A cryptic, microsomal-type arachidonate 12-lipoxygenase is tonically inactivated by oxidation-reduction conditions in cultured epithelial cells.

Cultured ovine tracheal epithelial cells converted arachidonic acid to prostaglandin E2 (PGE2), but microsome-containing subcellular fractions prepared from these cells under calcium-free conditions converted arachidonic acid to PGE2 and to 12-hydroxyeicosatetraenoic acid (12-HETE) at a high rate (2-4 nmol/mg of protein/15 min). Identification of the membrane-bound 12-HETE-forming activity as a 12-lipoxygenase included 12S-stereospecificity of product formation and trapping of 12-hydroperoxyeicosatetraenoic acid as a reaction product. The 12-lipoxygenase activity was extracted from cell membranes only with detergent (1% Triton X-100), and the activity (membrane-bound or detergent-solubilized) was completely inactivated by mixing with the cytosol-containing subcellular fraction. The inhibitory effect of the cytosolic fraction was reversed by treating the cytosol with GSH-depleting agents (2-cyclohexene-1-one or N-ethylmaleimide) or by mixing it with lipid hydroperoxide (13-hydroperoxyoctadecadienoic acid) at a concentration that had little direct effect on enzyme activity. Inhibition of 12-lipoxygenase activity could also be achieved by treatment of enzyme preparations with GSH at levels (0.1-10 mM) found in epithelial cell cytosol. In addition, treatment of cultured epithelial cells with a GSH-depleting agent (buthionine sulfoximine) and lipid hydroperoxide restored cellular 12-lipoxygenase activity. Little or no detectable 12-lipoxygenase activity was found in freshly isolated ovine tracheal epithelial cells, but the cytosolic 12-lipoxygenase found in freshly isolated bovine tracheal epithelial cells was relatively insensitive to regulation by GSH or lipid hydroperoxide. These observations indicate that a 12-lipoxygenase is expressed in a cryptic, microsomal-type form in primary-culture epithelial cells and that this form of the enzyme may be selectively regulated by changes in cellular oxidation-reduction conditions dependent on cytosolic levels of GSH versus lipid hydroperoxide.

A Cryptic, Microsomal-type Arachidonate 12-Lipoxygenase Is Tonically Inactivated by Oxidation-Reduction Conditions in Cultured Epithelial Cells* (Received for publication, August 25, 1992) Laurie P. Shornick and Michael J. HoltzmanS Cultured ovine tracheal epithelial cells converted arachidonic acid to prostaglandin E2 (PGE2), but microsome-containing subcellular fractions prepared from these cells under calcium-free conditions converted arachidonic acid to PGE2 and to 12-hydroxyeicosatetraenoic acid (12-HETE) at a high rate (2)(3)(4) nmol/mg of protein/lB min). Identification of the membrane-bound 12-HETE-forming activity as a 12-lipoxygenase included 12s-stereospecificity of product formation and trapping of 12-hydroperoxyeicosatetraenoic acid as a reaction product. The 12-lipoxygenase activity was extracted from cell membranes only with detergent (1% Triton X-loo), and the activity (membrane-bound or detergent-solubilized) was completely inactivated by mixing with the cytosol-containing subcellular fraction. The inhibitory effect of the cytosolic fraction was reversed by treating the cytosol with GSH-depleting agents (2-cyclohexene-1-one or Nethylmaleimide) or by mixing it with lipid hydroperoxide (13-hydroperoxyoctadecadienoic acid) at a concentration that had little direct effect on enzyme activity. Inhibition of 12-lipoxygenase activity could also be achieved by treatment of enzyme preparations with GSH at levels (0.1-10 mM) found in epithelial cell cytosol. In addition, treatment of cultured epithelial cells with a GSH-depleting agent (buthionine sulfoximine) and lipid hydroperoxide restored cellular 12lipoxygenase activity. Little or no detectable 12-lipoxygenase activity was found in freshly isolated ovine tracheal epithelial cells, but the cytosolic 12-lipoxygenase found in freshly isolated bovine tracheal epithelial cells was relatively insensitive to regulation by GSH or lipid hydroperoxide. These observations indicate that a 12-lipoxygenase is expressed in a cryptic, microsomal-type form in primary-culture epithelial cells and that this form of the enzyme may be selectively regulated by changes in cellular oxidation-reduction conditions dependent on cytosolic levels of GSH versus lipid hydroperoxide.
