Protease-nexin I as an androgen-dependent secretory - Europe PMC

0 downloads 0 Views 3MB Size Report
Feb 11, 1993 - presence of uPA inhibitors (PAls) in the tissue extracts. In two samples ..... and of the start codon (ATG) are indicated. in fibroblasts, glia and ...
The EMBO Journal vol. 1 2 no. 5 pp. 1 871 - 1878, 1 993

Protease-nexin I as an androgen-dependent secretory product of the murine seminal vesicle

Jean-Dominique Vassalli, Joaquin Huarte, Domenico Bosco, Andr6-Pascal Sappino', Nahid Sappino1, Alfredo Velardi1, Annelise Wohlwend, Henrik ErnO3, Denis Monard3 and Dominique BeIin2 Institute of Histology and Embryology, IDivision of Oncohematology and 2Department of Pathology, University of Geneva Medical School, 1211 Geneva 4 and 3Friedrich Miescher Institute, 4002 Basel, Switzerland Communicated by D.Monard

A search for inhibitors of urokinase-type plasminogen activator (uPA) in the male and female murine genital tracts revealed high levels of a uPA ligand in the seminal vesicle. This ligand is functionally, biochemically and immunologically indistinguishable from protease-nexin I (PN-I), a serpin ligand of thrombin and uPA previously detected only in mesenchymal cells and astrocytes. A survey of murine tissues indicates that PN-I mRNA is most abundant in seminal vesicles, where it represents 0.2-0.4% of the mRNAs. PN-I is synthesized in the epithelium of the seminal vesicle, as determined by in situ hybridization, and is secreted in the lumen of the gland. PN-I levels are much lower in immature animals, and strongly decreased upon castration. Testosterone treatment of castrated males rapidly restores PN-I mRNA levels, indicating that PN-I gene expression is under androgen control. Key words: protease inhibitors/protease-nexin I/seminal vesicles/urokinase-type plasminogen activator

Introduction Protease inhibitors are abundant in plasma and in most extracellular fluids, where they are believed to play an important part in limiting, both spatially and temporally, the activity of proteolytic enzymes. A number of protease inhibitors have been identified in male genital tract organs and seminal plasma of mammals (Fritz et al., 1975; Zaneveld et al., 1975). Those that have been best characterized belong to the Kazal and the Kunitz families of low molecular weight antiproteases (Laskowski and Kato, 1980), and they inhibit the catalytic activities of tryptic enzymes, such as acrosin, leucocyte elastase or plasmin (Mills et al., 1987; Fritz, 1988; Lai et al., 1991; Moritz et al., 1991). The major plasma protease inhibitors, i.e. oal-protease inhibitor, a2-antiplasmin and antithrombin III, belong to a different class of antiproteases, the serpins (Huber and Carrell, 1989). Recently, the synthesis of the serpin protein C inhibitor (PCI) by the human male reproductive system has been described (Espania et al., 1991; Laurell et al., 1992). The production of other serpins by male genital tract organs has not yet been reported.

A number of serine proteases are synthesized in the male genital tract, including acrosin, the major protease stored

in the head of spermatozoa. Proteases are also secreted in the lumen of the male genital tract (Zaneveld et al., 1975; Watt et al., 1986; Clements et al., 1988), and the synthesis of the urokinase-type plasminogen activator (uPA) by the male genital tract has been documented: in mice and rats, uPA is synthesized by epithelial cells of the cauda epididymis, the vas deferens and the seminal vesicle (Larsson et al., 1984; Huarte et al., 1987). uPA is a serine protease that converts plasminogen into plasmin, a neutral protease of broad specificity (Vassalli et al., 1991). Plasminogen, which is abundant in plasma and other extracellular fluids, is also synthesized in seminiferous tubules (Saksela and Vihko, 1986), suggesting that it is present throughout the lumen of the male genital tract. While the precise role of uPA in the biology of the reproductive system has not been elucidated, the enzyme appears to be secreted around the time of ejaculation, and to bind, possibly through interaction with a specific cell surface receptor, to the head region of the murine spermatozoon (Huarte et al., 1987). The presence of uPA in the epithelium of male genital organs, together with the frequent concomitant production of PAs and their inhibitors (PAIs) by a variety of cell types (Busso et al., 1987; Wohlwend et al., 1987; Pepper et al., 1990), prompted a search for PA inhibitors in these organs. Serpins form covalent complexes with their target protease(s), and they can therefore be revealed by a highly sensitive functional assay (Baker et al., 1980). Using radiolabelled uPA, we have detected high levels of a uPA ligand in extracts of adult mouse and rat seminal vesicles. We have identified this ligand as protease nexin I (PN-I), a serpin released by human fibroblasts (Baker et al., 1980) and identical to glia-derived nexin (GDN) (Gloor et al., 1986; Sommer et al., 1987; McGrogan et al., 1988), first detected as a factor promoting neurite outgrowth in vitro (Guenther et al., 1985) that is found in different structures of the nervous system (Reinhard et al., 1988; Meier et al., 1989). PN-I reacts with and inhibits thrombin, uPA, tissuetype PA (tPA), trypsin and plasmin (Baker et al., 1980; Eaton and Baker, 1983; Guenther et al., 1985; Stone et al., 1987). In addition, we have shown that the abundance of PN-I protein and mnRNA in the mouse seminal vesicle is under androgen control. These results thus define a novel and major site of synthesis of PN-I, previously identified as a product of fibroblasts and glial cells; they also indicate that steroid hormones may play an important part in controlling the production of PN-I.

Results A ligand for uPA in the mouse seminal vesicle

To search for serpin-class PA inhibitors, 125I-labelled uPA was added to extracts of different male and female mouse genital tract tissues, and the fate of the radioactive protein

1871

J.-D.Vassalli et al.

