Isolation and Characterization of a Dibasic Selective ...

4 downloads 0 Views 3MB Size Report
Jan 21, 2016 - expressed as the percent of activity remaining, taking the incubation ... the stage of the seminiferous cycle, these immunoreactive cells.
JOURNAL OF BIOUXICAL CHEMISTRY 0 1994 by The American Soeiety for Biochemistry and Molecular Biology, Inc.

Vol. 269, No. 3, Issue of January 21, pp. 20562061, 1994 Printed in U.S.A.

THE

Isolation and Characterization of a Dibasic Selective Metalloendopeptidase from Rat Testes That Cleavesat the Amino Terminus of Arginine Residues* (Received for publication, August 2, 1993, and in revised form, September 21, 1993)

Valhrie Chesneau, AdrianR. PierottiS, Nicole Barr6, Christophe Crhminon8, Claude Tougardn, and Paul Cohenll From the Laboratoire de Biochimie des Signaux Regulateurs Cellulaires et Moleculaires, Universite Pierre et Marie Curie, Unit6 de Recherches Associee au Centre National de la Recherche Scientifique 1682, 96 Boulevard Raspail, 75006 Paris, the $Service de Pharmacologie et Clmmunologie, Departement de Recherches en Imagerie, Pharmacologie et Physiologie du Centre &Etudes de Saclay, 91191 Gif-sur-Yvette Ceder, and the IGroupe de Biologie de la Cellule Neuroendocrine, Unite de Recherches Assocse au Centre National de la Recherche Scientifique 1115, College de France, 75231 Paris Cedex 05, France

A metalloendopeptidase that selectively cleaves doublets of basic amino acids on the amino-terminalside of arginine residues was purified to homogeneity fromrat testes and analyzed further. ’ l k o catalytically active forms with apparent relative molecularmasses of 110,000 and 140,000 Da,respectively, werepresent in the purified preparation of the enzyme. Antibodies raised against the purified testis endopeptidase revealed by immunoblot both the 110-and 140-kDa formsin both rat testis and brain cortex extracts. The isolated enzyme was inhibited by metal chelators and divalent cations. Its activity, lost after preincubation with EDTA, was restored bylow concentrations of Zn2+ and Mn2+, thus demonstrating the metallopeptidase nature of the enzyme. This endopeptidase also exhibited a high sensitivity to amastatin (100%inhibition at 20 p d , an aminopeptidase inhibitor. A substrate specificity study using physiologically important or synthetic peptides containing a processing dibasic site indicated that cleavage occurred selectively at the amino-terminalside of an arginine residue, independent of the nature of the basic doublet. The enzyme produced such a cleavage at the A r g u s doublet of somatostatin 28 (K, = 43 p d , at the ArgArg doublet of dynorphin A (K, = 6.45 p d and atrial and at the Lys-Arg dounatriuretic factor (K, = 6.25 p), blet of preproneurotensin-(154-170) (K, = 17.3 p d . Moreover, cleavage efficiency was found to be higher for the larger substrates. The distinctive properties of this endopeptidase imply that this protein is a member of a novel class of proteolytic enzymes that may be involved in the endoproteolytic maturation of hormonal precursors.

