High Expression of Bovine Glutathione S-Transferase (GSTA1, GSTA2 ...

0013-7227/99/$03.00/0 Endocrinology Copyright © 1999 by The Endocrine Society

Vol. 140, No. 8 Printed in U.S.A.

High Expression of Bovine a Glutathione S-Transferase (GSTA1, GSTA2) Subunits Is Mainly Associated with Steroidogenically Active Cells and Regulated by Gonadotropins in Bovine Ovarian Follicles* ˆ LE ´ , JEAN SIROIS, JEAN-FRANC FLORA RABAHI, SOPHIE BRU ¸ OIS BECKERS, DAVID W. SILVERSIDES, AND JACQUES G. LUSSIER Centre de recherche en reproduction animale (F.R., S.B., J.S., D.W.S., J.G.L.), Faculte´ de me´decine ve´te´rinaire, Universite´ de Montre´al, St-Hyacinthe, Que´bec, J2S 7C6, Canada; and Faculte´ de me´decine ve´te´rinaire (J.-F.B.), Universite´ de Lie`ge, B4000 Sart Tilman, Lie`ge, Belgium ABSTRACT We have previously shown that a major group of 28 –30 kDa proteins decreases after the LH surge in bovine granulosa cells (GC). In the present study, we have characterized two proteins in this group in search of factors that may intervene in folliculogenesis and oocyte maturation. Polyclonal antibodies raised against 28 kDa or 29 kDa bovine GC proteins were used to screen a complementary DNA (cDNA) expression library. This resulted in the characterization of two isoenzyme subunits for a class glutathione S-transferase, named bGSTA1 and bGSTA2. Both bGSTA1 (25.4 kDa, pI 8.9; 791 bp cDNA; GenBank Accession No. BTU49179) and bGSTA2 (25.6 kDa, pI 7.2; 959 bp cDNA; GenBank Accession No. AF027386) have 222 amino acids. The deduced amino acid sequences were compared and showed 82% (bGSTA1) and 74% (bGSTA2) identity to human GSTA1, whereas bGSTA1 and bGSTA2 are 81% identical to each other. The bGSTA2 represents a novel GSTA subunit because it harbors a specific 16 amino acid sequence not found in any other species and GST classes. Northern blots showed that bGSTA1 and bGSTA2 are coexpressed and are tissue specific with single transcripts of 1.2 kb and

G

RANULOSA cells constitute a heterogeneous cell population within an ovarian follicle. They have divergent morphological and functional properties depending on their location and consist of mural/antral (within the wall of the follicle) or cumulus (surrounding the oocyte) granulosa cells. Granulosa cells influence oocyte maturation throughout follicular development, and at the time of ovulation, normal oocyte maturation depends on signal transmission between cumulus cells and the oocyte. The importance of the cumulus cells for the final maturation and developmental competence of immature oocytes has been well established (1–3). The relationship that exists in vivo between the mural granulosa cells and the cumulus-oocyte complexes (COCs) Received November 19, 1998. Address all correspondence and requests for reprints to: Centre de recherche en reproduction animale, Faculte´ de me´decine ve´te´rinaire, Universite´ de Montre´al, P.O. Box 5000, St-Hyacinthe, Que´bec, Canada, J2S 7C6. E-mail: [email protected]. * This work was supported by the Natural Sciences and Engineering Research Council of Canada (NSERC), the Fonds de la recherche scientifique me´dicale (FRSM) of Belgium, the Fonds pour la Formation de Chercheurs et l’Aide a` la Recherche (FCAR) of Que´bec. Dr. F. Rabahi was supported by the Agence Francophone des Universite´s pour l’Enseignement Supe´rieur et la Recherche (AUPELF-UREF).

1.4 kb, respectively for bGSTA1 and bGSTA2. The messenger RNA (mRNA) were detected in GC, corpus luteum, adrenal gland, testis, liver, lung, thyroid, kidney and cotyledon, and the relative abundance of their mRNA varied. Ratios of bGSTA1/bGSTA2 mRNA vary between tisssues, indicating that expression of these genes is controlled differently. Immunohistochemistry observations revealed that expression of GSTA is cell specific, being associated with GC and theca cells, small luteal cells, Leydig cells, hepatocytes, adrenal cortex, specific chromaffin cells in the adrenal medulla, renal proximal convoluted tubular cells, and Clara cells in the bronchioles. Studies in vivo showed that levels of mRNA for bGSTA1 were elevated in follicular wall of preovulatory follicles before hCG treatment, but decreased by 77% 12 h after hCG injection. However, in FSH stimulated preovulatory follicles, the decrease in mRNA for both GSTAs was only 21% at 24 h following hCG injection. We concluded that bGSTA1 and bGSTA2 expression is tissue- and cell-specific, is associated with steroidogenically active cells, and is hormonally regulated by gonadotropins in the bovine ovarian follicle. (Endocrinology 140: 3507– 3517, 1999)

has been studied throughout the preovulatory period in cattle (4). Three major protein groups of approximately 76 kDa, 56 –58 kDa, and 28 –30 kDa were identified by de novo 35Smetabolic labeling and SDS-PAGE gel analysis. The pattern of proteins synthesized in mural/antral granulosa cells was shown to be modified in response to the LH surge. Rabahi et al. (1991) have shown by SDS-PAGE analysis that the de novo synthesized 28 –30 kDa protein group was decreased relative to total granulosa cell proteins after the LH surge. Thus, we have pursued the characterization of these proteins synthesized by granulosa cells that could modulate follicular development and/or COC maturation in the bovine species. In the present study, we have partially purified and raised polyclonal antibodies against two proteins contributing to the 28 kDa granulosa cell protein group that were initially named p28 and p29. We report herein their characterization by complementary DNA (cDNA) immunoscreening of a luteal cell expression library and reveal that these cDNAs encode for two uncharacterized bovine a glutathione S-transferase subunits (GSTA). We named these subunits bGSTA1 and bGSTA2, based on the proposed human nomenclature (5). The GSTs are a multigene family of related proteins that have been divided into five evolutionary classes of a, m, p,

3507

3508

GSTA EXPRESSION IN STEROIDOGENICALLY ACTIVE CELLS

s, and u on the basis of their biochemical and immunological characteristics. Within each class, GST isoenzymes were characterized and their number varies between species (6 – 8). These five classes of GST were reported to be localized in the cytoplasm, whereas the cell membrane associated GST, the microsomial GST and the C4 leukotriene synthetase, which differs in structure to the cytosolic GST, were classified differently. The described biological actions of GSTs are to provide protection against cellular oxidative stress through neutralization of a wide range of hydrophobic and electrophilic endogenous compounds or xenobiotics by catalyzing their conjugation to reduced glutathione. Their biological functions are extended to include binding protein of steroids (initially described as ligandin), bilirubin, carcinogens, organic anions; and peroxidase and [/d5-ketosteroid isomerase enzymatic activities (6, 8, 9). Materials and Methods Reagents All chemical reagents used were obtained from Sigma Chemical Co. (St. Louis, MO) if not otherwise stated.

