4a an 2, 18-208, 1g Inefficient Membrane Targeting, Translocation, and Proteolytic Processing by Signal Peptidase of a Mutant Preproparathyroid Hormone Protein* (Received for publication, September 27, 1994, and in revised form, November 7, 1994) Andrew C. Karaplist, Sung-Kil Lim$, Hisamitsu Babat, Andrew Arnoldi, and Henry M. Kronenberg From the Endocrine Unit, Massachusetts General Hospital and Harvard Medical School, Boston, Massachusetts 02114 A preproparathyroid hormone allele from a patient with familial isolated hypoparathyroidism was shown to have a single point mutation in the hydrophobic core of the signal sequence. This mutation, changing a cysteine to an arginine codon at the ~8 position of the signal peptide, was associated with deleterious effects on the processing of preproparathyroid hormone to propara- thyroid hormone in vitro. To examine the biochemical consequence(s) of this mutation, proteins produced by cell-free translation of wild-type and mutant RNAs ‘were used in assays that reconstitute the early steps of the secretory pathway. We find that the mutation im- Pairs interaction of the nascent protein with signal rec- ognition particle and the translocation machinery. Moreover, cleavage of the mutant signal sequence by solubilized signal peptidase is ineffective. The conse- quence of this mutation on processing and secretion of parathyroid hormone is confirmed in intact cells by pulse-chase experiments following transient expression of the mutant protein in COS-7 cells. The inability of the mutant signal sequence, however, to interfere with the targeting and processing of other secreted proteins does not support obstruction of the translocation apparatus the mechanism underlying the dominant mode of in- heritance of hypoparathyroidism in this family. Proteins destined for residence within membranes or for secretion contain hydrophobic amino-terminal sequences re- ferred to as signal sequences (1). These sequences direct the nascent polypoptides bound to ribosomes to form a functional junetion with rough endoplasmic reticwum (RER)' mem- bbranes, thereby assuring the translocation of the growing This work was supported in part by Nationel Institutes of Health Grant DKI1784. The costs of publication of this article were defrayed in part by the payment of page charges, This article must therefore be Thereby marked “advertisement” in accordance with 18 U.S.C. Section 1a golly to indeate thie act + Supported by Medical Research Council of Canada Centennial Fellowship Award. To whom correspondence shouldbe addrested: Diy of Endocrinology, Sir Mortimer B. Davis Jewish General Hospital, ‘McGill University, Lady Davie Inet for Medical Research, Ra. 628, ‘8155 Cote Ste Catherine Ra, Montréal, Québec, Canada HS 1E2. Tel {516-300-8260 (ext. 4007); Fax: 514.940.7573, ‘Present address: Dept. of Internal Medicine, Collegeof Medicine, ‘Yonsei University, C.P-O: Box 8044, Seoul, Korea. “TPresent addres: Phd Divison of internal Medicine, Kobe Univer- sity School of Medicin, Kobe 650, Japs. I Present ‘address: Laboratory of Endocrine Oncology, GRA 1021, Massachusetts General Hospital and Harvard Medial Schoo, Boston, aa oziig * The abbreviations used are: RER, rough endoplasmic reticulum: SRP, signal recognition particle; PTH, parathyroid hormone, PAGE, polyserylamide gel electrophoresis, polypeptide chain into the lumen of the endoplasmic reticulum and ite subsequent cleavage by the luminal-localized signal peptidase enzyme (for review, see Ref. 2). ‘The nascent secretory protein i localized to the endoplasmic reticulum via a targeting apparatus consisting of the signal ‘recognition particle (SRP) and its membrane-bound receptor on RER. The SRP binds to the signal sequence as it emerges from the large ribosomal subunit, This results in a transient delay or even arrest of translation (3) aimed in preventing premature folding of the precursor protein. When the SRP-ribosome com: plex encounters the SRP receptor or “docking protein,” a series of reactions take place that result in the insertion of the nas- cent chain into the translocation site, release of SRP, resump- tion of translation, and initiation of translocation. GTP binding and its hydrolysis are required for these events to take place (4-6). As the nascent polypeptide transverses the transloca tion channel (7, 8) and emerges into the Iumen of the endoplas- ‘mie reticulum, it is modified further by the signal peptidase enzyme complex that catalyzes the endoproteolytic cleavage of the signal sequence. Only a small number of natural mutations in human signal sequences have been reported to have direct correlation with defective secretion and associated pathological states (9-11). We have described such a mutation in the signal peptide of one allele of preproparathyroid hormone (prepro-PTH) gene from a kindred with a form of familial isolated hypoparathyroidism ). This is an inherited metabolic disorder characterized by hypocalcemia and hyperphosphatemia resulting from lack of biologically active cireulating PTH, the major ealcium-regul ing peptide. In this family, the disorder was inherited as an autosomal dominant trait (12). The single point (T to C) muta- tion changed the codon at position ~8 (signal peptide residues are numbered negatively starting from the site of cleavage toward the amino terminus) of the signal peptide of prepro- PTH from eysteine to arginine, thereby disrupting the bydro phobic core ofthe signal sequence (Fig. 1). Associated with this change was a dramatic impairment in the processing of the in vitro translated mutant prepro-PTH protein to pro-PTH by ‘microsomal membranes (9). ‘Which step(s) ofthe early secretory process is affected by this mutation is not readily evident. Conceivably it could preferen- tially affect one or all ofthe steps involved, such as binding to SRP, targeting to the RER, translocation through the mem- brane, and proteolytic processing by signal peptidase. In this report, we have systematically examined each of these steps ‘using tautant and wild-type forms of in vitro translated prepro- PTH proteins by assaying their interaction with components of these various processes. Moreover, the consequence of this ‘mutation on processing and secretion of PTH was examined in intact cells by transient expression of the mutant protein in COS-7 cells. Finally, we have used co-transfection assay to 16291680 2s “a7 Met Tle Pro Alo Lys Asp Met Alo Lys Met Tle Pro Alo Lys Asp Met Alo Lys Met Tle Pro Ala Lys Asp Met Ala Lys Wid type Single mutant Double mutant -16 -10 3 a Le Cis Phe Lev Thr tys sor Tle Val Het Leu Ala fle Arg Phe Leu Thr Lys Ser Te Val Wet Leu Arg Ile Arg Phe Leu Thr Lys Ser Val vet Val vet pe Asp Gly 7 asp Gly — asp Gly - Lys Ser Val Lys Lys Arg ~ PTHCL-B4) LYs Ser Val Lys Lys Arg - PTH(I-84) Lys Ser Val Lys Lys Arg ~ PTHCI-B4) Fic. 1, Signal peptide sequence of human prepro-PTH. Amine acids ~25 tg =I constituto the signal peptide of wild-type and mutant (ingle and double) forms of human propro-PTH, The 6 residues fallow: Ing the signal peptidase cleavage site (arrowhead) malke up the pro sequenee, and the 84 amino acids of mature PTH follow, Residues 16 fo "5 lundertined) comprise the Rydrophabic core ofthe signel peptide ‘The described patients mutation (Cys — Arg atthe —8 position, singe mutant) and the additional substitution at the 10 positon of the ‘Sgnal peptide (Ale — Arg: double mutant) are indicated by boldfhee ‘Numbers above the amino acids indicate their position relative to the Signal peptidase cleavage sie define the mechanism by which this specific mutation could cause clinical hypoparathyroidism in the presence of a second, apparently normal (9), PTH allele. [EXPERIMENTAL PROCEDURES Materiale—Restrition endonucleases, Klenow fragment of Bscher- ghia eali DNA polymerase I, and Té DNA ligase were purchaved from New England Biolabs (Beverly, MA). Endoglyeosidase HY (endo-f-N- acetyglucosaminidace H) waa from Boehringer Mannheim, Rabbit re ticuloeyte lysate, wheat germ extract, and canine pancreatic rough ‘microsomes were purchased from Promega Corp. SRPs prepared from: freshly excised canine pancreas and eluted from an_aminopentyl ‘agarose column (13) were a generous gift from Resa Gilmore, University of Massachusetts, Worcester, MA. Construction of Plasmids Plasmid pWTS4 was prepared by intert- Ing a DdeFXal fragment of human preproPTH cDNA encoding propro-PTHA-8t) within the HindllLXbal cloning sites of the mam ‘malian expression vector pCDMB(9). The single paint mutation T to C) ‘was introduced at tho ~8 position ofthe signal peptide of propro-PTH EDNA sing olizonscleotide-directed mutagenosis (14), resulting in plasmid pSM&4 (single mutant; se Fig. 1) Plasmid pSMB4 was then Further modified by two additional point mutations introduced into the “10 position af the signal peptide (GCA to CGA; Ala to Arg resulting in plasmid pDMB4 (double mutants Fig. Plasmics pWTSS, PSMES, and pDMSS, all containing propro-PTHA. 488) sequences without the termination eadon, were constructed by Subeloning a Hind ll-Pell fagment containing the propre sequences from plasmids pWPS4, pSMB, and pDMSi, respectively, into plasmid SP-PTHOtbal BalS1/Clal#6) comprising mature PTH1~83) sequences ‘without the termination eodoe (15). Imorder to constract «plasmid encoding » protein ofthe same size as prepro-PTH but lacking 2 functional signal sequence, Neo linkers were introduced into the Smal site in the polylinker of SP-PTHXbal Bats Clal¥7) (5), and a 225-base pair Heel Hindllt fragment derived fom the same plasmid waa ligated into the Neol site. This introduced an ATG initiation codon followed by PTH sequences (5i-82)-Ala in place ‘fhe signal sequence. This plasmid was then restricted with Clal and Hindill aod ligated to @ DpnL-Hindll fragment ielated from plasmid 'SP-PTH (15), containing sequences encoding PTHI1-84). The resulting plasmid (PRM, for random mutant signal sequence), encoded Mot- PTHIS1-82)-Aln'Ser-PTHI1~B, ‘To introduce an Nslinked glycosylation site into the PTH-coding sequence of plasmids pWT8s and pSMB4, synthetic oligonucleotides, encoding Ser-Asn-Gly-Ser-Gly-Ciu-Gly-Val.Gla-Ser, were ligated inte the unique TagT site of he prepro-PTH coding sequence the underlined sequence indicates Uhe consensus sequence for Neglycosylation). The eaulting plasmids, pWTS4G) ond pSMB4(G), fered from pWT34 and DSMB4, respectively, only by the insertion of 9 amino acids between residues Sor 17 and Mot-18 ofthe