THE JOURNAL OF BIOLWICAL CHEMISTRY 0 1994 hy The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 269, No. 27, Issue of July 8, pp. 17899-17904,1994 Printed in U.S.A.

Purification, Characterization, and Localization of Follipsin, a Novel Serine Proteinase from the Fluid of Porcine Ovarian Follicles*

(Received for publication, March 15, 1994, and in revised form, April 11, 1994)

Takashi HamabataS, Hikari OkimuraS, Naoki Yokoyamas, Takayuki Takahashih, and Kenji Takahashi3 From the $L)iuision of Biological Sciences, Graduate School of Science, Hokkaido University, Sapporo 060, Japan and t he §Department of Biophysics a n d Biochemistry, Faculty of Science, University of Tokyo, Bunkyo-ku, Tokyo 113, J a p a n

Follipsin, an enzyme that accumulates in the follicu- lar fluid of porcine ovaries during follicular matura- tion, was purified to apparent homogeneity. The puri- fied enzyme consists of two different polypeptide chains having M, = 45,000 and 32,000 each, associated cova- lently. The enzyme activity was strongly inhibited by diisopropyl fluorophosphate, benzamidine, leupeptin, and antipain, indicating that follipsin is a serine pro- teinase. Using synthetic peptide substrates containing 4-methylcoumaryl-7-amide, follipsin was shown to pref- erentially hydrolyze Arg-X bonds but not Lys-X bonds. The NH,-terminal amino acid sequences of the 45- and 32-kDa polypeptides were highly homologous with those of the heavy and light chains, respectively, of human plasma kallikrein and human factor XI,. Immunological analyses and substrate specificity studies, together with other existing evidence, indicated that follipsin is dis- tinct from kallikrein and factor XI,, thus being a novel type of serine proteinase. Follipsin is immunohisto- chemically localized in follicular fluid as well as in stroma cells of porcine ovaries. The results strongly sug- gest that follipsin originates from interstitial cells of the ovarian stroma.

~~~~~~ ~ ~ ~~

The ovaries are female reproductive glands containing germ cells. The cells grow to maturity in individual follicles and are eventually liberated under hormonal controls. A characterist ic fea ture of matured ovarian follicles in most mammals is the accumulation of fluid in the follicular space. This follicular fluid is known to consist mainly of transudates of plasma, a l though it also contains products synthesized in and secreted f rom fol- licle cells (1). Previous studies on the protein composition of follicular fluid revealed the presence of various kinds of en- zymes in addi t ion to serum proteins. It is particularly interest- ing that the fluid contains proteolytic enzymes that may pos- sibly be involved in degradative changes leading to follicular rup tu re (2-4). The ovulatory process has thus far been consid- ered to be mediated by p lasmin generated from plasminogen by plasminogen activator, which can decrease the tensile strength of the follicle wall and degrade the basement membrane (5-8).

In the previous study (91, we found that the follicular fluid of porcine ovaries contains an enzyme that preferably hydrolyzes

research from the Ministry of Education and Culture, Japan, and a *This study was supported in part by grants-in-aid for scientific

research grant from Nissan Science Foundation. The costs of publica- tion of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solelv to indicate this fact.

Fax: 81-11-717-9394. 11 To whom correspondence should be addresseh. Tel.: 81-11-706-2748;

synthetic, arginine-containing peptide 4-methylcoumaryl-7- amide (MCA)’ substrates and that its activity increases several times as the follicles grow, thus suggesting its biological role in follicular maturation and/or ovulation. We have now purified the enzyme, hereaf ter referred to as follipsin, from the follicu- lar fluid and investigated its properties. The results showed that follipsin is a novel type of serine proteinase dist inct from any other endopept idases hitherto described. The present s tudy a l so demonst ra tes its localization in the interstitial cells of the s t roma of porcine ovary cortex.