Arachidonate 12-lipoxygenase was the first lipoxygenase to * This research was supported by National Institutes of Health Grants HL-40078 and DK-38111 and the Schering Career Investigator Award of the American Lung Association. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
$To whom correspondence and reprint requests should be ad- be discovered in animal or human tissues. The enzyme was described initially in platelets and subsequently in leukocytes, and the two cellular forms have been distinguished catalytically (1) and structurally (2,3). The 12-lipoxygenase has also been detected at a high level in several nonhematopoietic cell types, including epithelial cells in tracheal and colonic mucosa (4-6), endocrine cells from anterior pituitary (7), adrenal glomerulosa (8), and pancreatic islets (9), and in neurons and perineural cells ( 5 , 10). In some species (including humans), a closely related 15-lipoxygenase may be expressed in some of the same tissues, e.g. in the tracheal epithelium (5,11,12). Immunochemical characterization of the epithelial 12-and 15-lipoxygenases indicates that they are similar to the leukocyte form of the enzyme (5, 6,12,13). Molecular cloning from cDNA libraries generated from leukocytes, epithelial cells, and reticulocytes indicates that the known members of the extended family of 12-and 15-lipoxygenases found in hematopoietic and nonhematopoietic cells are highly homologous with each other (2,14,15).
Determinants for the selective expression and regulation of the 12-and 15-lipoxygenases in epithelial cells (or in other types of cells) are uncertain. The absence of data is due in part to the difficulty in establishing cell culture systems which maintain the lipoxygenase phenotype of the original tissue. For example, cultured epithelial cells (in contrast to freshly isolated cells) often exhibit little or no evidence of 12-or 15lipoxygenase activities under a variety of tissue culture conditions (16)(17)(18)(19). The studies described here indicate that cultured ovine tracheal epithelial cells may express a 12lipoxygenase, but the enzyme may be bound to microsomal membranes (by a non-calcium-dependent mechanism) in a form which is inactive unless it is dissociated from the cytoplasm. In addition, the cytosolic inhibition of the microsomaltype 12-lipoxygenase is due a t least in part to the selective sensitivity of the enzyme to cellular levels of GSH versus hydroperoxide. The pattern of sensitivity is distinct from the cytosolic 12-lipoxygenase or PGH synthase/isomerase pathways found in epithelial cells and may suggest that the microsomal-type 12-lipoxygenase is actively linked to the chain of enzymes that regulate cellular oxidation-reduction conditions.
(Plymouth Meeting, PA) and Cayman Chemical Co. (Ann Arbor, Cell Purification and Culture-Animal tissues were obtained from a local abattoir, and epithelial cells were isolated from tracheal mucosal strips by enzymatic dissociation with 0.02% type I collagenase and 5 mM dithiothreitol for 2 h at 37 "C or 0.1% Pronase at 4 "C for 16 h in Joklik's minimum essential medium. Cells were cultured in LHC-8e medium on flasks coated with collagen/albumin as described previously (19) and were generally plated at densities (4-8 X 10' cells/cm2) to achieve cellular subconfluence at 4-5 days in culture. Experiments varying the degree of confluence and the plating density were carried out to characterize possible changes in oxygenation activity at different times in culture and demonstrated that these conditions resulted in maximal activity. Flasks processed for electron microscopy revealed cell morphologic features similar to those reported previously (11,21), and cells subjected to immunostaining were uniformly positive for keratin.