mature PTH protein ‘in Varo Transcription, Translation, and Analysis of Producte—Plas side wore linearted and sense RNA strands were transeribed using Processing and Secretion of Mutant Prepro-PTH sither7 or SP6 RNA polymerase (15. Transaton reaction in rabbit Fevculoste lsat and in wheat germ extract were porto acorn {othe manufactorers Promega) procedure, Translation preducts were Immunaprotated using afty-puried goat ant-human PTHCL 34} antnrut and mueted to 18-20% contnanoe gratin! SDS-poly acjlamide pl shtrophoens (PAGE). Astradagraphy was per fered after treating the gel with ENHANCE (Dalont NEN? ax deserbedproviuay (15) The iantty of PTIEeltedpplen Was Cetaliched ty amo-ersnal radjoeeguenee algae 16) Postronslatonel Membrane Binding-Troneed RNA ising tho termination codon were tanaribed fom plasm BTR, pS8 fn pDMES and trae in the rabbit reenter ‘stem for sin st 24°C. Asrintcarbnsic acd was ded 00 1 {iit tanalaton ination, and fer another imi eatin, ‘matin war add (1m Sina conseatration) to block popige long {ion Incubation was continued for anather 15 mins et egrets {sloth orginal rough micrwrne preparation that bas ben jute to'a concentration of Ay iieo ane Ref 17) of mcrnomal mmembranc'S lof tranalaion minturs were ddd, and incubation Contd fr 18 min at 24°. Translation prs were then ine Sted at 0 for 1D mi. Imerton nto the membranes was sceoed by Centrifugation othe membranes though ether a pyle elt oF Sn EDTA euro stop susbion 1,19. Superantants containing mom irene peptides snd pelts were inmunoprectsted and subjected {8 SDSPAGE analy, ee docrbod 15) Proteone Protection Assy astay fr c-tranlationa trnaac: tion wide and singletatant cRNA wore translated fn rabbi retcloyie Hate cll fee eaten in the presence a cnn panereas ‘iowa moribranes Aor a Iban inecostnn at 24°C, peters (30 gin nal concentration) wan ed tote ranaltion reaction ‘nntues either alms or in combination with Ton X00 (0), and incubation was coined fora further 0 min one, Aer inactivation ofthe proense nth phorsimethyxatons! Arie final concentration Sih radiaboled proteins wire sutpcted to immunoprecpaton and SDS PAGE analgee ‘Brdogijeosiane H Digestion —Trancrited products of pWT84G) and pSM4G) wore tarsated inthe absent or prone ef mi tome membrane, imiinopreipate, and tented with endoglen dame H Go miluit ae dened previousy 9) Following overnight digetion, bovin sera albunin was aed 0 a carrer al O58), fn the samples were precipitated hy kecol trichlooncetc aid 13 final concentration. Felling conefgaton, th precipitate wore fenupended in 04 M Teie HCl, 7, and prepared for SDS-PAGE ingts as uml ‘Signal Pepidase Assey EDTA stripped, nacease-treated ragh mi cronmee wore prapared fom satin pancross ov dvsibed (2). Al ute of rogh iiroomes mere restapended by bemogeniation in ‘Send butter 20 mu HEPES pH 7.8, 80 mu Nac) toa na concer {tation of 50 Aye unital Saableaton of signal peptigane with tedium densyekcate war performed. as decribed. previously (2) Bey, oe salen of 10% (wh) soz donycoate was mixed with 19 volumes of membrane roapesion. The enltant eer ation wat centrifged at 100000 % g forth and the supernatant as fone Siquld nitrogen. Wild type and singlemutant Props PTH and prope DPIHIG) cRNA wore transite n'a whest gern cles aay and fransation pris were ned an subetrte or Storgent solid Siena! peptcace 20) Atypical 90 pottrsoaatonal cleavage aay Conte 20 ofthe detergent erat 20 yo wheat germ real {fon mintre, and 10 pl of water, Poteranlaional cleavage by signa Dertldate was llowed to proved 2 25° fr 80 ni. Pulse Chase Anapee “COST cals ctured io 1-well 2-0 di ante plats tn Dulbocc’s model Basics medin supplemented ‘ith 1og fetal calf sro and anti mere trantoted with Pas This pWT64, pOMB«, nnd pDMO4 tsing the DEAE dostran proton flowed by ioe nth! elon shock (21) Five das inter, cle tree ringed tice with Dulteca's minded Eagle’ median mithot Tethionine and supplemented wth 2 dined fetal calf sera 9a then incabated for {5 min at 97° inthe sme medio consinng P*Simethionne (0 wCiwe. The ells were then washed once with Int of Delbeers modified Eagle's mediam ‘with methionine, supe Inented with 10% fetal ovine sem), nd then inetd farther the ‘amo medium forthe Umer indicated. Conditnnedsnedia were aved iErimmunopreipitaton ead celle wore sed with O ml of beibaer Got mn Tre Hch pt Fa, 1% Toten X00, 04 SDS, ¥e sodom Slonsholte, 5 mit DTA: 0.02% son aie, 0 156 NSC) with 50 SF honyimethysulfonl five sta ial eoncentration of 500 pp DNAs the el pete were shored witha Zn ec and te trate wore brifyconrfged prior inmmannpreiptaion.Processing and Secretion of Mutant Prepro-PTH 1631 Single Double Fic. 2. Processing of normal and i om mutant prepro-PTH. Autoradiogram of mulaltPe arcane ea [Pilleuine labeled proteins derived from the translation of eRNAs. transcribed Membrane equivalents 0 4 8 12 0 4 8 12 04 8 12 from plasmids ‘containing wildtype (pWr84}, single mutant (pSMBA, and double, mutant (pDMBA)" prepro-PTHL eDNA. Translation was performed inthe PreproeT#—— rabbit reticulocyte Isat cllfree system Prop — in the absence (0 eq) or presence of in- ‘aan an creasing amounts (4, 8, and 12 eq) of nine pancreatic microsomal membranes. Competition Stusies—COS-7 cells were transfected with pWT84 or SMB# either alone o in combination with pWTS2 encoding for human repro-PTH(.