_ .... QU-wnhg. ' a .. m 2 f). >~o‘-‘~. *‘- ‘HW ~ g 4'. ‘a‘ I, n ‘v '7 H [—258 - he... I...» My This is to certify that the thesis entitled SUBCELLULAR LOCALIZATION OF PGH SYNTHASE presented by Thomas Edmund Rollins has been accepted towards fulfillment of the requirements for M.S. degreein Biochemistry [Lg/{Claw /. £1124 Major professor Date4zm/ g x45 / 0-7639 MSU LIBRARIES .—_—. RETURNING MATERIALS: Place in bodk drop to remove this checkout from your record. FINES will be charged if book is returned after the date stamped below. © 1981 THOMAS EDMUND ROLLINS All Rights Reserved SUBCELLULAR LOCALIZATION OF PGH SYNTHASE By Thomas Edmund Rollins A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of Biochemistry 1981 ABSTRACT SUBCELLULAR LOCALIZATION OF PGH SYNTHASE By Thomas Edmund Rollins There were two phases to these studies. In the first phase the sub- cellular location of PGH synthase in Swiss mouse 3T3 cells was determined by electron microscopic immunocytochemistry using specific anti-PGH syn— thase IgG in the peroxidase anti-peroxidase staining procedure. Electron dense deposits were found associated with the endoplasmic reticulum and nuclear membrane in cells stained using immune but not preimmune IgG. In the second phase the transverse orientation of PGH synthase in the microsomal membranes of sheep vesicular glands was deduced on the basis of the susceptibility of the synthase to protease digestion and the ability of the enzyme to interact with Specific monoclonal antibodies. Protease treatment of intact sheep vesicular gland microsomes caused the destruction of cyclooxygenase activity. When the active site of PGH syn- thase was labelled with [3H]-acetylsalicylic acid and the microsomes incubated with protease, 90% of the tritium label was cleaved from the membrane. Three monoclonal antibodies which interact with different determinants in the pure PGH synthase were found to interact with the PGH synthase present in intact microsomes. The antigenic reactivity of the microsomal enzyme was unaffected by protease digestion. My experiments indicate (a) that the PGH synthase is localized on the cytoplasmic surface of the end0plasmic reticulum and nuclear memb- ranes, (b) that an active site fragment of PGH synthase can be cleaved from the membrane by treatment with proteinase K and (c) that three anti- genic sites on the PGH synthase are resistant to protease digestion. To Grae 11' ACKNOWLEDGEMENTS I would like to express my appreciation of the financial support and professional guidance given me by my major professor, Dr. William Smith. His ability to patiently guide me through my tenure here has made my stay enjoyable. I am grateful for the stimulating discussions and constructive criticism from my lab friends Frank Grenier, Arlyn Garcia Perez, Dave Dewitt, Jeff Day and Martha Kabalin. TABLE OF CONTENTS DEDICATION O O ...... O O O O O O O O O 0 ACKNOWLEDGMENTS ...... . . . . . . . . . . . TABLE OF CONTENTS 0 O O O O O O O O O O O O O 0 LIST OF FIGURES O ....... O O O O O O O O 0 LIST OF TABLES O O O O O O O O O O O O O O O O O ABBREVIATIONS . . . . . . . . . . . . ; . . . . INTRODUCTION 0 O O O O O O O 000000 O O 0 LITERATURE REVIEW . ...... . . . . . . . . Prostaglandin synthesis . . . . . . . . . . The conversion of arachidonic acid to PGH2 Metabolic regulation of PGH synthase . . . Pharmacological regulation of PGH synthase Conversion of PGHZ to other prostaglandins Prostaglandin catabolism . . . . . . . . . Mechanism of action of prostaglandins . . . . . sumary O O O O O O O O O O O O O O O O O O Membranes as functional boundaries . . . . . Components . . . . . . . . . . . . . . . . The endoplasmic reticulum . . . . . . . . . Topology in the transverse plane . . . . . Binding of membrane proteins ..... . . Page ii iv vii ix «30000 10 11 12 13 14 14 15 16 17 Topology in the lateral plane and lipid-protein interrelationships . . . . . . . . . . . . . . . Assembly of proteins into membrane . . . . . . The Membrane Trigger Hypothesis . . . . . . . . The Signal Hypothesis . . . . . . . . . . . . . Differences between the two models . . . . . . MATERIALS AND METHODS . . . . . . . . . . . . . . . Reagents . . . . . . . . . . . . . . . . . . . Microsomal preparation . . . . . . . . . . . . Cyclooxygenase assay . . . . . . . . . . . . . Antisera . . . . . . . . . . . . . . . . . . . Cell culture . . . . . . . . . . . . . . . . . Immunocytochemistry . . . . . . . . . . . . . . Precipitation of microsomal cyclooxygenase with Staphylococcus aureus-antibody complexes . . . Protease digestion of PGH synthase from sheep vesicular gland microsomes . . . . . . . . . . Protease digestion of 3sheep vesicular gland micro- somes labelled with [3 H]acetyl salicylic acid . Immunoradiometric assay of PGH synthase . . . . Determining microsomal integrity by the nannose-6- phosphatase assay . . . . . . . . . . . . . . . RESULTS . . . . . . . . . . . . . . . . . . . . . . Specificity of Anti-PGH synthase IgG . . . . . Immunocytochemistry . . . . . . . . . . . . . . Precipitation of microsomal PGH synthase with S. aureus-antibody complexes . . . . . . . . . . . Protease inactivation of PGH synthase from sheep vesicular gland microsomes . . . . . . . . . . 17 18 18 19 19 21 21 21 22 23 23 24 25 26 27 28 3O 32 32 35 42 42 Solubilization of a [3H]acetyl-labelled fragment by protease treatment of aspirin labelled microsomes Sensitivity of PGH synthase antigenic determinants to protease digestion . . . . . . . . . . . . . . . . Determination of microsomal integrity . . . . . . . . DISCUSSION 0 O O O O O O O O O O O O O O O O O O O O O 0 LIST OF REFERENCES vi 51 51 62 71 10 11 12 LIST OF FIGURES Prostaglandin metabolic pathway . . . . . . . . . Immunoradiometric assay for quantitating PGH syn- thase I O O O O O O I O O O O O O O O O O O O O 0 Mannose-6-phosphate assay scheme . . . . . . . . . Ouchterlony double-diffusion analysis of anti-PGH synthase serum and anti-PGH synthase 196 with pure sheep vesicular gland . . . . . . . . . . . . Ouchterlony double-diffusion analysis of IgG isolated from rabbit preimmune and anti-PGH synthase sera C O O O O O O O O O I O 0 O O O O O Photomicrograph of mouse 3T3 fibroblasts stained using anti-PGH synthase IgG . . . . . . . . . . . Electron micrograph of mouse 3T3 fibroblast using (A) anti-PGH synthase IgG and (B) rabbit pre- imune IgG 0 O I O O O I I O O O O O I O O I O O 0 Immunoprecipitation of PGH synthase from solu- bilized and intact sheep vesicular gland micro- somes by mouse monoclonal antibody7§. aureus complexes . . . . . . . . . . . . . . . . . . . . PGH synthase activities in microsomal and super- natant fractions following protease digestion of sheep vesicular gland microsomes . . . . . . . . . Release of 3H label from sheep vesicular gland microsomes preincubated with [3H] acetylsalicy- lic acid and then treated with proteinase K . . . Immunoradiometric assay of protease-digested sheep vesicular gland microsomes . . . . . . . . . Immunoradiometric assay of PGH synthase in super- natants obtained from protease- and albumin- treated sheep vesicular gland microsomes . . . . . vii 28 31 34 34 38 4O 44 48 50 53 55 14 15 P319: PGH synthase in membrane fractions obtained from immunoradiometric assay of protease- and albumin- treated sheep vesicular gland microsomes . . . . . . . . . 57 Mannose-6-phosphatase activity in mixed rat liv- er/sheep vesicular gland microsomes . . . . . . . . . . . 60 Model depicting the arrangement of PGH synthase in the endoplasmic reticulum . . . . . . . . . . . . . . . 