Partial purification of the p28 and p29 and polyclonal antibody production Follicles, obtained from cycling cows killed at a local slaughterhouse, were immediately transported to the laboratory in PBS (0.6 mm KH2PO4, 4.4 mm Na2HPO4, 0.15 m NaCl, pH 7.4) on ice. Granulosa cells were recovered after follicular puncture and washed in PBS by four centrifugations (500 3 g, 10 min, 4 C). Cells were frozen, lyophylized, then resuspended in water (30 C) and centrifuged (50,000 3 g, 1 h, 4 C) to collect the supernatant soluble proteins. Granulosa cell protein extracts were precipitated by 25% saturated (NH4)2SO4 (2 to 4 h at 4 C), then centrifuged (7,000 3 g, 1 h, 4 C). The supernatant was chromatographed on a phenyl Sepharose CL4B column (Amersham Pharmacia Biotech, Baie d’Urfe´, Que´bec, Canada) previously equilibrated with glycine buffer (20 mm, pH 9) with 25% saturated (NH4)2SO4. The column was developed in four steps with glycine buffer of decreasing ionic strength buffer (25%, 12%, 0% saturated (NH4)2SO4) and with 50% ethylene glycol (vol/vol). The four eluted fractions were dialyzed against 5 mm NH4HCO3 pH 8, and then lyophylized before SDS-PAGE analysis. Each fraction was subjected to one dimensional SDS-PAGE on a 9 –15% gradient slab gel in reducing buffer conditions as described previously (10). After electrophoresis, the proteins were either visualized in the gel by Coomassie blue staining or transferred to a nitrocellulose membrane (11), which was then stained using red-S-Ponceau. The stained p28 and p29 bands were cut and used to raise polyclonal antibodies in rabbits (12).

One dimensional SDS-PAGE and immunoblotting of p28 and p29 Total protein extracts from bovine granulosa cells or corpus luteum were obtained following homogenization in ice-cold Tris buffer (60 mm Tris-HCl, 5 mm EDTA, pH 6.8) in the presence of protease inhibitors (1 mm phenylmethylsulfonylfluoride (PMSF), 2.5 mg/ml leupeptin, 2.5 mg/ml aprotinin, 1.2 mg/ml pepstatin-A, 2 mm benzamidine). Following centrifugation (28,000 3 g, 20 min, 4 C), the supernatant was recovered. Protein concentration was determined using the Bradford assay (BioRad Laboratories, Inc., Hercules, CA) with BSA as standard (Fraction V, Sigma Chemical Co.). Protein extracts (100 mg/well) were incubated at 100 C for 5 min in reducing conditions then size-fractionated on a one-dimensional discontinuous SDS-PAGE slab gel (10). After electrophoresis, the proteins were either visualized after Coomassie blue staining or transferred overnight at 4 C onto nitrocellulose membrane (0.45 mm Hybond-C, Amersham) and stained with red-S-Ponceau. Immunological detection was performed according to Harlow and Lane (13). Briefly, nonspecific binding sites were blocked by incubation of the

Endo • 1999 Vol 140 • No 8

nitrocellulose membrane in a TNT buffer (10 mm Tris-HCl, pH 8, 150 mm NaCl, 0.1% Tween 20) with 20% normal calf serum (NCS). The membrane was then incubated with the rabbit polyclonal anti-p28 or anti-p29 antibodies diluted at 1:1,000 in a solution of TNT with 10% NCS. Antigen and first antibody complexes were revealed following incubation with a second antibody (goat antirabbit IgG linked to horseradish peroxidase, H & L chains; human serum adsorbed; Life Technologies, Inc., Burlington, Ontario, Canada) in TNT buffer with 10% NCS and developed with 4-chloro-1-naphtol (CN) in TN buffer (20 mm Tris-HCl, 150 mm NaCl, pH 7.5).

Immunological screening of the corpus luteum cDNA library A bovine corpus luteum cDNA expression library was established to allow the characterization of p28 and p29 since these proteins were detected in corpus luteum protein extracts. The cDNA library was plated following commercial protocols (Stratagene, La Jolla, CA), and the nitrocellulose membranes were then treated with anti-p28 and anti-p29 as described in immunoblotting section. Positive clones were purified and screened until all clones were positive. Positive recombinant phagemid clones were sequenced by the double stranded dideoxychain termination (14). The consensus nucleotide sequence of two independent clones (double-stranded sequenced) is reported. Amino acid sequence was deduced from the nucleotide sequence and analyzed for sequence homology, hydrophobicity, isoelectric point and conserved amino acid motifs (15) with programs provided by the Genetics Computer Group software (version 8, 1994, Madison, WI).

RNA isolation and Northern analysis of p28 and p29 Northern blot analyses were performed with different bovine tissues collected at a slaughterhouse as previously described (14). The RNA molecular weight markers from 9 kb to 0.25 kb (Life Technologies, Inc.) were used to estimate sizes of transcripts. Total RNA (20 mg) was size-fractionated on agarose gel, transferred onto a nylon membrane, and hybridized (14) to the full-length 744 bp (bGSTA1) [32P]-labeled cDNA probe. The membranes were stripped then rehybridized with the 918 bp (bGSTA2) [32P]-labeled cDNA probe. The rat elongation factor Tu cDNA (EFTu; 16) was used as a control gene for RNA expression. The membranes were exposed for 1 day to photographic film (Biomax-MR, Eastman Kodak Co., Rochester, NY) at 270 C with an intensifying screen. The film images were digitized and the intensity of specific transcripts expressed as ratios of EFTu expression to account for procedural losses as previously described (14).

Effect of hCG on GSTA1 and GSTA2 expression in bovine ovarian follicles Holstein heifers (2 to 3 yr old) were injected on day 7 of the estrous cycle with 25 mg PGF2a (Lutalyse, Pharmacia & Upjohn, Inc., Kalamazoo, MI) to induce corpus luteum regression and initiate a follicular phase. Bovine preovulatory follicles were isolated from eight cows at 0, 6, 12, 18, 20, 22, 24, and 26 h after the injection of an ovulatory dose of hCG as previously described (17). Two synchronized heifers received an ovarian stimulation treatment on day 9 or 10 of the estrous cycle with 320 mg FSH (Follitropin-V) given in 8 decreasing doses at 12 h interval over 4 days as described previously (18). They were treated with PGF2a at the fifth and sixth FSH injection. As for the previous group of heifers, hCG was injected 36 h following the first PGF2a treatment. The ovaries were collected at time 0 and 24 h following the hCG treatment. The preovulatory follicles were dissected from the surrounding ovarian tissue and the follicular wall recovered and stored individually at 270 C until extracted for RNA.