I/ 1)

iT 9 I l() TI

1:: .i;.

a..

80

,

60

*_ CO

Amak.mmm-

Awl

'IWW ..mIw IN

AMMLI

a) .k

-W

.

40

I::.e

20 Fig. 1. Detection of uPA ligands in mouse genital tract tissues. Detergent extracts of indicated organs were prepared and 10 ,ig of protein per tissue extract were mixed with 1 ng of [1251]uPA. After 1 h at 4°C, the samples were analysed by SDS-PAGE and autoradiography of the gel. Lane 1, no tissue extract; lanes 2 and 3, ovary and uterus from a virgin mouse; lanes 4 and 5, ovary and uterus from a term pregnant mouse; lane 6, term placenta; lane 7, testis; lane 8, epididymis; lane 9, vas deferens; lane 10, seminal vesicle; lane 11, coagulating gland; lane 12, prostate, samples 7- 12 were from an adult mouse. Unreacted [1251]uPA migrates as a single band of apparent Mr 33 000, while SDS-resistant complexes formed between the enzyme and ligands in the tissue extracts migrate with higher apparent Mr. The position of migration of protein standards of indicated Mr is shown on the right. ;4

;

)

t)

$3

(

f

Mr

^

x

i

(3~~~~~~1

~1'

O

0

0

.5

1

2

5

10

50

100

seminal vesicle extract (p1) Fig. 3. Neurite promoting activity of partially purified mouse seminal vesicle extract. Ability to promote neurite growth was tested as previously described (Schurch Rathgeb and Monard, 1978). A semipurified uPA ligand from seminal vesicle (pooled fractions eluted from heparin-Sepharose around 0.5 M NaCl) was tested at different dilutions in cultures of mouse neuroblastoma cells. After 4 h of incubation at 37°C, cells showing three or more neurites of length equivalent to at least the diameter of the cell body were scored. The proportion of cells exhibiting neurites (positive cells) versus the amount of semi-purified seminal vesicle extract is represented in the stippled columns, open columns: buffer control.

placenta (Feinberg et al., 1989), we identified the uPA ligand in the extract from murine term placenta as PAI- I (data not shown). The uPA ligand abundant in seminal vesicle extracts did not react with these antibodies (see below), and further experiments were performed for its identification.

..~~~~~~~~~~~~~~~~~~~~~~~~

PA

Fig.

2. The

thrombin.

(Mr

mouse

seminal vesicle contains

['1251]uPA

(Mr

34 000) (lanes 4-6)

33 or

000) (lanes

a

ligand

-3),

[1251]chymotrypsin

for uPA and

[1251]thrombin (Mr

25 000) (lanes

(lanes 1. 4 and 7), to seminal vesicle extract (lanes 2, 5 and 8) or to partially purified rat gliomaderived PN-I (lanes 3. 6 and 9). After 1 h at 4°C. samples were analysed as in Figure 1. 7-9)

were

added

to

buffer

a

mouse

determined by SDS-PAGE and autoradiography (Figure 1). In several cases, we observed that a fraction of the labelled uPA migrated with an apparent molecular size larger than that of the Mr 33 000 free enzyme (lane 1). In view of the known ability of uPA to form SDS-resistant complexes with different serpin-class antiproteases (Baker et al., 1980; Vassalli et al., 1984), these results indicate the presence of uPA inhibitors (PAls) in the tissue extracts. In two samples, placenta (lane 6) and seminal vesicle (lane 10), the amount of radiolabelled complexes was particularly striking. The complex formed with the extract from a term placenta migrated according to an apparent Mr of 72 000, and that formed with the seminal vesicle extract to an apparent Mr of -78 000. This difference in Mr, which is most probably due to a difference in the molecular size of the inhibitor component of the complexes, suggested that term placenta and seminal vesicle contained a different PAI. Using specific antibodies directed against PAI- and PAI-2, two uPA inhibitors previously detected in human was

-

1872

Mouse seminal vesicle contains PN-1 In addition to PAI- I and PAI-2, a third serpin that has high affinity for uPA is PN-I. PN-I forms 1: 1 complexes with thrombin, plasmin, trypsin, tPA and uPA. To determine whether the seminal vesicle uPA ligand is related to PN-J, functional and immunological tests were performed. Incubation of a seminal vesicle extract with 125I-labelled thrombin resulted in the formation of an SDS-resistant complex with an apparent Mr of -80 000 (Figure 2, lane 5), which comigrated with a complex formed with purified rat PN-I (Guenther et al., 1985) (lane 6). Similarly, incubation of [1251]uPA with the seminal vesicle extract (lane 2) or purified rat PN-I (lane 3) yielded complexes of identical migration. Neither sample contained a ligand reacting with 1251-labelled chymotrypsin (lanes 8 and 9). In addition, preincubation of the seminal vesicle extract with a 10-fold molar excess of unlabelled thrombin prevented complex formation with [1251]uPA, and preincubation with an excess of unlabelled uPA prevented complex formation with [1251]thrombin (data not shown). Thus, like PN-I, the uPA ligand in seminal vesicle extract is also a ligand for thrombin. PN-I has a high affinity for heparin (Baker et al., 1980; Guenther et al., 1985; Stone et al., 1987). Likewise, the uPA ligand in seminal vesicle could be quantitatively adsorbed on heparin - Sepharose, and eluted using a salt gradient at 0.5 M NaCl (not shown). Heparin Sepharosepurified seminal vesicle extract was also tested for its capacity to induce neurite formation in cultures of neuroblastoma cells (Schurch Rathgeb and Monard, 1978), a property of PN-I that is related to its capacity to inhibit thrombin-like proteases -

-

Protease-nexin I in murine seminal vesicle

I

:2

fi

)

;

I

.

