Inhibition of Plasminogen Activator Inhibitor-1 by 11-Keto-9(E),12(E)-octadecadienoic Acid, a Novel Fatty Acid Produced by Trichoderma sp

VOL. 52 NO. 9, SEPT. Inhibition 1999 THE JOURNAL OF ANTIBIOTICS of Plasminogen Activator Inhibitor-1 pp.797 by l l-Keto-9(£),12(£)-octadecadienoic _ 802 Acid, a Novel Fatty Acid Produced by Trichoderma sp. Toshihiro Chikanishi, Chikara Shinohara, Tadashi Kikuchi, Akira Endo and Keiji Hasumi* Department of Applied Biological Science, Tokyo Noko University, Fuchu, Tokyo 183-8509, Japan We have recently found a novel fatty (KOD), that enhances fibrinolytic KOD has been investigated. acid, ll-keto-9(£),12(iT)-octadecadienoic activity of endothelial KOD increased 2-fold acid cells. The mechanism of action of the plasmin activity of bovine aortic endothelial cells at 250 jam. The stimulation was dependent on plasminogen and was inhibited by anti-urokinase, whereas KODdid not enhance the urokinase-catalyzed plasminogen activation and the resulting plasmin activity in a cell-free system. Neither the production of urokinase nor the conversion of pro-urokinase to the active, two-chain form was elevated by KOD, but it decreased plasminogen activator inhibitor- 1 (PAI- 1) activity of cells and inactivated PAI-1 irreversibly in a purified system. These results demonstrated that the KODenhancement of endothelial fibrinolytic activity was mediated, at least in part, by inactivation of PAI-1. Endothelial cells synthesize and secrete plasminogen activators, such as urokinase (UPA) and tissue plasminogen activator (tPA), as well as plasminogen activator inhibitors. The balance of production between plasminogen activators and inhibitors by endothelial cells mainly regulates flbrinolysis in the blood vessel. In patients with atherosclerosis and thromboembolism, plasminogen activator inhibitor type 1 (PAI-1) may be dominant in this balance1?2). PAI-1 belongs to the serpin superfamily3). Humanmature PAI- 1 is a 48-kDa single-chain glycoprotein consisting of 379 amino acids3)4). PAI-1 is produced as an active molecule, but it spontaneously converts to an inactive, latent form5). Latent PAI-1 can be re-activated by exposure to denaturing agents, followed by refolding6). Like other serpins, PAI-1 is thought to associate with the target protease through a bait residue in a reactive site loop. The reactive site loop is partly inserted between strands 3 and 5 thereby, enhanced UPA-mediated plasminogen activation in BAEC. Materials and Methods Materials Chemicals and proteins were obtained from the following sources: human plasminogen (TV-terminal glutamic acid form) from Enzyme Research Laboratories (South Bend, IN, USA); human UPA and rabbit anti-UPA serum from JCR Pharmaceuticals (Kobe, Japan); goat antitPA IgG from Biopool (Umea, Sweden); bovine fibrinogen, human thrombin, nitroanilide), stearic hydroxyhexadecanoic S-2251 (H-D-valyl-leucyl-lysine-/?acid, oleic acid, linoleic acid and 3acid from Sigma (St. Louis, MO, USA); Spectrozyme UK(carbobenzoxy-L-7-glutamyl butoxy)-glycyl-arginine-p-nitroanilide) (a-t- American of the j8-sheet A of the molecule, and this insertion appears to be required for inhibition7). In the course of identifying agents that enhance fibrinolytic activity of vascular endothelial cells, we have isolated a novel fatty acid, ll-keto-9(£),12(£)octadecadienoic acid (KOD), from a culture of Trichoderma sp. F55948). This report deals with the Diagnostica mechanism of the KODenhancement of fibrinolytic method of Lawrence et al.9\ KODwas isolated from of bovine aortic endothelial cells (BAEC). activity The results demonstrated that KODirreversibly inactivated PAI- 1 and, Inc. (Greenwich, from CT, USA); [35S]EXPRESSá" Protein Labeling Mix (73% L-[35S]methionine, 22% l[35S]cysteine) from NEN; protein A-Sepharose from Pharmacia Biotech. PAI-1 was from two sources: one from the human melanoma cell line MJZJ (American Diagnostica) and the other from human HT1080 cells, which was purified and re-activated according to the cultures of Trichoderma sp. F55948). The compositions of media and buffers were: medium A, Eagle's minimum 798 essential mediumsupplemented with 10%fetal calf serum, 100units/ml