68 viii Table 1 LIST OF TABLES Effect of protease digestion on PGH synthase activity in sheep vesicular gland microsomes . . ix PG PGH 20:4 BSA DDC PI ABBREVIATIONS prostaglandin prostaglandin endoperoxide arachidonic acid bovine serum albumin diethyl dithiocarbamate phOSphatidyl inositol INTRODUCTION The study of prostaglandin metabolism has expanded tremendously since the early 1960's. Competitive and noncompetitive inhibitors have been developed for the major prostaglandin (PG) biosynthetic enzymes and new, more specific inhibitors are being sought. A new class of compounds, the leukotrienes, which are related biosynthetically to the prostaglandins, have recently been identified. Still we know relatively little about the properties of the enzymes involved in prostaglandin synthesis or how bio- synthesis is regulated in the cell. The studies presented in this thesis have focused on determining the subcellular location of the PGH syn- thase. PGH synthase is the first enzyme of prostaglandin biosynthesis. The enzyme contains both cyclooxygenase and peroxidase activities (1-4). The cyclooxygenase catalyzes the oxygenation of fatty acids, such as arachi- donic acid, leading to the formation of the cyclic endoperoxide, PGGZ (5), Figure 1. The peroxidase activity uses PGGZ as a substrate reduc- ing the hydroperoxy group, at position 15, to the hydroxyl group of PGH2(3,4). The prostaglandin endoperoxide, PGHZ, is then converted by other enzymes in the pathway to the corresponding prostaglandins and thromboxanes. The cyclooxygenase activity has a limited substrate spe- cificity while the peroxidase catalyzes the reduction of a variety of structurally diverse hydrOperoxides, Figure 1. Figure 1. Prostaglandin biosynthetic pathway. STIWLUS PGH synthase is a membrane-bound, heme-containing glycoprotein with a subunit molecular weight of about 72,000 (1,2,4). The enzyme exists as a dimer with identical subunits. Each subunit binds one heme molecule which functions in both the oxygenase and peroxidase reactions. Mn++- heme can substitute for Fe++ heme in the oxygenase but not the perox- idase reaction (6). These findings are consistent with observations which indicate that specific inhibitors of PGH synthase block only the cyclooxygenase activity without affecting the peroxidase activity (1,7). Acetylsalicylic acid is a specific, irreversible inhibitor of PGH synthase and acts by acetylating an active site serine hydroxyl group (1). In this work I will demonstrate that after labelling the enzyme in sheep vesicular gland microsomes with [3H]acetylsalicylic acid that most of the label is solubilized by proteinase K treatment. There are essentially two approaches which can be used to determine the subcellular location of an enzyme. The first is to separate cellular organelles by differential centrifugation and then determine with which fraction the activity is associated (8,9). Contamination of one fraction of organelles with another fraction is a common problem in trying to localize enzymatic activity using this approach. A second approach is to use antibodies directed against specific antigens to localize cellular proteins and enzymes (10-18). In this thesis, I will show the Specifi- city of IgG against anti-PGH synthase and demonstrate its use in localiz- ing the PGH synthase by electron microscopy. The advantages of the im- munocytochemical approach are that it is not necessary to separate cellu- lar organelles to elucidate the location of a specific component and that if the antibody is monospecific there is little question of cross contam- ination. Monoclonal antibodies are Specific for single antigenic determinants (19). To generate monoclonal antibodies, Splenic lymphocytes from immu- nized mice are fused with myeloma cells to produce long-lived hydridoma cell lines. These lines produce large amounts of monospecific antibodies against the antigen of interest. Monoclonal antibodies can be used to identify the presence of a specific cell type or cell membrane character- istic (20), or to localize specific antigens intra- and extracellularly. Monoclonal antibodies may also be used in immunoaffinity chromatography to purify RNA or proteins. In this thesis I will describe the use of four monoclonal antibodies directed against the PGH synthase in localiz- ing and quantitating the enzyme. The studies in this thesis address three questions regarding the PGH synthase. (a) Where in the cell is the PGH synthase located? (b) On which side of the membrane is the enzyme located? and (c) What are the spatial relationships between the active site and the antigenic determin- ants on the PGH synthase molecule. LITERATURE REVIEW Prostaglandin Synthesis Prostaglandins are a complex group of oxygenated fatty acids which have been detected in nearly all mammalian tissues. Prostaglandins are not stored in tissues but are synthesized in response to stimuli that cause the release of free fatty acids. A variety of factors can induce fatty acid release including inflammatory stimuli (21,22), calcium iono- phores (23), tumor promoters (24), hormones (25), and mechanical agita- tion (26). The fatty acid released in response to specific stimuli (e.g. hormones) is generally 5,8,11,14-eicosatetraenoic acid, arachidonic acid. Free arachidonic acid acts as a substrate for PGH synthase, the first enzyme in the prostaglandin biosynthetic pathway. This enzyme converts arachidonic acid to the endoperoxide PGHZ, which is then converted to a variety of biologically active products, the nature of which is deter- mined by the enzyme content of the tissue in question. For example, P612 is found to be synthesized by endothelial cells while thromboxane A2 is synthesized by platelets. The pathways for formation of various prostaglandin derivatives appears in Figure 1. One can visualize prosta- glandin production as occurring in three steps. The first is stimulus and release of arachidonic acid. The second is the conversion of arachi- donic acid to P662 and PGHZ via the PGH synthase reaction. And the third phase is the conversion of PGH2 to the biologically active compounds. The release of arachidonate is believed to be the major con- trol point for prostaglandin synthesis. Two mechanisms exist for the release of free arachidonate from phos- phoglycerides. The first mechanism involves release of arachidonate from its normal esterified form at the 2-position of cellular phosphoglycer- ides through the action of a phosholipase A2. It is believed that a stimulus acts upon the cell surface to produce a second messenger which stimulates the action of the phOSpholipase A2 (27). It has been shown that activation of phospholipase A2 in response to a toxin results in a selective release of arachidonic acid in 3T3 mouse fibroblasts (28). A second mechanism has been pr0posed which involves thromboxane synthesis by platelets; the importance of this mechanism has not been established in other cell types (29,30). In this alternate pathway two enzymes work sequentially to release free arachidonate from phosphatidylinositol. The first enzyme, phospholipase C, cleaves phosphatidylinositol specifically to yield the diglyceride and inositol phosphate. The next enzyme, digly- ceride lipase, releases the fatty acid from the 2-position (29). In platelets it has been shown that nearly 90% of phosphatidyl inositol exists as I-stearoyl 2-arachidonoyl phosphatidylinositol. Therefore, the sequential action of these two enzymes results in a selective release of arachidonic acid by platelets (30). It has been shown recently that platelets stimulated with thrombin selectively release arachidonate as the major free fatty acid. Following release of 20:4, phOSphatidyl ino- sitol is resynthesized. This newly formed phosphatidylinositol has a fatty acid composition different from the parent molecule but then under- goes a deacylation-reacylation process which results in the unique I-stearoyl 2-arachidonoyl Species (30). How phospholipase C is stimulated by thrombin or other agents is unresolved, but mobilization of intracellular Ca++ is apparently involved (31,32). The conversion of arachidonic acid to PGH? The PGH synthase which contains two enzymatic activities converts arachidonic acid to P662 via a cyclooxygenase (33) and then to PGHz by a nonspecific peroxidase. This enzyme has been purified to electro- phoretic homogeneity (3). Most of the information to date comes from studies of the sheep vesicular gland enzyme, although the rabbit kidney enzyme has similar kinetic and antigenic properties (34). The first step in the synthesis of PGH2 involves a stereospecific protium abstraction from C13 of arachidonic acid (Figure 