560
Bioconjugate Chem. 1993, 4, 560-567
Biotinylated Isocoumarins, New Inhibitors and Reagents for Detection, Localization, and Isolation of Serine Proteases Chih-Min Kam, Ahmed S. Abuelyaman, Zhaozhao Li, Dorothy Hudig,+and James C. Powers; The School of Chemistry and Biochemistry, Georgia Institute of Technology, Atlanta, Georgia 30332-0400, and The School of Medicine and The College of Agriculture, Howard Building, University of Nevada, Reno, Nevada 89557-0046. Received July 29, 1993”
Eight new biotinylated, mechanism-based isocoumarin serine protease inhibitors have been designed and synthesized to detect, localize, and isolate serine proteases. Isocoumarins that contain a 4-chloro group, a biotinylated substituent at the 7-position, and different 3-alkoxy groups are inhibitors of various serine proteases including human leukocyte elastase (HLE),porcine pancreatic elastase (PPE), trypsin, human recombinant granzyme A, chymotrypsin, and cathepsin G. Insertion of spacers between the isocoumarin moiety and the biotin moiety enhanced enzyme inhibitory potency and may also promote binding of the enzyme-inhibitor complex to avidin. The 3-alkoxy groups conferred selectivity toward different serine proteases with chymotrypsin being inhibited effectively by compounds with 3-phenylethoxy groups while derivatives with 3-methoxy, ethoxy, or propoxy groups were potent inhibitors of HLE and moderate inhibitors of PPE. Full enzymatic activity was regained after the immediate addition of hydroxylamine to the inactivated chymotrypsin and PPE derivatives, which indicated that a simple acyl enzyme derivative is formed initially in the inhibition reaction. Egg avidin did not effect the rate of spontaneous enzyme reactivation rate while streptavidin accelerated the reactivation reaction. PPE inhibited by 7-[ [6-(biotinylamino)caproyllaminol-4-chloro-3-ethoxyisocoumarin (BIC 5) or 7-[ [6[ [6-(biotinylamino)caproyllamino]caproyllamino]-4-chloro-3-methoxyisocoumarin (BIC 7) was bound to immobilized avidin columns. Most of inhibited PPE could be eluted from the monomeric or tetrameric avidin columns but only a portion (40-70 % ) of the enzyme was active due to the partial formation of a stable alkylated enzyme derivative during the isolation process. Under ideal circumstances and formation of a simple acyl enzyme derivative, BICs should be useful for the isolation of new active serine proteases. Biotinylated isocoumarins should also have wide applicability for the detection, quantitation, and histochemical localizationof stable biotinylated serine protease derivatives in a variety of physiological situations.
INTRODUCTION
Serine proteases play important roles in a number of physiological processes including digestion, blood coagulation, complement activation, fibrinolysis, reproduction, development, and the release of physiologically active peptides (Neurath, 1984). Uncontrolled proteolysis by serine proteases is also related to diseases such as pulmonary emphysema, adult respiratory distress syndrome, and pancreatitis. Thus, there is considerable interest in characterizing new serine proteases, determining their sites of storage and biological activity, and studying their roles in various normal and abnormal biological processes. Synthetic inhibitors tagged with biotin have frequently been utilized to detect, localize,and recover proteases from pathological tissues. For example, the chloromethyl ketone affinity label biotin-Arg-CHzC1 has been utilized to detect trypsin and thrombin in electrophoreticgels using Western blotting and a streptavididalkalinephosphatase detecting system (Kay et al., 1992; Walker et al., 1992). The diazomethyl ketone cysteine protease inhibitor biotinPhe-Ala-CHNz has been used to detect a cathepsin B-like precursor produced by breast-tumor cells in culture (Cullen et al., 1992). Biotin-Phe-Pro-Arg-CHzCl has been used as a specific probe for the detection of factor Xa and factor VIIa by blotting techniques and has also been used to remove traces of active enzymes from zymogen preparations in the coagulation and fibrinolytic pathways (Wil-
* Author to whom correspondence
should be addressed. of Nevada. Abstract published in Advance ACS Abstracts, November 1, 1993. + University @
liams et al., 1989). Two biotinylated derivatives of the ACE’ inhibitor lisinopril, have been synthesized (Berstein et al., 1990) and one derivative with a sizable spacer was used to isolate ACE from a crude mixture of proteins. Biotin with its high affinity (& = M) toward avidin and streptavidin has been covalently linked to numerous peptides and proteins, and the complexes have been used for the detection of hormone-receptor complexes, the isolation of peptide and protein receptors, and the study of protein-protein interactions (Billingsley et al., 1987; Finn et al., 1984). Biotinylated antibodies, which are detected by avidin-linked chromogens or fluorophores, are commonly used for the immunolocalization of antigens (Towbin and Gordon, 1984). Thus, biotinylated serine proteases would also be expected to have numerous applications. Effective inhibitors of serine proteases include both peptide derivatives and heterocyclic structures such as isocoumarins (Powers and Harper, 1986). 3,CDichloroisocoumarin is a general serine protease inhibitor and 3-alkoxy-4-chloro-7-substituted isocoumarins are potent and more specific inhibitors for serine proteases such as Aca, 6-aminocaproic acid; ACE, angiotensin-converting enzyme; BIC, biotinylated isocoumarin; Boc-Gly, tert- (butyloxycarbony1)glycine; Cat G, cathepsin G; CDI, 1,l’-carbonyldiimidazole; DCC, N,N’-dicyclohexylcarbodiimide;DMF, dimethylformamide; EtsN, triethylamine; Hepes, N-(2-hydroxyethyl)piperazine-N’-2-ethanesulfonicacid; HLE, human leukocyte elastase; HOBt, 1-hydroxybenzotriazole;HR granzyme A, human recombinant granzyme A; pNA, p-nitroanilide; PPE, porcine pancreatic elastase; SBzl, thiobenzyl ester; SUC,succinyl; 2, benzyloxy carbonyl.