Table I. PN-I mRNA levels in murine tissues

High Intermediate

Low

Fig. 4. Immunological characterization of the uPA ligand in seminal vesicle extract. Aliquots of a seminal vesicle extract were incubated with [125I]uPA and analysed before (lane 2) or after immunoabsorption with anti-rat PN-I (lane 3), anti-bovine PAI-I (lane 5), anti-human PAI-2 (lane 6), or irrelevant antibodies (lane 4). Immune complexes were eluted from the following Saureus pellets: anti-rat PN-I (lane 7), anti-bovine PAI-I (lane 9), anti-human PAI-2 (lane 10), or irrelevant antibodies (lane 8). Lane 1, unreacted [125I]uPA. 1

-E

f

>

-

4P

_

.>] 2

_

C

-_

Reproductive

Non-reproductive

Seminal vesicle (100%) Cauda epididymis (20%) Ovary (20%) Prostate (5 %) Placenta 18.5 days (4%) Uterus (3.5%) Vas deferens (2 %)

Bone (16%) Brain (11%)

Caput epididymis (1 %) Oviduct (1 %) Testis (0. 3 %)

Not detectable

PN-I mRNA levels were determined by RNase protection using a 32plabelled mouse PN-I cRNA probe. In each experiment a sample of seminal vesicle RNA was included as standard. Gels were exposed to storage screens and quantified using a Phosphorimager (Molecular Dynamics). Results are expressed as % of the abundance in adult seminal vesicle, which contains 40-80 pg of PN-I mRNA/Ag total RNA.

1

ww

-_

_-w

m-

Lung (8%) Stomach (4%) Spleen (3%) Heart (3 %) Kidney (3%) Intestine (1.5%) Skin (1 %) Thymus (0.7%) Bladder (0.5 %) Muscle (0.2%) Colon (0.2%) Pancreas (0.1%) Liver

?)

ol

-("

H

() I ( I

) . -o_ 1 t3

IiI

2f (:.O .t.;

;,; .1s. 1- t. II;

Ad'iU t.

Fig. 5. Quantification of PN-I content in seminal vesicles. Total seminal vesicle extract from adult and juvenile mice, tissue from an adult gland that had been emptied of its secretion, and secretion obtained by gently milking an adult gland, were prepared and diluted to the same concentration of total proteins (10 ug/sample); in the seminal vesicle preparation used in this experiment, protein in the secretion accounted for approximately three-quarters of the total protein of the gland. [1251]uPA was added to dilutions of the samples, and the samples were processed as for Figure 1.

(Monard et al., 1983; Gurwitz and Cunningham, 1988); a concentration-dependent effect of the partially purified material was observed (Figure 3). Finally, antibodies directed against three serpins that react with uPA, i.e. PAI-1, PAI-2 and PN-I, were tested for their capacity to immunoprecipitate complexes formed by addition of [1251]uPA to seminal vesicle extracts. Addition of anti-rat PN-I antibody resulted in complete adsorption of the [125I] uPA-ligand complex (Figure 4, lane 3), which was recovered in the immune pellet (ane 7); anti-PAI-l and anti-PAI-2 antibodies did not react with the complex Oanes 5 and 6). Thus, according to all criteria tested, the uPA ligand described here is similar to PN-I. The mouse seminal vesicle is, to our knowledge, the first genital tract tissue in which this protein has been detected in vivo. Analysis of extracts of rat seminal vesicles revealed the presence of two ligands forming complexes with [1251]uPA; one of these was identified as PN-I on the basis of its reactivity with [125I]thrombin, binding to heparin-Sepharose, reactivity

,-AA

I0(

10(

Fig. 6. RNase protection analysis of PN-I mRNA in male genital tract Total RNA from testis (lane 3, 10 jg), caput epididymis (lanes 4 and 5, 10 1tg), cauda epididymis (lanes 6 and 7, 10 Ag), seminal vesicle (lanes 8 and 9, 1 ,ug of tissue RNA with 9 14g of tRNA) and prostate (lanes 10 and 11, 10 ,ug) was hybridized to a 310 nt 32plabelled mouse PN-I cRNA probe (lane 1); tRNA (10 ,.g) was used to verify the complete digestion of the unhybridized probe (lane 2). After digestion with RNase A, the hybrids were electrophoresed on a 6% polyacrylamide-urea gel. organs.

with anti-rat PN-I antibodies, and capacity to induce neurite formation in cultured neuroblastoma cells (data not shown). PN-l is secreted by the adult mouse seminal vesicle Cultured human fibroblasts and rat glioma cells release PN-I into the medium (Baker et al., 1980; Guenther et al., 1985). To determine whether PN-I is present in the seminal vesicle secretory product, the amounts of PN-I in an extract of

glandular tissue (Figure 5, lanes 5 and 6) and in the liquid secretion (lanes 7 and 8) were compared with that in an unfractionated extract (lanes 2-4). Densitometric scanning of the Mr 78 000 [1251]uPA-PN-I complex indicated that the amount of PN-I per mg total protein in the secretion was 3-fold higher than that in the tissue; overall, we calculated that some 90% of the inhibitor in a total glandular extract is in the secretory liquid. In the same experiment, an unfractionated extract of seminal vesicle from a 5 week-old juvenile mouse was also analysed; a densitometric evaluation of the amount of [125I]uPA-PN-I complex indicated that -

1873

J.-D.Vassalli at a/.

A

sv VDL-

.f( ~~~~~~~~~~~~~~~~~~~~~...' .1ii; ...... . . .