1). The abstraction is followed by rearrangement of a double bond between C11 and C12 to yield a carbon centered radical at C11 (34,35). Molecular oxygen then adds to the radical to yield an 11 hydroperoxy rad- ical. This radical attacks C9 and a series of rearrangements occur leading to the formation of a cyclopentane ring encompassing Cg through C12 and a new carbon radical at C15. A second molecule of oxygen reacts with the radical at C15 to yield a 15-hydroperoxy radical. The H atom on the hydroperoxyl group is initially from the protium abstraction at C13. The peroxidase activity along with a reducing agent acts upon this hydroperoxyl group at C15 of PGGZ to yield the PGHZ (36). P662 and PGH2 are short lived intermediates but nay function in vivo since both have biological activity (37,38). Metabolic regulation of PGH synthase Synthesis, degradation, feedback inhibition and availability of substrates are a few ways the PGH synthase is regulated. Availability of arachidonic acid is the rate limiting step. Heme is required for both the cyclooxygenase and peroxidase reactions. Therefore, the rate of for- mation of PGH2 is directly dependent upon how effectively PGH synthase competes for heme with other proteins. After synthesizing a limited number of endoperoxides, the enzyme itself becomes irreversibly inactivated (33). This will eventually lead to a decreased rate of production of prostaglandins unless a cell synthe- sizes new PGH synthase. The irreversible inactivation probably stems from the generation of free radicals during the formation of P662 and PGHZ; the oxidant generated may cause protein degradation. The resyn- thesis of PGH synthase is affected by various hormones and other stimuli (34). This self-catalyzed destruction of PGH synthase may be one mechan- ism by which PGH synthase levels are regulated. Pharmacological regulation of PGH synthase Nonsteroidal anti-inflammatory drugs selectively inhibit the PGH syn- thase. The action of nonsteroidal anti-inflammatory drugs on the cyclo- oxygenase necessarily regulates the synthesis of all subsequent com- pounds. The initial abstraction of the hydrogen atom from C13 of arachidonic acid is apparently inhibited by these drugs. Two distinct classes of inhibitors exist. The first class includes ibuprofen, flufen- amic and mefenamic aicd which are relatively weak, reversible inhibitors. Indomethacin, flurbiprofen and meclofenamic acid are more potent irrever- sible compounds as compared to the first class (35). All irreversible inhibitors cause a time-dependent inactivation of cyclooxygenase activ- ity. Phenolic compounds reduce the inhibitory capacity of irreversible 10 inhibitors but enhance inhibition by the reversible inhibitors (35). Effective removal of the reversible inhibitors by dilution, metabolism, or increases in fatty acid substrate concentrations restore cyclooxygen- ase activity. To recover PGH synthase activity following exposure to irreversible inhibitors requires new synthesis of the enzyme. The mech- anism of irreversible cyclooxygenase inhibition has not been resolved. Indomethacin and flurbiprofen, although irreversible inhibitors, cause no covalent modification of the cyclooxygenase. These agents must cause structural changes in the cyclooxygenase which may or may not be reversi- ble in vivo. Acetylsalicylic acid, another irreversible inhibitor does acetylate an internal serine hydroxyl on the enzyme (1). This covalent binding, however, does not strictly parallel enzyme inactivation (37). Conversion of PGH? to other prostaglandins PGFga, PGDZ, PGEZ, TxAz, and P612 are the major physio- logically active prostaglandins. All of these products with the excep- tion of PGan can be derived from PGHZ in a single enzymatic step. P602 is synthesized via a PGH-PGD isomerase (38,39), PGIz by prosta- cyclin synthase (40), thromboxane A2 by TxAz synthase and PGE2 via a PGH-PGE isomerase. It is not clear whether PGan is synthesized directly by reduction of PGH2 or indirectly from PGEZ through a 9-keto PGEZ reductase (41). All these enzymes are membrane bound with the exception of the PGH- PGDZ isomerase and the 9-keto PGEZ reductase. These membrane-bound enzymes probably occur in close association with the PGH synthase (42, 43). Only the PGH synthase and the PGH-PGD isomerase have been purified to electrophoretic homogeniety (3,38). 11 The PGH-PGD isomerase has a molecular weight of approximately 85,000. It is a cytosolic enzyme and is stabilized by thiol compounds, but thiols are not required for enzymatic activity (38,39). Distinct cell types seem to synthesize only one major type of prosta- glandin. For example, human platelets appear to form mainly TXAZ (44), bovine endothelial cells produce prostacyclin, P612 (45), and rabbit kidney collecting tubule cells produce mainly PGEZ (56). The reason why differentiated cells produce a single prostaglandin is not known. Although different stimuli can cause prostaglandin release by one given cell type, changing the stimulus does not affect the type of prostaglan- din formed. For example, endothelial cells stimulated by thrombin, tryp- sin, or Ca++ ionophore release only PGIZ. This suggests that endo- thelial cells have only PGIZ synthase and lack other enzymes that use PGHZ as a substrate. Apparently, cells are programmed during develop- ment to produce specific prostaglandin biosynthetic enzymes. In any giv- en cell the specific activity of enzymes which use PGHZ as a substrate are substantially higher than the specific activity of PGHZ synthase. Therefore, most of the PGHZ synthesized is probably transformed enzy- matically to a specific product in differentiated cells. Prostaglandin catabolism Prostaglandins are catabolized rapidly in 1139. TxAz is hydrolyzed nonenzymatically to Tsz with a t1/2 = 30 sec at 37°. This hydro- lysis is probably the major mechanism for inactivation of TxAz. Tsz has no known biological activity. There are Type I and Type II 15-hydroxy prostaglandin dehydrogenases which oxidize the hydroxyl group at position 15 of PGE2, PGDg, 12 PGan and P612 thereby converting these prostaglandins to inactive 15-keto forms. Type I utilizes NAD+ and type II utilizes NADP+ as the hydride acceptor (47). The type I dehydrogenase has been partially purified from swine kidney cortex while type II enzyme has been purified to homogeneity from swine kidney medulla. The 9-keto prostaglandin reductase is associated with the type II dehydrogenase (48). Another enzyme of prostaglandin catabolism is the 15-keto prostaglan- din DA13 reductase which reduces the double bond between C13 and 014. This enzyme has been purified from bovine lung (49). There also exists a 9-hydroxy prostaglandin dehydrogenase which oxidizes the hydroxy group at position 9 (50). Mechanism of action of prostaglandins All prostaglandins are probably effective within or near the cells in which they are synthesized. P612 is the only known prostaglandin that may act as a circulating hormone although this concept has been disputed (51). Prostaglandins act on different cell types to stimulate the action of adenylate cyclase to release cAMP (75). It is not clear, however, if prostaglandins can sometimes cause a direct effect or always act via a secondary messenger. One physiological model that has been proposed involves prostaglandin synthesis as one of the participants in the first line of defense in hemostasis. Platelets from the bloodstream and endothelial cells which line the vasculature release thromboxane A2 and P612, respectively, in response to blood vessel damage. Thromboxane A2 has a half life of approximately 30 seconds at 37° and is the most potent of the eicosanoids in contracting aortic tissue and triggering platelet aggregation. In 13 contrast, P612 is a vasodilator and prevents platelet aggregation. In response to a damaged vessel wall, blood platelets release a variety of products one of which is PGH2. PGHZ is thought to migrate to nearby endothelial cells which then convert PGHz to PGIZ. P612 increases adenylate cyclase activity in platelets raising intracellular CAMP lev- els; this, in turn, prevents further TxAZ production by inhibiting arachidonic acid release. This interplay between TxAz and P612 per- mits platelets to aggregate at injured sites on arterial walls, without occluding the vessel. In this situation one sees two different prosta- glandins from two distinct cell sources affecting a biological response. In many biochemical mechanisms it is not the effect of one prostaglandin which controls a response but the ratio of two compounds involved in the process. Summar Prostaglandins are “local" hormones which affect cells near their site of synthesis. The overall synthesis is controlled primarily at the level of release of arachidonic acid. Individual cell types capable of prostaglandin synthesis produce only one major type of prostaglandin in response to different stimuli. Degradation of prostaglandins leads to inactive biological compounds. Two types of nonsteroidal anti-inflamma- tory drugs now exist for the cyclooxygenase and new inhibitors of P612 synthase and TxAz synthase are being developed. These new inhibitors should help elucidate the biochemical mechanisms underlying the actions of these two hormones. 