-52) (22). After the indicated times. posttransfetion, ‘ells were labeled with [*SImethionine for 15 min at 97 *C,as deseribed above. Cells were then lysed and processed for immunoprecipitation fand SDS-PAGE analysis, RESULTS Processing of Wild-type and Mutant prepro-PTH—To assess the processing of wild-type and mutant forms of prepro-PTH to pro-PTH, transcribed sense RNA strands were translated in a rabbit reticulocyte lysate cell-free system in the absence or presence of increasing concentrations of canine pancreatic crosomal membranes. In this co-translational assay, substitu: tion of Cys — Arg at the =8 position of the prepro-PTH signal peptide (single mutant) resulted in impaired processing to pro- PTH as compared with the wild-type form (Fig. 2). The addition of a second changed residue within the hydrophobic core of the signal peptide (Ala Arg at the ~10 position, double mutant) completely abolished its proteolytic processing by rough micro- somes. Maximal processing of wild-type and single-mutant pre- ‘cursors occurred when mierosomal membranes were added to a final concentration of 4 eq/26 ul of translation mixture. Interaction with SRP—The impaired processing of the mu- tant prepro-PTH proteins by microsomal membranes may re- sult from the inability of SRP to recognize efficiently the al- tered signal peptides. Binding of SRP to the signal sequence of nascent secretory proteins induces a site-specific elongation arrest of translation (8, 23). We, therefore, assessed the inter- action of the mutant prepro-PTH signal sequences with SRP by ‘examining the effect of exogenous SRP on inhibition of trans- lation in a wheat germ cell-free system (Fig. 8). The mRNAs for pWTS4, pSM84, pPDMS4, pRMB4, and rabbit globin were trans- lated in the presence (final eoneentration 0.02 and 0.04 Aran ‘units/mD or absence of exogenous SRP. The translation of the wild-type protein was inhibited significantly more than the translation of the mutant peptides by exogenous SRP. More- over, translation of the single-mutant form was impaired to a greater extend than that of the double mutant. The addition of SRP did not affect the synthesis of globin (eytoplasmie protein with no signal peptide) and PTH-like protein with a random ‘mutant signal peptide. These results suggested that the ob- served inhibition in processing of the mutant precursors i, at least, partly due to the inability of their signal sequences to interact effectively with SRP. ‘Translocation-competent Binding—To assess binding of the ‘mutant prepro-PTH proteins to microsomal membrane compo- nents, truncated cRNAs for wild-type and mutant forms of prepro-PTH missing the termination codon were transcribed from plasmids pWTS3, pSMS3, and pDMS3 and translated in the rabbit reticulocyte lysate cell-free system, thereby “freez- ing” the naseent proteins on ribosomes. After translation was complete, rough microsomes were added, and insertion into ‘membranes was assessed following centrifugation of the reac- tion mixture through either a physiological salt- or an EDTA- sucrose step gradient. Only nascent chains tightly bound to the translocation apparatus pellet with the membranes in the pres- cence of high concentrations of EDTA (18). When membranes were added to the reaction mixture post-translationally, the propeptide (wild-type and single mutant) as well as the pre- propeptide bands were seen in the precipitates (Fig. 4A). More- ‘over, peptides pelleting with the rough microsomes were resist: ant to extraction with EDTA, suggesting that translocation competent binding of the ribosome-nascent chain complex to the membranes had taken place. This post-translational assay suggested that the mutant proteins can be targeted to micro- somal membranes and bind in a translocation-eompetent man- ner. Although the mutant forms were able to engage the trans- location apparatus, they did so less efficiently than the wild: type form, Equal loading of the various reaction mixtures was verified by examining corresponding supernatant fractions us- ing SDS-PAGE analysis (Fig. 4B). Thus, the amount of nascent protein that sedimented with rough microsomes paralleled the affinity of SRP for the respective peptides (wild-type > single ‘mutant > double mutant) likely reflecting their degree of in- teraction with SRP. As shown in Fig. 44, processing of the ‘mutant forms to pro-PTH was again markedly impaired when compared with the wild-type protein, even though transloca tion-competent binding had been achieved. These results sug gested that additional processes in the early steps of the secre tory pathway, i. translocation and/or interaction with signal peptidase, may be impaired by the mutant signal sequences, Protease Sensitivity of Membrane-bound Nascent Chains— ‘To determine the location of the naseent chains within the rough microsomes, we examined their sensitivity to digest with proteinase K either in the absence or presence of Triton X-100. Protein products that are translocated to the lumen of ‘the microsomal vesicles would be protected from digestion by proteolytic enzymes that are unable to enter these vesicles. AS shown in Fig. 5, the