EXPERIMENTAL PROCEDURES Materials-Fresh porcine ovaries were obtained from Teikoku Hor-

mone Manufacturing Co. (Tokyo, Japan). Synthetic peptide MCA sub- strates, E-64, leupeptin, antipain, bestatin, and pepstatin A were pur- chased from the Peptide Institute (Osaka, Japan). DEAE-cellulose (DE- 52) and CM-cellulose (CM-52) were obtained from Whatman. Benzamidine-Sepharose 6B and Sephacryl S-200 were from Pharmacia Biotech Inc. Fluorescamine, diisopropyl fluorophosphate, phenylmeth- anesulfonyl fluoride, N”-p-tosyl-L-lysine chloromethyl ketone, Nu-p-to- syl-L-phenylalanine chloromethyl ketone, p-chloromercuribenzoic acid, and o-phenanthroline were purchased from Sigma. Crude porcine plasma kallikrein was obtained from Canadian Bioclinical Ltd. (Canada). A mixture of molecular weight marker proteins for sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE) was from Bio-Rad. Bicinchoninic acid protein assay kit was from Pierce Chemical Co. Vectastain ABC kit for detection of antigens was obtained from Vector Laboratories, Inc. Rabbit anti-human plasma kallikrein antiserum was from Protogen AG (Switzerland), goat anti-human factor XI antiserum from Nordic Immunological Laboratories (Netherlands), and alkaline phosphatase-conjugated goat anti-mouse IgG polyclonal antibody from Immunotech S.A. Immusorbin was obtained from Wako Pure Chemical Industries, Ltd. Polyvinylidene difluoride transfer mem- brane was purchased from Millipore Corp. Other reagents used were of the highest grade available.

Enzyme Assuy-Enzyme activities toward peptide MCA substrates were assayed in 0.1 M “is-HC1 (pH 8.0) as described (10) except that the substrate concentration was raised from 50 to 100 PM. Enzyme activi- ties toward histones, bovine serum albumin, ovalbumin, and casein were assayed by the fluorescamine method as described (11).

Polyacrylamide Gel Electrophoresis-Slab gel electrophoresis in the presence and absence of SDS was performed by the method of Laemmli (12).

Protein Assuy-Protein was determined by the method of Smith et al. (13) using the bicinchoninic acid reagent.

NH,-terminal Amino Acid Sequence-The purified enzyme prepara- tion was subjected to SDS-PAGE under reducing conditions. The polypeptide bands were electroblotted onto polyvinylidene difluoride membrane and were analyzed directly for NH,-terminal sequences as described by Matsudaira (14) using an Applied Biosystems model 477A sequenator with an on-line model 120A phenylthiohydantoin-derivative analyzer.

The abbreviations used are: MCA, 4-methylcoumaryl-7-amide; PAGE, polyacrylamide gel electrophoresis; Boc, t-butyloxycarbonyl; Z, benzyloxycarbonyl; TBS, Tris-buffered saline.

17899

17900 Follipsin from Porcine Ovary Preparation of Antiserum-Specific antiserum against porcine ovary

follipsin was raised by injecting mice (B&B/c strain, 7 weeks old) with 100 pg each of the purified enzyme emulsified with Freund's complete adjuvant. Boosting was started 4 weeks later by injecting each at 2-week intervals with 100 pg ofthe antigen emulsified with incomplete Freund's adjuvant. The antisera obtained from the blood collected after three booster injections were found to specifically recognize the 32-kDa, but not 45-kDa, polypeptide band of follipsin.

W e s ~ e ~ n Blotting and Protein Detection-Proteins separated on SDS- PAGE gel were transferred to pol~nyl idene difluoride membrane us- ing the Towhin transfer buffer (15). The blotted membrane was incu- bated with anti-follipsin antiserum at 1:2000 dilution and subsequently with biotinylated anti-mouse IgG antibody. The membrane was then incubated with avidin conjugated with horseradish peroxidase and stained with hydrogen peroxide and diaminobenzidine, according to the manufacturer's instructions.

Inlmilnopreei~itat~~n-En~yme samples were mixed with various amounts of antiserum or control serum in 50 mM phosphate buffer (pH 7.5) containing 0.2 M NaCI and incubated at 4 "C for 15 h. One hundred pl of lmmusorbin (10% v/v, binding capacity of 4 mg o f human XgG/ml of suspension) was added, and the mixtures were incubated at 4 "C for 2 h. After centrifuging at 10,000 x g for 10 min, the superna- tants were assayed for activity.

Im,nunohistachemistry-Fresh porcine ovaries were cut into small blocks and fixed by immersion in Zamboni's fucative (16) for 6 h at room temperature. The fured blocks were dehydrated, cleared, embedded in paraffin, and cut into 5-pm sections. The sections were placed on clean slides, deparaffinized in xylene, and rehydrated through a graded alco- hol series. After rinsing in 0.1 M %is-HC1 (pH 7.5) containing 0.15 M NaCl (Tris-buffered saline (TBS)), the sections were incubated with 1.5% normal goat serum in TBS for 30 min and then with 1% anti- follipsin antiserum for 30 min. In control expriments, mouse normal serum or the supernatant of antiserum preabsorbed with follipsin was used instead of anti-follipsin antiserum. After several washes in TBS, the sections were incubated in TBS with 0.5% alkaline phosphat~se" con jug at^ goat anti-mouse IgG po~yclonal antibody for 30 min. After washing with TBS, the sections were stained for alkaline phosphatase by incubating in 0.1 M Tris-HC1 (pH 9.5) containing nitroblue tetrazo- lium (0.34 mg/ml), 5-bromo~-chloro-3-indolyl phosphate (0.17 mglml), 10 mM levamisole, and 50 m MgCI, for 30 min in the dark. The endo- genous alkaline phosphatase activity was quenched by the presence of levamisole during the color development (17). Slides were then washed in water and counterstained in Mayer's hematoxylin for 10 s and then washed in water again and mounted.