penicillin G and lOO^g/ml streptomycin; medium B, medium A devoid of NaHCO3but supplemented with 20mM HEPES, pH 7.4; buffer A, 50mM Tris-HCl and lOOmMNaCl, pH 7.4; buffer B, 150mM NaCl and 20mMsodium phosphate, pH 7.4; buffer C, 10mM Tris-HCl, 150mM NaCl, 0.1% SDS, 1% Triton X100, 0.5% sodium deoxycholate and 1 mMEDTA, pH 7.5; buffer D, buffer B containing 0.05% Triton X-100; buffer E, 150mM NaCl, 50mM Tris-HCl, pH 7.5, 100/xg/ml bovine serum albumin and 0.01% Tween 80. Cell Culture BAECwere isolated from bovine aorta and subcultured in mediumA for approximately 6 passages. For assays, cells SEPT. THE JOURNAL OF ANTIBIOTICS were seeded into 96-well 104cells/50jA 1.5X106cells/1.5ml tissue culture plates (5X per well) or 35-mm culture dishes (1.0 to per dish) and grown for 24hours beforeuse. Determination of Fibrinolytic Activity of BAEC BAECgrown in 96-well plates were washed twice with medium B and preincubated at 37°C for 6 hours in 50jA of medium B with or without KOD. At the end of the incubation, cells were washed with buffer A and then received 100 /^1 of buffer A containing 0. 1 jiu plasminogen and 0.1mMS-2251. After incubation at 37°C for up to 4hours, the release of /?-nitroaniline was determined by measuring the change in absorbance at 405 nm. 1999 SDS, placed onto a fibrin-agar indicator geln) containing 0.2unit/ml UPA, and incubated at 37 °C for 3-5 hours. The indicator gel was stained with 0.1% amide black in 30% methanol and 10% acetic acid and was destained in 30% methanol and 10% acetic acid. Determination of PAI- 1 Activity PAI-1 activity was determined as the inhibitory activity against UPAas follows. Active humanmelanomaPAI-1 (4.21jUg/ml) was preincubated with or without KODin 50Art of buffer D at 37°C for 15minutes. Subsequently, 25 jul of UPA (8 units/ml) in buffer D containing 20mg/ml bovine serum albumin were added and the mixture was incubated at 37°C for 30minutes. Then, remaining UPA activity was determined by adding 25ji\ of 0.4mM Spectrozyme UK(in buffer D) to the mixture, followed by further incubating at 37°C to measure changes in absorbance at 405nm. In the experiment by which reversibility of KODinhibition was examined, re-activated PAI-1 from human HT1080 cells (2.45/zg/ml) preincubated in buffer E at 37°C for 30minutes was in the absence or presence of KOD.The KOD-treated PAI-1 was assayed for UPAinhibition either directly or after dialysis against 4 m guanidine HC1 in phosphate-buffered saline, pH 7.3 containing 0. 1 mMdithiothreitol and 0.01% Tween 80 at 37°C for 4.5hours and then against 50mMsodium phosphate, pH 6.6, 500 mMNaCl, 0.1 mMdithiothreitol and 0.01% Tween 80 at 4°C for 16hours9). For UPAinhibition assay, aliquots of the treated PAI-1 (20ji\) was incubated with 10Art of UPA (75units/ml in buffer E) at 37°C for Immunoprecipitation of UPA BAECwere washed with buffer B and labeled at 37 °C 15minutes. Subsequently, the mixture received 20Art of Spectrozyme UK (0.25mM in buffer E) and changes in for 6hours in the presence of 35S protein labeling mixture absorbance at 405 nm were measured at 37°C. (50 /iCi/ml) in methionine-free mediumA. After washing with buffer B, cells were scraped in buffer C containing 10 /ig/ml aprotinin and disrupted by sonication. Using the supernatant of the cell lysate (approximately 1.3 X 107 cpm), immunoprecipitation was carried out as described10). Reverse Fibrin Zymography BAECgrown in 35-mmdishes were washed twice with buffer B and incubated in 750/xl of medium A at 37°C for 2 hours in the absence or presence of KOD.After washing, cells were removed by scraping in 1ml of buffer B and centrifuged. The resulting cell pellet was dissolved in 100^1 of buffer B containing 0.5% Triton X-100. The cell lysates were subjected to SDS-polyacrylamide gel electrophoresis (SDS-PAGE) under reducing conditions. After electrophoresis, the gel was washed twice with 250ml of 2.5% Triton X-100 for 45minutes to remove Results The fibrinolytic activity of BAECwas determined by incubating the cells at 37°C for 6 hours in the absence or presence of KOD, followed by washing and further incubating the cells with plasminogen and a chromogenic plasmin substrate, S-2251. KODstimulated