Immunohistochemistry Immunohistochemistry was performed on fixed (PBS buffered 10% formalin) bovine tissues collected at the slaughterhouse. Paraffin-embedded tissues were cut at 7 to 8 mm thickness, mounted on Poly-l-lysine slides, deparaffined through graded xylene and alcohol series, and rehydrated. Tissues sections were rinsed in TBS (150 mm NaCl, 0. 1 m Tris, pH 7.5), immersed in 0.1 m glycine in TBS for 10 min, and the antigen


3509

epitope sites were revealed by pressure cooker treatment as previously described (19). Tissue sections were incubated in TBS with 1% BSA and 1% fat free skim milk to block nonspecific binding sites; then they were incubated for 16 h at 4 C with the rabbit anti-p28 or anti-p29 antibody at a dilution of 1:500 in TBS including 1% normal cow serum. Negative control tissue sections were incubated similarly with a nonimmune rabbit serum. The first antibody bound to p28 or p29 was revealed with a mouse monoclonal antirabbit IgG conjugated to alkaline phosphatase (Sigma Chemical Co.) at a dilution of 1:100 for 2 h at room temperature, followed by several washes in TBS. Visualization was performed using the NBT/BCIP (Roche Molecular Biochemicals, Laval, Que´bec, Canada) substrate and photographs taken under bright field illumination.

Results Specificity of anti-p28 and anti-p29 antibodies and immunoscreening of the cDNA library

The p28 and p29 proteins were isolated from the soluble fraction of bovine granulosa cells and eluted predominantly from the phenyl Sepharose chromatography in the third chromatographic step gradient with glycine buffer (20 mm, pH 9) without (NH4)2SO4, as detected by the 9 –15% gradient SDS-PAGE gel (Fig. 1A). This chromatographic step represented almost 45% of the total amount of proteins recovered after phenyl Sepharose chromatography. Following electrophoresis and transfer onto nitrocellulose, protein bands corresponding to p28 and p29 were cut and used to raise polyclonal antibodies in rabbits. These antibodies were shown to bind to proteins of 28 kDa (p28) and 29 kDa (p29) in supernatant fraction (28,000 3 g) of total protein extracts obtained from bovine granulosa cells and corpus luteum (Fig. 1B). Attempts to microsequence the NH2-termini of each protein failed because their N termini were blocked. Therefore, the characterization of the granulosa cell p28 and p29 was achieved by screening a l ZAP bovine corpus luteum cDNA library. The first immunoscreening of 1.5 3 105 pfu with the anti-p28 or the anti-p29 antibodies revealed a similar number of positive clones (0.04%). Two independent clones for p28 and p29 were subsequently purified to homogeneity, followed by phagemid in vivo excision and further characterization. Two cDNAs of approximately 850 bp and two cDNAs of approximately 1.0 kb were obtained for p28 and p29, respectively. cDNA sequence of p28 and p29

Complete sequencing of the two clones for p28 by doublestranded sequencing revealed a cDNA of 791 bp. Analysis of the sequence showed a 59 untranslated (UTR) region composed of 47 bp, an open reading frame (ORF) of 666 bp, and a 39-UTR of 78 bp that included one polyadenylation signal followed by a polyA1 tail (Fig. 2A). The ORF codes for a protein of 222 amino acids of an estimated molecular weight of 25.4 kDa and an isoelectric point of 8.9. Sequencing the two clones for p29 revealed a cDNA of 959 bp (Fig. 2B). Analysis of the sequence showed a 59 untranslated (UTR) region composed of 41 bp, an open reading frame (ORF) of 666 bp, and a 39-UTR of 252 bp that included two polyadenylation signals followed by a polyA1 tail. The ORF codes for a protein of 222 amino acids of an estimated molecular weight of 25.6 kDa and an isoelectric point of 7.2. An amino acid homology search in GenBank for p28 and p29 revealed an overall 82% or 74% identity to human a class glutathione S-transferase

FIG. 1. A, Purification of bovine granulosa cells p28 and p29. Soluble protein fraction of granulosa cells was applied on a phenyl Sepharose CL4B column equilibrated with glycine buffer (20 mM, pH9) with 25% saturated (NH4)2SO4. The column was developed in four steps with glycine buffer of decreasing ionic strenght (25%, 12%, 0% saturated (NH4)2SO4) and with 50% ethylene glycol (vol/ vol). The p28 and p29 were eluted in the third chromatographic step gradient (20 mM glycine, pH9 without (NH4)2SO4). Following transfer onto nitrocellulose, protein bands corresponding to p28 (black star) and p29 (black circle) were cut and used to raise polyclonal antibodies. B, Specificity of polyclonal antibodies raised against the bovine granulosa cell p28 and p29. Protein extracts (100 mg) from granulosa cells of unstimulated follicles or from mature corpus luteum were separated on a denaturing 10% SDS-PAGE. The gel was transferred onto nitrocellulose and incubated with the respective polyclonal antibody, anti-p28 or anti-p29 (1:1,000 dilution). The complex was revealed with a peroxidase-coupled second antibody and CN substrate. Protein bands of either 28 kDa or 29 kDa were found in granulosa cells and corpus luteum. Molecular weight standards are shown at the right.

3510


Endo • 1999 Vol 140 • No 8

(hGSTA; 20), respectively, and 87% or 78% identity to porcine GSTA (21; Fig. 3). Because these sequences are the first bovine GSTAs to be characterized, they are named as bGSTA1 and bGSTA2 subunits according to the proposed human glutathione S-transferase nomenclature (5). Bovine GSTA1 and GSTA2 are 81% identical to each other and structural differences are mainly located in a 16 amino acid stretch found between 63M and 81L of bGSTA1 (Fig. 3). Interestingly, bGSTA1 shows the main specific traits of the a class: conserved amino acid required to form a glutathione binding site (9Y, 15R,45K or 45R, 67Q, 68T, 101D, 131R), amino acids involved in subunit dimerization (52F, 69R, 73N, 82Y, 89R, 94M, 97E, 98G, 135 A, 136F, 139V, 155R), and the motif 68TRAIL72 that are also present in the human and porcine a class GST. However, bGSTA2 lacks the TRAIL motif that is replaced by a specific 16 amino acid stretch (64S to 79H) not present in human and porcine GSTA sequences (Fig. 3). No amino acid match was found for this specific 16 amino acid GSTA2 peptide when compared by BLAST analysis in GenBank, underlying the specificity of this new bovine GSTA isoform subunit. Amino acid consensus motif analysis was performed on both bGSTAs and showed the presence of three potential serine and threonine protein kinase C and casein kinase II phosphorylation sites at position 129TNR, 154SR(K)AD and 202SQR. The potential phosphorylation site 154SR(K)AD is conserved also in human and porcine GSTA. Analysis of bGSTA1 and bGSTA2 expression