0 1993 American Chemical Society 1Q43-18Q2/93/29Q4-Q56Q$Q4.QQ/Q
Bloconlugste Chem., Vol. 4, No. 6, 1993 561
Blotlnylated Isocoumarins
C1S.0.25Hz0: C, 59.39; H, 5.22, N, 7.70. Found: C, 59.08; H, 5.37; N, 7.94. 7-(Biotinylamino)-4-chloro-3-propoxyisocoumarin (BIC 2). This compound was prepared from 7-amino-4-chloro-3-propoxyisocoumarin using the procedure described for BIC 1: yield 20%; mp 127-131 "C; lH NMR (DMSO) 6 10.3 (8, lH), 8.50 (d, lH), 8.0 (dd, lH), 7.60(d,lH),6.3-6.5(d,2H),4.3(t,2H),4.15,(m,2H),3.10 (m, lH), 2.8 (m, lH), 2.6 (m, lH), 2.35 (m, 2H), 1.7-1.8 (m, 2H), 1.2-1.7 (m, 6H), 1.0 (t,3H); MS (FAB+)mle 502 (M+ Na 1). Anal. Calcd for C22H26N305ClS-Hz0: C,53.06; H, 5.67; N, 8.44; C1, 7.12. Found: C, 53.30; H, 5.67; N, 8.49; C1, 7.03. 74[6-(Biotinylamino)caproyl]amino]-4-chloro-3-(2pheny1ethoxy)isocoumarin (BIC3). 64Biotinylamino)caproic acid was prepared from the N-hydroxysuccinimide ester of biotin (Jasiewicz et al., 1976)and methyl 6-aminocaproate acid hydrochloride by a previously described method (Hofmann et al., 1984). 6-(Biotinylamino)caproic acid chloride was prepared by incubating 6-(biotiny1amino)caproicacid (0.36g, 1mmol) in 4 mL of thionyl chloride at 25-35 "C for 1h and removing excess thionyl chloride in uacuo. The residue was dissolved in a few milliliters of DMF, 7-amino-4-chloro-3-(2-phenylethoxy)isocoumarin (0.34 g, 1 mmol) and Et3N (0.1 g, 1 mmol) were added, and the reaction mixture was stirred at room temperature overnight. The product was purified by column chromatography and eluted with CH2CldMeOH = 1O:l: yield 28%; mp 163-167 "C; 'H NMR (DMSO) 6 10.3 (8, lH), 8.50 (d, lH), 8.0 (dd, lH), 7.75 (m, lH), 7.60 EXPERIMENTAL PROCEDURES (d, lH), 7.20-7.40 (m, 5H), 6.40 (d, 2H), 4.55 (t,2H),4.104.30 (m, 2H), 3.0-3.10 (m, 5H), 2.80 (m, lH), 2.60 (m, lH), Reagents. Biotin was obtained from Chemical Dy2.35 (t, 2H), 2.0 (m, 2H), 1.20-1.60 (m, 12H); MS (FAB+) namics Corp., South Plainfield, NJ. Avidin, streptavidin, mle 655 (M+ + 1). Anal. Calcd for C33H39N406ClS.H20: bovine trypsin, and bovine chymotrypsin were obtained C, 58.87; H, 6.14; N, 8.32; C1, 5.27. Found: C, 58.72; H, from Sigma Chemical Co., St Louis, MO. Tetrameric 6.22; N, 8.90; C1, 5.50. avidin agarose was obtained from Vector Laboratories, 7-[ [ 6-(Biotinylamino)caproyl]amino]-4-chloro-3Inc., Burlington, CA. Immobilized monomeric avidin was propoxyisocoumarin (BIC 4). This compound was obtained from Pierce, Rockford, IL. Porcine pancreatic prepared from 7-amino-4-chloro-3-propoxyisocoumarin elastase (PPE) was obtained from United States Biousing the procedure described for BIC 3: yield 25%; mp chemical Corp., Cleveland, OH. Human leukocyte elastase 141-145 "C; 1H NMR (DMSO) 6 10.3 (9, lH), 8.50 (d, lH), (HLE) and cathepsin G (Cat G )were obtained from Athens 8.0 (dd, IH), 7.75 (b, lH), 7.65 (d, lH), 6.3-6.4 (d, 2H), 4.3 Research and Technology, Inc., Athens, GA. Human (m, 3H), 4.1 (m, lH), 2.95-3.15 (m, 3H), 2.8 (m, lH), 2.6 recombinant granzyme A was obtained from Dr. Duke (m, lH), 2.3 (m, 2H), 2.0 (m, 2H), 1.9 (m, 2H), 1.2-1.7 (m, Virca at Immunex Corp., Seattle, WA. 6-Aminocaproic 12H), 1.0 (t, 3H); MS (FAB+) mle 593 (M++l). Anal. acid and N-hydroxysuccinimide were purchased from Calcd for CzsH37N406ClS.0.5HzO: C, 55.14; H, 6.13; N, Aldrich Chemical Co., Inc., Milwaukee, WI. Hepes was 9.53; C1,6.04. Found: C, 54.81; H, 6.26; N, 9.47; C1, 5.93. purchased from Research Organics, Inc., Cleveland, OH. 74[6-(Biotinylamino)caproyl]amino]-4-chloro-37-Amino-4-chloro-3-(2-phenylethoxy)isocoumarin,7-amiethoxyisocoumarin (BIC5). This compound was preno-4-chloro-3-ethoxyisocoumarin, 7-amino-4-chloro-3pared from 7-amino-4-chloro-3-ethoxyisocoumarin using methoxyisocoumarin (Harper and Powers, 1985), and the procedure described for BIC 3: yield 15%; mp 1567-amino-4-chloro-3-propoxyisocoumarin (Powers et al., 162 "C; 1H NMR (DMSO) 6 10.3 (s, lH), 8.50 (d, lH), 8.0 1990) were prepared as previously described. 