Assay of Oxygenation Activity in Intact Epithelial Cells-Replicate flasks of cultured cells (1-4 X lo6 cells/flask/3 ml) or replicate aliquots of cell suspensions (4 X lo6 cells/2 ml) were incubated in Hepesbuffered Hanks' balanced salt solution containing arachidonic acid in ethanol. Final concentrations of 0.3-320 p~ [3H]arachidonic acid and 0.5% ethanol were tested for 2-45 min at 37 "C to define conditions for maximal product generation. Preliminary experiments showed that incubating cultured cells with arachidonic acid in flasks or in suspension after trypsinization from flasks yielded similar values for metabolite generation. Reagents incubated with boiled cells (100 "C for 15 min) or with cell-free media demonstrated insignificant generation of metabolites. In some experiments, the cells were assayed for oxygenation activity after treatment with 13-HPODE (to increase hydroperoxide levels) or with huthionine sulfoximine (to decrease GSH levels via y-glutamylcysteine synthase inhibition) (22). Subcellular Fractionation of Epithelial Cells-Arachidonic acid oxygenation was also studied in preparations of cells disrupted by sonication and in the soluble and membrane-containing subcellular fractions obtained after separation by differential centrifugation at 4 "C. Cultured epithelial cells were collected by centrifugation after dissociation with 0.05% trypsin, 0.02% EDTA for 10 min at 37 "C, and trypsin neutralization with LHC basal medium containing 10% fetal bovine serum and 0.5% soybean trypsin inhibitor. Cell suspensions (5 X lo7 cells/ml) were homogenized for 30 s and then sonicated at the microtip limit (105 watts) for 200 s in a disruption buffer of 50 mM Tris-HC1, pH 7.4, containing 5 mM EDTA and 1 mM EGTA. These conditions were designed to maximize recovery of lipoxygenase activity but also preserve PGH synthase and PGE isomerase activities (13,19). The sonicated cell mixtures were centrifuged at 400 X g for 10 min, and the resulting supernatants were centrifuged at 12,000 X g for 20 min and 100,000 X g for 1 h to yield subcellular fractions. Electron microscopy demonstrated that microsomes were contained in the 12,000 and 100,000 X g pellets and that subcellular organelle (e.g. mitochondria) were confined to the 12,000 X g pellet. Thus, the 12,000 X g pellet was designated as the mitochondrial fraction, the 100,000 X g pellet as the microsomal fraction, and the 100,000 X g supernatant as the cytosolic fraction. For detergent extraction, the mitochondrial and microsomal fractions were each resuspended in 50 mM Tris-HC1, pH 7.4, containing 5 mM EDTA, 1 mM EGTA, and 0.01-5% Triton X-100, and stirred for 1 h at 4 "C. The mixtures were centrifuged at 100,000 X g for 1 h, and the resulting supernatants (Triton-soluble) and pellets (Triton-insoluble) were assayed for oxygenase activity. Protein in each fraction was determined by the Bradford method or by the bicinchoninic acid method (for Triton- The abbreviations used are: HETE, hydroxyeicosatetraenoic acid HEpETrE, hydroxyepoxyeicosatrienoic acid; HHT, hydroxyheptadecatrienoic acid; HPETE, hydroperoxyeicosatetraenoic acid HPLC, high pressure liquid chromatography; HPODE, hydroperoxyoctadecadienoic acid PG, prostaglandin. treated samples) using bovine serum albumin as a standard.
Assay of Oxygenation Activity in Disrupted Cells and Subcellular Fractions-Each fraction (400 X g supernatant, 12,000 X g pellet, and 100,000 X g supernatant and pellet) was assayed for oxygenation activity. The pellets obtained by centrifugation at 12,000 and 100,000 X g were resuspended in the original volume of buffer for assay of activity, and the 400 X g and 100,000 X g supernatants were assayed directly. All fractions (100-250 pl with 1-5 mg protein/ml) were diluted in an assay buffer of 50 mM Tris-HC1, pH 7.4, containing 5 mM EDTA, 1 mM EGTA, 5-80 p~ [3H]arachidonic acid, and 1% ethanol (total volume of 2 ml). Assays were carried out at 37 "C for 2-45 min after addition of enzyme. Preliminary experiments indicated that maximal lipoxygenase activity was achieved by incubation with 10-20 pM arachidonic acid and that the activity increased linearly over the range of total protein concentrations from 0.1-0.8 mg/ml. In some experiments, the microsome-containing or cytosolic subcellular fractions were assayed after treatment with 13-HPODE or with Nethylmaleimide or 2-cyclohexene-1-one (GSH-derivatizing agents) at concentrations that provide maximal activity (23-25). In comparative experiments, freshly isolated ovine or bovine tracheal epithelial cells were prepared, subjected to fractionation, and assayed for oxygenation activity using identical procedures.