Expression of bGSTA1 and bGSTA2 transcripts was studied by Northern blots on total RNA isolated from different bovine tissues. A single transcript corresponding to 1.2 kb or 1.4 kb was observed for bGSTA1 or bGSTA2, respectively (Fig. 4). Analysis of steady-state levels of bGSTA1 and

FIG. 2. A, The cDNA and predicted amino acid sequence of the bovine GSTA1 (GenBank accession number: BTU49179). Nucleotides and amino acids are numbered at the right and left. The cDNA consists of 791 nucleotides, an open reading frame of 666 bases, a 59-UTR of 47 bases, and a 39-UTR of 78 bases followed by a polyA1 tail. Amino acid numbering begins at the first methionine codon of the ORF, which encodes for 222 amino acids representing a protein of 25.4 kDa with a pI of 8.9. B, The cDNA and predicted amino acid sequences of the bovine GSTA2 (GenBank accession number: AF027386). The cDNA consists of 959 nucleotides, an ORF of 666 bases, a 59-UTR of 41 bases, and a 39-UTR of 252 bases followed by the polyA1 tail. The ORF encodes for 222 amino acids, which yields a protein of 25.6 kDa with a pI of 7.2. Polyadenylation signals found in the 39 untranslated region are underlined.

FIG. 3. Comparison of bGSTA1 and bGSTA2 amino acid sequences to human and porcine GST a class. The bovine GSTA1 and GSTA2 sequences are 82% and 74% identical to the human GSTA subunit sequence (GenBank: M14777), respectively, and 87% and 78% identical to the porcine GSTA subunit sequence (GenBank: Z69586). Dots represent identical amino acid residues and spaces (-) were introduced in the bGSTA2 sequences to allow for maximum matching. The amino acid segment that is the least conserved between the bGSTA2 protein and other a class GST is underlined. The TRAIL motif known to be conserved in all a class is overlined. Amino acids required for the glutathione binding site are indicated with a star and the one involved in subunit dimerization with a box.


3511

FIG. 4. Expression of bGSTA1 and bGSTA2 mRNA in different bovine tissues. Total RNA (20 mg/well) was analyzed by northern blots as described in Materials and Methods. The presence of major transcripts of 1.2 kb (bGSTA1) or 1.4 kb (bGSTA2) were expressed at high levels in the adult testis, 7-month-old fetal testis, FSH treated granulosa cells, corpus luteum, liver, and adrenal gland. Lower levels of expression were observed in the lung, thyroid, and kidney, whereas much lower levels were detected in cotyledon. A minor 3.3 kb transcript was observed only in FSHtreated granulosa cells and in adrenal gland. The EFTu gene expression was used as RNA control.

bGSTA2 messenger RNA (mRNA) showed that some tissues expressed GST at high levels: adult testis, 7-month old fetal testis, FSH-treated granulosa cells, corpus luteum, liver, and adrenal gland. Lower levels of bGSTA1 and bGSTA2 expression were observed in lung, thyroid, and kidney, whereas very low levels were detected in the cotyledon. No expression was found in epididymis, 3-month old fetal ovary (data not shown), uterus, heart, spleen, proximal part of duodenum, pyloric region of the stomach, and the pituitary. A minor mRNA transcript of 3.3 kb for both genes were observed only in FSH-stimulated granulosa cells, and at a lower level in the adrenal gland (Figs. 4 and 5B). Results show that bGSTA1 and bGSTA2 are differentially expressed, and their level of expression varies tremendously between the tissues analyzed. The ratio of bGSTA1/bGSTA2 mRNA levels was above 1 in adult (1.22) and fetal (1.04) testis, granulosa cells (1.2), corpus luteum (1.15), cotyledon (1.33), adrenal gland (1.31), and kidney (1.28). However, bGSTA1/bGSTA2 ratios were lower than 1 in the liver (0.74), the lung (0.55), and the thyroid (0.86), indicating that levels of expression of the two genes are controlled differently within different tissues.

Effect of gonadotropins on bGSTA1 and bGSTA2 expression in bovine ovarian follicles

Expression of bGSTA1 was studied in individual preovulatory follicles isolated from eight cows at different time after the injection of hCG. We observed that bGSTA1 expression decreased by 77% 12 h following hCG treatment, and remained low until 26 h (Fig. 5A). The expression of bGSTA1 and bGSTA2 in FSH-stimulated preovulatory follicles was studied in relation to hCG treatment. Results show a high level of bGSTA1 (1.2 and 3.3 kb) and bGSTA2 (1.4 and 3.3 kb) transcripts in FSH treated follicular wall before hCG (Fig. 5B). Twenty-four hours after the hCG injection, the mRNA levels of bGSTA1 and bGSTA2 were slightly reduced by 22.2% and 20.2%, respectively, when compared with follicular wall extract before hCG injection. The high expression of bGSTA1 and bGSTA2 in FSH stimulated follicles 24 h following the hCG injection (Fig. 5B) contrasts with the 77% decrease for bGSTA1 obtained in nonFSH-stimulated preovulatory follicles (Fig. 5A) 24 h following the hCG injection. An LH inducible gene, PGHS-2, was used as positive control and