';i!!~ ~ ~~ ~

;~ ~ ~~ ~~~ ~~ ~~~ ~~ ~~ ~~ ~~ ~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~. ..'..'....2

Fig. 7. Detection of PN-I mRNA by in situ hybridization. 32P-labelled (A and B) and 3H-labelled (C and D) PN-I cRNAs were hybridized to frozen sections from seminal vesicle and vas deferens, and PN-I mRNA was revealed by autoradiography. A and C are bright-field micrographs; B is the autoradiograph of a section corresponding to A; D is the dark-field micrograph of section C. SV = seminal vesicle, VD = vas deferens, L =

lumen, E

=

epithelium, S

=

stroma.

the concentration (w/w) of PN-I in the immature gland (lanes 9 and 10) was -20-fold less than that in the gland of an adult mouse (lanes 2-4). By cutting and counting the radioactivity in the relevant regions of electrophoretic gels, it was also possible to quantify PN-I in total seminal vesicle extracts. For three different adult mice, we determined the amount of PN-I to be 5.8, 12.1 and 6.4 jig per gland, or 0.6, 1.2 and 0.8 ,tg per mg total protein, respectively. Thus, PN-I amounts to

1874

-0.1 % of seminal vesicle proteins, and is not one of the major secretory proteins that constitute the structural elements of the copulatory plug. PN-I mRNA in murine tissues Total RNAs from different organs of the male mouse genital tract were analysed for PN-I mRNA by Northern blot hybridization using a rat PN-I cRNA probe. A single major transcript, of a size comparable to that of rat PN-I mRNA

Protease-nexin I in murine seminal vesicle

MOUSE

AGTGCAGTGGTTGCACGGGAGTGCGGGCTGCACGCGTCACCGTCACCGCCGCCTGTCCCCCACCGCCGCGCAGCG ****************

RAT

****

**********

********

***********

CTAGTGCAGTGGTTGCACCGGAGCGCGGGCTGCAAGCGTCACC------ACCGCCTGTCCCCCACCGCCGCG-----

V

MOUSE

CCGATCTCCCTCCCGGTTTCGGCCGCCACCTGGGGATCCAAGCGAGGACGGGCTGTCCTTGTTGGAAGGAACCATG

RAT

CCGACCTCCTTCGCCGCTCCGGCCGCCACCTGGGGATCCAAGCGAGGACGGGCTGTCCTTGTTGGAAGGAACCATG

****

****

**

*

*

*

*********************************************************

I

Fig. 8. Alignment of the 5' unstranslated region of PN-I mRNA in mouse seminal vesicle and rat glial cells. The murine cDNA was isolated by RT/PCR as described in Materials and methods. The nucleotide sequence of the rat mRNA is derived from genomic clones, and primer extension and RNase protection analysis of glial cells mRNA (H.Ern0 and D.Monard, manuscript in preparation). The positions of the first intron (arrowheads) and of the start codon (ATG) are indicated.

in fibroblasts, glia and brain, was revealed in seminal vesicle and cauda epididymis RNA (not shown). Using two oligonucleotides corresponding to regions conserved between rat and human PN-I (Sommer et al., 1987), a 1 kb DNA fragment was obtained by PCR amplification of murine seminal vesicle mRNA. The sequence of this fragment is respectively 93 and 86% identical to those of the rat and human PN-I cDNAs. Using this probe in RNase protection assays, we have measured the PN-I mRNA content in several murine tissues (Table I and Figure 6). High levels of PN-I mRNA were observed in RNAs from seminal vesicle and cauda epididymis (Figure 6, lanes 6-9). Lower levels of PN-I mRNA were observed in RNAs from testis (lane 3), caput epididymis (lanes 4 and 5) and prostate (lanes 10 and 11). This analysis confirmed that the seminal vesicle is a major site of PN-I gene expression in the adult mouse. To localize the cellular sites of PN-I synthesis in the male mouse genital tract, in situ hybridization studies were performed using the mouse PN-I cRNA probe. Specific hybridization was detected in the epithelium of the seminal vesicle (Figure 7, panels B and D) and the cauda epididymis (data not shown) but not in the vas deferens (Figure 7, panel B); the sub-epithelial connective tissue of the seminal vesicle did not contain detectable levels of PN-I mRNA. The structure of the human (McGrogan et al., 1990) and rat (H.Ern0 and D.Monard, unpublished) PN-I genes have been recently reported. In glial cells and fibroblasts, the 5' untranslated region of PN-I mRNA is encoded by two exons separated by at least 12 kbp. It was therefore possible that the PN-I mRNA expressed in epithelial cells is directed by a separate promoter. The 5' portion of the PN-I mRNA in murine seminal vesicle was isolated by reverse transcription and 5 '-anchored PCR. A comparison of the 5' untranslated sequence of mouse PN-I mRNA in seminal vesicle with that of rat PN-I mRNA in glial cells is shown in Figure 8. The high degree of identity between these sequences strongly suggests that the same promoter directs PN-I expression in glial cells and seminal vesicle epithelial cells.