HPLC Analysis of Oxygenation Products-Cell supernatants and enzyme mixtures were extracted with 2-propanol/acetic acid/chloroform or diethyl ether/acetic acid using PCB, as an internal standard for product recovery as described previously (26, 27). Extracts were reconstituted in chromatographic solvent and analyzed by reversephase and by chiral-phase HPLC as described previously (27). For reverse-phase HPLC, the chromatograph was fitted with a 4.6 X 100 mm analytical column packed with 3-pm octadecylsilane-coated particles and was run at a flow rate of 1.0 ml/min with two solvents (A and B) set at 34% B for 0-9 min, 60% B for 10-35 min, and 90% B for 36-43 min where A was water/acetic acid (1OO:O.Ol; v/v) and B was acetonitrile/acetic acid (1OO:O.Ol). Chiral-phase HPLC was performed with a pair of 4.6 X 250-mm columns in series which were packed with 5-pm aminopropyl silica particles ionically bonded to dinitrobenzoylphenylglycine at a flow rate of 0.8 ml/min with two solvents (A and B) used in an isocratic program of 8% B where A was hexane and B was hexane/2-propanol(lOO4). The HPLC eluate was monitored using a diode array spectrophotometer set at 270 nm for conjugated trienes (including PCB,) and 235 nm for conjugated dienes (including 12-HETE). The outflow from the spectrophotometer was routed to a radioactivity detector for concurrent measurement of 3H or 14C. Compounds were quantified using standard molar absorption coefficients and measurements of specific activity (26).
TLC Analysis of Products-Enzyme incubations with [14C]arachidonic acid were stopped with the addition of 2.25 volumes of diethyl ether, methanol, 1 M citric acid (30:4:1), the mixture was centrifuged at 1500 X g, and the upper phase was spotted on 20 X 20-cm Silica Gel 60 plates which were preactivated for 1 h at 110 "C. Plates were developed in petroleum ether/diethyl ether/acetic acid (15:85:0.1, v/ v) for 50 min at -10 "C and then subjected to autoradiography for detection of products. TLC reference standards for 12-HETE and 12-HPETE were generated by incubation of ['4C]arachidonic acid with bovine epithelial 12-lipoxygenase as described previously (13).
Assay of GSH Leuel-Cellular GSH content was determined by methods developed by Tietze (28) and Griffith (23) and later modified by Liang et al. (29). Briefly, 100 pl of cell cytosol was mixed with 800 p1 of 0.3 mM NADPH in 125 mM sodium phosphate, pH 7.5, with 6.3 mM EDTA, 100 pl of 6 mM 5,5'-dithiobis(nitrobenzoic acid), and 20 p1 of 25 units/ml GSH reductase. The mixture was incubated at 30 "C, and the change in absorbance at 412 nm was used to calculate total GSH concentration by comparison to a standard curve. Oxidized GSH (GSSH) was determined using the same method after first derivatizing endogenous GSH with 4-vinylpyridine. Cytosolic GSH content was calculated as the difference between total and oxidized GSH levels.

RESULTS AND DISCUSSION
Arachidonate Oxygenation Products from Cultured Tracheal Epithelial Cells versus Subcellular Fractions from the Cells-Cultured ovine tracheal epithelial cells converted arachidonic acid to PGE2 (Fig. lA), as described previously (16, 19); however, the microsome-containing subcellular fractions derived from these same cells converted arachidonic acid to a metabolite identified as 12-HETE (Fig. 1B) In ( A ) , 2 X lo6 cells were incubated with 10 p~ [3H]arachidonic acid (2.2 X lo5 dpm/ml) in 3 ml of Hepesbuffered Hanks' balanced salt solution for 15 min at 37 "C. In ( B ) , 5 X lo7 cells were first disrupted in 50 mM Tris-HC1, pH 7.4, containing 5 mM EDTA and 1 mM EGTA; then mixtures were subjected to differential centrifugation and the resulting 100,000 X g pellet resuspended in the same Tris-HC1 buffer with 10 PM [3H]arachidonic acid for 15 min at 37 "C. Products were extracted and resolved using an acetonitrile (ACN)/water/acetic acid solvent system. Major peaks of radioactivity coeluted with PGE, (E2), 12-HETE (12H), and arachidonic acid ( A A ) .