their maturationis the excision of peptide sequencesby limited endoproteolysis (1-3). These processing events, initiated in the Golgi apparatus (4,5),occur at basic residues mainly arranged as doublets or singlets (1-3). Depending on the selectivity of the cleavages either in single or distinct cell types, the same precursor can generate various bioactive fragments (6). Similarly, some biologically active peptidesthat constitute the products of larger precursor maturation, such as dynorphin A, atrial natriuretic factor (ANF-(1-28)),l or somatostatin 28, may themselves be processed further to produce smaller activefragments. Attention hasfocused on the enzymes responsiblefor proteolytic processing of prohormones. Because of the low representation of these enzymes in cells, only a few endoproteases exhibiting a selectivity for dibasic processing sites inpropeptides and proteins have been proven to be involved in theproteolytic maturation process. Thebestcharacterizedexample is the yeast Saccharomycescerevisiae KEX2 gene product, a Ca2+dependent “subtilisin-like” endoprotease(7,8). Recently, mammalian KEX2 homologs have been identified by cDNA cloning (9-14). PC2 and PClPC3have been implicated in prohormone processing in endocrine and neuroendocrine cells (11,15); furin seems to be involved in a more general manner in proprotein maturation (16, 17). Many proteolytic enzymes, such as neutral endopeptidase (neprilysin,EC 3.4.24. l l ) , angiotensin converting enzyme (peptidy1 dipeptidase, EC 3.4.15.11, endopeptidase 24.15 (thimet oligopeptidase, EC 3.4.24. E ) , insulin-degrading enzyme (insulinase, EC 3.4.99.451, and PC4, a member of the KEx2 gene family (13, 18-20), have been found to be present in the testes in elevated quantities as measured by enzyme assay or by Northern blot analysis. Additionally, a number of different neuPeptide hormones and neuropeptides are synthesized as ropeptides and hormones have been shown to be present in the larger inactive precursors and then submitted to a series of testes by immunological and genetic techniques. The testes post-translational modifications along the secretory pathway to would thus seem to be a good source tissue for the isolation of produce bioactive peptides. One of the most important steps in a novel endoprotease. Previous work in thislaboratory has led to theidentification, * This work was supported in part by grants from the Ministhe de in rat braincortex, of an endopeptidase called “somatostatin 28 YEnseignement Sup6rieuret de la Recherche (to theUniversit6 Pierre convertase,” which was proposed as being involved in somatoet Marie Curie) andby the Centre National de la Recherche Scientifique (to Unit4 de Recherches 1682). The costs of publication of this article statin precursor processing (21-23). This and a number of other were defrayed inpart by the payment of page charges. Thisarticle must activities have been isolated by biochemical techniques and “aduertisenzent”inaccordancewith 18 proposed as putative processing enzymes, but none have yet therefore beherebymarked U.S.C. Section 1734 solely to indicate this fact. been cloned (3). $ Recipient of a research fellowship from the program “Stimulation” of the European Economic Community and a position of “Chercheur The abbreviations used are: ANF-(1-28), atrial natriuretic factorAssociB” from the Centre Nationalde la Recherche Scientifique. ( 1 To whom correspondence should be addressed. Tel.: 33-1-42228185; (1-28); HPLC, high performance liquid chromatography; PAGE, polyacrylamide gel electrophoresis;5-28, somatostatin 28. Fax: 33-1-42221398.