Testosterone-induced accumulation of PN-I The lower levels of PN-I in seminal vesicles of young animals suggested that PN-I gene expression may be under androgen control. In agreement with this possibility, we observed that the relative concentration of PN-I in the seminal vesicle was markedly decreased after castration (Figure 9). This decrease was more rapid and more marked in the tissue itself (lanes 2-4) than in the secretory liquid (lanes 5-7). RNA was extracted from seminal vesicles after castration, and at different times after administration of testosterone to castrated animals. The yield of total RNA recovered per

1

2 2)

;(5

7

MrxlO -~94

iv

.- 67

4t3

.11 ..0.1IJ FiI ssi

>.ec.reLion

Fig. 9. Castration-induced decrease in seminal vesicle PN-I content. Tissue (lanes 2-4) and secretion (lanes 5-7) were obtained from control mice (lanes 2 and 5), and from mice 1 week (lanes 3 and 6) or 2 weeks (lanes 4 and 7) after castration. Samples were diluted to the same concentration of total protein, [1251]uPA was added and the samples were processed as for Figure 1. Lane 1: unreacted [1251]uPA. 250 a)

200

EL.. 0 0

150

c

a) =3

100

CZ

0

0o

50 0

0 hour

12 36 ; testosterone treatment

110

Fig. 10. Quantification of total RNA, PN-I mRNA and uPA mRNA in seminal vesicles of castrated mice following testosterone injection. Total RNA was measured by UV absorbance at 260 nm. The RNA yields for both animals from each time point were within 10-20% of the average shown in the figure. PN-I mRNA and uPA mRNA content were estimated by laser densitometry of autoradiograms of Northern blots. The values for PN-I mRNA were obtained from several exposures of the Northern blot shown in Figure 11. Values in control animals (non-castrated and non-treated) were given a value of 100%. White columns, total RNA; black, PN-I mRNA; dotted, uPA mRNA.

gland was markedly reduced after castration, returned to values around control levels after 2 days of testosterone treatment, and further increased under testosterone stimulation (Figure 10, open columns). Changes in PN-I -

1875

J.-D.Vassalli et al. e"

E` j

ANOL Alik.

v

;,

sm.

i.w

:... (

12h as6t td

st. ( ) ";t .

Contr

Ca st~raled(

Fig. 11. Northern blot analysis of PN-I mRNA in seminal vesicle of castrated mice, and following testosterone injection. Seminal vesicles were dissected from two control mice (lanes 1 and 2), two castrated mice (lanes 3 and 4); and six castrated mice injected daily with testosterone starting 2 weeks after castration (lanes 5-10). Organs from testosterone-treated animals were dissected 12 h (lanes 5 and 6), 36 h (lanes 7 and 8) or 110 h (lanes 9 and 10) after the first hormone injection. Total RNA was extracted from both glands of individual mice and 10 yg of RNA was loaded on each lane. A long exposure time was chosen to visualize the residual PN-I mRNA content in castrated mice.

mRNA abundance were expressed as relative amount per of total RNA (Figure 10, filled columns), and were determined by densitometric analysis after Northern blot hybridization (Figure 11). Three weeks after castration, PN-I mRNA abundance was reduced at least 10-fold; 12 h after injecting testosterone into castrated animals, the abundance of PN-I mRNA had returned to levels comparable to those in non-castrated animals. By contrast, the relative abundance of the mRNA for uPA, another secretory product of the mouse seminal vesicle (Huarte et al., 1987), was only slightly decreased upon castration, and was not affected by testosterone treatment (Figure 10, stippled columns). Taken together, these results suggest that the absolute and relative amounts of PN-I mRNA are highly sensitive indices of the response of the seminal vesicle to the androgen status of the animal.

/.g

Discussion Using a sensitive assay for serpins with specificity for uPA, we have tested a number of organ extracts from male and female mice for the presence of PAls. The seminal vesicle proved to be the organ with the highest concentration of a functionally detectable PAI. This seminal vesicle serpin is indistinguishable from PN-I: (i) it reacts with thrombin and uPA; (ii) it binds to heparin; (iii) it induces neurite formation in cultured neuroblastoma cells, and (iv) it is immunologically related to rat PN-I. Furthermore, using a rat PN-I cRNA probe, as well as RT/PCR cloning and sequencing, we have demonstrated the presence of PN-I mRNA in the mouse seminal vesicle. Taken together, these results provide strong evidence that the mouse seminal vesicle PAI is PN-I. Consistent with the high PN-I content of this gland, PN-I mRNA was abundant in the seminal vesicle. By contrast, in the caudal part of the epididymis and in the ovary, we also found high levels of PN-I mRNA, but only little functional PN-I in tissue extracts. This suggests a posttranscriptional difference in PN-I expression between the seminal vesicle and these other organs. For instance, PN-I

1876

may be secreted in all tissues, but stored only in the seminal vesicle, thus accounting for the accumulation of high levels of the protein in this gland. Another possibility is that PN-I reacts with a target protease in the other organs, since the functional assay of PAIs only detects unreacted inhibitor. In accord with the first hypothesis, we found that PN-I mRNA was present in epithelial cells of the seminal vesicle, and that most of the PN-I in seminal vesicle extracts could be separated from the bulk of the tissue by gently milking the gland. We conclude that PN-I is a secretory product of the seminal vesicle that accumulates in the lumen of the gland. PN-I was first identified as a secretory product of cultured human foreskin fibroblasts (Baker et al., 1980). In separate studies, a rat glioma-derived 43 kDa protein, which had been identified on the basis of its activity in promoting neurite outgrowth in mouse neuroblastoma cells, was shown to be the rat equivalent of human PN-I (Gloor et al., 1986; Sommer et al., 1987). More recently, PN-I has been identified in human platelets, smooth and striated muscle, rat brain astrocytes, regenerating rat peripheral nerve and rat olfactory system (Rosenblatt et al., 1987; Reinhard et al., 1988; Gronke et al., 1989; Meier et al., 1989; Festoff et al., 1990); in the latter two cases, PN-I probably originates primarily from the non-neuronal component of the tissues. The present work indicates that PN-I is more broadly distributed than previously suspected, and should thus lead to a further exploration of the possible function of PN-I in various tissues expressing this gene. Although PN-I is a powerful antiprotease, it is not yet clear what function it plays in vivo. A receptor for PN-I- protease complexes has been described; it binds the PN-I moiety of such complexes, and this leads to their rapid internalization and degradation (Low et al., 1981). Thus, PN-I has a number of properties in common with other antiproteases. The neurite-promoting activity of PN-I is related to the antithrombin function of the molecule (Monard et al., 1983; Gurwitz and Cunningham, 1988), and it is likely that the biological effects of PN-I in other tissues such as the male genital tract are also due to protease inhibition. It should be noted here that PN-I is an Arg-serpin with a relatively broad specificity: it inhibits thrombin, trypsin and plasmin in addition to PAs; the possible inhibitory activity of PN-I towards other trypsin-like proteases, including acrosin, remains to be tested. We suggest that PN-I may play a part in controlling proteolysis in the lumen of the seminal vesicle, or in semen. Maintenance of the copulatory plug, for instance, may require that the high levels of uPA synthesized by the vas deferens and the seminal vesicle, and present in semen and on ejaculated spermatozoa, be subjected to a tight inhibitory control, possibly through the activity of PN-I. In addition, PCI, a serpin distinct from PN-I but which is also an inhibitor of uPA, tPA and thrombin, has recently been identified in human seminal plasma and seminal vesicles (Espafia et al., 1991; Laurell et al., 1992); whether PN-I is present in the human genital tract, or whether PCI may play in humans a role similar to that of PN-I in rodents, is at present not known. Recent studies have shown that the synthesis of many members of the serpin family is under hormonal control. For instance, the production of PAI-1 and PAI-2 is modulated by a variety of hormones and growth factors (Andreasen et al., 1990). There is also abundant evidence