14 Membranes as functional boundaries The phOSpholipid bilayer is thought to be the basic component of which virtually all biological membranes are made. A number of theories exist to describe how proteins are associated with and assembled into the membrane. It is clear that there is an intimate relation between the membrane and the protein in terms of structure and function. This sec- tion of the review will discuss (a) how proteins are situated in the endoplasmic reticulum, (b) the interrelationships of membrane lipids and proteins and (c) current theories of how proteins are assembled into bio- logical membranes. Eukaryotic cells contain a variety of membranous organelles which may have arisen from prokaryotic ancestors (52). These membranes apparently exist to compartmentalize functions and are thereby play an important role in metabolic control (63). For example, the nuclear membrane of eucaryotic cells separates the nucleus from the cytoplasm and thus tran- scription from translation. Intracellular membranes also contain pro- teins of the electron tranSport mechanism and enzymes of steroid and phospholipid metabolism. Other membrane proteins may act to control electrochemical gradients across intracellular membranes (108). Components Biological membranes are composed of glycerophospholipids and choles- terol derivatives arranged in bilayers with polar head groups at the two surfaces. The lipid comprising the membrane seems to have no special structural relation to the membrane proteins (53). Two general types of membrane proteins occur. The peripheral proteins are hydr0philic but contain small nonpolar regions which bind to the membrane. These 15 proteins can be dissociated by treatment of membranes with high concen- trations of salt. In contrast, integral membrane proteins are more tightly associated with the membrane sometimes spanning the bilayer. This high degree of association is due mainly to the large number of hydrophobic sequences found in these proteins. Most lipids diffuse rather rapidly in the plane of the bilayer (54), while the movement of proteins is less rapid (55). There is little evidence that either phos- pholipids or proteins rotate transversely (flip-flop) in the membrane (56-58). Using purified specific phOSpholipases it has been found that phos- pholipid species are distributed assymetrically in the transverse plane of the membrane of microsomes. It is interesting to note here that in the rat liver endoplasmic reticulum most of the phosphatidyl inositol, PI, which is a prostaglandin precursor in platelets is found on the luminal side of the end0plasmic reticulum (62). In platelets, a Specific phospholipase C, utilizes PI along with a diglyceride lipase to release 20:4 for PGH synthase. Phosphatidylcholine, phosphatidylethanolamine, and triacylglycerol are all synthesized on the cytoplasmic surface of the endOplasmic reticulun (63). The phospholipids are then integrated into the membrane of the endoplasmic reticulum. The endoplasmic reticulum Endoplasmic reticulum is an intracellular network of tubules, vesi- cles, and lamellae (63). The functions of this organelle include pro- tein synthesis and transport, synthesis of phospholipids, cholesterol and triglycerides, and metabolism of xenobiotics. In a cell devoted to pro- tein synthesis, 19% of the total protein, 48% of the total phospholipid 16 and 58% of the total RNA is associated with the endoplasmic reticulum. The membrane itself is approximately 60-70% protein and 30-40% phospho- lipid by weight (64,65). At least 34 polypeptides have been identified with the endoplasmic reticulum of rat liver (66). Phospholipid composi- tion varies with cell type. For example, platelets contain relatively large amounts of phosphatidyl inositol while rat hepatocytes contain very little. The endoplasmic reticulum is easily and extensively disrupted by gen- tle homogenization. This disruption causes the endoplasmic reticulum to be pinched off into spheres of membrane called microsomes. These micro- somes are formed with their cyt0plasmic side out, luminal side in (67). Most proteins are located on only one side of the microsomal sphere although there are a few examples of proteins which traverse the bilayer. It is relatively easy to study proteins present on the cytoplasmic side of the microsomes. Microsomes are impermeable to exogenous proteins and even small charged unlecules. Topology in the transverse plane To ascertain the transverse distribution of an enzyme in the endo- plasmic reticulum one can determine whether that protein, as it exists in intact microsomes, is accessible to proteases, antibodies and other probes to which the membrane is impermeable (61). This impermeability should never be assumed, but tested with each probe to be used. Microsomes have been shown to be impermeable to uncharged molecules greater than 600 molecular weight and to charged substances with molecu- lar weights as low as 90. Treatment of intact microsomes with proteases does not destroy the permeability of the membrane (68). All microsomal 17 enzymes that have been investigated have been f0und to have their cata- lytic activity associated with only one side of the membrane. This sug- gests that all proteins are arranged in the membrane with specific orien- tations. Binding of membrane proteins The membrane-bound protein which has been investigated most exten- sively is cytochrome b5. It is an integral membrane protein of the endoplasmic reticulum (69-72). In studies involving protease treatment of rat liver microsomes, the cytochrome b5 has been feund to have a catalytically active, hydr0philic head and a noncatalytically active hydrophobic tail. Detergent-solubilized cytochrme b5 has a molecular weight of 16,700 while a protease-derived cytochrome b5 has a molecular weight of 11,000. The difference in size is due to a relatively hydro- phobic amino terminus having 44 amino acid residues (73). The hydropho- bic segment of the cytochrome b5 is thought to anchor the protein in the endoplasmic reticulum. The other portion of the molecule extends into the cytoplasm where it can interact with substrates. NADH-cyto- chrome b5 reductase (70) and NADPH cytochrome C reductase also bind to the endoplasmic reticulum in a manner similar to that found for cyto- chrome b5 (59). Topology in the lateralgplane and lipid-protein interrelationships Techniques of subfractionation (59), freeze fracture electron micros- copy (60), and reconstitution of isolated components have been used to study enzyme topology in the lateral plane of the membrane bilayer. 