processed peptides (pro-PTH) were pro- tected by the membranes from proteolysis by proteinase K. The protection ofthese forms must have resulted from their seques- tration into the lumen of the microsomal vesicles. This was confirmed by the addition of Triton X-100 to the reaction mix- ‘ture, thereby, permeabilizing the membrane bilayer and allow= ~itig the protease to gain access to all protected polypeptides. ‘The minor unprocessed single-mutant product that is protected from proteolysis may represent unprocessed nascent protein, ‘that has not fully translocated, yet is protected from proteolysis, by the ribosomes and the tight ribosome-membrane junction required for translocation. N-Glycosylation of a Modified PTH Sequence—Since the mu- tant signal sequence might be a poor substrate for signal pep- tidase, cleavage alone is an insufficient criterion for the local- ization of PTH peptides. The fact that a fraction of the mutant prepro-PTH was protected from proteolysis raises the question of how far into the translocation process the uneleaved precur- sor has progressed. Because glycosylation is restricted to the Tumen of the endoplasmic reticulum, the addition of earbohy-1632 Fic, 3. Effect of SRP on translation, ‘The cRNAs for wild-type (pWTSH), single ‘mutant (pSMB4), double mutant (pbMS), Fandom mutant (pRMB4), and rabbit glo: bin were translated in a wheat germ cell free system inthe absence (0) or presence ff exogenous SRP (inal concentration 1.0211 end 0.04 (2) A280 unite Prysilogia ‘olsen EDTA Solution 2 2 g22 g 32 33 24 2 5 8 a 8 A es e-- a = 8 roe oa iia ore Fic. 4 Translocation-competent binding to microsomal mem- branes. Truneated eRNAs missing the termination codon were tran serbed from plasmids encoding wild-type, single mutant, end double mutant cDNAs and translated in rabbit reticulocyte Iyaate system, ‘Translocationcompetent binding to microsomal membranes was. a sessed by centrifugation ofthe membranes through either a physiclog- {cal slt-or an EDTA-sucrose step cushion. Radiolabeled proteins inthe pellets (A) and corresponding supernatanis (B) are displayed. Single Wild type Mutant Membranes tee tae Proteinase K ean 28 Triton X-100 -- -- + PreproPT—e ProPri—— Fic.5, Protease sensitivity of membrane-bound nascent chains. Wild type (pWTS4) and single mutant (pSM84) eRNAs wore ‘translated inthe rabbit reticulocyte lysate system in the presence (=) of ‘dog pancreas microsomal membranes. After tranaltion wae completed, reactions were treated with proteinase K (20 ag/ml) with (+) or without ("yt addition of 1% Triton X-100, Radiolabeled translation products ‘were immunoprecipitated and analyzed by SDS-PAGE. drate to the PTH sequence by microsomal membranes would constitute further evidence of translocation and would not di- rectly require the presence of a suitable substrate for signal peptidase (19), Because the mature PTH protein does not have an N-linked glycosylation site, we engineered such a site 23, amino acid residues distal tothe signal peptidase cleavage site. Fig. 6 shows that upon translation of pWTS4(G)-transeribed cRNA in reactions supplemented with microsomal membranes, ‘two translational products with PTH immunoreactivity were seen that were not present in the absence of membranes. The smaller product migrated with an apparent molecular weight slightly greater than that of authentic pro-PTH and was there- Giobin Cp Cine Oni Oler sence cltbed Processing and Secretion of Mutant Prepro-PTH Double Mutant. Random Mutant wid Single ‘Type Mutant Single Wild type Mutant a — £4 Ef Gly-proPTics) PreproPTicc) = ProPTHG) “35s Fic. 6 Glycosylation of wild-type and mutant prepro-PTHG). Plasmids pWT84(G) and pSMBA(G) were tranteribed in vitro and trans lated in the rabbit reticulocyte lysate eclfree system in the absence (CD or presence (M) of canine microsomal membranes, Translation producte were immunoprecipitated and treated with (+) or without (TE) endoglyeosidase H fore felt to be pro-PTH(G). The second product, which migrated ‘more slowly than prepro-PTH(G), was believed to be the glyco- sylated form of pro-PTH(G). This was confirmed by treatment with endoglycosidase H which, by removing carbohydrate on this peptide, shifted its position on an SDS-PAGE gel to that of pro-PTH(G). Interestingly, prepro-PTH(G) appeared to be proc- ‘essed much more efficiently by canine microsomal membranes ‘than the unmodified form of the protein (compare Figs. 2 and 6), and this may simply reflect the influence of the extended length or the specific structure of the naseent chain. ‘Translation of pSM84(G)-transcribed eRNA in the presence ‘of membranes resulted in the appearance of three PTH-immu- noreactive products, two of which eo-migrated with pro-PTHG) and its glycosylated form, while the third was consistent with the unprocessed prepro-PTH(G) form. Again, the addition of 9 amino acids to the mature PTH molecule resulted in a more efficient cleavage of the signal sequence by mierosomal mem branes as compared with the unmodified form (see Figs. 2 and 6). Yet, the mutant signal sequence was once again processed less efficiently than the wild-type sequence, as indicated by the persistence of the unprocessed mutant prepro-PTH(G) form in the immunoprecipitated