TABLE I Subcellular fractionation and detergent extraction
of epithelial cell 12-lipoxygenme activity Cell suspensions (5 X IO7 cells/ml) were disrupted by sonication in 50 mM Tris-HC1, pH 7.4, containing 5 mM EDTA and 1 mM EGTA; the mixture was subjected to differential centrifugation at 400 X g, 12,000 X g, and 100,000 X g; the resulting pellets were resuspended in the same volume of the same buffer, and each fraction was assayed for 12-lipoxygenase activity as described in Fig. 1. The microsomal membrane-containing 12,000 and 100,000 X g pellets were also treated with 1% Triton X-100 for 1 h at 4 "C, then recentrifuged, and the resulting supernatants (Triton-soluble) and pellets (Triton-insoluble) were assayed for activity. Values represent the mean for two experiments.  (19). Subcellular fractionation experiments indicated that the 12-HETE-forming activity was found predominantly in the 100,000 X g pellet and to a lesser extent in the 12,000 X g pellet and was undetectable in the 100,000 x g supernatant (Table I). Because the 100,000 x gpellet contains only microsomes, the distribution of 12-HETE-forming activity suggests that the activity is microsomal; because the 12,000 X g pellet (under the present conditions for cell disruption) contains microsomes and other cellular organelles, localization of 12-HETE-forming activity in other cellular membranes (e.g. mitochondrial) cannot be excluded.
Additional experiments using increasing concentrations of detergent (0.01-5% Triton X-100) to extract the membranebound enzyme indicated that the 12-HETE-forming activity was completely extracted from epithelial cell membranes only with 1% Triton X-100 and that progressive increases in activity were accompanied by corresponding increases in extracted total protein (not shown). These results suggest that the 12-HETE-forming enzyme is an integral membrane protein. In addition, the fact that the cell disruption and fractionation experiments and the enzyme assays were performed in the presence of divalent cation chelation suggests that the enzyme responsible for 12-HETE formation is bound to the microsomal membrane by a non-calcium-dependent mechanism.
In contrast to the results with cultured ovine tracheal epithelial cells, experiments with freshly isolated ovine tracheal epithelial cells showed no evidence of a membranebound 12-HETE-forming activity (and only low and often undetectable levels of cytosolic 12-HETE-forming activity). It is therefore likely that the cell culture conditions used for the present experiments were responsible for induction of the microsomal 12-HETE-synthetic activity in cultured ovine tracheal epithelial cells (see below).

Evidence for a Lipoxygenase Mechanism for 12-HETE Formation in Membranes from Cultured Epithelial
Cells-The type of enzyme responsible for 12-HETE formation by epithelial cell membranes was established by determining the degree of enzymatic stereospecificity (detection of 12s-versus 12R-HETE) and the requirement for hydroperoxide generation (detection of 12-HPETE). Absolute stereochemistry of 12-HETE was assigned by coelution of epithelial cell-derived 12-HETE with authentic, stereochemically pure 12-HETE enantiomeric standards during chiral-phase HPLC (13, 27): analysis of the stereochemistry of the 12-HETE generated from epithelial cell membranes demonstrated that it consisted entirely of the 12s isomer (Fig. 2 A ) . Formation of 12-HPETE was determined by comigration of epithelial cell-derived 12-HPETE with authentic standards under assay conditions (incubation for 15 min at 25 "C and analysis of products by TLC at 4 "C) which permit detection of the labile hydroperoxide intermediate (13): analysis of radiolabeled arachidonate metabolites generated from epithelial cell membranes demonstrated that arachidonate was converted to 12-HPETE (Fig. 2B). The initial 12-lipoxygenase-catalyzed abstraction of the 10s-hydrogen of arachidonic acid followed by typical insertion of molecular oxygen on the opposite face of the substrate molecule at carbon-12 (antarafacial addition) leads to the formation of 12s-HPETE (30). Thus, our findings indicate that 12-HETE derived from cultured ovine tracheal epithelial cells is formed by the activity of an arachidonate 12-lipoxygenase and not by other HETE-forming enzymes such as cytochrome P-450 monooxygenases that typically generate a mixture of R-and S-enantiomers and do not form an obligatory hydroperoxide intermediate (27).