18 These techniques allow one to investigate Specific associations between membrane proteins and the phospholipids surrounding them. Prior to 1978 phospholipids were thought to be relatively inert, serving as a support matrix for membrane-bound protein. Hirata and Axel- rod showed that phosphatidyl ethanolamine is methylated on the inside of the membrane of erythrocyte ghosts by a methyl transferase I and trans- ported to the outside of the membrane by methyl transferase II, which also methylates CH3-phosphatidyl ethanolamine to dimethyl-PE and final- ly to phOSphatidyl choline. They proposed that a specific interaction of phospholipid and protein is essential fer these processes to occur (74). Assembly of proteins into membrane After mRNA is transcribed in the nucleus, it is processed and moves to the cytoplasm to associate with ribosomes for translation into a unique protein. Two theories have been proposed which try to explain the process of translation of mRNA into membrane bound proteins. Wickner's theory is known as the membrane trigger hypothesis (76), whereas Blobel and Milstein's theory is called the signal hypothesis (77). The Membrane Trigger Hypothesis Using this hypothesis Wickner emphasizes the ability of a membrane lipid bilayer to trigger the folding of a polypeptide into a conformation that spans the bilayer or is at least associated with it. Information encoded in the N-terminal region of the nascent peptide is thought to activate the protein for membrane assembly by altering the fOTding path- way. Membrane binding to the nascent protein triggers the protein to expose hydrophobic residues to the bilayer. This may or may not occur 19 before the protein is completely synthesized. Finally the N-terminal sequence is removed proteolytically. The Signal Hypothesis In the model proposed by Blobel and Milstein, a protein destined to be membrane bound has a hydrophobic N-terminal signal sequence that causes it and the ribosome to which it is bound to bind to a specific protein transport channel. The force of polypeptide chain elongation drives the chain through the bilayer. Once the N-terminal sequence has traversed the bilayer it is then cleaved proteolytically leaving a mem- brane-bound peptide. Differences between the two models The membrane trigger hypothesis differs significantly from the signal hypothesis. The trigger hypothesis does not require a protein transport channel or concommittant protein synthesis and transport or even that the ribosomes be bound to the endoplasmic reticulum during protein synthesis. However, according to Blobel's model, protein is synthesized and inte- grated into the membrane at the same time. The signal hypothesis has been useful in describing the synthesis of secretory proteins (78-80). However, the model appears to be of limited value in explaining the synthesis of membrane-bound proteins. It has been shown in the case of the M13 coat protein of bacteriophage that translocation of a peptide does not have to occur during protein synthe- sis (81). Also some proteins are pleiotropic, spanning the bilayer sev- eral times, as is the case for bacteriorhodOpsin (82). It is hard to 20 imagine a protein transport channel which allows a protein to cross the bilayer several times. The membrane trigger hypothesis states that protein transport chan- nels do not exist but that interaction of the nascent sequence with the bilayer is required for folding the protein into a conformation that eas- ily binds the membrane. The way to resolve the differences between the two models is to determine if protein transport channels exist. This can be done geneti- cally by selecting various temperature selective mutants whose nembranes do not effectively transport or bind nascent proteins. One should then be able to genetically map the deleted gene responsible for the protein channel (if indeed it exists), to prove the existence of the protein responsible for transporting nascent peptides through the membrane. MATERIALS AND METHODS Reagents Flurbiprofen was obtained from the UPJOHN Company. Flufenamic acid was obtained from Aldrich. Acetylsalicylic acid, trypsin, bovine serum albumin (99% fatty acid free), polyoxyethylene sorbitan monolaurate (Tween 20), sodium diethyldithiocarbamate, mannose-6-phosphate, 2-mercap- toethanol, 3,3'-diaminobenzidine and lysine were obtained from Sigma Chemical Co., St. Louis, MO. Proteinase K was purchased from E.M. Bio- chemicals. Collagenase was purchased from Worthington Biochemicals and thermolysin was obtained from Boehringer-Manheim. Arachidonic acid was obtained from Nu-Check prep. Goat anti-rabbit whole serum, goat anti- rabbit IgG, and peroxidase anti-peroxidase (rabbit) were purchased from Miles Research Products. Paraformaldehyde, Araldite 502, dodecenyl suc- cinic anhydride, 2,4,6-tri-(dimethylaminomethyl)phenol, osmium tetroxide, Epon 812, and propylene oxide were from Electron Microscopy Sciences. Dulbecco's modified Eagle media, fetal calf serum, antibiotic-antimycotic (100x), and glutamine were obtained from Grand Island Biological Co. Protein A-Sepharose was from Pharmacia Fine Chemicals. Other chemicals were reagent grade obtained from common commercial sources. Microsomal preparation Sheep vesicular glands and rat liver were obtained from freshly slaughtered animals, frozen as whole tissue on dry ice and stored at 21 22 -80°C. No appreciable loss in cyclooxygenase activity occurred over time. Both sheep vesicular gland and rat liver were handled in the same way. Preparation of microsomes was performed at 4°. Tissue was weighed and then homogenized in 0.1 M tris-chloride, pH 7.4, containing 20 mM diethyldithiocarbamate at a tissue to buffer ratio of 1:10 (w/v). Tis- sues were homogenized using a Polyton PCV 1 for 3 minutes at full Speed. An initial centrifugation was performed at 12,000 x g for 15 minutes. The resulting supernatant was decanted and centrifuged at 200,000 x g for 35 minutes. The pellet was resuspended to a concentration of 10 mg pro- tein/ml in starting buffer. Protein concentrations were determined using a Coomassie blue assay (83). Cyclooxygenase assay Cyclooxygenase activity was measured at 37° on a YSI oxygen monitor with a voltage offset control described by Smith and Lands (84). Ali- quots of a suspension of sheep vesicular gland microsomes were added to oxygen electrode chambers containing 300 nmoles of arachidonic acid, 3 nmoles of phenol and 2 nmoles of bovine hemoglobin in a final volume of 3 ml of 0.1 M tris-chloride, pH 8.0. Negative controls were performed by measuring the rates of oxygen uptake in the presence of two specific inhibitors, Flurbiprofen, or flufenamic acid (85), (10'4 M). Further confirmation that oxygen uptake was a direct measure of PGH synthase comes from the observation of the self-catalyzed destruction phenomenon characteristic of the enzyme (1). One unit of cyclooxygenase is defined as that amount of enzyme that catalyzes the uptake of 1 nmoles of 02 per minute per ml of assay solution at 37°. 23 Antisera Rabbit preimmune and rabbit anti-cyclooxygenase sera were prepared as described previously (87). Rabbit IgG was isolated from both immune and preimmune sera using column chromatography on Protein A-Sepharose (88). Serum (5 ml) was applied to a column (1.1 x 4 cm) equilibrated at 4°C with 0.1 M sodium phosphate, pH 7.0. The column was washed with 20 vol- umes of equilibration buffer. The IgG fraction (20 to 35 mg) was then eluted with I'M acetic acid, and the eluant was neutralized with concen- trated NH40H. Isolated IgG was dialyzed overnight at 4° against 100 volumes of 0.1 M sodium phosphate, pH 7.0, diluted to a concentration of 2 mg/ml, and stored at this concentration in 0.1 sodium phosphate, pH 7.0, at -20°C. The protein concentration was based on an €E%§0= 1.45 for IgG (86). Prior to use for immunocytochemistry, IgG was routinely diluted with 0.1 M sodium phosphate, pH 7.2, containing (0.9% NaCl phos- phate-buffered saline) to a concentration of 5 to 50 pg/ml. Ouchterlony double diffusion analyses were performed in petri dishes containing 1.5% Bacto-agar. Cell Culture Swiss mouse 3T3 fibroblasts (ATCC CCL 92) obtained from American Type Culture Collection were grown in Dulbecco's modified Eagle media at 37° under a water-saturated 10% 002 atmosphere. Sterile transfers were performed following detachment of cells from flasks with 1% trypsin (GIBCO, 1/250) and 0.02% EDTA in phosphate-buffered saline solution, pH 7.2, and washing in Dulbecco's modified Eagle media containing 10% fetal calf serum. Cells used for staining were routinely grown on glass micro- scope slides or in polystyrene culture dishes (35 x 100 mm). 