products. Once cleaved, however, pro- PTH(G) was glycosylated appropriately as confirmed by treat ‘ment of the reaction products with endoglycosidase H. The addition of carbohydrate to this moiety provides direct and independent evidence for its translocation, although signifi- cantly impaired, across the endoplasmic reticulum membranes. Furthermore, no larger glycosylated product was found that might have represented a protein that was translocated, gly- cosylated, and yet not cleaved by signal peptidase. Therefore, the uncleaved mutant prepro-PTH(G) was not delivered to the ‘slycosylation machinery on the inner surface of the microsomal, membranes. Processing by Solubilized Signal Peptidase—To determine whether the single-mutant form of prepro-PTH is an unsuit- able substrate for signal peptidase, we assessed proteolytic processing of the nascent protein in a translocation-independ- ent assay, Following translation of wild-type and single-mu- tant prepro-PTH eRNAs in the wheat germ cell-free system, signal peptidase assays were performed by mixing aliquots ofProcessing and Secretion of Mutant Prepro-PTH = $8 = S8 e a3 BaF ‘Signal Peptidase Fic. 7. Signal peptidase assay. Following translation of either pWTS4- and pSMB4- or pWTS4G) and pSMS4(G}-transribed RNAS In a wheat germ extract, signal peptidase assays were performed by ‘xing aliquot of translation mixture and signal peptidase prepared by directly solubilizing eanine pancreatic rough microsomes. translation mixture and signal peptidase prepared by directly solubilizing canine panereatie rough microsomes, In this post- translational assay, no eleavage was detected for the mutant precursor, while a minor amount of processing was observed for ‘the wild-type form of prepro-PTH (Fig. 7). Since the wild-type prepro-PTH is also an inefficient substrate for solubilized sig- ral peptidase, an attempt was made to inerease the sensitivity of the assay using as substrate the modified form of the prepro- PTH molecule containing an engineered glycosylation target site. RNAs encoding the wild-type and single-mutant form of prepro-PTH(G) were translated, and solubilized signal pepti dase was added posttranslationally (Fig. 7). As found with intact membranes, wild-type and mutant forms ofthis modified PTH protein were processed more efficiently than their normal ‘counterparts. Indeed, the mutant prepro-PTH(G) was proc- cessed by solubilized signal peptidase unlike its unmodified form, but the extent of cleavage was significantly less than that of the wild-type prepro-PTH(G). These results suggested that the mutation in the signal peptide makes the protein a less suitable substrate for signal peptidase. Processing and Secretion of Prepro-PTH in COS-7 Cells—To confirm the cell-free data outlined above in intact cells, we examined the processing and secretion of mutant and wild-type forms of prepro-PTH from COS-7 cells transfected transiently with the corresponding expression plasmids. Five days follo ing transfection, cells were pulse-labeled for 15 min wit [P°Slmethionine and then chased for the times indicated in Fig, 8 with medium containing cold methionine. The PTH species in cell extracts and conditioned media were immunoprecipitated and then resolved by SDS-PAGE. At the earliest time point examined (0 min), the predominant product immunoprecipi- tated in wild-type PTH-transfected cells was pro-PTH, while only barely detectable levels of prepro-PTH precursor were observed (Fig. 84). Moreover, immunoreactive PTH was seen in the culture medium of these cells after 30 min of chase (Fig. 8B), The rate of processing seen with the wild-type sequence was in striking contrast to that observed with the mutant forms. In lysates from cells transfected with either mutant, the predominant band at the earlier time point corresponded with the prepro-PTH precursor with proteolytic cleavage to pro-PTH being dramatically reduced in efficiency. With the single mu- tant, only a small amount of pro-PTH was detected; no pro- PTH was found in cells expressing the double mutant. With the single-mutant form, a band corresponding to prepro-PTH was detectable even after 120 min of chase, consistent with the in vitro observed inefficient cleavage of this precursor to pro-PTH by microsomal membranes. Immunoreactive PTH was detect- able in culture media of these cells after 30 min of chase but at ‘substantially reduced levels as compared with the wild-type form. The double mutant form of prepro-PTH, although ef ciently translated, was not processed to pro-PTH, nor was 1633 immunoreactive PTH detectable in the culture media of eells transfected with this plasmid (data not shown), These results demonstrate in intact eells the inefficient processing of the ‘mutant prepro-PTH molecules as compared with the wild-type form. The single-mutant prepro-PTH allows a small amount of normal processing, consistent with the results in the cell-free studies that showed that the single mutant molecule can en- gage the translocation apparatus, although less efficiently than normal. The double mutant prepro-PTH almost totally fails to engage the translocation apparatus in cell-free extracts and is not processed at all in intact cells. ‘Competition Experiments—A co-transfection assay in COS cells was used to address the issue of dominant transmission of hhypoparathyroidism in the family carrying the single-mutant prepro-PTH allele. The objective was to determine whether the inefficient processing of the mutant precursor can result i obstruction of the translocation apparatus and, thereby, to global defects in protein processing. For this study, an expr sion vector containing sequences encoding a truneated version of prepro-PTH, missing the last 32 residues (WT52), was used to provide a readily distinguishable varient of prepro-PTH with ‘normal signal sequence. 