7-(Biotinylamino)-4-chloro-3-(2-phenylethoxy)iso- (dd, lH), 7.75 (b, lH), 7.65 (d, lH), 6.3-6.45 (d, 2H), 4.4 (9,2H), 4.2-4.3 (m, 2H), 2.95-3.15 (m, 3H), 2.8 (m, lH), coumarin (BIC1). The acid chloride derived from biotin 2.6 (m, lH), 2.3 (t,2H), 2.0 (m, 2H), 1.2-1.7 (m, 15H); MS was prepared by incubating biotin (0.4 g, 1.6 mmol) in 6 (FAB+)mle 601 (M+ + Na + 1). Anal. Calcd for C27H36' mL of thionyl chloride at 25-35 "C for 1 h, followed by N406ClS.HzO: C, 55.03; H, 6.43; N, 9.17; C1,5.80. Found: removal of excess thionyl chloride in uacuo. The residue C, 54.81; H, 6.26; N, 9.23; C1, 5.69. was dissolved in a few milliliters of DMF, 7-aminO-4-ChlOrO7-[[6 4[6 4Biotinylamino)caproyl]amino]caproyl]3-(2-phenylethoxy)isocoumarin (0.26 g, 0.8 mmol) and amino]-4-chloro-3-(2-phenylethoxy)isocoumarin (BIC EtsN (0.082 g, 0.8 mmol) were added and the reaction 6). 6- [[6-(Biotinylamino)caproyllaminolcaproic acid was mixture was stirred at room temperature overnight. The prepared by coupling biotin sequentially to two units of product was purified by column chromatography and 6-aminocaproic acid. Biotin (2.27 g, 9.3 mmol) was diseluted with CHZCldMeOH = 151: yield 34%; mp 182solved in 70 mL of DMF at 100 "C, and CDI (1.5 g, 9.3 185 "C; IH NMR (DMSO) 6 10.3 (s, lH), 8.50 (d, lH), 8.0 mmol) was added with continuous stirring. After 0.5 h, (dd, lH), 7.65 (d, lH), 7.20-7.40 (m, 5H), 6.30-6.50 (d, heating was terminated and stirring was continued at room 2H),4.55 (t, 2H), 4.15-4.13 (m, 2H), 3.0-3.20 (m, 3H), 2.80 temperature for 2 h. Methyl 6-aminocaproate (1.6 g, 9.3 (m, lH), 2.60 (m, lH), 2.30 (t,2H), 1.30-1.70 (m, 6H); MS mmol) and Et3N (1.3 mL, 9.3 mmol) were added, and the (FAB+) mle 542 (M+ + 1). Anal. Calcd for C27H28N305-
PPE, HLE, chymotrypsin, and the trypsin-like enzymes that initiate blood coagulation and the complement cascade (Harper and Powers, 1985;Kam et al., 1988; Powers et al., 1990; Kam et al., 1992). These reagents are mechanism-based or "suicide" inhibitors where the isocoumarin ring is opened by a serine protease to unmask a new, potent reactive structure which can further inactivate the enzyme by covalent bond formation. Isocoumarins have several potential advantages in the inhibition process due to the possible formation of either an acyl enzyme derivative or an alkylated enzyme derivative. The acyl enzyme can be reactivated by hydrolysis or treatment with a nucleophile, while the alkylated enzyme is stable and not reactivatable. In this paper we report the synthesis and characterization of several isocoumarin-based serine protease inhibitors that incorporate a biotin tag. The biotin was attached to the isocoumarin either directly or through a spacer of different lengths. These derivatives have the ability to inhibit serine proteases with the inhibition specificity being controlled by the 3-substituent on the isocoumarin ring. After coupling to serine proteases, the biotin of the inhibitor is accessible to avidin. We have probed the structures of the enzyme-inhibitor complexes with nucleophile hydroxylamine to distinguish between reactivatable simple acyl enzyme derivatives and nonreactivatable alkylated enzyme derivatives. We also demonstrate with representative serine proteases that these biotinylated isocoumarins can be used to isolate and recover inactivated serine proteases.