2056

Rat Testis Dibasic Selective Endopeptidase

2057

Determination of Substrate Specificity-A 39 w final concentration of peptide was incubated with an aliquot of the post-DEAE-TrisacrylM enzyme pool in 250 l ~ Ms-HC1 l ~ buffer (pH 7.5) for 1,3, or 15 h at 37 "C. The reaction mixture was then analyzed byHPLC (gondapak CI, column (3.9 x 300 mm), Waters) using the following gradients of acetonitrile inbuffer A (1g/liter l-hexanesulfonic acid, 20 m acetic acid) at a flow rate of 1mumin: 20-28.2% in 12 min for [Ala17,Ty1.201S-28-(1020)-NH,, [Ala"]S-28-(~19)-NHz, [Ar~~a171S-28-(1-19)-NHz, and prosomatostatin-(5&68)-NHz; 2635% in 30 min for a-neoendorphin and dynorphin B; 2048% in 20 min for dynorphin A and related peptides, bovine adrenal medulla dodecapeptide, and proenkephalin4128EXPERIMENTALPROCEDURES 140); and 2040% in 20 min for preproneurotensin-(154-170) and proMaterials ocytocin/neurophysin-(1-20). ANF-(1-28) was analyzed using an Peptides were purchased from Neosystem, Sigma, and Bachem or isocratic gradient of 29% acetonitrile in buffer A. Somatostatin 28 was analyzed in 25% acetonitrile in 100 m phosphate/triethanolamine synthesized by solid-phase procedures and analyzed byHPLC and buffer (pH 3.0). amino acid composition (21). Absorbance was monitored a t 215 nm. Fragments were identified by Methods their retention times compared to standard peptides and by amino acid Enzyme Preparation-Purification steps were carried out at 4 "C. composition. Frozen testes from Wistar rats (10 testes, 13 g; Pel-Freez Biologicals) Determination of Kinetic Constants-K,,, and V,, values were deterwere defrosted; homogenized in 50 m phosphate buffer (pH 7.4), 100 mined frominitial velocity measurements a t various substrate concenm KC1 (20%, wh); and centrifuged at 2000 x g for 15 min. The super- trations (six or seven points in a range from five times below to five natant was precipitated at pH 4.7 with 1 M ammonium acet5te buffer times above the K,,, were taken to draw plots). Enzyme assays were (pH 4.5), 5 m p-mercaptoethanol and centrifuged at 7800 x g for 30 analyzed by HPLC as described above.K,,, values were calculated from min. The pH of the supernatant was readjusted to 7.5. The enzyme Hanes-Woolf representations by a linear regression program. Since inpreparation was concentrated to one-tenth of its initial volume in an cubation conditions varied, pseudo-k,,, values were evaluated by taking ultrafiltration cell (Diaflomembrane Y"30, Amicon Corp.); submitted arbitrarily 1 unit of enzyme = 4.5 ng of enzyme preparatiodpl. to a Sephadex G-150 superfine column (Pharmacia LKB Biotechnology EndopeptidaseAntibodies-Polyclonal antibodies were raised in rabInc.); and eluted in 250 m Tris-HC1 buffer (pH 7.5), 5 t l l ~p-mercap- bits by multiple subcutaneous injections of 50 pg of purified endopeptoethanol. The active fractions (diluted five times with deionized water) tidase emulsified in Freund's complete adjuvant (24). The first booster were applied to a DEAE-TrisacrylM ion-exchange column (100-ml bed injection was given 6 weeks later and then monthly thereafter, and volume; Industrie Biologique FranGaise) pre-equilibrated with 50 mM rabbits were bled weekly. The titers of the different antisera were tested Tris-HC1 buffer (pH 7.5) and eluted by two successive linear salt gra- with an enzyme immunoassay using biotinylated endopeptidase and dients from 0 to 150 m KC1 in 400 ml and from 150 to 400 m KC1 in acetylcholinesterase-labeledavidin (25). 400 ml.Fractions (3 ml) were collectedand tested for enzymeactivity as PAGE and Immunoblots-300 ng of purified testis enzyme wereanadescribed below. Fractions within the central part of the activity peak lyzed under denaturing conditions on 10-15% Phast System gradient were pooled and submitted to a hydroxylapatite-ultrogelcolumn (30-ml electrophoresisgels (Pharmacia). Separation conditions werethose recbed volume; IBF) pre-equilibrated with 50 m Tris-HC1 buffer(pH 7.5), ommended by the supplier. Proteins were revealed by silver staining. 