Evidence for 12-Lipoxygenase Sensitivity to Cytosolic Znhibition That Zs GSH-dependent-The results of subcellular fractionation studies (Table I) suggested the presence of a cytosolic inhibitor of the epithelial microsomal 12-lipoxygenase. Verification of a cytosolic inhibitory effect was obtained when recombining the microsome-containing subcellular fractions (12,000 or 100,000 X gpellets) with the cytosolic fraction A " , " (12,000 and 100,000 X g pellets) prepared from cultured tracheal epithelial cells were incubated with 20 p M ['4C]arachidonic acid for 15 min at 37 "C; products were extracted, converted to corresponding methyl esters, and analyzed using a hexane/2-propanol solvent system. Major peaks of radioactivity coeluted with 12s-HETE ( 1 2 s ) with a smaller amount of 15s-HETE (15s) and no detectable 12R-HETE (12R). In ( B ) , the microsome-containing fractions from cultured epithelial cells were incubated in increasing amounts (60, 15, and 4 pg of protein in lanes 1,2, and 3 ) in 100 mM Tris-HCI with 25 PM ["C]arachidonic acid for 15 min a t 25 "C. Products were extracted into acidified diethyl ether, and analyzed by TLC at -10 "C using a petroleum ether/diethyl ether/acetic acid solvent system. Radioactive bands correspond to 12-HETE (12H), 12-HPETE (12HP), and arachidonic acid ( A A ) .

Direct effects of cytosol, GSH, and 13-HPODE on membrane-bound 12-lipoxygenase, PGH synthnse/PGE isomerase, and cytosolic 12-lipoxygenase activities
The microsomal membrane-containing fractions (12,000 and 100,000 X g pellets) were prepared from epithelial cells and then were mixed with: (i) an amount of the cytosolic fraction equivalent to the proportions in starting material (l:l, v/v) or with half the amount of cytosol (1:2) or twice the amount of cytosol (2:l) as the starting material; (ii) GSH; or (iii) 13-HPODE. The resulting mixtures were assayed for 12-lipoxygenase and PGH synthase activities by incubation with 10 PM ['Hlarachidonic acid for 30 min a t 37 "C and products detected by reverse-phase HPLC. a Abbreviations are m12-lox and PG syn/isom (membrane-bound 12-lipoxygenase and PGH synthase/PGE isomerase from cultured ovine tracheal epithelial cells) and cl2-lox (cytosolic 12-lipoxygenase from freshly isolated bovine tracheal epithelial cells).
Combining the microsome-containing fractions with the cytosolic fraction in the same proportions as those isolated from the cell suspension (1:1, v/v) resulted in significant inhibition of 12-lipoxygenase activity, and mixing the membrane-bound 12-lipoxygenase with varying amounts of the cytosolic fraction resulted in a concentration-dependent alteration of activity (Table 11). The same inhibitory effect of cytosol on 12lipoxygenase activity was observed when detergent-solubilized enzyme (instead of resuspended microsomal membranes) was tested (not shown). Inhibition of 12-lipoxygenase activity was also achieved by treatment of membrane-bound 12-lipoxygenase with GSH (Table 11) in a range of concentrations (0.1-10 mM) equivalent to those found in cultured ovine tracheal epithelial cells (see below) as well as a variety of other cell types and tissues (23)(24)(25)29,31,32). Interestingly, the PGH synthase/PGE isomerase activities of cultured tracheal epithelial cells were increased (rather than decreased) by GSH and were markedly decreased (not insensitive) to lipid hydroperoxide treatment (Table 11). Thus, the net effect of cytosol on the PGH synthase/PGE isomerase pathway was to increase activity, in contrast to the cytosolic inhibition of microsomal 12-lipoxygenase activity. To test whether the microsomal 12-lipoxygenase sensitivity to GSH was distinct, we performed comparative experiments with the cytosolic 12-lipoxygenase which is expressed in freshly isolated tracheal epithelial cells from several animal species (33). Because freshly isolated ovine tracheal epithelial cells were a poor source of cytosolic 12lipoxygenase activity (as noted above), we utilized freshly isolated bovine tracheal epithelial cells to prepare cytosolic 12-lipoxygenase as described previously (13). The cytosolic 12-lipoxygenase exhibited only slight inhibition by the highest concentrations of GSH or cytosol prepared from cultured ovine tracheal epithelial cells (Table 11). Taken together, these findings imply that the membrane-bound 12-lipoxygenase expressed in cultured ovine tracheal epithelial cells is selectively susceptible to inhibition by a GSH-dependent mechanism.