3512


Endo • 1999 Vol 140 • No 8

lection of bovine tissues analyzed was based on previously observed mRNA expression. In all bovine tissues analyzed, the use of anti-p28 or anti-p29 gave identical labeling results. In follicles, GSTA expression was clearly localized in granulosa and theca cells (Fig. 6, A–H). The signal for GSTA was localized in the cytoplasm of granulosa cells but not in the nuclei. Granulosa cells of primordial to preovulatory follicles stained positively. However, heterogeneity in GSTA staining was observed in granulosa cells of primordial follicles; some follicles did not stain, whereas others showed a positive signal (Fig. 6, A and B). In theca cells, a strong GSTA signal was observed in late preantral, in preovulatory and in late atretic follicles (Fig. 6, C–H). Follicular fluid of large antral follicles showed a positive reaction. Oocytes from primordial to late antral stages did not express GSTA. In the corpus luteum, all luteal cells stained positively, but the intensity of labeling was not uniform. The small luteal cells showed stronger GSTA signal compared with large luteal cells (Fig. 7, A and B). In the testis, intense cytoplasmic GSTA staining was associated with Leydig cells. No staining was observed in cells located within the seminiferous tubules (Fig. 7C). In the kidney, GSTA expression was mainly localized in epithelial tubular cells with a strong signal in proximal convoluted tubules and the thick loop of Henle, whereas no staining was observed in the glomerulus, the distal tubules, and collecting ducts (Fig. 7D). In the adrenal gland, GSTA was localized in the cortex. Expression of GSTA was more pronounced in the zona glomerulosa compared with the zona fasciculata and reticularis (Fig. 7, E and F). The adrenal medulla did not show expression of GSTA except for groups of chromaffin cells located near the center of the medulla (Fig. 7, G and H). The expression of GSTA in the lung was localized in the epithelial cells of bronchioles with strong staining associated with C-cells cells and negative staining in the alveoli and stroma (Fig. 7, I and J). In the liver, the hepatocytes stained uniformly for GSTA, whereas follicular thyroid cells and cotyledon derived from a 7-month old fetus, showed a weak GSTA expression (data not shown). Discussion FIG. 5. Effects of gonadotropins on GSTA expression in bovine ovarian follicles. A, Follicular walls were collected from individual preovulatory follicles isolated from 8 cows in relation to the time of hCG injection. A 77% decrease in relative levels of bGSTA1 mRNA was observed from 12 h to 26 h following the hCG injection when compared with nonhCG treated preovulatory follicles. B, Follicular walls were collected and pooled from animals primed with FSH at time 0 or 24 h after the injection of hCG. Relative levels of mRNA expression for bGSTA1 and bGSTA2 decreased by 22.2% and 20.2%, respectively, 24 h following the hCG treatment compared with nonhCG FSH primed preovulatory follicles. Expression of PGHS-2 in granulosa cells as an LH inducible gene was detected 24 h posthCG injection. The EFTu gene expression was used as RNA control.

showed expression only at 24 h following the hCG injection (Fig. 5B). Immunohistochemistry

Immunohistochemical studies were undertaken to identify specific cell types contributing to GSTA expression. Se-

A group of 28 kDa proteins was shown to constitute one of the conspicuous groups of proteins in bovine granulosa cells (4). In this latter study, it was shown that the expression of the 28 kDa protein group was reduced when granulosa cells from bovine preovulatory follicles were treated with LH. In the present study, we have partially purified and raised polyclonal antibodies against two proteins contributing to the 28 kDa granulosa cell protein group, that were named p28 and p29. We report herein their characterization by cDNA immunoscreening of a luteal cell expression library, and reveal that these cDNAs encode two uncharacterized bovine a glutathione S-transferase subunits (GSTA). We named these subunits bGSTA1 and bGSTA2, based on the human nomenclature (5). The bGSTA1 and bGSTA2 subunits are 82% and 74% identical to the human GSTA1 subunit (20). The bGSTA1 subunit shares the 68TRAIL72 amino acid strech, which is conserved in all known a GST class proteins, and possesses the specific amino acid residues (9Y, 13R, 14G, 20 R, 69R, 187R), which are conserved in the a, m, and p GSTs.


3513

FIG. 6. Immunohistochemical localization of GSTA in bovine ovarian follicles. Paraffin embedded tissues were incubated with anti-p28 antibody (1:1,000 dilution) and the complex was revealed with a mouse monoclonal antibody coupled to alkaline phosphatase, and NBT/BCIP used as substrate. A, Primordial follicles showing negative (PFN) and positive (PFP) staining in flattened granulosa cells (6003); B, positively stained granulosa (G) cells of primary follicle (PRF) with negative GSTA stained primordial follicles and oocyte (O; 6003); C, late preantral follicle; note positive staining in granulosa and theca (T) cells whereas oocyte and stroma are negative (4003); D, early antral follicle with positive staining in granulosa cells and well formed theca layer without staining in oocyte and stroma (2003); E, a large antral follicle with positively stained granulosa and theca cells, and follicular fluid (FF; 403); F, follicular wall of two large healthy antral follicles with labeling in granulosa and theca cells, and in follicular fluid (2003); G) follicular wall of two large antral atretic follicles (1003) and H) follicle in advanced stage of atresia (1003) showing positive staining in theca cells and follicular fluid. Control sections were negatively stained when anti-p28 was substituted with normal rabbit serum.

3514


Endo • 1999 Vol 140 • No 8

FIG. 7. Immunohistochemical localization of GSTA in different bovine tissues. Paraffin embedded tissues were incubated with anti-p28 antibody (1:1,000 dilution), and the complex was revealed with a mouse monoclonal antibody coupled to alkaline phosphatase and NBT/BCIP used as substrate. A, Corpus luteum showing heterogeneity for GSTA staining (1003); B), corpus luteum showing more intense labeling in small (SLC) compared with large (LLC) luteal cells (2003); C) testis showing strong staining in Leydig cells (LE) whereas cells within seminiferous tubules (ST) are negative (1003); D) kidney showing GSTA expression localized in epithelial tubular cells with strong signal in proximal convoluted tubules (PT) and thick loop of Henle, whereas the glomerulus (GL), the epithelial cells of the distal tubule (DT) and collecting duct are negative (2003); E) all three zona of the adrenal cortex (CO) are stained for GSTA but the adrenal medulla (ME) is negative (403); F) adrenal cortex showing stronger signal in the zona glomerulosa (ZG) compared with the zona fasciculata (ZF; 1003); G and H) Some groups of strongly stained chromaffin cells (MCC) located near the center of the adrenal medulla (1003; 6003); I) GSTA staining in epithelial cells of bronchioles (ECB), whereas alveoli cells and stroma are negative (2003); J) strong staining in respiratory epithelium is localized in Clara cells (CC; 6003). Control sections were negatively stained when anti-p28 was substituted with normal rabbit serum.