RESULTS Purification of Follipsin-All procedures were performed at

4 "C. Ovaries of 40 female pigs were carefully sliced with a sharp razor blade in 500 ml of cold phosphate-buffered saline. The materials were then centrifuged at 1,000 x g for 10 min to obtain the supernatant. The precipitated tissue was suspended in 250 ml of cold phosphate-buffered saline, and the suspension was centrifuged again at 1,000 x g for 10 min. Both superna- tants were combined and centrifuged again at 10,000 x g for 20 min to remove turbid materials. The resulting supernatant was referred to as "follicular fluid." To the follicular fluid fraction was added solid ammonium sulfate to 40% saturation, and the mixture was centrifuged at 10,000 x g for 20 min, The precipi- tate was dissolved in 250 ml of cold distilled water, and the sample was dialyzed extensively against 20 mM Tris-HC1 (pH 8.5). After removing insoluble materials by centrifugation, the clear supernatant was applied to a DE-52 column (3.2 x 30 cm) previously equilibrated with 20 I~IM Tris-HCl (pH 8.5). The retained materials were eluted with a linear gradient of NaCl concentration (Fig. la). The fractions having the enzyme activ- ity were pooled and the proteins were precipitated with ammo- nium sulfate at 80% saturation. After solution in 40 mf of cold distilled water, the sample was dialyzed against 4 liters of 50 mM sodium acetate (pH 6.0) overnight and applied to a CM-52 column (1.7 x 32 cm) that had been equilibrated with the same buffer. The retained proteins were eluted with a linear gradient of NaCl concentration (Fig. lb), and the active fractions were pooled. In order to concentrate the sample, it was subjected to

h

i v

E C 0 Eo N

c m

u a C

5 m

0 u) a

L"

d

Fraction number

FIG. 1. C ~ o m a ~ ~ p ~ i c p u ~ ~ ~ a t i o n of porcine ovary folfip-

the column (3.2 x 30 crn). The column was washed extensively and was sin. a, chromatography on a DE-52 column. The sample was applied to

eluted with a linear gradient of NaCl using 1 liter each of 20 R ~ M Tris-HC1 (pH 8.5) and the same buffer containing 0.2 M NaCI. Flow rate was 60 myh, and fractions of 10 ml were collected. The fractions were assayed for enzyme activity toward Boc-Gin-kg-kg-MCA. A ~ o r ~ z o ~ ~ a ~ 6csI. indicates the fractions pooled. b, chromatography on a CM-52 col- umn. The sample was applied to the column (1.7 x 32 em). The retained proteins were eluted with a linear gradient of NaCl using 250 ml each of 50 ma;f sodium acetate (pH 6.0) and the same buffer containing 0.2 M NaCl. Flow rate was 45 mllh, and fractions of 5 ml were collected. c, chromatography on a benzamidine-Sepharose 6B column. The sample was applied to the column (1 x 4.5 cm). After unretained proteins were washed away, the column was eluted with 50 mhp benzamidine in 0.1 M Tris-HC1 (pH 7.5) containing 0.2 M NaCI. Flow rate was 18 mYh, and fractions of 1 ml were collected. Enzyme activity of the pooled sample increased about 60% after dialysis, presumably due to removal of the inhibitor benzamidine.

CM-52 column chromatography again under conditions as above except that a column of smaller size and stepwise elution with 0.2 M NaCl were used. The concentrated enzyme sample, which was obtained in 50 w sodium acetate (pH 6.0) contain- ing 0.2 M NaCl, was mixed with 0.25 volume of 0.5 M Tris-HC1 (pH 7.5) containing 0.2 M NaCl and applied to a benzamidine- Sepharose column (1 x 4.5 cm) previously equilibrated with 0.1 M Tris-HCI (pH 7.5) containing 0.2 M NaCI. The enzyme was eluted with the same bufter containing 50 mx benzamidine (Fig. IC). The purification of follipsin is summarized in Table I. With this procedure, an approximately 12,400-fold puriftcation with about 42% yield was achieved. The enzyme obtained was pure as shown in the following section.