+ +
502
Bioconjugere Chem., Vol. 4, No. 6, 1993
reaction mixture was stirred overnight. Any solid that formed was removed by filtration and the solvent was evaporated in uucuo. A solution of 1N NaOH (24 mL) was added, followed by a few drops of MeOH until a clear solution was obtained. The alkaline solution was stirred for 2 h. Most of the solvent was then removed in uucuo. The solution was cooled in ice and acidified with concentrated HCl. During the acidification, a white precipitate formed, was collected, and recrystallized from boiling water to give a white product, 6-(biotinylamino)caproic acid: yield 2.6 g, 79%; mp 221-222 "C; TLC R f = 0.3 (CHCL/MeOH/HzO = 8:3:1). This compound was then coupled to methyl 6-aminocaproate to give 6- [[6-(biotinylamino)caproyllaminolcaproicacid using the same procedure: yield 75 % ;mp 197-199 "C. The NMR spectra of the two intermediates are consistent with their proposed structures. 6-[[6-(Biotinylamino)caproyllaminolcaproic acid (0.5 g, 1.1mmol) was dissolved in 15 mL of DMF at 70 "C and cooled to 40 "C, then 7-amino-4-chloro-3-(2pheny1ethoxy)isocoumarin (0.4 g, 1.3 mmol) was added, followed by the addition of HOBt (0.172 g, 1.3 mmol) and diisopropylcarbodiimide (0.16 g, 1.3 mmol). The mixture was stirred at room temperature overnight and the solvent was removed by evaporation. The crude product was purified by silica gel chromatography using CHCldMeOH/ HOAc = 65:10:3 as an eluant. The eluted product contained HOBt, which was removed by washing several times with 1N HCl. The final product was obtained as a yellow solid: yield 33%; mp 163-165 "C; lH NMR (DMSO) 6 10.30 (s, lH), 8.50 (d, lH), 8.00 (dd, lH), 7.85-7.68 (bs, 2H),7.63 (d, lH), 7.35-7.15, (m, 5H), 6.35 (d, 2H), 4.55 (t, 2H),4.35-4.25and4.15-4.05 (mandm,2H),3.13-2.90(m, 7H), 2.85 and 2.75 (dd and d, 2H), 2.35 (t, 2H), 2.00 (9, 4H), 1.70-1.11 (m, 18H); MS (FAB+) mle 768 (M + 1). Anal. Calcd for C ~ ~ H N N S O ~C, C ~60.96; S : H, 6.56; N, 9.11; C1, 4.60. Found C, 60.72; H, 6.60; N, 9.04; C1, 4.67. 7 4 [6 4 [6 4 Biotinylamino)caproyl]amino]caproyl]amino]-4-chloro-3-methoxyisocoumarin(BIC 7). 6-[[6(Biotinylamino)caproyllamino]caproicacid (0.65 g, 1.4 mmol) was dissolved in 15mL of DMF at 70 "C and cooled to 40 "C, then 7-amino-4-chloro-3-methoxyisocoumarin (0.37 g, 1.7 mmol) was added, followed by the addition of HOBt (0.22 g, 1.7 mmol) and DCC (0.34 g, 1.7 mmol). The mixture was stirred at room temperature overnight, and the solvent was removed by evaporation. The crude product was purified by silica gel chromatography using CHCl3/MeOH/HOAc = 65:10:2 as an eluant. The eluted product contained HOBt which was removed by washing several times with 1N HC1; 'H NMR (DMSO) 6 10.30 (s, lH), 8.55 (d, lH), 8.00 (dd, lH), 7.80-7.70 (bs, 2H), 7.65 (d, lH), 6.40 (d, 2H), 4.35-4.25 and 4.15-4.08 (m and m, 2H),4.03 (s,3H),3.15-2.95 (m, 5H), 2.80and 2.55 (ddand d, 2H), 2.35 (t, 2H), 2.05 (q,4H), 1.70-1.15 (m, 18H); MS (FAB) mle 677 (M+). Anal. Calcd for C3~H44N50,ClS.0.5HzO: C, 55.93; H, 6.59, N, 5.15; Cl, 10.19. Found: C, 55.55; H, 6.57; N, 5.15; C1, 10.06. 7-[[[ [ [[2-(Biotinylamino)ethyl]amino]carbonyl]methyl]carbamoyl]amino]-4-chloro-3-(2-phenylethoxy)isocoumarin (BIC 8). Biotin (1 g, 4.1 mmol) was dissolved in 20 mL of DMF at 70 "C and cooled to 40 "C. CDI (0.97 g, 6 mmol) in 3 mL of DMF was then added and a white precipitate appeared. After stirring at room temperature for 2 h, ethylenediamine (1.34 mL, 20 mmol) in 10mL of DMF was added and the reaction mixture was stirred for another 3 h. After the DMF was removed by evaporation, the semisolid residue was dissolved in 50 mL of refluxing methanol and the unreacted biotin was removed by filtration. The solution was evaporated to
Kam et al.