the activity at a concentration higher than 30 /im, and the activity doubled at -250/iM (Fig. 1A). The elimination of plasminogen in the second incubation abolished the effect of KOD (data not shown), indicating that its effect was mediated by plasminogen activation. Stearic, oleic, linoleic and 3- hydroxyhexadecanoic acids failed to enhance, rather were inhibitory to, fibrinolytic activity at a concentration of 250~260/iM' (Fig. IB), suggesting an essential role of the VOL.52 NO.9 THE JOURNAL OF ANTIBIOTICS 799 Fig. 1. Effects ofKODand other fatty acids on fibrinolytic activity ofBAEC. BAECwere preincubated at 37°C for 6 hours in the presence of the indicated concentrations of KOD(A) or the indicated fatty acids: 270//m KOD, 250/zm stearic acid, 250jllm oleic acid, 250/iM linoleic acid or 260/iM 3- hydroxyhexadecanoic acid (HHD) (B). After washing, cells received plasminogen and S-2251 and were further incubated at 37°C for 4hours to measure changes in absorbance at 405 nm. Each value represents the mean±S.D. from triplicate determinations. a,/?-unsaturated carbonyl function in the KODmolecule. To determine the contribution of plasminogen activators in the KODeffect, fibrinolytic activity was determined in the presence of anti-UPA and anti-tPA antibodies at concentrations inhibiting >95% of respective enzyme in a purified system. In control cells, activity was markedly reduced by anti-UPA but slightly by anti-tPA (Fig. 2), indicating that UPA was the predominant plasminogen activator in BAEC under the present experimental conditions. In KOD-treated cells, anti-tPA caused slight inhibition, while anti-UPA reduced activity to a level comparable to that in control cells. KODfailed to enhance UPA-catalyzed plasminogen activation and the resulting plasmin activity in a cell free system (data not shown); this excluded the possibility of a direct activation of UPAand/or plasmin by KOD.Thus, these results suggested that the KODeffect was mediated by a change in the amount of the active species of UPAon BAEC.However, the level ofuPA in BAEC was unaffected by KOD, as revealed by immunoprecipitation of UPAthat had been metabolically labeled with 35S in the presence and absence of KOD. Further, the conversion of the [35S]pro-UPA (single-chain form) to the two-chain form was not enhanced in cells incubated with KOD(data not shown). To test for the possibility that the KOD effect is associated with changes in PAI activity, the level of PAI in BAECwas determined by reverse fibrin zymographyafter SDS-PAGE under reducing conditions, which enable the detection of PAI-1 but not PAI-212). In the zymogram, cell lysate from untreated control BAECgave a prominent lysisresistant area that indicated the presence of PAI-1. On the other hand, lysate from KOD-treated cells produced a rather faint band, showing that PAI-1 activity in KOD-treated BAECwas significantly reduced (Fig. 3A). Furthermore, when lysate from untreated BAECwas incubated with KOD in a cell-free conditions, the intensity of a lysis- resistant band was markedly reduced as compared to that observed with lysate not incubated with KOD(Fig. 3B), indicating a direct inhibition ofPAI- 1 by KOD. Next, the effect of KODon human PAI-1 activity was examinedin a cell-free system using purified materials (Fig. 4). In the first experiment, UPAwas incubated with human PAI-1 that had been preincubated with or without KOD, following which residual UPA activity was 800 SEPT. THE JOURNAL OF ANTIBIOTICS 1999 Fig. 3. Effect ofKOD on PAI activity in BAECas visualized by reverse fibrin zymography. Fig. 2. Effects of KODon fibrinolytic activity of BAECin the absence and presence of anti-tPA or anti-UPA antibody. (A) BAECwere incubated at 37°C for 2hours in the absence or presence of KOD(300 fiM) in medium A. Subsequently, cell lysate (50 jig protein) was subjected to SDS-PAGE on a 10% gel under reducing conditions, and the gel was processed for reverse fibrin zymography. Lane 1, lysate from control cells; lane 2, lysate from KOD-treated cells. (B) Lysate prepared from untreated BAEC(250 ^g/ml) was incubated at BAECwere incubated at 37°C for 6hours in the absence or presence of 270 /iM KOD.After