One note should be made on the apparent increase (about 60%) in total activity at the step of a~monium sulfate precipi- tation. It is not known at present whether this activation is caused by the removal of i ~ i b i t o ~ s ) that are complexed with follipsin in the crude fluid or by proteolytic conversion of an inactive, latent form to the active enzyme.

Purity and Molecular Weight-The purified enzyme showed a single polypeptide band (M, = 85,000) in SDS-PAGE under

Follipsin from Porcine Ovary TABLE I

Purification of folliosin from porcine ovarian follicular fluid

17901

protein Total enzyme

Total

activity Specific activity Yield Purification

mg nmol I min nmollminlmg protein % .fold Follicular fluid 11,220 5,927 0.528 100 1 Ammonium sulfate, 40% 5,033 9,670 1.92 163 3.6 DE-52 487 5,699 11.7 96.2 22.2 CM-52 43.4 3,225 74.3 54.4 14 1 Benzamidine-Sepharose 0.38 2.483 6,534 41.9 12,375

nonreducing conditions, whereas it showed two polypeptide bands (M, = 45,000 and M, = 32,000) in SDS-PAGE under reducing conditions (Fig. 2). The apparent molecular weight was estimated to be approximately 80,000 by gel filtration on Sephacryl S-200 (data not shown). It is concluded accordingly that the enzyme has a two-chain structure cross-linked by in- terchain disulfide bond(s).

The molecular weight of the enzyme was estimated to be approximately 350,000 in a preliminary study (9). This was obtained with the partially purified enzyme preparation by the electrophoretic method ofAndersson et al. (18) using a gradient gel without SDS. As suggested by its chromatographic behavior on ionic exchange columns, follipsin is probably a basic protein and should migrate rather slowly to the anode, thus deviating from the standard calibration curve. We believe that a molecu- lar weight of 80,000 for follipsin is correct.

Effects of Inhibitors-Effects of various proteinase inhibitors were examined (Table 11). Diisopropyl fluorophosphate, benz- amidine, leupeptin, and antipain strongly inhibited the activity of follipsin. These results clearly indicate that the enzyme is a serine proteinase.

Substrate Specificity toward Various MCA Derivatives of Peptides-Table I11 shows the activities toward various MCA- containing substrates. A common feature of the substrates hy- drolyzed by follipsin was the presence of an arginine residue at the P, position, although the hydrolytic rates were variable with different substrates. Little or no hydrolysis occurred when the substrates contained Lys-MCAbonds. In addition, the MCA derivative of a single arginine residue did not serve as a sub- strate of the enzyme.

Activity toward Protein Substrates-The activities of follip- sin toward denatured casein, bovine serum albumin, ovalbu- min, and histone were measured. The enzyme hydrolyzed his- tone most effectively. The values relative to that for histone (taken as 100%) were 18.5, and 3% for bovine serum albumin, ovalbumin, and casein, respectively.

NH,-terminal Amino Acid Sequences-The NH,-terminal amino acid sequences of 18 and 14 residues were determined for the 32- and 45-kDa polypeptides, respectively. As shown in Fig. 3, the sequence of the 32-kDa polypeptide was highly ho- mologous with the NH,-terminal sequences of the light chains of human plasma kallikrein (20) and factor XI, (21) and, to a lesser extent, with those of other serine proteinases (22-24). Similarly, the homology was observed between the NH,-termi- nal sequences of the 45-kDa polypeptide and the heavy chains of human plasma kallikrein (20) and factor XI, (21).

Immunological Characterization-The purified enzyme was immunoprecipitated with rabbit anti-human plasma kallikrein antiserum but not with goat anti-human factor XI, antiserum (Fig. 4), suggesting that follipsin shares structural homology with kallikrein, but not with factor XI,. The antiserum against porcine ovary follipsin raised in mice was capable of immuno- precipitating plasma kallikrein only when used in large amounts (Fig. 5).

Comparison with Plasma Kallikrein-In order to compare the chromatographic behaviors of porcine follipsin with plasma

4 5 - , p -45 - -32 31-

21- Y

FIG. 2. SDS-PAGE analysis of follipsin. The enzyme was electro- phoresed on a 10% polyacrylamide gel in the presence of SDS under reducing (middle lane) or nonreducing (right lane) conditions. Values shown are relative molecular mass (kDa) obtained from the mobilities of standard marker proteins (left lane). PME, P-mercaptoethanol.