GSH and GSH peroxidase are capable of reducing hydrogen peroxide and organic hydroperoxides, but direct evidence for regulation of oxidation-reduction metabolism by GSH in vivo is often lacking (22). Accordingly, we first tested the effects of the lipid hydroperoxide 13-HPODE and glutathione-depleting agents on the cytosolic inhibition of epithelial microsomal 12-lipoxygenase activity. Treatment of the cytosolic fraction with 13-HPODE, at a concentration that abolished the lag phase of the 12-lipoxygenase reaction (13,34) and had little direct effect on the final level of 12-lipoxygenase activity (Table I1 and Fig. 3A), was capable of markedly reversing the cytosolic inhibition of the activity (Table I11 and Fig. 3B). It was also possible to mimic the 13-HPODE-dependent reversal of cytosolic inhibitory activity by treating the cytosol with the glutathione-derivatizing agents 2-cyclohexene-1-one or Nethylmaleimide (Table 111). The effect of N-ethylmaleimide (which may be less specific than 2-cyclohexene-1-one) (24, 25) was due to derivatization of GSH, because the subsequent addition of 10 mM GSH to the assay buffer reversed the capacity of N-ethylmaleimide to restore 12-lipoxygenase activity (Table 111). These findings suggested that the cytosolic inhibition of microsomal 12-lipoxygenase activity depended on the endogenous balance between formation of lipid hydroperoxides (which stimulate the 12-lipoxygenase activity) and GSH-dependent depletion of the hydroperoxides (which leads to inhibition of the 12-lipoxygenase). GSH depletion or 13-HPODE treatment of the cytosolic fraction from cultured  cell sonicates ( B ) . Microsomal fractions or cell sonicates prepared from cultured tracheal epithelial cells were incubated with 10 p~ ['HI arachidonic acid for 2-45 min at 37 "C in the presence ( 0 ) or absence (0) of 1 p~ 13-HPODE, and the resulting metabolites were analyzed and quantified using reverse-phase HPLC as in Fig. 1. Values represent the sum of measurements for 12-lipoxygenase products (HETE and 13-HEpETrE); each value represents mean for two experiments. The 13-HPODE-dependent increases in 12-lipoxygtmase activity in cell sonicates were also observed in preparations of microsomecontaining fractions recombined with cytosolic fractions.

Effects of lipid hydroperoxide and GSH-depleting agents o n cytosolic inhibition of membrane-bound 12-lipoxygenase activity
The cytosolic fraction from cultured ovine tracheal epithelial cells was treated with 13-HPODE, 2-cyclohexene-l-one, or N-ethylmaleimide (with or without subsequent GSH). The treated (or untreated) cytosol was then mixed with epithelial cell membrane fractions (12,000 and 100,000 x g pellets) in a ratio of 2:1 (v/v, cytosol/ membrane fraction), and the resulting mixtures were assayed for 12lipoxygenase activity as in Table 11. Abbreviations as defined in Table 11. 13-HPODE (1 p M ) 2.82 0.14 * Values represent the mean for two experiments. Control values for epithelial cell membranes for these experiments were 2.45 and 0.07 nmol product/mg protein/30 min for ml2-lox and PG syn/isom, respectively. epithelial cells resulted in a concomitant increase in cytosolic 12-lipoxygenase activity (final ratio of cytosolic/microsomal 12-lipoxygenase specific activities of 1-21), indicating that the epithelial 12-lipoxygenase may be expressed in membranebound and soluble-cytosolic forms.