The GSTA subunit is composed of two domains. Domain 1 spans from the NH2 terminal to amino acid 86 and from 192 to 222. It contains the binding site for glutathione, called the G-site, which is defined by the following conserved amino acids: 9Y, 15R,45K or 45R, 67Q, 68T, 101D, 131R (Fig. 3; 9, 22–26). Domain 2 is less conserved between GSTs and spans from amino acids 86 to 190. It is involved in the binding of hydrophobic substrates and is also called the H-site (23, 25). Purification of GSTAs from bovine adrenal cortex was previously achieved by affinity chromatography on S-hexylglutathione-Sepharose (S-hex-G-Ag) and their electrophoretic mobility analyzed by SDS-PAGE (27, 28). However, their molecular characterization and pattern of expression associated with cell types in different tissues were not reported. GSTAs were found to be major cytosolic proteins in the bovine adrenal cortex (28). Similarly, we have observed a major group of 28 kDa proteins as demonstrated herein and a previous study (4). In the bovine adrenal gland study (28), two GSTAs were observed: a major protein band at 25.9 kDa estimated to contribute to 1.3% of the cytosolic protein content, and a minor protein band at 26.5 kDa. We have shown that bGSTA1 possesses the conserved characteristics of a class GST known to allow interaction with S-hex-G-Ag as reported for other species (22, 29), which suggest that bGSTA1 may correspond to the 25.9 kDa bovine GSTA previously reported (28). The bovine adrenal gland 26.5 kDa GST (28) did not interact to S-hex-G-Ag but was purified on glutathione-Sepharose (GSH-Ag) affinity column from material not binding to S-hex-G-Ag matrix. Interestingly, we have shown that the bGSTA2 subunit presents characteristics not previously found in a class GST: it lacks the TRAIL motif known to be conserved in GSTA of all species. The TRAIL motif is replaced by a specific 16 amino acid sequence (64S to 79H) when compared with bGSTA1, human (20) or porcine a class GST (21). This change in bGSTA2 results in the lost of four recognized amino acids known to contribute to the G-site and intervene in subunit dimerization : 67Q and 68T of bGSTA1 contribute to the G-site by interacting with the g-glutamyl residues of glutathione, whereas 69R and 73N are known to contribute to subunit dimerization. This 16 amino acid peptide (64S to 79H) is specific to bGSTA2 because no amino acid match was found when compared with available protein sequences in GenBank, and thus represents a novel GSTA isoenzyme. Furthermore, characterization of enzymatic substrate for the two bovine adrenal GSTs showed a high metabolism of 4-hydroxyalkenal activity when compared with their human counterpart (28). It was deduced that the 26.5 kDa bovine GST was responsible for the high 4-hydroxyalkenal activity. The 4-hydroxynonemals, the major endogenously produced form of 4-hydroxyalkenals, are derived from cellular lipid peroxidation and are genotoxic (30 –32). Thus, it is reasonable to hypothesize that the specific amino acid characteristics presented by bGSTA2 could be responsible for the high 4-hydroxyalkenal processing in the bovine adrenal gland study (28). Whether bGST2 isoenzyme mediates this difference remains an interesting and, as yet, open question. Bovine GSTA1 and GSTA2 mRNA were coexpressed in many bovine tissues as single transcripts of 1.2 kb and 1.4 kb,

3515

respectively. Tissues with the highest levels of GSTA mRNA were the ovarian follicle, the corpus luteum, the testis, the liver, and the adrenal gland. Only FSH-treated granulosa cells and the adrenal gland showed the larger 3.3 kb transcript that was observed for both GST isoenzymes. This larger size transcript was found only in tissues with the highest level of GSTA expression. Bovine GSTA1/GSTA2 mRNA ratios varied in different tissues. Bovine GSTA2 mRNA expression was higher than bGSTA1 only in the liver, the lung, and the thyroid. Different steady-state levels of mRNA for bGSTA1 and bGSTA2 indicates tissue-specific and different control of their expression. The cellular localization of GSTA in different tissues was studied by immunohistochemistry for the first time in the bovine species. In the ovary, GSTA was immunolocalized in the granulosa and theca cells of preantral and antral follicles. Furthermore, GSTA labeling was observed in follicular fluid of large antral follicles, which suggests secretion of GSTA by granulosa cells in the follicular fluid. Heterogeneity in GSTA staining was noticeable in flattened granulosa cells of primordial follicles, which could relate to differences in the physiological stage of follicles (33). GSTA positive follicles could represent the primordial follicles that have entered in the growing pool, whereas GSTA negative primordial follicles could be quiescent follicles. In agreement with our observations, Rahilly et al. (1991) found GSTA expression in human primary follicles, but contrast with the low levels of staining reported for granulosa cells in antral follicles. In the pig (34), no expression of GSTA was observed in granulosa cells contrasting with the bovine and human observations and the recent cloning of two GSTA isoenzymes from an FSH-treated porcine granulosa cell substractive hybridization cDNA library (21). Different levels of GSTA expression reported for granulosa cells may reflect species specific differences, effects of gonadotropins, or differences in the developmental stage at which the follicles were studied. Bovine theca interna cells expressed GSTA. Strong staining was observed concomitant with the beginning of theca interna formation in late preantral follicles up to advanced stages of atresia in large antral follicles. Strong GSTA immunostaining in theca interna cells was also reported in humans (35) and pigs (34). Bovine oocytes at all stages of follicular development showed no expression of GSTA. A significant reduction (77%) in the expression of GSTA mRNA was observed in preovulatory follicles isolated 12 h post-hCG injection, which agrees with the observed decrease synthesis of the 28 –30 kDa protein group in granulosa cells after the LH surge (4). Using the same animal model, it was previously shown that estradiol concentrations were reduced in follicular fluid following hCG injection (17, 18). Therefore, reduction in GSTA mRNA expression was concomitant with the reduction of estradiol concentrations. However, when we pretreated follicles with FSH, they responded differently to hCG with only a slight decrease (21%) of expression for GSTAs mRNA at 24 h post-hCG injection. It is well known that ovarian stimulation with FSH in cattle increases steroid synthesis as found in follicular fluid (36). Corroborating our observations, gonadotropin treatment of luteinizing porcine granulosa cells (34) or rat ovary (37) was shown to increased GSTA expression. In the bovine model,