dryness, the residue was washed with CHCl3 to remove imidazole, dissolved in 6 mL of water, acidified to pH 3.0 with 1N HCl, and evaporated to dryness. The residue was crystallized from methanol to give 1.04 g of biotinylethylenediamine hydrochloride: yield 79 % ; TLC Rf = 0.21 (butanol/acetic acid/HzO = 4:l:l); mp 241-242 "C; lH NMR (DMSO) 6 8.05 (bs,lH), 7.86 (bs, 3H), 6.42 (bd, 2H), 4.31 (t, lH), 4.12 (t,lH), 3.25 (m, 2H), 3.09 (m, lH), 2.83 (m, 3H), 2.56 (d, lH), 2.05 (t, 2H), 1.53-1.28 (m, 6H). B ~ O ~ ~ ~ ~ ~ - N H ( C H ~ ) Z N H - was COC prepared H~NH-BO~ by the reaction of biotinylethylenediamine hydrochloride, DCC, and Boc-Gly in the presence of triethylamine in DMF: yield 53%; TLC Rf = 0.55 (butanol/acetic acid/ HzO = 4:1:1), mp 136-139 "C; NMR is consistent with assigned structure; MS (FAB+) mle 444 (M 1). Deblocking the Boc group from biotin-NH(CH2)2NH-COCHr NH-Boc with trifuoroacetic acid at 0 "C and the addition of saturated HC1 in ethyl acetate to the residue gave biotinyl-NH(CHz)zNHCOCHzNH*HCl: yield 81% ;TLC Rf = 0.32 (butanoVacetic acid/HzO = 4:l:l); mp 146-148 "C; 1H NMR (DMSO) 6 8.43 (be, lH), 8.03 (bs, 3H), 7.90 (bs, lH), 6.41 (bs, 2H), 4.30 (t, lH), 4.13 (m, lH), 3.49 (m, 2H), 3.14 (m, 5H), 2.79 (dd, lH), 2.55 (d, lH), 2.06 (t, 2H), 1.52-1.28 (m, 6H). BIC 8 was synthesized by the reaction of 7-amino-4chloro-3-(2-phenylethoxy)isocoumarin (0.43 g, 1.35 mmol) with CDI (0.24 g, 1.48 mmol) in 8 mL of DMF at 0 "C. After the isocoumarin solution was stirred at room temperature for 4 h, a solution of biotin-NH(CH2)zNHCOCHzNH.HCl(O.47 g, 1.23 mmol) and triethylamine (0.17 ml, 1.23 mmol) in DMF was added. The reaction mixture was stirred at room temperature for 24 h, decolorized,filtered, and evaporated to give a dark greenish residue. The residue was washed with water and 0.5 N HC1, and purified by silica gel chromatography using CHCldMeOH = 5:l as an eluant. The final product was obtained as yellowish green solid: yield 13% ; TLC Rf = 0.46 (CHCldMeOH = 51); mp 192-193 "C dec; lH NMR (DMSO) 6 9.28 (d, lH), 8.31 (8, lH), 8.03 (d, 2H), 7.80 and 7.54 (m, 3H), 7.31-7.20 (m, 5H),6.39 (d, 2H), 4.52 and4.39 (d o f t , 2H), 4.28 (t, lH), 4.11 (t, lH), 3.72 (d, 2H), 3.10 (m, 4H), 3.04 (m, 2H), 2.93 (t,lH), 2.78 (m, lH), 2.55 (d, lH), 2.03 (t, 2H), 1.60-1.22 (m, 6H); MS (FAB+)mle 686 (M+l). Anal. Calcdfor C ~ Z H ~ ~ N G O C,~56.09; C ~ SH, : 5.44; N, 12.27. Found: C, 55.92; H, 5.46; N, 12.15. Substrate Kinetics. Enzymatic hydrolysis rates of peptide p-nitroanilides catalyzed by chymotrypsin, PPE, HLE, Cat G, and trypsin were measured at 410 nm (€410 = 8800 M-l cm-'; Erlanger et al., 1961). A 0.1 M Hepes, 0.5 M NaC1, pH 7.5 buffer containing 5-1076 MezSO was used for all enzymes except for trypsin, where a 0.1 M Hepes, 0.01 M CaC1-2, pH 7.5 buffer was used. Stock solutions of the substrates were prepared in Me2SO and stored at -20 "C. Chymotrypsin and Cat G were assayed with Suc-Val-Pro-Phe-pNA (0.48mM; Tanakaet al., 1985). HLE and PPE were assayed with MeO-Suc-Ala-Ala-ProVal-pNA (0.24-0.47 mM; Nakajima et al., 1979) and SucAla-Ala-Ala-pNA (0.29-0.48 mM; Bieth et al., 1974), respectively. Trypsin was assayed with Z-Glu-Phe-ArgpNA (0.032 mM; Cho et al., 1984). The initial rates were measured at 410 nm using a Beckman 35 spectrophotometer when 10-25 pL of an enzyme stock solution was added to a cuvette containing 2.0 mL of buffer and 25 pL of substrate. HR granzyme A was assayed with Z-Arg-SBzl (0.120 mM; Odake et al., 1991) in the presence of 4,4'dithiodipyridine (Grassetti and Murray, 1967). The initial rates of thiobenzyl ester hydrolysis were measured at 324 nm ( E 324 = 19 800 M-' cm-l) when a 10-25 pL of an enzyme
+
Bloconjugate Chem., Vol. 4, No. 6, 1993 563
Blotlnylated Isocoumarlns
Table I. Inhibition of Serine Proteases by Biotinyl
Ieocoumarin Derivatives' kObJII1 (M-I s-*) Cat
inhibitors
chymotrypsinb
BIC 1 BIC 2 BIC3 BIC 4 BIC 6 BIC6 BIC 7 BIC 8
330 65 1,100 260 260 1,100 170 640
NIh
GC 6.7 13%' 3.3 59
NI 2.2 19%'
HLEd
PPEe 740 NI
20,000 670 77,000 96,000 230 38,000 6.6
470
NI 350 520
trypsinf
HR granzyme A8
1.1 2.2 1.0 10.6 16.6
NI NI 26%'
BIC 1: Spacer = None, OR = O(CH2)2-Ph
NI
NI
BIC 2: Spacer = None, OR = O(CH2)2CH3
7,100
5%' 18%'
BIC 3: Spacer = Aca, OR = O(CH2)2-Ph
Inhibition was measured in 0.1 M Hepes, 0.5 M NaCl (or 0.01 M CaClZ), pH 7.5 buffer, 510% MeZSO and at 25 OC. Suc-ValPro-Phe-pNA (0.48 mM) was used as the substrate for chymotrypsin (0.24-0.47 mM) and Sucand Cat G. MeO-Suc-Ala-Ala-Pro-Val-pNA Ala-Ala-Ala-pNA (0.29-0.48 mM) were used as the substrates for HLE and PPE, respectively. Z-Glu-Phe-Arg-pNA (0.032 mM) was the substrate for trypsin. Z-Arg-SBzl (0.120 mM) was a substrate for human recombinant (HR) granzymeA. b Inhibitor concentrations were 2Cb400 pM. C Inhibitor concentrations were 75-400 pM. Inhibitor concentrations were 1.6-210 pM. e Inhibitor concentrations were 38-78pM. f Inhibitor concentrations were 41-42 pM. 8 Inhibitor concentrations were 160-210 pM. No inhibition after 10 min of incubation of inhibitor and enzyme. Percent inhibition after incubation of inhibitor and enzyme for 10 min. a