washing, cells were incubated at 37°C for 10 minutes with buffer A containing none, anti-tPA IgG (2.85 /zg/ml) or antiuPA serum (215 /ig/ml). Subsequently each culture received plasminogen and S-2251. After 37°C for 2hours in the absence or presence of KOD (300 ^m), and a portion of the mixture (5 /ig protein) was subjected to SDS-PAGE on a 10% gel under reducing conditions. Subsequently, the gel was processed for reverse fibrin zymography. Lane 1, control; lane 2, KOD-treated. Arrowhead denotes the further incubation at 37°C for 4 hours, changes in absorbance at 405nmwas measured. Each value represents the mean± S.D. from triplicate determinations. determined (Fig. 4A). The residual UPA activity position of lysis-resistant was band. Discussion reduced to 15%by PAI-1 that had not been treated with KOD.On the other hand, KOD-treated PAI-1 became less inhibitory. 50 and The residual UPA activity was increased to 67% by PAI-1 preincubated with KOD at concentrations of 300 and 600 jjm, respectively. KODdid not affect UPAactivity in the absence ofPAI-1. To examine whether KODinhibition of PAI-1 involved an inactivation or of an elevated conversion of active PAI-1 to latent analogue, active PAI-1 was incubated with KODand the KODincreases plasmin activity of BAEC.Although the KODeffect is not apparent in the absence of plasminogen and is blocked by anti-UPA, it enhances neither production nor activation of UPAin the cells. Further, KODis not stimulatory to urokinase-catalyzed plasminogen activation and the resulting plasmin activity in a purified system. On the other hand, PAI-1 activities both in BAECand in a purified system are markedly inhibited by KOD.From these treated PAI-1 was assayed for UPAinhibition before and after dialysis against guanidine HC1, which re-activates latent PAI-1. As shown in Fig. 4B, activity of KODtreated PAI-1 to inhibit UPAwas not restored even after dialysis against the denaturant; this demonstrated that an results, irreversible treatment with guanidine HC1. Similarly, the KOD-treated inactivation of PAI- 1 occurred. it was concluded that KOD enhances fibrinolytic activity of BAEC, at least in part, by inhibiting PAI-1. PAI1 is spontaneously converted to the latent form5), which can be re-activated by exposure to denaturing agents6). The activity of KOD-inhibited PAI-1 is not restored by a BAEC lysate exerts much less PAI-1 activity electrophoresis even after under the denaturing conditions, which can VOL.52 NO.9 801 THE JOURNAL OF ANTIBIOTICS Fig. 4. Inhibition ofPAI-1 by KODin apurified system. (A) The indicated concentrations of KODwere preincubated with 0 (å ) or 4.21 jUg/ml of human melanoma PAI1 (à") at 37°C for 15 minutes. Subsequently, inhibition of UPA activity was determined in duplicate. (B) PAI-1 from HT1080 cells (2.45/xg/ml) was preincubated with 0 or 200/im KODat 37°C for 30minutes. Aliquots of the preincubated PAI-1 were assayed for UPAinhibition either directly (before denaturation) or after denaturation by guanidine HC1. Each value represents the mean±S.D. from triplicate determinations. restore activity of latent PAI-1. These results demonstrate that the KOD inhibition of PAI-1 is caused by an irreversible inactivation, rather than by an increased conversion into the latent form. The a,/3-unsaturated carbonyl function of the KOD molecule appears to be essential to PAI-1 inactivation. a,/J-Unsaturated carbonyl compounds may cause protein modification at sulfhydryl, imidazolyl and amino groups13?14). The PAI-1 molecule does not contain cysteine, and histidine and lysine are not present in the essential Acknowledgments Wethank Shigemitsu Nakashima for technical assistance and Yoshikazu Kitano for discussion. This work was supported in part by grants from the Ministry of Health and Welfare, Japan and the Ministry of Education, Science, Sports and Culture, Japan. References 1) Aznar, J. & A. Estelles: Role ofplasminogen activator region (P15 to Pr) of the mobile reactive site loop4), which plays a role by being partially inserted into the ^8-sheet A inhibitor type 1 in the pathogenesis of coronary artery diseases. Haemostasis 24: 243-251, 1994 2) Wiman, B.: Plasminogen activator inhibitor 1 (PAI-1) in plasma: its role1995in thrombotic disease. Thromb. Haemost. 