3516


prostaglandin endoperoxidase synthetase-2 (PGHS-2) mRNA was induced between 18 to 26 h post-hCG injection (17, 18). PGHS-2 is the first rate limiting enzyme in the prostaglandin biosynthesis pathway, and prostaglandins are required for the ovulatory process. Collectively, these results suggest that GSTA mRNA expression is not temporally related to PGHS-2 expression but is modulated by gonadotropins and intimately linked to estradiol synthesis. The heterogeneity of immunohistochemical localization in bovine corpora lutea was related to a stronger expression of GSTA in small compared with large luteal cells, which confirms observations in human and porcine corpus luteum (34, 35). In cattle, granulosa cells differentiate into large luteal cells, whereas theca interna cells differentiate into small luteal cells (38). The large luteal cells produce high basal concentrations of progesterone but are not responsive to LH, whereas small luteal cells produce low basal levels of progesterone but are stimulated by LH. The significance of high GSTA expression in small luteal cells is unknown but suggests a potential association with steroidogenesis and LH. In the bovine testis, strong staining for GSTA was found in the testosterone-producing Leydig cells, and no expression was detected in all cells of the seminiferous tubule. These observations agree with those in humans (39). We did not observe GSTA1 nor GSTA2 mRNA expression in the bovine gastrointestinal tract (stomach and intestine) except for expression in the liver, which contradicts observations in human gastrointestinal tissues (39). Bovine adrenal cortex showed strong expression of GSTA in the zona glomerulosa compared with the zona fasciculata and reticularis, in contrast to very low levels observed in human zona glomerulosa and high levels in the zona reticularis (39). Staining for GSTA was observed in the zona glomerulosa and fasciculata of the rat adrenal gland (40). Negative staining was observed in most of the bovine adrenal medulla except for islands of chromaffin cells, which showed a strong signal, confirming observations in rat adrenal medulla (40). In the bovine kidney, GSTA labeling was associated only with epithelial cells in proximal convoluted tubules and the thick descending loop of Henle, agreeing with observations reported in human (39). In agreement with data in cattle (41) and humans (42), we found GSTA mRNA expression in bovine lungs. These observations contradict previous reports that found no GSTA subunit expression in bovine lung (28). Moreover, we found strong GSTA immunostaining in bovine bronchioles that was associated exclusively with Clara cells. Our observations in the bovine species differ from the immunostaining for GSTA in humans where intensity of staining was low but widely distributed, extending from the epithelium of the bronchi to the type 1 and type 2 alveolar epithelium, and to vascular smooth muscle (42). The differences observed in the expression of GSTA in the bovine and human species may be associated with cross-reaction of antibodies used in the human study or to species specific difference in GSTA expression. The common biochemical link shared by the different cell types that supports the expression of GTSAs remains speculative. The gonadal, adrenal, and liver cells contain mitochondrial and microsomal associated cytochrome P450 enzymes involved in steroidogenesis (43). Renal epithelial

Endo • 1999 Vol 140 • No 8

tubular cells are known to be involved in the synthesis of vitamin D (43). The human or rat lung were shown to have the capacity to metabolize steroid hormones (44, 45), and Clara cells express specific cytochrome P450 involved in the metabolism of xenobiotics (46). Steroid hormones synthesis and metabolism was shown to release free radicals (47). These free radicals may thereby stimulate GSTA expression (31, 32). Expression of GSTA isoenzymes in bovine species may thus be linked to cell types involved in steroid synthesis or metabolism, although validation of this hypothesis will require future work. In conclusion, this report is the first to demonstrate in the bovine species that expression of GSTA isoenzymes is tissue and cell specific and is modulated by gonadotropins in the developing bovine ovarian follicle. We have also shown that the highest level of GSTA expression was localized to steroidogenically active cells. The specific biological role of GSTA may be linked to cellular oxidative stress induced by reactive oxygen species generated during steroidogenesis. Acknowledgments The authors would like to thank Mr. Benoit Remy for help during the course of protein purification and antibody production.

References 1. Salustri A, Hascall VC, Camaioni A, Yanagishita M 1993 Oocyte-granulosa cell interactions. In: Adashi EY, Leung PPK (eds) The Ovary. Raven Press, New York, pp 209 –225 2. Wassarman PM, Albertini DF 1994 The mammalian ovum. In: Knobil E, Neill JD (eds) The Physiology of Reproduction. Raven Press, New York, vol 1:79 –122 3. Eppig JJ 1996 Mammalian oocyte growth and development in vitro. Mol Reprod Dev 44:260 –273 4. Rabahi F, Monniaux D, Pisselet C, Chupin D, Durand P 1991 Qualitative and quantitative changes in protein synthesis of bovine follicular cells during the preovulatory period. Mol Reprod Dev 30:265–274 5. Mannervik B, Awasthi YC, Board PG, Hayes JD, Di Ilio C, Ketterer B, Listowsky I, Morgenstern R, Muramatsu M, Pearson WR, Pickett CB, Sato K, Widersten M, Wolf CR 1992 Nomenclature for human glutathione transferases. Biochem J 282:305–308 6. Hayes JD, Strange RC 1995 Potential contribution of the glutathione S-transferase supergene family to resistance to oxidative stress. Free Radic Res 22:193–207 7. Board PG, Johnston PN, Ross VL, Webb GC, Coggan M, Suzki T 1991 Molecular genetics of the human glutathione S-transferase. In: Ernster L (ed) Xenobiotics and Cancer. 21st Symposia Princess Takamatsu. Japan Sci Soc Press, Tokyo/Tatlor & Francis Ltd., London, pp 199 –211 8. Mannervik B, Danielson UH 1988 Glutathione transferases. Structure and catalytic activity. Crit Rev Biochem Mol Biol 23:283–337 9. Hayes JD, Pulford DJ 1995 The glutathione S-transferase supergene family: regulation of GST and the contribution of the isoenzymes to cancer chemoprotection and drug resistance. Crit Rev Biochem Mol Biol 30:445– 600 10. Laemmli UK 1970 Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227:680 – 685 11. Towbin H, Staehelin T, Gordon J 1979 Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: procedure and some applications. Proc Natl Acad Sci USA 76:4350 – 4354 12. Knudsen KA 1985 Proteins transferred to nitrocellulose for use as immunogens. Anal Biochem 147:285–288 13. Harlow E, Lane D 1988 Antibodies. A Laboratory Manual. Cold Spring Harbor Laboratory Press, New York 14. Houde A, Lambert A, Saumande J, Silversides DW, Lussier JG 1994 Structure of the bovine follicle-stimulating hormone receptor complementary DNA and expression in bovine tissues. Mol Reprod Dev 39:127–135 15. Bairoch A, Bucher P, Hofmann K 1996 The PROSITE database, its status in 1995. Nucleic Acids Res 24:189 –196 16. Levine RA, Serdy M, Guo L, Holzschu D 1993 Elongation factor TU as a control gene for mRNA analysis of lung development and other differentiation and growth regulated systems. Nucleic Acids Res 21:4426 17. Sirois J 1994 Induction of prostaglandin endoperoxide synthetase-2 by human chorionic gonadotropin in bovine preovulatory follicles in vivo. Endocrinology 135:841– 848 18. Liu J, Sirois J 1998 Follicle size-dependent induction of prostaglandin G/H synthase-2 during superovulation in cattle. Biol Reprod 58:1527–1532