TABLE I1 Effects of inhibitors on the activity

Inhibitor Concentration Inhibition

DFP" PMSFb Benzamidine TLCK' TPCKd p-Chloromercuribenzoic acid o-Phenanthroline

Leupeptin Antipain Bestatin PeDstatin

E-64

mM 1 1 1 0.1 0.25 0.2 1 0.1 0.02 0.02 0.02 0.02

%

95 21 87 23 9 8

10 15 93 88 5 2

" Diisopropyl fluorophosphate. Phenylmethanesulfonyl fluoride. N"-p-tosyl-L-lysine chloromethyl ketone. N"-p-tosyl-L-phenylalanine chloromethyl ketone.

kallikrein, partial purification of the kallikrein was conducted using a synthetic substrate, Z-Phe-Arg-MCA (25). Crude por- cine plasma kallikrein was fractionated sequentially on col- umns of DE-52, CM-52, and benzamidine-Sepharose 6B under conditions similar to those for follipsin. I t was found that kal- likrein was eluted from the DE-52 column at 0.25 M NaC1, whereas the enzyme did not adsorb on the CM-52 column. In addition, the kallikrein activity was only slightly retarded and no retained activity was detected in the affinity column chro- matography. Thus, the chromatographic behaviors of kallikrein were entirely different from those of follipsin.

The kallikrein sample thus prepared was tested for various types of proteinase inhibitors using four different MCA sub- strates: Boc-Gln-Arg-Arg-MCA, Z-Phe-Arg-MCA, Boc-Leu-Thr- Arg-MCA, and Boc-Phe-Ser-Arg-MCA. The inhibition profiles obtained with each substrate were basically the same (data not shown). In addition, gel filtration of the enzyme sample on a TSK 3000SW column gave rise to a single symmetric activity peak a t a position corresponding to approximately M, = 80,000 as assayed with Z-Phe-Arg-MCA (data not shown). The same

17902 Follipsin from Porcine Ovary

FOL45K bB$a FXIa-HC C V T LLKDTCFEG PKAL-HC CLTEL EN FFRG

FIG. 3. NH,-terminal sequences of the 32- and 45-kDa polypep- tides of follipsin and their homology to other serine proteinases. The NH,-terminal 18-residue sequence of the follipsin 32-kDa polypep- tide (FOL32K) was compared with those of the human plasma kal- likrein light chain ( P K 4 I “ C ) ( Z O ) , the human factor XI, light chain (FXla-LC) (21), bovine trypsin (TRP) (22), bovine chymotrypsin ( C H Y ) (231, and porcine elastase (ELA) (24), and the NH,-terminal 14-residue sequence of the 45-kDa polypeptide (FOL45K) with those of the human kallikrein heavy chain (PKAL-HC) (20) and the human factor XI, heavy chain (Fxla-HC) (21). The sequences are shown in one-letter codes, and identical amino acid residues are enclosed. The second residue of the 45-kDa polypeptide was not identified and is shown as X .

Antiserum added (PO

FIG. 4. Immunoprecipitation test with anti-human plasma kal- likrein antiserum and anti-human factor XI, antiserum. Aliquots (Boc-Gln-Arg-Arg-MCA-hydrolyzing activity, 0.6 nmol/min) of a purified follipsin preparation were incubated in 1 volume of 40 pl with the indicated amounts of rabbit anti-human plasma kallikrein antiserum (0) or goat anti-human factor XI, antiserum (x). The enzyme was also incubated with mouse anti-follipsin antiserum (0) prepared in the pres- ent study. The mixtures were further processed as described under “Experimental Procedures.” Note that the follipsin activity was par- tially immunoprecipitated by the anti-human kallikrein antiserum.

Antiserum added(p1)

FIG. 5. Immunoprecipitation test with anti-follipsin anti- serum. Indicated amounts of anti-follipsin antiserum were incubated in 1 volume of 40 p1 with purified follipsin (0) or partially purified porcine plasma kallikrein (0). The enzyme activities in all tubes were adjusted to 0.3 nmol/min using Boc-Gln-Arg-Arg-MCA. The mixtures were further treated as described under “Experimental Procedures.” The activities of the supernatants were assayed toward the same substrate.

results were obtained in assays for enzyme activities of the fractions toward Boc-Gln-Arg-Arg-MCA and Boc-Phe-Ser-Arg- MCA. Furthermore, the substrate specificity of the present en- zyme preparation examined with various MCA substrates is very similar to that reported for bovine plasma kallikrein (19).

TABLE I11 Enzyme activities toward various MCA substrates

Enzyme activities were determined at pH 8.0 with 0.1 mM substrates as described under “Experimental Procedures” and are expressed as percentage of activity toward Boc-Gln-Arg-Arg-MCA.