'WTS2 was co-transfected into COS-7 cells with either wild type or mutant prepro-PTH expression vector. Over-expression ‘of the mutant precursor for up to 10 days did not interfere with the processing of prepro-PTH(1-52) (Fig. 9). Although these results may simply reflect lack of sensitivity ofthe system, they do not support global interference with protein processing at the microsomal membrane level as a consequence of overex- pression of the mutant prepro-PTH form, DISCUSSION ‘Translocation from the eytoplasm into the endoplasmic re- ticulum, is a multistep process requiring: a functional amino- terminal signal peptide. A signal sequence must perform effec tively several distinct functions required for the efficient translocation of secreted proteins. These subfunetions include its recognition and binding to SRP, its interaction with mem- brane-bound components of the export machinery, opening the protein-conducting channels to initiate translocation, and ap- propriate presentation to the signal peptidase for cleavage. ‘Three domains have been identified as a common feature of ‘eukaryotic signal sequences, and considered to be necessary for carrying out these functions: a positively charged NH, termi- nus, a central hydrophobic core of 10-15 amino acid residues, ‘and a polar COOH-terminal region (24, 25). While the COOH- ‘terminal region influences the efficiency and fidelity of signal peptidase cleavage, intactness of the hydrophobic region is indispensable for initiating translocation, "Two reported inherited mutations in the signal sequence of Jhuman secreted proteins, namely preprovasopressin (10), and preprofactor X (11), involve the COOH-terminal region of the respective signal peptides. Thus, a point mutation resulting in ‘substitution of Arg for Gly at the ~3 position of the factor X signal peptide (Factor X sents Domingy blocks cleavage by signal peptidase but does not interfere with targeting and transloca- tion to the RER (11). Similarly, a naturally oceurring substitu: tion of Thr for Ala at the ~1 position of the signal peptide of preprovasopressin results in central diabetes insipidus (10) ‘This mutant protein, similar to the Factor Xsunto Domina»: undergoes inefficient cleavage by signal peptidase, although targeting and translocation to the RER are not measurably affected. ‘The present study is the first to examine the effect of @ naturally occurring substitution at the hydrophobie core of a signal peptide that results in human disease. Prepro-PTH, the precursor of PTH, contains a typical 25-residue amino-terminal1634 A Time (ain) Fio.8. Expression of wild-type and mutant forms of prepro-PTH. in 7 COS-7 cells. Plasmids pWTS4, pSMBi, Cadi fand pDMB4 were transiently transfected into 'COS-7 ells, Five days, following transfection, the cella were pulse-labeled with (Shmethionine for 15'min. At the Indicated times following pulselabeling, the media were removed, and cell extracts ‘were prepared. Both ell extracts (A) and ‘media () were immanopresipitated = ing a PTH.apecife antibody pricr to SDS- PAGE analysis, sample not processed Time 4 7 swt52 st swt52-45m84 wis? 4584 im wet smi sz wes ye sss 52-98 eae rms) preproPTis2} roms 77 rins2) signal sequence followed by a 6-residue prospecitic peptide and the mature hormone (residues 1-84; see Fig. 1). The hydropho- bie core ofthe human prepro-PTH signal peptide is composed of 12 contiguous uncharged amino acids (residues ~5 to ~16 of the signal peptide). Tn the present study, we have demonstrated that substitu- tion ofa charged amine acid, Arg for Cys, in the signal peptide hydrophobic cre of prepro-PTH impairs co-ranslational trans- location as wel as posttranslational cleavage by isolated signal peptidase. The impairment was even more evident when two charged residues were introduced in the hydrophobie core of ‘the signal sequence. In contrast to deletion mutants (26), this change interfores not only with translocation and cleavage by signal peptidase but also with binding to SRP. Since substitu- tion of a hydrophobie amino acid, leucine, for eysteine o its deletion was ineffective in modifying translocation and proc- ‘essing (26), the present findings would suggest that a change in ‘the hydrophobicity of the core is responsible for the observed slobal disruption inthe processing ofthe mutant prepro-PTH. Similarly, single charged amino acids introduced in the hydro- phobic core of the E.coli maltose binding protein signal peptide impair secretion of the protein into the external periplasmic compartment of the cell (27). Itis rather intriguing that three distinct reports of inherited ‘mutations in the signal sequence of human seereted proteins, namely prepro-PTH (8), preprovasopressin (10), and preprofae- tor X (11), demonstrate inheritance ofthe associated disorder in an autosomal dominant fashion. As is the ease for the other ‘two reported disorders, however, it remains unclear why indi- viduals with a mutated PTH allele have hypoparathyroidism. From transfection studies in COS-7 cells, it would appear that * met See Processing and Secretion of Mutant Prepro-PTH. Wild type © 15 30° 60120 Single Mutant © 15 30 60120 Doubte Mutant: 0 15 30 60120 se 8 ~ Wild type 15 30 60 120 ail Single Mutant 15 30 60120 ag Time (win) oo post co-transfection (days) 10 a Ro, 9, Co-transfection of wild-type and mutant prepro-PTH. Plasmids WT and pSMBt wore transfected into 08-7 celseither alone orn combination ‘with pWTS2. ARer the indicated number of days posttransfection, cells were pulse Tabeled with ('Shmethionine for 19 min, Cell extracts were then prepared and im ‘munoprecipitated prior to SDS-PAGE analysis wes ames se 52-408 iS2-4sm88 expression of the mutant allele does lead to secretion of PTH, albeit inefficiently. Moreover, the normal allele would be ex: pected to produce sulficient circulating PTH to maintain eal- cium homeostasis. The only other reported ease of familial isolated hypoparathyroidism segregating with a mutation in the PTH gene, involved a point mutation affecting intron splic- ing and was associated with autosomal recessive inheritance of the disorder (28). Since heterozygous individuals for this mu: tant allele were unaffected, it would appear that one normal PTH allele is sufficient for maintaining calcium homeostasis. ‘Hypopathyroidism in the presence of one normal PTH allele would therefore suggest that the mutant gene product exerts a dominant negative effect in vivo. The mutant protein might interfere with the normal targeting and processing of other secreted proteins, including the normal PTH precursor. Such interference might even lead to destruction of parathyroid tis- sue in affected individuals; unfortunately, this is difficult to evaluate because the tissue is not readily accessible. Export incompatibility, however, has been observed in E.coli express- ing transport-defective P-galactosidase leading to lethal jam- ‘ming of the cellular export machinery (29). ‘The phenotype of the single-mutant prepro-PTH suggests that it might have dominant negative effects under appropriate conditions. The mutation allows a fraction of the prepro-PTH precursor to enter the translocation machinery, but the mutant protein is then cleaved inefficiently by signal peptidase. The inability of the uncleaved protein to reach the glycosylation ‘machinery (assessed using the precursor modified by inclusion of a glycosylation signal) suggests that the precursor is not ‘transported fully across the microsomal membrane. Such a protein that engages the translocation apparatus but fails toProcessing and Secretion of Mutant Prepro-PTH move through the apparatus efficiently might well have domi- nant negative effects. The competition experiment in Fig. 9 failed to demonstrate such an effect; perhaps higher levels of. protein expression or a longer term experiment are needed. Nevertheless, the observation that all three reported human signal eoquence mutations involve proteins that partly engage the secretory apparatus and appear to have dominant effects suggests that in vivo these matant proteins cause domainant secretory dysfunction REFERENCES 1 Walter, P, Giinare, and Hobe, (1084) Ca 38, 5-8 Cir Ct Sh Walter and Blbe, (1961) Calf Bis Consliy Fond Gore, (1939) Call 62, Rapisk Bend Gor, R. 1909 JCal Ba 11, 499-609 Mller Witboim Hl, Graze, In, Gamore, tnd Waker, P. (19) ‘Natare 268, 351-364 Sinem 5 Mand Boel 1991) Cell 5, 211-980, Simen 8M: and Blobel. 1992) Cl 0, 077-084 ‘Aral, Ac Horst. SA. Gardella, 7. 3, Baba, H, Levine, MA. nd TKrenbere HM. (1990) Clin Tass 86 084-i08? 10, Ito, Ov, ara. Konda, Salt H, Chinas Rach, M, and ‘ety, O° (19031 Clive Bh 68-2671 1635 21, Rach, ML, Wats, HHL High, KA. and Lively, MO. (1085) 4. Bil. Chom, 12, Aha, T-G., Antonarokie, 8, B., Kronenberg, H.M., Igarashi, and Levine, M "960 fede 66, 7-81 18. Walter P lbrabins Led Blebe, C1951). Cat Bi. 9, 545-650, 1. Zoller M3 and Smt M, (1988) Methods Enzymol, 100, 468-200 18 Be, Kats AG Wie, KM Reuao,HTaod Kroeber. 16, Winen, KM, Potts, JT, Jr, and Kronenberg, H.M. (1988) J. Bial. Chem, 263, 1077 1g777 12, Wolter, Py and Rubel, 98%) Methods Bneymal 96, 84-28 18 Berar By and Lngappe,V. Mat J Cl Bit TOL, 202-2901 20 Sacage, Cand bel 11927) Pre Da ead SoU SA. 74 BS ee 21. Lathan, Hand Meason, (198) Noclete Acide Ret. 11 1285-1907, 22 Lim SX, Candala Td. aba, H, Numbour, Sev and Kronenberg HM ‘92) Badoerinaigy iat, nas. 2890 128, Meyer, Ds Kraus, and Dobbertete, (982) Nature 297, 647-660, 34 Petnan, D’ and Hlaneron tO” (108) Mol Rll 167, 917100 35 von elie 988) J. fal. Bist 1B4, 9-108 8. Gath JR Alen, Ke, ively, M- Or and Kerper, B, (1989). Bil. Chem. 127, Bedovelle, H., Bassford, P. 4, Jr, Fowler, A. V.,Zabin, L, Beckwith, J, nd Hoang, (1990 Nasr Da 78-81 28, Parkinson, By ae Pale, RV 1982) Nature Gone. y 149-162 28, Basso, PJ, dr Shery. TJ. and Beckwith J 1979) J. Bact 138,