24 Immunocytochemistry Mouse 3T3 cells were seeded at a concentration of 106 cells/150 mm2 and grown in culture dishes for 24 h. Cells were then rinsed with phosphate-buffered saline to remove excess media and fixed for 4 h at 24°C in a freshly prepared solution of 10 EM Na104, 75 EM lysine, 2% paraformaldehyde, and 37 mM sodium phosphate, pH 7.4 (89), containing 0 to 0.1% Tween 20. Fixed cells were subjected to three 15 min washes in phosphate-buffered saline pH 7.0, and then overlayed with IgG isolated from either rabbit anti-cyclooxygenase or preimmune sera at a final con- centration of IgG of 5 to 50 ug/ml. All washes and dilutions of antisera were performed with 0.1_M sodium phosphate, pH 7.2, containing 0.9% NaCl. After incubation for 2.5 h at 24°, the cells were washed four times, 15 min each time. The cells were then incubated for 2 h at 24° with goat anti-rabbit IgG (1:10 dilution) and subsequently washed for 15 min four times. Peroxidase rabbit anti-peroxidase-soluble complex (1:50 dilution) was added to the cells which were then incubated for 2 h at 24°C and washed. Washed cells were incubated for 10 min with a solution contain- ing 0.3 EM 3,3'-diaminobenzidine and 0.3 EU H202 in 0.05 M_tris- chloride, pH 8.0 (90). The stained cells were subjected to four 15-min washes and then post- fixed in 1% 0504 for 2 h at 24°. The cells were washed and then dehy- drated by sequential exposure to 50%, 70%, 80%, 90% and 100% (three times) solutions of ethanol. After the third treatment with 100% ethan- ol, the cells were removed from culture dishes by scraping with a rubber policeman, transferred to 4-dram screw top vials, and collected by cen- trifugation at 500 x g for 2 min. Cell pellets were resuspended by agi- tation with propylene oxide and then mixed with an equivalent volume of 25 Epon-Araldite resin (Epon 812:Araldite 502:dodecenyl succinic anhydride, 25:20:60, v/v/v) and agitated for 4 h at 24°. Extra resin was then added to provide a ratio of resin to propylene oxide of 2 and the mixture agi- tated overnight. The cells were then collected by centrifugation and resuspended in Epon-Araldite resin to which had been added 0.024 volume of 2,4,6-tri-(dimethylaminomethyl)phenol. The samples were placed in Beem capsules and incubated for 72 h at 60°. The capsules were sectioned and sections examined and photographed using a Phillips model 201 trans- mission electron microscope. The sections were not counterstained with heavy metals. Kodak EM4463 film was used for photography. Light photo- microscopy of cells cultured, fixed, and stained on glass slides was per- formed with a Leitz Orthoplan microscope using Kodak TriX Pan film (ASA 28). Precipitation of microsomal cyclooxygenase with Staphylococcus aureus- antibody complexes Microsomes were prepared as described above. Attenuated Staphylococ- cus aureus (Cowen I strain) were prepared essentially as described by Kessler (109) and stored as 10% suspensions at -80°. Immediately prior to use S. apggpg cells were washed twice in 0.1 M tris-chloride, pH 7.4 containing 5% BSA and 1% Tween 20, centrifuged at 500 x g for 5 minutes after each wash and resuspended in 0.1 M tris-chloride, pH 7.4. The cells were centrifuged again and resuspended in 0.1 M tris-chloride, pH 7.4. Anti-PGH synthase monoclonal IgG molecules secreted by clones Exp-1,.gyp-5,Igyp-7 or control 23-3 IgG were linked noncovalently through the FC region of the molecule to the Fc binding Protein A which is present on S. aureus cell surface. The antibody S. aureus complexes were 26 then mixed with either intact or detergent-solbuilized (1% Tween 20) microsomes prepared from sheep vesicular gland, and the mixture centri- fuged to pellet S. pppgpg cells. Both the supernatant and pellet then were assayed for cyclooxygenase activity. The rationale underlying these experiments is that anti-PGH synthase IgG molecules will interact with antigenic sites exposed to the outside but not on the inside of microsom- al Spheres and that microsomes formed during homogenization of the endo- plasmic reticulum are formed such that the cyt0plasmic surfaces faces outward (91). Thus, precipitation of the cyclooxygenase activity of intact microsomes only can occur if antigenic sites on the enzyme face the cytoplasmic side of the endoplasmic reticulum. Experiments performed with solubilized enzyme serve as positive controls. Protease digestion of PGH synthase from sheep vesicular gland microsomes In these experiments, I determined if the cyclooxygenase as it exists in intact microsomes, was affected by digestion with various proteases. Microsomes were prepared in the usual manner. Cyclooxygenase activity was determined as described above. Proteinase K, collagenase, thermoly- sin, trypsin and a control, bovine serum albumin (BSA), were dissolved at a concentration of 1 mg/ml in 0.1 M tris-chloride, pH 8.0, containing 1 .mM_CaCl2. A sample (50 pg) of each of these solutions was added to 1.0 ml aliquots sheep vesicular gland microsomes at 24° and incubated for 45 minutes. Cyclooxygenase activity was determined at zero and 45 min. After 45 min, preparations were cooled to 4° and subjected to centrifuga- tion at 50,000 x g for 75 min. The resulting pellet was resuspended in 1.0 ml of 0.1 M_tri-chloride, pH 7.4, containing 20 EM 27 diethyldithiocarbamate (DDC) and both supernatant and pellet fractions were assayed for cyclooxygenase activity. Protease digestion of sheep vesicular gland microsomes labelled with L3Hjacetyl salicylic acid [3H]Acetylsalicylic acid, labelled in the acetyl moiety (50 Ci/mole) (92), was added at a concentration of 0-1.Efl.t0 0.1 ml of sheep vesicular gland microsomes (8-12 mg/ml protein), and allowed to incubate 15 minutes at room temperature. The microsomes were then cooled to 4° and unlabelled acetylsalicylic acid added in excess (1 mM) to prevent further incorporation of radiolabel. The mixture was centrifuged at 50,000 x g for 75 minutes to precipitate the microsomes. The supernatant was discarded and the pellet resuspended in 0.1 ml of 0.1 M_tris-chlor- ide, pH 7.4, containing 20 pM DDC, centrifuged again as before and resus- pended in its original volume. Proteinase K (100 pg) was added to 0.5 ml of sheep vesicular gland microsomes (10 mg/ml), which had been prelabeled with [3H]acetylsalicylic acid. The digestion was stopped at 0,5,10,45 and 90 minutes following addition of protease by cooling the samples to 4°. Cyclooxygenase activity was determined in a parallel sample pre- treated without acetylsalicylic acid. As expected, the aspirin-acetylat- ed microsomes showed no cyclooxygenase activity (1). Samples from each time point from protease digestion were centrifuged at 50,000 x g for 75 min. The pellet was carefully separated from the supernatant and resus- pended in its original volume. All pellets and supernatants were placed in liquid scintillation vials along with 7.0 ml of Bray's solution, vor- texed and counted on a Searle Analytical liquid scintillation counter. [3HJAcetyl salicylic acid was checked for purity by thin layer 28 chromatography. Selectivity of aspirin labelling of microsomes was as- sessed by determining if Flurbiprofen, flufenamic acid, or salicylic acid (10'4 M) prevented the incorporation of tritium into the microsomes. Immunoradiometric assay of PGH synthase Hybridoma line pypr3 secretes an 1961 which interacts with PGH syn- thase but does not bind S, ppggpg cells. This IgGl was isolated from culture media, labelled with 125I-Bolton-Hunter reagent (93) and used in an immunoradiometric assay for quantitating PGH synthase as fol- lows (Figure 2): fixed amounts of 1251-1961 were incubated with fixed amounts of a precipitating complex prepared by affixing Ing secreted by either gyp-l, Exp-5, 219‘7 or 2973 to attenuated S. apggpg cells. Antigen (PGH synthase) was added at various concentrations to generate a standard curve. The amount of precipitated (cell bound) 1251 was determined by gamma counting in a Beckman Biogamma. A lin- ear relation between precipitated 1251 and added sheep vesicular gland microsomes exists over a range of 0.005-0.005 units (0.17-1.7 ng), when using pyp-S or gyp-I S. aureus complexes. protein A S. aureus e—l’f’ ._, s“~ synthase IgG] Ing (cyo-3, 125I) 1. 