GSTA EXPRESSION IN STEROIDOGENICALLY ACTIVE CELLS 19. Se´vigny J, Levesque FP, Grondin G, Beaudoin AR 1997 Purification of the blood vessel ATP diphosphohydrolase, identification and localisation by immunological techniques. Biochim Biophys Acta 1334:73–78 20. Tu CP, Qian B 1986 Human liver glutathione S-transferases: complete primary sequence of an Ha subunit cDNA. Biochem Biophys Res Commun 141:229 –237 21. Tosser-Klopp G, Benne F, Bonnet A, Mulsant P, Gasser F, Hatey F 1997 A first catalog of genes involved in pig ovarian follicular differentiation. Mamm Genome 8:250 –254 22. Stenberg G, Board PG, Calberg I, Mannervik B 1991 Effects of directed mutagenesis on conserved arginine residues in a human class alpha glutathione transferase. Biochem J 274:549 –555 23. Sinning I, Kleywegt GJ, Cowan SW, Reinemer P, Dirr HW, Huber R, Gilliland GL, Armstrong RN, Ji X, Board PG, Olin B, Mannervik B, Jones TA 1993 Structure determination and refinement of human alpha class glutathione transferase A1–1, and a comparison with the Mu and Pi class enzymes. J Mol Biol 232:192–212 24. Atkins WM, Wang RW, Bird AW, Newton DJ, Lu AY 1993 The catalytic mechanism of glutathione S-transferase (GST). Spectroscopic determination of the pKa of Tyr-9 in rat a 1–1 GST. J Biol Chem 268:19188 –19191 25. Cameron AD, Sinning I, L’Hermite G, Olin B, Board PG, Mannervik B, Jones TA 1995 Structural analysis of human a-class glutathione transferase A1–1 in the apo-form and in complexes with ethacrynic acid and its glutathione conjugate. Structure 3:717–727 26. Bjornestedt R, Stenberg G, Widersten M, Board PG, Sinning I, Jones TA, Mannervik B 1995 Functional significance of arginine 15 in the active site of human class a glutathione transferase A1–1. J Mol Biol 247:765–773 27. Hayes JD, Milner SW, Walker SW 1989 Expression of glyoxalase, glutathione peroxidase and glutathione S-transferase isoenzymes in different bovine tissues. Biochim Biophys Acta 994:21–29 28. Meikle I, Hayes JD, Walker SW 1992 Expression of an abundant a-class glutathione S-transferase in bovine and human adrenal cortex tissues. J Endocrinol 132:83–92 29. Bogaards JJ, van Ommen B, van Bladeren PJ 1992 Purification and characterization of eight glutathione S-transferase isoenzymes of hamster. Comparison of subunit composition of enzymes from liver, kidney, testis, pancreas and trachea. Biochem J 286:383–388 30. Ålin P, Danielson UH, Mannervik D 1985 4-Hydroxy-2-enals are substrates for glutathione transferase. FEBS Lett 179:267–270 31. Poli G, Dianzani MU, Cheeseman KH, Slater TF, Lang J, Esterbauer H 1985 Separation and characterization of the aldehydic products of lipid peroxidation stimulated by carbon tetrachloride or ADP-iron isolated hepatocytes and rat liver microsomal suspensions. Biochem J 227:629 – 638 32. Kaur K, Salomon RG, O’Neil J, Hoff HF 1997 (Carboxyalkyl) pyrroles in human plasma and oxidized low-density lipoproteins. Chem Res Toxicol 10:1387–1396

3517

33. McNatty KP, Heath DA, Lundy T, Fidler AE, Quirke LD, Stent VC, O’Connell A, Smith PR, Groome N, Tisdall DJ, Control of early ovarian follicular development. J Reprod Fertil Suppl, in press 34. Keira M, Nishihira J, Ishibahi T, Tanaka T, Fujimoto S 1994 Identification of a molecular species in porcine ovarian luteal glutathione S-transferase and its hormonal regulation by pituitary gonadotropins. Arch Biochem Biophys 308:126 –132 35. Rahilly M, Carder PJ, Al Nafussi A, Harrison DJ 1991 Distribution of glutathione S-transferase isoenzymes in human ovary. J Reprod Fertil 93:303–311 36. Fortune JE, Hansel W 1985 Concentrations of steroids and gonadotropins in follicular fluid from normal heifers and heifers primed for superovulation. Biol Reprod 32:1069 –1079 37. Toft E, Becedas L, Soderstrom M, Lundqvist A, Depierre JW 1997 Glutathione transferase isoenzyme patterns in the rat ovary. Chem-Biol Interact 108:79 –93 38. Alila HW, Dowd JP, Corradino RA, Harris WV, Hansel W 1988 Control of progesterone production in small and large bovine luteal cells separated by flow cytometry. J Reprod Fertil 82:645– 655 39. Sundberg AGM, Nilsson R, Appelkvist E-L, Dallner G 1993 Immunohistochemical localization of a and p class glutathione transferases in normal human tissues. Pharmacol Toxicol 72:321–331 40. Mankowitz L, DePierre JW, Mannervik B, Hansson H-A 1991 Immunohistochemical distribution of isoenzymes of glutathione transferase in adult rat adrenal gland before and after hypophysectomy. Biochem J 280:399 – 405 41. Aceto A, Sacchetta P, Dragani B, Bucciarelli T, Angelucci S, Longo V, Gervasi GP, Martini F, Carmine DI 1993 Glutathione transferase isoenzymes in olfactory and respiratory epithelium of cattle. Biochem Pharmacol 46:2127–2133 42. Anttila S, Hirvonen A, Vaino H, Husgafvel-Pursiainen K, Hayes JD, Ketterer B 1993 Immunohistochemical localization of glutathione S-transferase in human lung. Cancer Res 53:5643–5648 43. Wilson JD, Foster DW, Kronengerg HM, Larsen PR 1998 Williams Textbook of Endocrinology, ed 9. WB Saunders Co, Philadelphia 44. Honkakoski P, Maenpaa J, Leikola J, Pasanen M, Juvonen R, Lang MA, Pelkonen O, Raunio H 1993 Cytochrome P450 2A-mediated coumarin 7-hydroxylation and testosterone hydroxylation in mouse and rat lung. Pharmacol Toxicol 72:L107–L112 45. Waxman DJ, Lapenson DP, Aoyama T, Gelboin HV, Gonzalez FJ, Korzekwa K 1991 Steroid hormone hydroxylase specificities of eleven cDNA-expressed human cytochrome P450s. Arch Biochem Biophys 290:160 –166 46. Ji CM, Cardoso WV, Gebremichael A, Philpot RM, Buckpitt AR, Plopper CG, Pinkerton KE 1995 Pulmonary cytochrome P-450 monooxygenase system and Clara cell differentiation in rats. Am J Physiol 269:L394 –L402 47. Rapoport R, Sklan D, Hanukoglu I 1995 Electron leakage from the adrenal cortex mitochondrial P450scc and P450c11 systems: NADPH and steroid dependence. Arch Biochem Biophys 317:412– 416