200 m KCI, 1m potassium phosphate (bufferA). Elution was carried Immunoblots were performed after electrotransfer onto nitrocelluout with a 200-ml linear gradient of buffer A to the same buffer with 250 lose membranes (0.45 pm; Schleicher & Schuell)using a Phast Transfer m potassium phosphate. The pool of active fractions was concentrated apparatus (Pharmacia). For these experiments, 340 ng of both purified and desalted on ultrafiltration cartridges (Centriflo membrane cones, brain and testis enzymes were loaded on the gel. Crude extracts were type CF-25, Amicon Corp.) and subjected to PAGE as described below. analyzed after theprecipitation step (see "Enzyme Preparation"); 2 and Endopeptidase Tissue Distribution-Tissues were rapidly dissected 7 pg of testis and brain extracts, respectively, were loadedon the gel. from Wistar male rats (135 g) and immediately frozen at -80 "C. Start- The polyclonal serum was used at a 1:2000 final dilution. Antigening with 5-9 g of frozen tissue (450 mg for adrenal glands), a partial antibody complexes were visualized using an anti-rabbit IgG alkaline purification of the endopeptidase was performed. The purification pro- phosphatase conjugate (Promega Biotec). cedure was followed to the DEAE-TrisacrylM column step (see "Enzyme I m m u ~ o ~ u o r e s c e n s tof e sadult Wistar rats (Iffa Credo) were Preparation"). Specific cleavage was checked using [Ala17,Ty?olS-28- fured with 4% paraformaldehyde in 0.1 M phosphate buffer (pH 7.4) for (10-20)-NHZpeptide substrate (see "Enzyme Assay" below).The same 6 h at 4 "C and washed overnight at 4 "C in the same buffer containing active elution fractions from the ion-exchange chromatography were 10% sucrose. Tissues were then impregnated in a graded series of suthen pooled, and specific activity was determined. crose solutions (10-18%) and embedded in Tissue Tek (Miles LaboratoEnzyme Assay--Routine enzyme assays were carried out by incubat- ries Inc.) before being frozen at -20 "C in the cryostat chamber. Sections ing 5 pg of [Ala17,Ty?01S-28-(10-20)-NHz (22,23) or 3.5 pg of dynorphin (5 pm) were mounted on glass slides, rinsed in 0.01 M phosphate-buffA with an aliquot of the enzyme preparation in a lOO-pl final volume of ered saline (pH 7.45), and incubated successively for30 min in 50 mM 250 m Tris-HC1 buffer(pH 7.5). The reaction was stopped by adding 10 NH,CI in phosphate-buffered saline and then in 0.2% gelatin in phospl of concentrated acetic acid and submitted to HPLC analysis (see phate-buffered saline before beingincubated with specific antibodies or "Determination of Substrate Specificity"). with normal rabbit serum as control for2 h at room temperature. They Determination of Optimum pH-% determine the effect of pH, buff- were washed for 30 min in phosphate-bufferedsaline and incubated for ers were prepared as follows: pH 5.5-7.75 = 250 m sodium phosphate 90 min with goat anti-rabbit IgG labeled with tetramethylrhodamine buffer; pH6.7-8.9 = 250 m Tris-HC1 buffer;and pH 8.6-10.6 = 250 m (Biosys). glycine/NaOH buffer. Activity was tested using [Ala17,Tyr20]S-28-(1020)-NHZas substrate(see "Enzyme Assay"). RESULTS Effect of Inhibitors and DiuaZent Cations-An aliquot of the enzyme preparation (post-hydroxylapatite pool) was preincubated with various Enzyme Preparation-The endoproteolytic activity was inhibitors or divalent cations for 30 min at 37 "C in 250 m Tris-HC1 monitored using [Ala17,W01S-28-(10-20)-NH2 (22, 23), which buffer (pH 7.5). The reaction was initiated by the addition of dynorphin A at a 6.5 w final concentration and carried out for 15 min at 37 "C. mimics the somatostatin 28 sequence around the maturation Results are expressed as the percent of remaining activity. 100% of dibasic site, as substrate. The results of a tissue distribution endopeptidase activity was measured by a control incubation without study are shown in Table I. The same active fractions pooled inhibitor or cation. from each tissue were found to be pure in terms of cleavage Reactivation by Divalent Cations after Inhibition byEDTA-After specificity after the DEAE-Trisacryl M purification step (see preincubation of the enzyme with 10 w EDTA for 30 minat 37 "C,the "Experimental Procedures"). As shown in Table I, all active substrate (dynorphinA at a 6.5 w final concentration) and thedivalent cation (at a 15 or 25 p~ final concentration) were added, and the reac- tissues exhibited a specific activity higher than that observed tion mixture was incubated for 30 min a t 37 "C. Results are expressed in brain cortex: 1.5- and 3.5-fold for heart and adrenalglands, as the percent of recovered activity. 100%activity was measured by a respectively, and >lO-fold for testes. This tissue was thus secontrol incubation with enzyme and substrate alone. lected as the enriched source for enzymepreparation.