212'7 2. pypeS 3..gyo-1 ' 4. 1973 Figure 2. Immunoradiometric assay for quantitating PGH synthase PGH 29 An experiment to determine if the antigenic activities of PGH syn- thase were destroyed by protease digestion was performed. Samples of microsomes (1.0 ml) prepared from sheep vesicular gland were treated as follows (a) proteinase K (50 pg) at 4° (b) BSA (50 pg) at 25° or (c) pro- teinase K (50 pg) at 25°. Cyclooxygenase activity was determined at 0, 45, 90 and 360 min. All samples were then diluted based on the original cyclooxygenase activity for measurement of antigenicity using the immuno- radiometric assay. S. ppggpp anti-PGH synthase antibody complexes were prepared as described above. Solubilized microsomes (1% Tween 20 v/v) from all 3 samples were diluted into 100 pl 0.1 M tris-chloride, pH 7.4, 1% Tween 20 to contain the equivalent of 0.0-0.05 units of cyclooxygenase activity. IS..apggps-Ig62 complex, (0.01 ml; sufficient to bind 3 units of cyclooxygenase), was then added, followed by 0.01 ml of 125I labelled 1961 (219-3), containing 30,000 cpm of 1251. The assay samples were incubated overnight at 4°. Pellets were collected by cen- trifugation, washed once using 0.2 ml of assay buffer and recentrifuged. The supernatants were removed by aspiration and tubes containing the pel- lets were inserted into vials and counted using a Beckman Biogamma gamma counter. A second experiment was performed in the same manner as the previous experiment except that after treatment of microsomes for 6 h the samples were centrifuged to prepare microsomal and supernatant fractions. The centrifugation was at 50,000 x g for 75 minutes at 4° and supernatant and microsomes carefully separated. The samples were diluted as described above and the immunoradiometric assay was performed on both fractions. 30 Determining microsomal integrity by the mannose-6-ph05phatase assay I found that sheep vesicular gland microsomes had no detectable man- nose-6 phosphatase activity. An alternative procedure outlined in Figure 3 was used to determine the permeability of the microsomes. Rat liver (2 g), was mixed with sheep vesicular gland (2 9) prior to preparing micro- somes. Mannose-6-phosphatase activity of the rat liver microsomes was used as marker for the luminal surface of the membrane. Mixed rat liver- sheep vesicular gland microsomes were diluted to 10 mg protein per ml based on Coomassie blue protein assays. Proteinase K (50 pg) was added to 1 ml of the mixed microsome preparation and incubated as in the previ- ous experiment. In a control sample BSA (50 pg) substituted for the pro- tease. The incubation was allowed to proceed for 6 h at 24° and then cooled to 4°. Mannose-6-phosphatase and cyclooxygenase activities were determined in each sample. Mannose-6-phosphatase was assayed as follows. The protease- and albumin-treated samples were each split into two frac- tions; one was solubilized with 1% sodium taurocholate and the other sam- ple was left untreated (Fig. 3). Mannose-6-ph05phatase was assayed at 24° by adding 50 pl of a microsomal preparation to 1.0 ml of 0.1 M tris- chloride, pH 6.6 containing 2‘ M mannose-6-phosphate and 10 my B-mercap- toethanol. The reactions were allowed to proceed for 0,5,10,15 and 20 minutes and then stOpped by addition of 0.25 ml of 10% trichloroacetic acid. After the assay was complete, 0.25 ml of 5 N_HZSO4, 0.5 ml of ammonium molybdate and 2.5 mg of reducing reagent (sodium sulfite, sodium bisulfite, 1-amino-2-napthol-4-sulfonic acid, 1.2/1.2/O.2, W/W/W)) were added and vortexed after each addition in a modified Fiske-Subbarow meth~ od (94). Absorbance was recorded on a Hitachi double beam spectrophoto- meter at 650 nm after ten min. 31 2.0 g rat liver mixed with 2.0 9 sheep vesicular gland MICROSOMES; as described in text. 10 mg/ml: protein/0.1 M tris-chloride, pH 7.4 20 mM DDC. separate into 2, 1.0 ml fractions u: up 1.0 ml microsomes 1.0 ml microsomes 50 pg proteinase K 50 pg BSA Digest 6 h at 24° and split both A, B into two fractions after centriguation at 50,000 xg for 75 minutes. Discard supernatant. Resuspend in original volume. B /\ / 0.5 ml microsomes 0.5 ml microsomes 0.5 ml microsomes 0.5 ml microsomes +proteinase K +proteinase K -proteinase K -proteinase K +1% Na tauro- -Na taurocholate +Na taurocholate -Na taurocholate °"°'“\ \ / / Mannose-6-phosphate assay Figure 3. Procedure for monitoring integrity of microsomal membranes. RESULTS Specificity of Anti-PGH synthase IgG We demonstrated previously that rabbit anti-PGH synthase serum is monospecific for the PGH synthase of sheep vesicular gland and that pre- immune serum does not react with the enzyme (87). Ouchterlony double diffusion analysis (Figure 4) shows that both anti-PGH synthase serum and IgG isolated from the serum by chromatography on Protein A-Sepharose give single lines of precipitation with the purified sheep vesicular gland PGH synthase; furthermore, the reaction of identity between the two lines indicates that both immune serum and immune IgG interact with the same set of antigenic determinants on the enzyme. As expected, no immunopre- cipitation lines were formed between the well containing the PGH synthase and wells containing either rabbit preimmune IgG or immune 190 which had been adsorbed with PGH synthase (Figure 4). Since the preparation of PGH synthase used to adsorb the anti-PGH synthase 196 is homogeneous by sodi- um dodecyl sulfate gel electrOphoretic criteria (3), adsorption of the immune IgG with the purified enzyme will remove only those IgG molecules which interact specifically with the PGH synthase. Therefore, preimmune IgG and enzyme-adsorbed hnnune IgG provide appropriate negative controls for immunocytochemical staining with anti-PGH synthase 198. The purity of the anti-PGH synthase IgG was assessed by double diffu- sion analysis by comparing the reactions of both anti-PGH synthase serum and IgG with goat anti-rabbit whole serum and with goat anti-rabbit 196 32 Figure 4. Figure 5. 33 Ouchterlony double diffusion analysis of the interaction of anti-PGH synthase serum and anti-PGH synthase IgG with sheep vesicular gland PGH synthase. Center well, purified PGH syn- thase (5 pg (3)); well 1, IgG purified from rabbit anti-PGH synthase serum by Protein A-Sepharose chromatography (138 pg); ygll_2, goat anti-rabbit 190 (1:1 dilution); well 3, anti-PGH synthase 196 (138 pg) mixed with purified PGH syn- thase (2 pg); well 4, partially purified IgG from rabbit pre- immune serum (50 pg); well 5, goat anti-rabbit whole serum (1:1 dilution); well 6, rabbit anti-PGH synthase serum (1:1 dilution). All dilutions were with 0.1 M sodium phosphate, pH 7.0. The initial volume in all wells was 25 pl. Ouchterlony double diffusion analysis of the purity of IgG isolated from rabbit preimmune and anti-PGH synthase sera. Center well, goat anti-rabbit whole serum (1:4 dilution); well 1, IgG isolated from rabbit anti-PGH synthase serum (50 pg); yell 2, rabbit anti-PGH synthase serum (1:4 dilution); wells 3 and 6, goat anti-rabbit IgG (1:4 dilution); ggl_;4, IgG isolated from rabbit preimmune serum (50 pg); well 5, rabbit preimmune serum. All dilutions were with 0.1 M sodium phosphate, pH 7.0. The initial volume in all wells was 25 pl. 35 (Figure 5). The anti-PGH synthase IgG gives a single line of immunopre- cipitation with both goat anti-rabbit whole serum and anti-rabbit IgG with a reaction of identity present at the intersection of the lines. In contrast, multiple lines of precipitation are apparent between the well containing anti-PGH synthase serum and that containing goat anti-rabbit whole serum. Thus, the only serum proteins present in the anti-PGH syn- thase IgG preparation are IgG molecules. Analogous results were obtained in testing the purity of the IgG isolated from preimmune serum (Figure 5). Immunocytochemistny Swiss mouse 3T3 cells grown on glass slides and then quick frozen in isopentane (-70°C) and air-dried could be stained for cyclooxygenase antigenicity with 196 (6.20 to 50 pg of IgG/ml) isolated from rabbit anti-PGH synthase serum using the peroxidase-anti-peroxidase procedure developed by Sternberger and co-workers (95). Cells processed in this manner were used initially to determine what fixation conditions could be employed to retain PGH synthase antigenicity for subsequent electron microscopy. Cells incubated for as long as 4 h at 24°C using the peri- odate/lysine/paraformaldehyde fixative developed by McLean and Nakane (89) stained positively for the PGH synthase. Although fixation with glutaraldehyde provided better retention of cellular ultrastructure, treatment of 3T3 cells with 0.05% glutaraldehyde in 0.1 M sodiwn phos- phate, pH 7.2, for 5 min at 4° completely abolished PGH synthase immuno- reactivity. Therefore, the periodate/lysine/paraformaldehyde fixative was used for all subsequent immunocytochemistry. 