stock solution was added to a cuvette containing 2.0 mL of buffer, 150 pL of 5 mM 4,4'-dithiodipyridine, and 25 pL of substrate. The same volumes of substrate and 4,4'dithiodipyridine were added to the reference cell in order to compensate for the background hydrolysis rate of the substrates. Enzyme Inactivation-Incubation Method. An aliquot of inhibitor (15-25 pL) in MezSO was added to 0.5 mL of a buffered enzyme solution (0.06-2.3 pM) to initiate the inactivation. Aliquots (15-25 pL) were withdrawn at various intervals, and the residual enzymatic activity was measured as described above. Pseudo-first-order inactivation rate constants (hobs) were obtained from plots of In ut/uo vs time, and the correlation coefficients were greater than 0.98. Reactivation Kinetics. The reactivation of enzymes inhibited by substituted biotinylated isocoumarins was studied by monitoring the recovery of enzymatic activity upon standing of the inhibited solution at 25 OC after removal of excess inhibitor. Excess inhibitor was removed by two cycles of dilution and concentration by centrifugation at 0 "C with Amicon Centricon-10 microconcentrators. Enzymatic activity of the inhibited solution was assayed at varying time intervals as described above. Reactivation was also measured after the addition of 3-4 units of avidin or streptavidin (1 unit is the amount required to bind 1pg of D-biotin)to 2 nmol of chymotrypsin or PPE which had been inhibited by a biotinylated isocoumarin. This amount of enzyme derivative contained 0.5 pg of biotin. Hydroxylamine-catalyzed reactivation was measured in the presence of excess inhibitor after the addition of buffered hydroxylamine (0.3-0.4 M).
RESULTS AND DISCUSSION Synthesis. Eight biotinylated isocoumarins (BICs)that contained various 3-alkoxysubstituents on the isocoumarin ring and various spacer groups were synthesized from the corresponding 7-aminoisocoumarins (Figure 1). The alkoxy groups on the isocoumarin ring including 3-methoxy, ethoxy, propoxy, and phenylethoxy groups are a major determinant of the specificity of the inhibitors. The biotin moiety was either directly attached to the isocoumarin or
BIC 4: Spacer = Aca , OR = O(CH2)2CH3 BIC 5: Spacer = Aca , OR = OCH2CH3 BIC 6: Spacer = Aca-Aca, OR = O(CH2)2-Ph BIC 7 : Spacer = Aca-Aca, OR = OCH3 BIC 8: Spacer = NH-(CH2)2NHCOCH2NHC0, OR = O(CH2)2-Ph Figure 1. Structures of biotinylated isocoumarin serine protease inhibitors.
Scheme I. Synthesis of Biotinylated Isocoumarins 1. CDI / DMF, Aca-OMe, Et,N;
2. OH-, Hi
Biotin
*
Biotin-Aca-Aca-OH
3. CDI / DMF, Aca-OMe, Et3N; 4. OH', H+
-
3-Alkoxy-7-amino-4-chloro ismoumarin HOBt, DCC / DMF
F! Biotin-NH(CH,),CONH(CH,),COHN
BIC 6, R = CH2CH2Ph
CI
BIC 7, R = CH,
various spacer groups including one or two 6-aminocaproyl (Aca) groups or a NHCHzCHzNHCOCHzNHCO group separating the 7-aminoisocoumarin and biotin. BIC 1and 2 were prepared from the acid chloride of biotin and the corresponding 3-alkoxy-7-amino-4-chloroisocoumarin in DMF using Et3N as a base. BICs 3-5 were prepared from 6-(biotiny1amino)caproylchloride and the corresponding 3-alkoxy-7-amino-4-chloroisocoumarin. BICs 6 and 7 were prepared from 6-[ [6-(biotinylamino)caproyllaminolcaproic acid and the corresponding 3-alkoxy-7-amino-4chloroisocoumarin using the DCC/HOBt coupling method (the synthetic scheme is shown in Scheme I). BIC 8 was prepared by reaction of 7-amino-4-chloro-3-(2-phenylethoxy)isocoumarin and CDI to yield the isocyanate derivative which was then coupled to biotin-NH(CH2)zNH-COCHzNHgHC1in the presence of Et3N. The final products were purified by column chromatography and characterized by NMR, mass spectra, and elemental analysis. Inhibition Kinetics. The biotinylated isocoumarins were tested as irreversible inhibitors of representative serine proteases including chymotrypsin, cathepsin G (Cat G ), porcine pancreatic elastase (PPE), human leukocyte elastase (HLE), trypsin, and HR granzyme A, and second order inhibition rates kod[Il are reported in Table I. The Substituents a t both the 3- and 7-positions of the isocoumarin ring profoundly affected the reactivity and specificity of BICs toward the various serine proteases tested. The specificity of many isocoumarin inhibitors toward various serine proteases has previously been shown to be determined by the nature of the 3-alkoxy group which usually binds to the primary specificity pocket of serine proteases such as chymotrypsin and elastases (Harper and Powers, 1985). As expected, similar trends were observed with BICs. For example, BICs 3,6,and 8, which contain
564
Bloconjugate Chem., Voi. 4, No. 6, 1993
Kam et al.