Substrate Follipsin kallikrein ‘yh$;:{ Plasma

(porcine)

Boc-Gln-Arg-Arg-MCA 100 lop 100 Boc-Leu-Lys-Arg-MCA 112 - 31 2-Phe-Arg-MCA 101 2,837 19 Boc-Gln-Gly-Arg-MCA 98.1 - 200 Boc-Gln-Ala-Arg-MCA 94.9 337 169 Boc-Leu-Thr-Arg-MCA 83.1 256 122 Boc-Glu(OBz1)-Ala-Arg-MCA‘ 71.8 - 614 Boc-Leu-Ser-Thr-Arg-MCA 27.8 393 79 Boc-Val-Pro-Arg-MCA 19.6 - 341 Boc-Phe-Ser-Arg-MCA 19.1 122 127

Boc-Val-Leu-Lys-MCA 4 Suc-Leu-Leu-Val-Tyr-MCA 0 0 - Bz-Arg-MCAd 0 118 2 Arg-MCA 0 0 -

a Cited from Ref. 19.

e OBzl, benzyl. -, not tested.

Bz, benzoyl.

Boc-Glu-Lys-Lys-MCA 2.5 - 0

- -

From these considerations, we judged that the enzyme prepa- ration derived from benzamidine-Sepharose column chroma- tography contained porcine plasma kallikrein with no serious contamination by other proteinases. The enzyme activities to- ward various MCA derivatives of peptides are shown in Table 111. The substrate specificity of porcine plasma kallikrein is apparently different from that of follipsin.

Absence of Follipsin in Porcine Plasma-Porcine plasma pro- teins were fractionated on a DE-52 column with a linear NaCl concentration gradient. No enzyme activity hydrolyzing Boc- Gln-kg-kg-MCA was detected at a position expected for fol- lipsin if present in the plasma. Instead, the enzyme activity was found at around 0.12 M NaCI. These active fractions were pooled and analyzed for immunoreactivity with mouse anti- porcine follipsin antiserum as described under “Experimental Procedures.” The antiserum failed to precipitate the enzyme activity. The results indicate that follipsin is not present in the plasma in an active state.

Localization of Follipsin in Porcine Ouary-Immunohisto- chemical examination of follipsin in ovaries showed that follip- sin is present in the follicular fluid, stroma, and blood vessels but not in granulosa cells, theca cells, and oocytes (Fig. 6). Control experiments performed with normal mouse serum or immunoneutralized antiserum showed no staining anywhere on the sections, clearly indicating that staining in follicular fluid and stroma was attributable to the anti-follipsin antibody.

DISCUSSION A number of proteins including enzymes are detected in the

follicular fluid of mammalian ovaries. They are thought to originate from either liver cells or the tissues of the ovary itself. Among the proteins of ovarian origin, enzymes known as in- tracellular proteins often occur in the fluid simply as a result of degeneration of follicle cells, whereas others are actively se- creted from the cells. The latter enzymes should be of biological importance and thus are worth investigating in detail. Our preliminary study suggested that follipsin is one such secretory enzyme found in the porcine ovary. In the present study we have purified the enzyme for detailed characterization.

Follipsin is a serine proteinase with a molecular weight of approximately 80,000 and consists of two different polypeptide chains ( M , = 45,000 and 32,000) linked covalently through di- sulfide bond(s). The molecular size, polypeptide chain arrange-

Follipsin from Porcine Ovary 17903

. ..

. . . .

. .

..

. ,

. ..

Fit;. 6. Immunohistochemical localization of follipsin in por- cine ovary. A, deep color stained with the 5-brorno-4-chloro-3-indolyl phosphatehitroblue tetrazolium system is shown in the follicular fluid (arrow), stroma (open arrow), and blood vessels (arrowhead). B , normal mouse serum was used instead of anti-follipsin antiserum. C, immu- noneutralized anti-follipsin antiserum was used instead of anti-follipsin antiserum. No positive staining is found in either B or C. Original magnification, x 80. Bar = 0.25 mm.

ment, and cleavage specificity for Arg-X bonds of the enzyme are similar to those of two well known blood-clotting protein- ases, plasma kallikrein and factor XI,. Striking homology was indeed observed for the NH,-terminal amino acid sequences of follipsin and human plasma kallikrein and factor XI,.