In thisreport, we describe the complete purification and the functional properties of a novel metalloendopeptidase from rat testes that exhibits strict selectivity for cleavage of dibasic moieties in peptides at theamino terminus of an arginine residue. Antibodies directed against the testis enzyme have detected an immunologically identical protein in brain extracts. Additionally, the cellular distribution of the endopeptidase in rat testes has been studied.

Rat Testis Dibasic Selective Endopeptidase

2058

a

TABLEI

b

c

d

e

lFssue distribution of the endopeptidase activity For each tissue, specific activity was analyzed after the DEAE-‘kisacryl M purification step using [Ala17,~01S-28-(10-20)-NHz subas strate. The same active elution fractions were then pooled, and their specific activity was determined. For further details, see “Experimental Procedures.” Tissue

activity

Specific pg product l~lg protein I h

Cerebral cortex Adrenal gland Testes Heart Intestinal muscle

0.085 0.189 0.979 0.145 0.000

The endopeptidase from rat testes was purified to homogeneity in a four-step procedure, with the DEAE-Trisacryl M ion-exchange step constituting themost efficient step of purification. Starting with12 g of testes, 250 pg of purified enzyme were recovered. PAGE analyses of the post-hydroxylapatite active pool under denaturing (Fig. 1, lanea ) and nondenaturing conditions showed two separate components with apparentmolecular masses of about 110,000 and 140,000 Da, respectively. Both proteins displayed activity after electroelution from nondenaturing gels. Peptide maps, obtained after digestion by endolysine C (performed by J. d‘Alayer, Institut Pasteur, Paris), showed a large degree of similarity between the two proteins. Furthermore, corresponding peptides from the two proteolytic digests have been sequenced and have proved identical (data not shown). Sinceno NHz-terminal sequence informationcould be obtained from the two enzyme forms, no conclusion could be deduced about their interrelationship. Functional Characterization-Optimal activity of the enzyme was found at basic pH, with a maximum at pH 8.85, as shown in Fig. 2. A minor peak at pH 6.5 was observed, which could be accounted for by the ionization of an enzyme group involved in long-range effects on the catalytic activity. Alternatively, this could reflect the pK, of a substrate-ionizable group. The ability of known protease inhibitorsto block activity was tested withdynorphin A as substrate (Table 11).Similar results were obtained using [Ala17,~0]S-28-(10-20)-NHz(data not shown). The high sensitivity to metal chelators such as EDTA (74% inhibition at 1w) and 1,lO-phenanthroline (79% inhibition at 100 w)clearly demonstrates themetallopeptidase character of this enzyme. N-Ethylmaleimidealso completely blocked activity at low concentration, whereas iodoacetamide and dithiothreitol had a negligible effect. Interestingly, two aminopeptidase-specific reagents, amastatin and, to a lesser extent, bestatin, drastically inhibited activity (100% inhibition at 20 w and 87% inhibition at 50 w, respectively). Serine protease and carboxypeptidase inhibitors had no significant effect (Table 11). Both cation inhibition (Table 111)and reactivation (Table IV) profiles confirmed the metallopeptidase nature of the enzyme. The enzyme was sensitive to Zn2+ ions, which exhibited both the most inhibitory and the most reactivatory effects. The endopeptidase did not exhibit theCa2+ dependence of the KEX2 gene product family. The substrate specificity of the enzyme was studied using synthetic peptides mimicking dibasic maturation sites of prohormonalsequences and physiologically important peptides that can constitute hormonal precursorsfor further maturation steps. Peptides were first tested as potential substrates under the same incubation conditions for 1-15 h. The cleavage products were identified by their retention timescompared to standard peptides on reverse-phase HPLC and by amino acid composition. Kinetic parameters were determined for each peptide substrate (Table V).

LT

2 0

y

2

30 20 -e 14 -e

Ftc. 1. Electrophoretic analysis under denaturing conditions of rat testis enzyme and ita relationship to brain cortex endoprotease. Left panel, molecular mass markers (Pharmacia). Lune a , purified testis enzyme. The gel was stained using the silver staining technique (Pharmacia). Center panel, immunoblot of purified testis enzyme (lane b ) and brain cortex protease (lane c ) . Right panel, immunoblot of the corresponding crude extracts from testes (2 pg of protein) (lane d ) and brain cortex (7 pg) (lane e). 100-

80a,

0

m> ma

-

60-

0

c

40-

5

6

7

8

9

10

pH FIG.2. pH dependence profileof endopeptidase. The percentage of peptide substrate cleavage was determined under the standard enzyme assay conditions at various pH values (5.5-10.0)using three different buffers: 0.25 M sodium phosphate (O),0.25 M Tris-HC1 (A), and 0.25 M glycindNa0H ( ) (see ’Experimental Procedures”). The extent of cleavage at the optimum pH (8.85)was taken as 100%.

Cleavage strictly occurred at the amino terminus of arginine residues in dibasic sites.This specificity was independent of the nature of the basic doublet: somatostatin-related peptides were cleaved upstream of the Arg-Lys moiety; the major cleavage of opioid substrates occurred in between the Arg-Arg site; and the preproneurotensin-(l54-170)Lys-Arg and ANF-(l-28) Arg-Arg doublets wereboth cleaved in themiddle. For some of the opioid peptides, a second minor cleavage was seen upstream of the Arg-Arg site. The only observed exception was a-neoendorphin. However, the K, value determined for this peptide suggests that it is unlikely that it is a specific substrate for the enzyme. The endopeptidase did not exhibitany amino- or carboxypeptidase activity: [Ty?O]S-28-( 13-20)-NHz and dynorphin A-( 1-7) were not substrates. Additionally, the prosomatostatin monobasic site and the single arginine residue present in the peptides were not cleaved, reinforcing the conclusion of the selectivity for dibasic sites. The endopeptidase displays higher affinity for larger peptides. Indeed, comparison of the somatostatin peptide series and the dynorphin A-related peptides shows that lower K,,, values and highercleavage efficiencies were obtained withthe larger substrates: 43 1.1~for somatostatin 28 and 6.45 p.ffor dynorphin A. The two other “better” substrates of the enzyme are also long peptides: preproneurotensin-(l54-170) (K, = 17.3 w) and ANF-(1-28) ( K , = 6.25 w). The particularly high cleavage efficiency of dynorphin A-( 1-13) was due to an overestimation of the V,, value resulting from incomplete HPLC