36 In order to stain unfrozen cells, it was necessary to include Tween 20 at concentrations of 0.03 to 0.1% in the fixation solution. This treatment permits penetration of immunolabeling reagents without causing extensive solubilization of integral membrane proteins (96-98). Tween 20 has no major effect on the catalytic or immunochemical properties of the PGH synthase (2,3). Figure 6 shows a photomicrograph of 3T3 cells which were subjected to immunocytochemical staining after fixation in peri- odate/lysine/paraformaldehyde solutions containing 0.05% Tween 20. PGH synthase antigenicity is apparent as dark rings around nuclei and can also be seen somewhat diffusely in the cyt0plasm. Significantly, no staining of the plasma membrane was observed. When IgG isolated from the preimune serum was substituted for the immune IgG at equivalent concen- trations, no staining was observed, nor did staining occur with immune 196 which had been preincubated with homogeneous (3) PGH synthase (100 pg of 190/15 pg of PGH synthase for 30 min at 24°). These latter two results confirm that the staining noted in Figure 6 is actually due to the selective interaction of anti-PGH synthase IgG with the enzyme. Our studies by light microscopy coupled with the results of previous subcel- lular localization studies which employed differential centrifugation techniques suggested that the PGH synthase is associated with the endo- plasmic reticulum and the nuclear membrane. To verify these interpretations, we extended our work to the ultra- structural level. Mouse 3T3 cells cultured in petri dishes were fixed in solutions containing 0.05% Tween 20 and stained using immune IgG, preim- mune IgG, or immune IgG adsorbed with PGH synthase as described above and then processed for elecron microscopy. In those 3T3 cells stained with IgG isolated from anti-PGH synthase serum (Figure 7A), electron-dense 37 Figure 6. Light photomicrograph of mouse 3T3 fibroblasts stained using anti-PGH synthase 196 as described in the text. NM, nuclear metorane. Bar line represents 42 pH. 38 39 Figure 7. Electron micrograph of a mouse 3T3 fibroblast stained using anti-PGH synthase 196 (A) and rabbit preimmune IgG (8) as described in the text. NM, nuclear membrane; ER, endoplasmic reticulum; M, mitochondria; P, plasma membrane; N, nucleus. Bar line represents 1 pM. 41 staining was found throughout the endoplasmic reticulum and on the nuclear membrane in 87% of 464 cells examined in five separate experi- ments. In contrast, electron-dense staining occurred only on limited areas of the endoplasmic reticulum of 11% of 382 cells examined after staining with preimmune IgG or immune IgG adsorbed with purified PGH syn- thase (Figure 7B). No PGH synthase-positive staining was associated with mitochondria, in agreement with the results of studies which have employed centrifugal techniques to localize PGH synthase. The diffuse distribution of PGH synthase antigen over the entire endoplasmic reticu- lum is similar to that seen with other reticular enzymes, including NADPH cytochrome 9 reductase and glucose-6-ph05phatase (63). Careful examination of 3T3 cells stained using anti-cyclooxygenase IgG failed to reveal any electron-dense staining associated with the plasma membrane. The lack of plasma membrane staining is not due to the existence of an unusual nonantigenic enzyme form because the cyclooxygen- ase activity solubilized from 3T3 cell microsomes could be precipitated quantitatively by anti-PGH synthase serum. It is also doubtful that any PGH synthase was solubilized from the plasma membrane selectively during fixation since even cells that were quick frozen and stained without pri- or fixation exhibited no cell surface staining under the light micro- scope. Thus, our results indicate that the enzyme is not distributed uniformly over the cell surface. It also seems unlikely that the PGH synthase is concentrated in discrete pockets on the plasma membrane as has been observed for low density lipoprotein receptors on human fibro- blasts (99). 42 Precipitation of microsomal PGH synthase with S. aureus-antibody com- plexes Three different monoclonal antibodies against the PGH synthase (pyp-1,'pyp-5 and pyp-7) and one control antibody (Sp-3) were affixed to S. ppggpg cells and these complexes mixed with intact and solubilized microsomes prepared from sheep vesicular gland (Figure 8). When intact microsomes prepared from sheep vesicular gland were mixed with S. Spgggg cells complexed to one of the anti-PGH synthase antibodies, complete pre- cipitation of cyclooxygenase activity occurred (i.e. all the enzyme activity was found in the S._gp[gg§ pellet). No appreciable precipita- tion resulted when a nonimmune monoclonal mouse Ing (gs-3) was used. As expected, similar results were obtained with control, detergent solu- bilized microsomes. Antibodies secreted by Exp-1, -5 and -7 interact with different antigenic sites on the PGH synthase, thus our results indicate that at least 3 antigenic determinants on the sheep vesicular gland cyclooxygenase are situated on the outer surface of microsomal spheres and thus on the cytOplasmic side of the endoplasmic reticulum. Protease inactivation of PGH synthase from sheep vesicular gland micro- EQEEE Four proteases were tested for their ability to degrade microsomal PGH synthase during a 45 min incubation (Table 1). Thermolysin, trypsin, and proteinase K caused significant losses in cyclooxygenase activity. The greatest degree of inactivation was observed with proteinase K which destroyed 45% of the starting activity. A second experiment was performed to determine if any cyclooxygenase activity could be solubilized by protease treatment of the microsomes. Figure 8. 43 Immunoprecipitation of microsomal sheep vesicular gland PGH synthase by mouse monoclonal antibody-S. aureus complexes. The experiment was performed as described in the text. Pel- lets (P) and supernatants (S) were obtained by centrifuging mixtures of various S. aureus- antibody complexes and micro- somal or solubilized preparations of PGH synthase at 500 x 9. Samples were then assayed for cyclooxygenase activity. 44 0. 3| §\\\\\\\\\\W "’ o '3 l ItL gal 7w “' .. l i _ O. 5 34 Mm 8 _ _ 3 gl L l“- o I”, 2:3. SP g .. a a an L L: E ET I j]: ET I 7]: SF 7.? <3 (mum-umuov asvwaoxxooioxo 45 use .omm an :_5 ma com umponaucw mew: mm_asmm .oe cw mm_QEmm asp maven—a an uwaaopm mew: mcovuoemc as» ._E\me HH mm: co_pmcp=mo:ou :_mpocq Fmsomoco_s _e:_mvco one m.¢¢. w.mm ¢.mH w.“ 0.0 o.n o.w w.oH ¢.NH ¢.mH ¢.mH cam—m «.mH ucmpm «.mH ucm_m ¢.mH ucm_m «.mfi awmcm _a o.H :_mqaeu m: ooH.m=Fa mmsomocowe cm_:u_mm> ammzm _E o.H :mepoEemcu a: ooH m=_a mmeomocu_e Lm_:u_mm> ammgm _E o.H mmecmmmp_ou m: ooH m:_a mmsomoLUFE cm_=uwmm> ammgm _E o.H m: sea m:_q mmeomocuwe cm—auwmm> ammzm _E o.H ARV Aup>_uom 4o mmom _e\ S x muwcav :o_Mmmmwc empem —5\ OH x mp_::v :o_mem_u mcommn »u_>_puc mmocmmxxoo_uxu >u_>wuum mmmcmmxxoo_uxu mmeomocu_5 v:m_m Lm_:owmm> ammcm mo mmmcucxm Ion ecu co cowpmmm_c mmmmuoca mo pumwmm H u4m