The differences between follipsin and porcine plasma kal- likrein are as follows. 1) The follipsin adsorbed on the DE-52 column was eluted at 0.04 M NaCl, but the plasma kallikrein was at a much higher concentration of NaCl. 2) Kallikrein was not retained on the CM-52 column under conditions that assure the binding of follipsin. These results strongly suggest that follipsin is more basic than plasma kallikrein. 3) Follipsin had affinity for the benzamidine-Sepharose column but kallikrein did not. Taking into consideration that benzamidine is a com- petitive inhibitor and binds to the active sites of the enzymes,

their different behaviors on the affinity columns possibly re- flect a difference in the microenvironment of the catalytic sites. This is further supported by the results of a substrate specific- ity study; porcine kallikrein exhibited the highest activity to- ward Z-Phe-Arg-MCA, Boc-Gln-Ala-Arg-MCA, Boc-Leu-Thr- Arg-MCA, and Boc-Leu-Ser-Thr-Arg-MCA were also well hydrolyzed by kallikrein in agreement with the previous find- ings of McRae et al. (26) and Kawabata et al. (19) that kal- likrein prefers aromatic residues (Phe or T r p ) as well as Thr in the P, position. In contrast, follipsin does not have such a preference, as shown in Table 111. Furthermore, it should be pointed out that Bz-Arg-MCA (where Bz is benzoyl), which is hydrolyzed by kallikrein to a considerable extent, is completely resistant to follipsin action. However, a dipeptide MCA sub- strate Z-Phe-Arg-MCA was found to be a substrate for follipsin, indicating that filling at least three binding sites (S2, SI, and S'l) seems to be a prerequisite for hydrolysis.

The substrate specificity of follipsin is also different from that of human factor XIt,. As shown in Table 111, relative activity values toward various MCA substrates vary over a wider range with factor XI, than with follipsin. Notable is the difference in the activities toward Boc-Glu(OBz1)-Ala-Arg-MCA (where OBzl is benzyl) and Boc-Val-Pro-Arg-MCA; in particular, the former, a substrate for factor XI, (19), is not the best substrate for follipsin.

It is well known that factor XI is a glycoprotein consisting of two identical polypeptide chains each of M , = 60,000 held to- gether by a disulfide bond(s). Upon activation to factor XI,, each subunit is cleaved at a specific peptide bond, generating a protein composed of two heavy chains and two light chains. The resulting four chains are also linked by disulfide bonds, and the molecular weight of the native factor XI, is 130,000 (27, 28). This is much greater than the estimated molecular weight (M, = 80,000) of native follipsin. The results, together with the distinct substrate specificity profiles discussed above, are con- sistent with the idea that follipsin is different from factor XI,. Absence of a reaction of follipsin with goat anti-human factor XI, antibody favors this idea.

Immunohistochemical examination demonstrated the local- ization of porcine follipsin in the fluid-filled antral cavity and the stroma of the ovarian cortex. The localization in the cavity is supported by our previous biochemical data (9). On the other hand, positive staining in the stroma was unexpected, because follicular fluid proteins of ovarian origin are generally thought to be synthesized and subsequently secreted by follicle-produc- ing cells such as granulosa cells or theca cells. The ovarian stroma contains clusters and cords of large polygonal epithe- lioid cells called the interstitial cells of the ovary. They are also presumed to have an active secretory function (29). Thus, it is possible that follipsin present in the follicular fluid originates from interstitial cells of the stroma of the ovarian cortex. The problem of how it reaches the follicular space from the site of synthesis remains to be answered, but the presence of intra- vascular materials, presumably the plasma, recognized by anti- follipsin antibody appears to indicate that the protein enters ovarian follicles via the circulatory system.

Follipsin has a two-chain structure and is thought to be synthesized as a single polypeptide chain precursor protein. Previously, we showed that follipsin activity is detected almost exclusively in the follicular fluid fraction and that little or no enzyme activity is detected in the tissue fraction of porcine ovary (9). On the other hand, the tissue-associated distribution of the protein was clearly demonstrated in the present study. This apparent conflict, however, can be explained such that the stroma cells contain inactive precursor follipsin, whereas fol- licular fluid contains the active enzyme. The absence of active

17904 Follipsin from Porcine Ovary

follipsin in the plasma strongly suggests that the precursor protein reaches the follicles and thereafter undergoes conver- sion to the active form. Studies on the mechanism of its proc- essing to the mature two-chain form are under way.

The present study clearly indicates that follipsin is a new protein that has not been reported before. The physiological role of this enzyme is not clear at present. Since the enzyme activity in the fluid progressively increases as follicles grow (9), it should be involved in the events closely relating to follicular maturation. The substrate specificity of the enzyme toward the Arg-X bond suggests that it may function as a processing or inactivation enzyme of some bioactive peptides. Indeed, mam- malian ovary follicles are known to be exposed to a number of peptide hormones. In order to elucidate its physiological role, the natural substrate(s) of the enzyme must be identified.

