (1988) 16515-16518. Fuller, R.S., Sterne, R.E. and Thorner, J.: Enzymes required for yeast prohormone processing. Annu. Rev. Physiol. 50 (1988) 345-362. Gantvik, K.M., Teig, V., Halvorsen, J.F., Arnesen, E., Myhre, L., H~imann, P. and Tollman, R.: Development of sequence specific radioimmunoassay of human parathyroid hormone and it,~use in the diagnosis of hyperparathyrodism. Scand. J. Clin. Lab. Invest. 39 (1979) 469-478. Gautvik, K.M.,Gordeladze, J.O., Jahnsen, T., Haug, E., Hansson, V. and Lystad, E.: Thyroliberin receptor binding and adenylyl cyclase activation in cultured prolactin-producing rat pituitary tumor cells (GH cells). J. Biol. Chem. 258 (1983) 10304-10311. Ga~vik, K.M., Gordeladze, J.O., Moxheim, E. and Gautvik, V.T.: Peripheral metabolism of parathyroid hormone in patients with primary hyperparathymidism as judged by immunological ~ d biological studies. Eur. Set& Res. 16 (Suppl. 2) (1984) 41-54. Green, R., Schaher° M.D., Shields, D. and Kramer, R.: Secretion of somatostatln by S e ~ y c e s ¢ertnt~e. Correct processing of an a-factor-somatmtatin hybrid. J. BinL Chart. 261 (1986) 7558-7565.

Hegset, A., Blingsmo, O.R., Gautvik, V.T., Seether, O., Jacobsen, P.B., Gordeladze, J.O., Alestmm, P. and Gautvik, K.M.: Expression of human parathyroid hormone in Escherichia coll. Biochem. Biophys. Res. Comm. 166 (1990) 50-60. Keutmann, H.T., Saner, M.M., Hendy, G.N., O'Riordan, J.L.H. and Potts, Jr., J.T.: Complete amino acid sequence of human parathyroid hormone. Biochemistry 17 (1978) 5723-5729. Kurjan, J. and Herskowitz, I.: Structure of a yeast pheromone gene (MF~): a putative ~-factor precursor contains four tandem copies of mature ~-factor. Cell 30 (1982) 933-943. Laemmli, U.K.: Cleavage of structural proteins during assembly of the head of bacteriophage T4. Nature 227 (1970) 680-685. Loison, G., Findeli, A., Bernard, S., Nguyen-Juilleret, M., Marquet, M., Riehl-Bellon, N., Carvallo, D., Guerra-Santos, L., Brown, S.W., Courtney, M., Roitsch, C. and Lemoine, Y.: Expression and secretion in S. ceremiae of biologically active leech hirudin. Bio/Technology 6 (1988) 72-77. Ma, H., Kunes, S., Schatz, P.J. and Botstein, D,: Plasmid construction by homologous recombination in yeast. Gene 58 (1987) 201-216. Maniatis, T., Fritsch, E.F. and Sambrook, J.: Molecular Cloning. A Laboratory Manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY, 1982. Matsudaira, P.: Sequence from picomole quantities of proteins electroblotted onto polyvinylidenedifiuoridemembranes. J. Biol. Chem'.262 (1987) 10035-10038. Morel, F.: Regulation ofkidney functions by hormones: a new approach. In Greep, R.O. (Ed.), Recent Progress in Hormone Research, Vol. 39. Academic Press, New York, 1983, pp. 271-304. Norman, A.W., Roth, J. and Orci, L.: The vitamin D endocrine system - steroid metabolism, hormone receptors, and biological respons (calcium binding proteins). Endocrinul. Rev. 3 (1982) 331-336. Rabbani, S.A., Yasuda, T., Bennett, H.PJ., Sung, W.L., Zahab, D.M., Tam, C.S., Goltzman, D. and Hendy, G.N.: Recombinant human parathyroid hormone synthesized in Escherichla coll. Purification and characterization. J. Biol, Chem. 263 (1988) 1307-1313. Reeve, J,, Meunier, P.J., Parsons, J.A., Bernat, M., Bijvoet, O.L.M,, Courpron, P., Edouard, C., Klenerman, L., Near, R,M., Renier0J.C., Slovik, D., Vismans, F.J.F.E. and Ports Jr., J.T.: Anaholic effect of human parathyroid hormone fragment on the trahecular hone in involutional osteoporosis: a multicentre trial. Br. Med. J. 280 (1980) 1340-1351. Slovik, D.M., Rosenthal, D.I., Doppelt, S.H., Ports Jr., J.T., Daly0M.A., Campbell, J.A. and Neer, R.M.: Restoration of spinal bone in osteoporotic men by treatment with human parathyroid hormone (1-34) and 1,25-dihydroxyvitaminD. J. Bone Min. Res. 1 (1986) 377-381. Thomas, G., Thome, B.A.,Thomas, L., Allen, R.G., Hruby, D,E., Fuller, R. and Thorner, J.: Yeast KBX2 endopeptidase correctly cleaves a neuroendocrine prohormone in mammalian cells. Science 241 (1988) 226-230. Towbin, H., Staehelin, T. and Gordon, J.: Electrophoretic transfer of proteins from pulyaerylamidegels to nitrocellulose sheets: procedure and some applications. Proc. Natl. Acad. Sci. USA. 76 (1979) 4350-4354. Vlasuk, G.P., Bencen, G.H., Scarborough, R.M., Tsai, P.-K., Whang, J.L., Maack, T., Camargo, MJ.F., Kirs|~'.,,r, S.W. and Abraham, J.A.: Expression and secretion of biologically active human atrial natriuretic peptide in Saccharomyces ¢em~s/ae.J. Biol.Chem. 261 (1986) 4789-4796. Wingender, E., Berez, G., BlOcker, H., Frank, R. and Mayer, H.: Expression ofhuman parathyroid hormone in EsckerickIa coll. J, Biol. Chem. 264 (1989) 4367-4373. Zsebo, K.M., Lu, H.-S., Fieschko, J.C., Goldstein, L., Davis, J., Duker, If., Suggs, S.V., Lai, P-H., and Bitter, G.: Protein secretion from Saa:&m~yezs m'e~s~e directud by the prepro-a-factor leader region~ J. Biol. Chem. 261 (1986) 5858-5865.