74: 71-76, acids 1 10- 145 region, which participates 3) upon binding binding15), to a serine protease5). contains a histidine However, the amino in target protease and three lysine residues4). Thus, it is suggested that KODmodification of one (or more) of these residues in the 110-145 region may be involved in PAI-1 inactivation. Recently, a,/3-unsaturated ketone-containing diketopiperazines have been identified as PAI-1 inhibitors16~18). The observation that such com- pounds bind the 1 10-145 region18) may support the above hypothesis. Van Meijer, M. & H. Pannekoek: Structure plasminogen activator inhibitor 1 (PAI-1) and function in fibrinolysis: an update. Fibrinolysis 263-276, of its 9: 1995 4) Ny, T.; M. Sawdey, D. Lawrence, J. L. Millan & D. J. Loskutoff: Cloning and sequence ofa CDNAcoding for the human /^-migrating gen activator 6776-6780, inhibitor. 1986 endothelial-cell-type plasmino- Proc. Natl. Acad. Sci. USA83: 5) Lindahl, T. L.; O. Sigurdardottir & B. Wiman: Stability of plasminogen activator inhibitor 1 (PAI-1). Thromb. 802 SEPT. THE JOURNAL OF ANTIBIOTICS Haemost. 62: 748-751, 1989 Hekman, C. M. & D. J. Loskutoff: Endothelial cells produce a latent inhibitor of plasminogen activators that can be activated by denaturants. J. Biol. Chem. 260: 11581-11587, 1985 Lawrence, D. A.; S. T. Olson, S. Palaniappan & D. Ginsburg: Serpin reactive center loop mobility is required for inhibitor function but not for enzyme recognition. X Biol. Chem. 269: 27657-27662, 1994 Shinohara, C; K. Hasumi, T. Chikanishi, T. Kikuchi & A. Endo: l l-Keto-9(£),12(£)-octadecadienoic acid, a novel fatty acid that enhances fibrinolytic activity of endothelial cells. J. Antibiotics 52: 171- 174, 1999 Lawrence, D.; L. Strandberg, T. Grundstrom & T. Ny: Purification of active human plasminogen activator inhibitor 1 from Escherichia coli. Comparison with natural and recombinant forms purified from eucaryotic cells. Eur. J. Biochem. 186: 523-533, 1989 1999 type 4-hydroxy-2-nonenal adducts in modified lowdensity lipoproteins: markers for atherosclerosis. Biochemistry 33: 12487-12494, 1994 Requena, J. R.; M. X. Fu, M. U. Ahmed, A. J. Jenkins, TJ. Lyons, J.W. Baynes & S.R. Thorpe: Quantification of malondialdehyde and 4-hydroxynonenal adducts to lysine residues in native and oxidized human low-density lipoprotein. Biochem. J. 322: 317-325, 1997 Keijer, B. I; M. Linders, A.-J. van Zonneveld, H. J. Ehrlich, J.-P. de Boer & H. Pannekoek: The interaction of plasminogen activator inhibitor 1 with plasminogen activators (tissue-type and urokinase-type) and fibrin: localization of interaction sites and physiologic relevance. Blood 78: 401-409, 1991 Bryans, J.; P. Charlton, I. Chicarelli-Robinson, M. Collins, R. Faint, C. Latham, I. Shaw & S. Trew: Inhibition of plasminogen activator inhibitor- 1 activity by two diketopiperazines, XR330 and XR334 produced Stoppelli, M. P.; C. Tacchetti, M. V Cubellis, A. Corti, V J. Hearing, G. Cassani, E. Appella & F. Blasi: Autocrine saturation of pro-urokinase receptors by Streptomyces sp. J. Antibiotics 49: 1014-1021, 1996 Charlton, P. A.; R. W. Faint, F. Bent, J. Bryans, I. Loskutoff, D. J. & R. Schleef: Plasminogen activators and their inhibitors. Methods Enzymol. 163: 293-302, of human plasminogen activator inhibitor-1 Thromb. Haemost. 7.5: 808-815, 1996 onhuman A431 cells. Cell 45: 675-684, 1986 1988 Miskin, R. & R. Abramovitz: One-phase reverse zymography after denaturing gel electrophoresis: high sensitivity detection of activity of plasminogen activator inhibitor 2 and1995other protease inhibitors. Fibrinolysis 9: 331-342, Uchida, K.; S. Toyokuni, K. Nishikawa, S. Kawakishi, H. Oda, H. Hiai & E. R. Stadtman: Michael addition- Chicarelli-Robinson, I. Mackie, S. Machin & P. Bevan: Evaluation of a low molecular weight modulator activity. Friederich, P. W; M. Levi, B. J. Biemond, P. Charlton, D. Templeton, A. J. van Zonneveld, P. Bevan, H. Pannekoek & J. W. ten Kate: Novel low-molecularweight inhibitor of PAI-1 (XR5118) promotes endogenous fibrinolysis and produces postthrombolysis thrombus growth in rabbits. Circulation 96: 916-921, 1997