Protease-nexin I in murine seminal vesicle

for estrogen, progesterone and glucocorticoid regulation of other serpin genes. As for PN-I itself, one study has reported on the modulation of its production, which was shown to be increased in human foreskin fibroblasts under the influence of epidermal growth factor and of the protein kinase C activator phorbol myristate acetate (Eaton and

Baker, 1983).

Castration caused a large absolute and relative decrease in seminal vesicle PN-I mRNA and protein levels, and testosterone injection into castrated animals resulted in a rapid increase in the relative abundance of PN-I mRNA, which had returned to control levels within 12 h. Similar results have been described for other androgen-dependent genes (Mills et al., 1987; Schifman et al., 1988). In contrast, restoration of the normal mRNA level for the androgendependent secretory protein IV of mouse seminal vesicle required 4 days of testosterone injection in castrated animals (Chen et al., 1987). Thus, PN-I production in the mouse seminal vesicle is highly sensitive to androgenic regulation, a conclusion that is reinforced by the observation that PN-I concentration in the gland of an immature mouse was markedly lower than in adult animals. By contrast, the relative abundance of uPA mRNA, which is also present in epithelial cells of the seminal vesicle (Huarte et al., 1987), was only slightly decreased by castration and was not significantly affected by hormone replacement, indicating that it is probably not under androgen control. As yet only limited information is available on the structure of the PN-I gene (McGrogan et al., 1990), and further studies will be required to determine whether it contains an androgenresponsive element that could allow direct regulation of transcription by the androgen receptor (Beato, 1989). Alternatively, PN-I transcription could be controlled by a trans-acting factor, itself under transcriptional regulation by androgens. Also, the possibility that PN-I mRNA processing or stability might be affected by androgens should be considered (Higgins and Hemingway, 1991). The rodent seminal vesicle has long been known to be exquisitely sensitive to the androgenic status of the animals, and a number of its secretory products have been shown to be under testosterone control (Ostrowski et al., 1982; Brooks et al., 1986; Chen et al., 1987; Mills et al., 1987; Schifman et al., 1988; Harris et al., 1990; Higgins and Hemingway, 1991). A comparison of the genes coding for these different products, which now include PN-I, should yield valuable information regarding the mechanism of androgen regulation in the male genital tract. The expression of PN-I may share another property with some of the androgen-dependent proteases (Clements et al., 1988) and protease inhibitors (Mills et al., 1987) secreted by the seminal vesicle in that PN-I is also produced by other cell types, in which its regulation is likely to differ.

Materials and methods Materials NMRI mice (BRL, Basel) were used in this study. Thrombin from human plasma and chymotrypsin were from Sigma. uPA was generously provided by Serono (Denens, Switzerland). Anti-rat PN-I IgG has been described (Reinhard et al., 1988). Rabbit antiserum to PAI-1 from bovine aortic endothelial cells was generously given by Dr D.J.Loskutoff; these antibodies also recognize human (WohWwend et al., 1987) and murine (A.Wohlwend, unpublished) PAI-1. Affinity-purified rabbit IgG raised against PAI-2 from human placenta was kindly provided by Dr T.C.Wun; these antibodies recognize both human and murine PAI-2 (Wohlwend et al., 1987).

Detection of enzyme -inhibitor complexes Radiolabelling of uPA, thrombin and chymotrypsin using lodogen (Pierce Chemical Co, Rockford, IL) and NaI25I (Amersham Ltd, Amersham, UK), and detection of enzyme-inhibitor complexes by SDS-PAGE and autoradiography were performed as described by Vassaili et al. (1984) and Wohlwend et al. (1987). Protein extracts were prepared by homogenizing tissues in a loose fitting Dounce in 0.1 M Tris-HCI pH 8.1, 0.25% Triton X-100. After 5 min centrifugation at 12 000 g, protein concentration in the supernatants was determined as described by Bradford (1976), using BSA as a standard. The extracts were kept frozen at -20°C until use. To separate the tissue and secretion from seminal vesicles, the glands were gently milked and washed in PBS prior to homogenization. The gland content was collected and rapidly diluted in 0.5 ml of PBS, centrifuged for 5 min at 12 000 g and stored at -20°C; insoluble material upon thawing was removed by centrifugation.