Acknowledgments-We wish to thank Drs. Takao Mori and Seiichiro Kawashima (Faculty of Science, The University of Tokyo) and Samuel H. Hori (Graduate School of Environmental Earth Science, Hokkaido University) for valuable suggestions and Yasuko Sakurai for her able technical assistance.

REFERENCES 1. McNatty, K. P. (1978) in The Vertebrate Ovary (Jones, R. E., ed) pp. 215-259,

2. Schochet, S. S. (1916) Anat. Rec. 10,447457 3. Espey, L. L. (1975) in Proteases and Biological Control (Reich, E., Rifkin, D. E.,

and Shaw, E., eds) pp. 767-776, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY

Plenum Press, New York

4. Espey, L. L., and Coons, P. J. (1976) Biol. Reprod. 14, 233-245 5. Beers, W. H. (1975) CeZE 6, 379-386 6. Beers, W. H., Strickland, S., and Reich, E. (1975) Cell 6,387-394

7. Canipari, R., and Strickland, S. (1985) J. Biol. Chem. 260,5121-5125 8. Liu, Y. X., Peng, X. R., and Ny, T. (1991) Eur. J . Biochem. 195,549-555 9. Takahashi, T., Tsuchiya, Y., Tamanoue, Y., Mori, T., Kawashima, S., and Ta-

10. Uchino, T., Sakurai, Y., Nishigai, M., Takahashi, T., Arakawa, H., Ikai, A,, and

11. Tamanoue, Y., Takahashi, T., and Takahashi, K. (1993) J. Biochem. (Tokyo)

12. Laemmli, U. K. (1970) Nature 227,680-685 13. Smith, P. K., Krohn, R. I., Hermanson, G. T., Mallia, A. K, Gartner, F. H.,

Provenzano, M. D., Fujimoto, E. K., Goeke, N. M., Olson, B. J., and Klenk, D. C. (1985) Anal. Biochem. 150,76435

kahashi, K. (1992) ZooZ. Sci. 9, 343347

Takahashi, K. (1993) J. Biol. Chem. 268, 527-533

113,229-235

14. Matsudaira, P. (1987) J. Bid. Chem. 262,10035-10038 15. Towbin, H., Staehelin, T., and Gordon, J. (1979) Proc. Natl. Acad. Sci. U. S. A.

16. Stefanini, M., De Martino, C., and Zamboni, L. (1967) Nature 216, 173-174 17. Borgers, M. (1973) J. Histochem. Cytochem. 21, 812-824 18. Andersson, L.-O., Borg, H., and Mikaelsson, M. (19721 FEBSLett. 20,199-202 19. Kawabata, S., Miura, T., Morita, T., Kato, H., Fujikawa, K., Iwanaga, S.,

Takada, K., Kimura, T., and Sakakibara, S. (1988) Eur J. Biochem. 172,

20. Chung, D. W., Fujikawa, K., McMullen, B., and Davie, E. W. (1986) Biochen- 17-25

21. Fujikawa, K., Chung, D. W., Hendrickson, L. E., and Davie, E. W. (19861 istry 25,2410-2417

76,4350-4354

Biochemistry 25, 2417-2424 22. Walsh, K. A,, and Neurath, H. (1964) Proc. Natl. Acad. Sci. U. S . A. 52,884-

23. Hartley, B. S. (1964) Nature 201, 1284-1287 24. Shotton, D. M., and Hartley, B. S. (1970) Nature 226, 802-806 25. Morita, T., Kato, H., Iwanaga, S., Takada, K. , Kimura, T., and Sakakibara, S.

26. McRae, B. J., Kurachi, K., Heimark, R. L., Fujikawa, K., Davie, E. W., and

27. Kurachi, K., and Davie, E. W. (1981) Methods Enzynol . 80,211-220 28. Barrett, A. J., and McDonald, J. K. (1980) in Mammalian Proteases (Barrett,

A. J., and McDonald, J. K., eds) Vol. 1, pp. 65-68, Academic Press, Orlando, FL

29. Peters, H., and McNatty, K. P. (1980) in The Ouary (Peters, H., and McNatty, K. P., eds) pp. 29-31, Granada Publishing, London

889

(1977) J. Biochem. (Tokyo) 82, 1495-1498

Powers, J. C. (1981) Biochemistry 20,7196-7206