2059

Rat Testis Dibasic Selective Endopeptidase TABLE I1 Inhibitor profile Enzyme (250 ng) was preincubated with the inhibitor in a total volume of 250 pl for 30 min a t 37 "C in 250 mM Tris-HC1 buffer(pH 7.5). The reaction was initiated by the addition of dynorphin A at a 6.5 w final concentration and was carried out for 15 min a t 37 "C. Results are expressed as the percent of activity remaining, taking the incubation control as 100% cleavage. Inhibitor

Activity

TABLEIV Cation reactivation profile Enzyme (420 ng) was preincubated for 30 min with 10 w EDTA at 37 "C in a volume of 250 pl. Dynorphin A (6.5 w final concentration) and the divalent cation (15 or 25 w final concentration) were then added, and the reaction mixture was incubated for 30 min at 37"C. Results are expressed as the percent of activity recovered, taking the incubation control as 100%cleavage. Cation %

% of control

100 control Activity EDTA Inhibition control 22 0 10 w CaCl, 26 1W EGTA 74 15 w 7 500 w 25 w 90 73 100 pM MgCL 1,lO-Phenanthroline 15 w 21 0 500 pm 25 w 43 COCl, 21 100 pM 87 Phosphoramidon, 10 p 15 w 74 120 Captopril, 10 1.1~ 25 w 74 MnCl, TPCKa 50 250 w 15 w 94 74 100 p M 25 w 114 cuso4 95 PMSF, 100 w 116 Leupeptin, 100 w 14 15 w 78 Dm, 100 w 25 w 18 Iodoacetamide, 1000 w 92 ZnSOl NEM 122 15 w 100 pM 0 25 w 90 63 10 w GEMSA, 100 w 100 rat testes and brain are shown in Fig. 1(lanes b-e). The same Amastatin 0 20 w two proteins and immunologically related forms of the endo78 10 w peptidase were found to be present in both testis and brain Bestatin extracts (lanes d and e). 13 50 w Partial characterization of the brain endopeptidase estab31 20 w 100 10 w lished its metalloprotease character andits cleavage specificity a TPCK, L-1-tosylamido-2-phenylethyl chloromethyl ketone; PMSF, for somatostatin-related peptides (data not shown). It can be phenylmethylsulfonyl fluoride; DTT, dithiothreitol; NEM, N-ethylma- assessed that the brain enzyme is catalytically and immunoleimide; GEMSA, (2-guanidinoethylmercapto)succinicacid. logically related to the testis endopeptidase. TABLE I11 Cation inhibition profile In these experiments, the same protocol as thatused in the inhibitor study (Table 11) was followed. Cations were used a t a 100 p final concentration. Results are expressed as the percent of activity remaining, taking the incubation control as 100% cleavage. Cations

Activity % of control

Ca2+(CaC1,) Mg2' (MgCIz) CU2' (CUSOI) C02' (COCl,) Mn2+(MnC1.J Zn2+(ZnS04)

112 105 44.5 36 57 11

separation of the NH2- and COOH-terminal fragments generated. The nature of the basic doublet does not seem to be a restrictive condition for enzymeactivity: the Lys-Arg doublet was not cleaved in preproenkephalin-(128-140) and pro-ocytocid neurophysin-(l-20), but was cleaved with high efficiency in preproneurotensin-(154-170). This peptide, which also contains an Arg-Arg moiety in the middle of the neurotensin sequence, was specifically cleaved at the prohormonal maturation site. These data suggest that secondary structure and/or specific subsite requirements are important for specific recognition or catalysis by the endopeptidase. Correlation between %tis and Cortex Endopeptidases "Polyclonal antibodies were raised against the purified testis enzyme preparations. Immunoblot analyses after SDS-PAGE of both the purified enzyme preparation and crude extracts from