Affinity chromatography 2 ml of protein extract of the seminal vesicle secretion (four glands) were loaded on 1 mil of heparin-Sepharose (Pharmacia), previously equilibrated with 0.2 M NaCl in 0.02 M sodium phosphate buffer (Na/PO4), pH 7. After extensive washing with the same buffer, bound proteins were eluted with a linear NaCl gradient (0.2-1 M NaCl in 0.02 M Na/PO4 pH 7). Successive 0.4 ml fractions were tested for the presence of uPA ligand as described above.

Immunoprecipitation of the uPA ligand 10 01 samples of protein extract from seminal vesicle secretion were incubated with 5 pd (2 ng) of [1251]uPA for 1 h at 4°C, then mixed with 4 sll of antirat PN-I IgG (6 mg/ml), 2 A1 of anti-PAI-I antisenum, 2 p1 of affinity-purified anti-PAI-2 IgG (0.6 mg/ml), or 8 1l of irrelevant IgG (1.35 mg/ml). After a 2 h incubation on ice, antigen-antibody complexes were precipitated using Staphylococcus aureus (Vassalli et ad., 1984), and uPA-inhibitor complexes were analysed by SDS-PAGE and autoradiography. RNA analysis Total RNA was isolated by centrifugation over CsCl cushions, and Northern blot hybridizations were performed as described by Busso et al. (1986) and Huarte et al. (1987). To obtain a cRNA probe for mouse PN-I, a 328 bp PstI fragment of a rat PN-I cDNA (Sommer et al., 1987) was subcloned in pGEMI. The plasmid was linearized with EcoRI and transcribed with SP6 RNA polymerase (Busso et al., 1986). The blots were hybridized at 58°C, and the three stringency washes were done at 760C. RNase protection assays and in situ hybridizations were performed as described by Belin et al. (1989) and Sappino et al. (1991). A 1 kbp murine PN-I probe, obtained by PCR amplification of seminal vesicle cDNA using two oligonucleotides corresponding to regions conserved between rat and human PN-I (Sommer et al., 1987), was cloned into pGEM-3Z and transcribed with SP6 RNA polymerase; the probe corresponds to the sequence coding for 11e42 to Ser366 (Sommer et al., 1987). The mouse PN-I probe was used to screen a mouse Balb/c genomic library in XEMBL3. A phage insert was shown to contain the second exon of the mouse gene, which includes 22 nt of the 5' untranslated region and codes for the first 86 amino acids of the protein. To isolate the 5' portion of PN-I mRNA, poly(A)+ RNA from seminal vesicle was reverse transcribed with an oligo corresponding to Arg46-Val53. The cDNA was dC-tailed and amplified with the same oligo and an anchor primer (5' RACE kit, Gibco-BRL). The PCR product was restricted, cloned in pUC18 and sequenced. The partial sequences of murine PN-I mRNA will appear in the EMBL nucleotide sequnece database (accession numbers X70296 and X70946).

Androgen regulation of PN-1 Three-month-old NMRI males were anaesthetized by an intraperitoneal injection of 0.4 mil of a 2.5 % tribromethanol solution, and through a single ventral incision both testes were dissected after ligature of blood vessels and ductus efferens. For the testosterone induction of PN-I, castrated animals were injected subcutaneously 2 weeks after the operation with a daily dose of 0.35 mg of testosterone propionate in corn oil (Chen et al., 1987). For each time point seminal vesicles from two animals were dissected and total RNA extracted.

Acknowledgements We thank D.Ducrest, C.Combepine and M.Khoshbeen for expert technical assistance, and B.Favri, J.P.Gerber and J.C.Rumbeli for photographic work. This work was supported by grants from the Swiss National Fund for Scientific Research (nos 31-30294.90 and 32-29289.90).

1877

J.-D.Vassalli et al.

References Andreasen,P.A., Georg,B., Lund,L.R., Riccio,A. and Stacey,S.N. (1990) Mol. Cell Endocrinol., 68, 1-19. Baker,J.B., Low,D.A., Simmer,R.L. and Cunningham,D.D. (1980) Cell, 21, 37-45. Beato,M. (1989) Cell, 56, 335-344. Belin,D., Wohlwend,A., Schleuning,W.D., Kruithof,E.K. and Vassalli,J.D. (1989) EMBO J., 8, 3287-3294. Bradford,M.M. (1976) Anal. Biochem., 72, 248-254. Brooks,D.E., Means,A.R., Wright,E.J., Singh,S.P. and Tiver,K.K. (1986) J. Biol. Chem., 261, 4956-4961. Busso,N., Belin,D., Failly-Crepin,C. and Vassalli,J.-D. (1986) J. Biol.

Chem., 261, 9309-9315. Busso,N., Belin,D., Failly-Crepin,C. and Vassalli,J.-D. (1987) Cancer Res.,