Direct evidence that modulation of cellular GSH and hydroperoxide levels influenced 12-lipoxygenase activity in vivo was finally provided when GSH depletion and lipid hydroperoxide supplementation of cultured epithelial cells markedly increased 12-HETE synthetic activity (Table IV). Treatment of cultured ovine tracheal epithelial cells with buthionine sulfoximine (an irreversible inhibitor of y-glutamylcysteine synthase) caused significant decreases in cellular GSH levels, increases in 12-HETE-forming activity, and magnified increases in activity evoked by lipid hydroperoxide. The pretreatment levels of GSH in cultured ovine tracheal epithelial cells (7.8 nmol GSH/107 cells) were in the middle of the range determined for other cell types and tissues (23)(24)(25)29,31,32).
T h e base-line value for GSH in cultured epithelial cells corresponds to an intracellular GSH concentration of 3.9 mM based on an intracellular water volume of 2 pl/107 cells (35) Cultured ovine tracheal epithelial cells were treated with huthionine sulfoximine for 24 h, 13-HPODE for 5 min, or with both reagents, and then assayed for 12-lipoxygenase activity as in Fig. 1A. BSO, buthionine sulfoximine; ND, not determined. See Table I1  and is well within the range of GSH concentrations used in the present experiments to inhibit microsomal 12-lipoxygenase activity (see above). The relatively small effect of GSH depletion alone (without hydroperoxide supplementation) on epithelial 12-lipoxygenase activity (Table IV) may reflect the difficulty in depleting intracellular stores of GSH (and other reducing agents) to a level that results in significant hydroperoxide tone (36). In comparative experiments, GSH depletion (by treatment with 2-cyclohexene-l-one, N-ethylmaleimide, or buthionine sulfoximine) of freshly isolated ovine tracheal epithelial cells caused no significant increase in 12-HETE-forming activity, and therefore (as noted above) provided no evidence for a cryptic 12-lipoxygenase (cytosolic or microsomal) in freshly isolated ovine cells. These findings again imply that the microsomal-type 12-lipoxygenase is selectively sensitive to GSH-dependent inhibition and selectively induced in cultured ovine tracheal epithelial cells.
In summary, the present characterization of arachidonate oxygenase activities in cultured ovine tracheal epithelial cells indicates that 12-HETE formation is catalyzed by a 12lipoxygenase. An unusual feature of this epithelial 12-lipoxygenase is that it may be membrane-bound. Thus, 12-lipoxygenase activity is more generally confined to the cytosolic fraction (e.g. in porcine or bovine granulocytes or in freshly isolated bovine tracheal epithelial cells) (1,5,13,34). Even when the 12-lipoxygenase is detected in membranes (e.g. in platelets), the localization may be highly dependent on increases in intracellular calcium concentration (37). The present observation for distribution of 12-lipoxygenase to the microsome-containing subcellular fractions under calciumfree conditions suggests an additional mechanism for enzyme localization in cultured epithelial cells. Our finding that the 12-lipoxygenase expressed in cultured tracheal epithelial cells is completely extracted from the membrane only with 1% Triton X-100 detergent suggests that this form of the enzyme may contain a hydrophobic domain which allows for alternative membrane localization. Purification studies are underway to determine the difference between this form of the epithelial Wlipoxygenase and the previous cytosolic forms described in tracheal epithelial cells, leukocytes, and platelets (33).
In addition to its unusual cellular location, the microsomaltype 12-lipoxygenase in cultured epithelial cells exhibits an unusual sensitivity to inactivation by the cytosolic fraction. A similar sensitivity to cytosolic inhibition was recently observed for a membrane-bound 12-lipoxygenase in A431 epidermoid carcinoma cells, but the mechanism for the inhibitory effect was not determined (38). The present findings indicate