Immunofluorescence-By immunofluorescence, the enzyme was localized within numerous seminiferous tubules. It was distributed in thecytoplasm of cells that corresponded to round (immature) or elongating (mature) spermatids. Depending on the stage of the seminiferous cycle, these immunoreactive cells were more or less abundant (Fig. 3, compare a-c with d). Control sections were completely devoid of immunolabeling (data not shown). DISCUSSION

In this report, we describe the complete purification and characterization of a novel metalloendopeptidase from rat testes that cleaves a t the amino terminus of arginine residues from basic maturation doublets. The pure testis enzyme is composed of by two active forms that exhibit high sequence homology according to peptide maps. The first hypothesis that could explain this observation is amino- or carboxyl-terminal proteolysis of the higher molecular mass form. This is supported by the presence of additional peptides in the endolysine C peptide map of the higher molecular mass form. A carboxyl-terminal autoprocessing event has been evoked for two other maturation proteases after theirexpression in heterologous cells, the Kex2 protein (26) and furin (27). We cannot exclude that thesmaller form is a degradation product. Both proteins have been shown to be functional, but comparison of their respective efficiency could not be achieved. Different post-translational modifications couldalso explain the size difference. Alternate splicing of the mRNA, particularly frequent in the testes, could account for this molecular mass heterogeneity as for PC4 (28). The presence of the same two related forms of the endopep-

Rat Testis Dibasic Selective Endopeptidase

2060

TABLE V Substrate dependence profile All peptides werechecked a s possible substrates by incubation at a 39 p~ final concentration withthe same quantity of enzyme for 1,3,and 15 h. The reaction mixtures were analyzed on a reverse-phase HPLC column. The kinetic parameters were determined from initial velocity measurements (duplicate points) a t various substrate concentrations and calculated from Hanes-Woolf plots by a linear regression program. The experiments were run twoto three times for each peptide. Values are given 210%.

uence

Peptide

Somatostatin peptides [T~?~]S-28-(13-20)-NH2 [Ala17,Ty1301S-28-(10-20)-NHZ [Ala'71S-28-(6-19)-NH2 [Are,Ala171S-28-(1-19)-NH2 Somatostatin 28

Prosomatostatin-(56-68)-NH2 Opioid peptides Dynorphin A Dynorphin A-(1-13) Dynorphin A41-10) Dynorphin A-(1-9) Dynorphin A-(1-7) Dynorphin B a-Neoendorphin BA"12P Preproenkephalin-(128-140) Miscellaneous neuropeptides Preproneurotensin-(154-170) ANF-(1-28)

pro-OT/Np-(l-2OP

Pseudo-k,,lK,,,

R PRE .1R PAMAPRE R RSANSNPAMAPRE R SANSNPAMAPRE R DEMRLELQRSANS-NH2 YGGFL YGGFL YGGFL YGGFL YGGFL YGGFL YGGFL YGGFM GGEVLA

K K K K K

AGAKNY-NH2 AGAKNY-NH, AGAKN-NH2 AGAKN-NH2 AGCKNFFWKTFTSC

NC,not cleaved.

e

bovine adrenal medulla dodecapeptide. pro-OT/Np-(l-20), pro-ocytocinIneurophysin4 1-20).

428

Pseudo-kc., pmo1lminlu.e.

p1lminlu.e.

NC" 400

0.93

NC 227 43

74.8 15.8

0.33 0.37

NC

R 1R RJR RJR RJR R R RJR R 1K RIR K R

IRPKLKWDNQ IRPKLK IRP IR

6.45 24 127 162

59.9 625 240 105

QFKWT YPL VGRPE YGGFM

209 974 103

101 92.2 365

9.3 26 1.89 0.65

NC

ENKPRRPYIL K 5 R ASYYY SL R J R SSCFGGRIDRIGAQSGLGCNSFRY CYIQNCPLGG K R AVLDLDVR

a

K,

0.48 0.095 3.54

NC 17.3 6.25

85.6 114

4.90 18.3

NC

* BA"12P,

FIG.3. Immunofluorescence of endopeptidase in testes. Shown are 5-pm thick sections of rat testesimmunocytochemically stained with antibodies against the endopeptidase. Thespecific immunofluorescence is distributed within seminiferous tubules(a