47, 364-370. Chen,Y.H., Pentecost,B.T., McLachlan,J.A. and Teng,C.T. (1987) Mol. Endocrinol., 1, 707-716. Clements,J.A., Matheson,B.A., Wines,D.R., Brady,J.M., MacDonald,R.J. and Funder,J.W. (1988) J. Biol. Chem., 263, 16132-16137. Eaton,D.L. and Baker,J.B. (1983) J. Cell Biol., 97, 323-328. Espania,F., Gilabert,J., Estelles,A., Romeu,A., Aznar,J. and Cabo,A. (1991) 7hromb. Res., 64, 309-320. Feinberg,R.F., Kao,L.C., Haimowitz,J.E., Queenan,J.T.,Jr, Wun,T.-C., Strauss,J.F. and Kliman,H.J. (1989) Lab. Invest., 61, 20-26. Festoff,B.W., Rao,J.S., Rayford,A. and Hantai,D. (1990) J. Cell Physiol., 144, 272-279. Fritz,H. (1988) Biol. Chem. Hoppe. Seyler, 369, 79-82. Fritz,H., Schiessler,H., Schill,W.B., Tschesche,H., Heimburger,N. and Wallner,O. (1975) In Reich,E., Riflin,D.B. and Shaw,E. (eds), Proteases and Biological Control. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, pp. 737-766. Gloor,S.M., Odink,K., Guenther,J., Nick,H. and Monard,D. (1986) Cell, 47, 687-693. Gronke,R.S., Knauer,D.J., Veeraraghavan,S. and Baker,J.B. (1989) Blood, 73, 472-478. Guenther,J., Nick,H. and Monard,D. (1985) EMBO J., 4, 1963-1966. Gurwitz,D. and Cunningham,D.D. (1988) Proc. Natl. Acad. Sci. USA, 85, 3440-3444. Harris,S.E., Harris,M.A., Johnson,C.M., Bean,M.F., Dodd,J.G., Matusik,R.J., Carr,S.A. and Crabb,J.W. (1990) J. Biol. Chem., 265, 9896-9903. Higgins,S.J. and Hemingway,A.L. (1991) Mol. Cell Endocrinol., 111, 55-61. Huarte,J., Belin,D., Bosco,D., Sappino,A.P. and Vassalli,J.-D. (1987) J.

Cell Biol., 104, 1281-1289. Huber,R. and Carrell,R.W. (1989) Biochemistry, 28, 8951-8966. Lai,M.L., Chen,S.W. and Chen,Y.H. (1991) Arch. Biochem. Biophys., 290, 265-271. Larsson,L.I., Skriver,L., Nielsen,L.S., Grondahl-Hansen,J., Kristensen,P. and Dano,K. (1984) J. Cell Biol., 98, 894-903. Laskowski,M.,Jr and Kato,I. (1980) Annu. Rev. Biochem., 49, 593 -626. Laurell,M., Christensson,A., Abrahamsson,P.-A., Stenflo,J. and Lilja,H. (1992) J. Clin. Invest., 89, 1094-1101. Low,D.A., Baker,J.B., Koonce,W.C. and Cunningham,D.D. (1981) Proc. Natl. Acad. Sci. USA, 78, 2340-2344. McGrogan,M., Kennedy,J., Li,M.P., Hsu,C., Scott,R.W., Simonsen,C.C. and Baker,J.B. (1988) BiolTechnol., 6, 172-177. McGrogan,M., Kennedy,J., Golini,F., Ashton,N., Dunn,F., Bell,K., Tate,E., Scott,R.W. and Simonsen,C.C. (1990) In Festoff,B.W. (ed.), Serine Proteases and their Serpin Inhibitors in the Nervous System. Plenum Press, New York, pp. 147-161. Meier,R., Spreyer,P., Ortmann,R., Harel,A. and Monard,D. (1989) Nature, 342, 548-550. Mills,J.S., Needham,M. and Parker,M.G. (1987) EMBO J., 6, 3711-3717. Monard,D., Niday,E., Limat,A. and Solomon,F. (1983) Prog. Brain Res., 58, 359-364. Moritz,A., Lilja,H. and Fink,E. (1991) FEBS Lett., 278, 127 -130. Ostrowski,M.C., Kistler,M.K. and Kistler,W.S. (1982) Biochemistry, 21, 3525-3529. Pepper,M.S., Belin,D., Montesano,R., Orci,L. and Vassalli,J.-D. (1990) J. Cell Biol., 111, 743-755. Reinhard,E., Meier,R., Halfter,W., Rovelli,G.F. and Monard,D. (1988) Neuron, 1, 387-394. Rosenblatt,D.E., Cotman,C.W., Nieto Sampedro,M., Rowe,J.W. and

Knauer,D.J. (1987) Brain Res., 415, 40-48. Saksela,O. and Vihko,K.K. (1986) FEBS Lett., 204, 193-197.

1878

Sappino,A.P., Huarte,J., Vassalli,J.-D. and Belin,D. (1991) J. Clin. Invest., 87, 962-970. Schifman,A.L., Mansson,P.E., Carter,D.B., Yamada,K., Harris,M.M. and Harris,S.E. (1988) Mol. Cell Endocrinol., 59, 57-65. Schurch Rathgeb,Y. and Monard,D. (1978) Nature, 273, 308-309. Sommer,J., Gloor,S.M., Rovelli,G.F., Hofsteenge,J., Nick,H., Meier,R. and Monard,D. (1987) Biochemistry, 26, 6407-6410. Stone,S.R., Nick,H., Hofsteenge,J. and Monard,D. (1987) Arch. Biochem. Biophys., 252, 237-244. Vassalli,J.-D., Dayer,J.M., Wohlwend,A. and Belin,D. (1984) J. Exp. Med., 159, 1653-1668. Vassalli,J.-D., Sappino,A.P. and Belin,D. (1991) J. Clin. Invest., 88, 1067-1072. Watt,K.W., Lee,P.J., M'Timkulu,T., Chan,W.P. and Loor,R. (1986) Proc. Natl. Acad. Sci. USA, 83, 3166-3170. Wohlwend,A., Belin,D. and Vassalli,J.-D. (1987) J. Immunol., 139, 1278-1284. Zaneveld,L.J., Polakoski,K.L. and Schumacher,G.F. (1975) In Reich,E., Rifkin,D.B. and Shaw,E. (eds), Proteases and Biological Control. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, pp. 683-706. Received on December 23, 1992; revised on February 11, 1993