Scheme 11. Mechanism of Inhibition of Serine Proteases by Biotinylated Isocoumarins and Mechanisms of Reactivation 0
H
00Ser195
Biotin-Spacer0
deacylation COpR
Biotin-SH pacerONm O
Active Enzyme
CI
/
10 (Simple Acyl Enzyme)
OR CI
i
9 H
0 Stable Toward Reactivation COpR
11 (Alkylated acyl enzyme)
a 3-phenylethoxy substituent, were good inhibitors for bovine chymotrypsin, which prefers aromatic amino acids at the PI site2 in good substrates, but these BICs were poor inhibitors for PPE which prefers PI amino acid residues with short alkyl groups in good peptide substrates. Comparison of the inhibition rates of compounds with an Aca spacer and varying 3-alkoxy substituents illustrates the preference of chymotrypsin for the 3-phenylethoxy group. BIC 3 (with a 3-phenylethoxy group) effectively inhibited chymotrypsin, with a k&,/[I] of 1100 M-' s-l, which was 4-fold faster than the inhibition rates of BICs 4 and 5, which had 3-ethoxy or 3-propoxy substituents, respectively. The reactivity of BIC 3 with chymotrypsin is consistent with the results observed in the inhibition of chymotrypsin by the non-biotin analog 7-amino-4-chloro3-(phenylethoxy)isocoumarin,which had a hob$ [I] of 1200 M-l s-1 (Harper and Powers, 1985). BICs with longer spacer groups inhibited serine proteases more potently than comparable BICs that had no spacer or shorter spacers. For example, BICs 3 , 6 , and 8 were 2-3-fold better inhibitors for chymotrypsin than BIC 1 (no spacer). A similar effect was observed in the inactivation rates of HLE, where BIC 4 had a better inhibition rate constant than BIC 2. However, there appeared to be little advantage to spacers longer than a single aminocaproyl group. The longer spacers with two repeated aminocaproyl groups (BIC 6) or the NHCH2CHzNHCOCHzNHCOgroup (BIC 8) afforded no increase in reactivity over BIC 3, which had only a single aminocaproyl spacer. Thus a suitable spacer is needed for good reactivity, but extended spacers appear unwarranted. The biotinylated inhibitors which contained spacers of varying lengths were as reactive or more reactive inhibitors than comparable nonbiotinylated isocoumarins. For example, BIC 5, which had an aminocaproyl spacer, had an irreversible inhibition rate constant of 520 M-l s-l with PPE. This is similar to the rate constant for the nonbiotinylated derivative 7-amino-4-chloro-3-ethoxyisocoumarin, for which kob$[I] = 700 M-l s-1 with PPE. In another comparison, the 3-methoxy derivative BIC 7, with a spacer of two 6-aminocaproyl groups, was actually more reactive than the nonbiotinylated 7-amino-4-chloro-3methoxyisocoumarin with PPE. The rate constants were respectively 7100 vs 1000 M-' s-l. BICs 5 and 7 also inhibited HLE more effectively than the corresponding nonbiotinylated isocoumarin inhibitors by 10- and 4-fold
* The nomenclature used for the individualamino acid residues
(PI, Pz,etc.) of a substrate and the subsites (SI, Sz,etc.) of the enzyme is that Schecter and Berger (1967).
respectively (96 000 vs 9400 M-l s-l; 38 000 vs 10 000 M-l s-l). The 3-phenylethoxy inhibitor BIC 3, which also has an aminocaproyl spacer, inhibited chymotrypsin with a rate constant of 1100 M-1 s-l, which is comparable to the rate constant for inhibition of chymotrypsin by 7-amino4-chloro-3-(phenylethoxy)isocoumarin(1200 M-l s-l). Thus, the inhibitory reactivity of isocoumarins was maintained or even enhanced when there were spacer groups present. This indicates that the spacer group or the biotin moiety may be interacting with the extended substrate binding site of the enzyme. Inhibition Mechanism and Reactivation Kinetics. The mechanism of inactivation of serine proteases by biotinylated isocoumarins is similar to the mechanism of inhibition by nonbiotinylated isocoumarin inhibitors (Scheme 11; Harper and Powers, 1985; Kam et al., 1988). The mechanism involves the initial formation of an acyl enzyme (10) by reaction of the isocoumarin carbonyl group with the active site Ser-195. This reaction unmasks a 4-aminobenzyl chloride functional group, which can eliminate chloride to give a quinone imine methide intermediate. This intermediate can react further with an enzyme nucleophile such as His-57 to give an alkylated enzyme (11) or with a solvent molecule to produce a new acyl enzyme that can regenerate active enzyme upon deacylation. The alkylated enzyme is stable toward reactivation, while active enzyme can be regenerated from the simple acyl enzyme. To assess the extent of alkylation of the active site histidine by BICs after the initial inhibition reaction and formation of the acyl enzyme derivative (lo), we determined the extent of reactivation of the inhibited serine proteases by spontaneous deacylation of the inhibited enzyme during storage at room temperature after removal of excess inhibitors and the extent of reactivation after treatment with the strong nucleophile hydroxylamine. The results obtained with three serine proteases inhibited by five biotinylated isocoumarins are shown in Table 11. Chymotrypsin inhibited by BICs 1 or 3 regained only 1015% of activity after removal of excess inhibitor with Amicon Centricon-10 microconcentrators and storage at room temperature at pH 7.5 for 2 days. In contrast, PPE, which had been inhibited by BICs 2,4, or 5, regained 3566 % of the original enzyme activity during the same 2-day incubation period. This indicates that PPE is undergoing a slow spontaneous deacylation reaction, while chymotrypsin is forming a more stable acyl enzyme derivative. Hydroxylaminewas quite effectiveat regenerating active enzyme from the acyl enzyme derivatives (10). Chymotrypsin regained 100% activity if NHzOH was added
Blotlnylated Isocoumarins
Table 11. Reactivation of Inhibited PPE, Chymotrypsin, and Trypsin by Biotinyl Isocoumarins in Buffer or after Hydroxylamine Treatment* enzyme activity recovered (%) PPE chymotrypsin inhibitors buffer NH,OH* buffer NHzOHb buffer 10 100 BIC 1 BIC 2 35 98 15 100 BIC 3 BIC 4 37 90 33 BIC 5 66 93 92 a Enzyme (1.7 pM) was incubated with inhibitor (40 pM) in 0.1 M Hepes, 0.5 M NaCl, pH 7.5 buffer, 10% MezSO at 25 OC for 40-50 min. The residual enzyme activity was then less than 10%. Recovered enzyme activity was then measured after removal of excess inhibitor by two cyclesof dilution and concentration by centrifugation at 0-5 OC with Centricon-10 microconcentrators and incubation at 25 "C for 2 days. Recovered enzyme activity wm determined 5-60 min after the addition of 0.4 M NHzOH to the inhibited enzyme solution.
immediately after inhibition by BICs 1 or 3. Similarly, PPE regained >90% activity when NHzOH was added immediately after inhibition with compounds 2, 4, or 5. Full recovery of enzyme activity with the addition of NH2OH indicates the formation of a reactivatable acyl enzyme derivative (10) in the case of chymotrypsin and PPE, and not the alkylated derivative 11. We have some data to indicate that there may be a slow conversion of the acyl enzyme derivative (10) to the alkylated form (11). When NHzOH was added after a few hours to an inhibited PPE derivative, less than the 90% of the potential enzyme activity was regained. These results are consistent with a slow conversion of 10 to 11. Effects of Avidin and Streptavidin on the Reactivation Reaction. We next evaluated the potential for egg white avidin or bacterial streptavidin to affect the reactivation of the inhibited enzyme derivatives. Avidin and streptavidin differ considerably in their net charges, with pls of -10 and -5, respectively. We were initially concerned that the interactions between the biotinylated, inhibited serine proteases and avidin or streptavidin might promote the deacylation reaction and thereby restrict the utility of the biotinylated protease derivatives for detection and/or recovery of biotinylated proteases. Reactivation of chymotrypsin and PPE inhibited by BICs 1-4 were measured in the presence of excess avidin or streptavidin (Table 111). The acidic bacterial protein streptavidin appeared to promote reactivation of chymotrypsin inactivated by BICs 1 and 3 and slightly increased reactivation of PPE inactivated by BICs 2 and 4. In contrast, interactions with an excess molar concentration of the basic protein egg white avidin caused no changes in enzyme recovery, even though both avidins have similar affinities for biotin. It is our conclusion that similar evaluations should be performed for each biotinylated, inactivated serine protease-avidin combination being considered for cytochemical use or for use in serine protease isolation. The spacers between biotin and the isocoumarin ring serve the purpose of increasing accessibility of avidin to the biotin associated with enzyme-inhibitor complexes. The spacer could also affect the stability of the avidinBIC-protease complex. For example, the insertion of a 6-aminocaproic acid spacer between the biotin and insulin exerted a stabilizing effect on the complexes with avidin, when biotinylated hormones were used as ligands to isolate the hormone receptors on avidin-Sepharose columns (Hofmann et al., 1984). With BICs, the lack of a spacer
Bloconlugete Chem., Vol. 4, No. 8, 1993 585
Table 111. Effect of Streptavidin and Avidin on Reactivation of Inhibited Chymotrypsin and PPE Derivatives* enzyme activity recovered ( % ) chymotrypsin PPE buffer streptavidin avidin buffer streptavidin avidin BIC1 10 27 10 BIC 2 35 51 32 BIC3 15 39 12 BIC 4 37 43 32 a Enzyme (1.7 pM) was incubated with inhibitor (40 pM) in 0.1 M Hepes, 0.5 M NaCl, pH 7.5 buffer, 10% MezSO at 25 OC for 40-50 min. The residual enzyme activity was then less than 10%. Recovered enzyme activity was then measured after removal of excess inhibitor by two cyclesof dilution and concentration by centrifugation at 0-5 OC with Centricon-10 microconcentrators. Then 3-4 units of avidin or streptavidin (6-8 pM)were added and enzymatic activity was monitored for 2 days.
decreased the reactivity of the inhibitors, but had little effect on the recovery of enzyme activity upon treatment with avidin or streptavidin. In the case of a BIC with no spacer, it is possible that the avidin is unable to interact with the biotin portion of the inhibited enzyme. However, in the case of enzyme derivatives with longer spacers, we have demonstrated interaction of the biotin with avidin by two methods. The first involves column chromatography (followingsection) and the second involves detection on protein blots. The biotin in chymotrypsin-inhibitor complexesand lymphocyte granzyme-inhibitor complexes formed with BICs 6 and 7 can be detected on protein blots developed with avidin-alkaline phosphatase (N. J. Allison and D. Hudig, unpublished results). Thus it is clear that the biotin in the complexes is accessible to avidin. Isolation of BIC-Inhibited PPE via an Avidin Column. One use of biotinylated isocoumarins would be the isolation of new serine proteases using the strong binding of biotin to avidin in BIC-protease complexes. If the enzyme-inhibitor complex remains in the form of an acyl enzyme (10) for a period of several hours, the serine protease could be released from the isolated complex by deacylation with NH20H treatment or an increase of the pH to >8. We have used PPE as a model enzyme to evaluate the potential utility of BICs for serine protease isolation and recovery. Experimental results are shown in Table IV. P P E was inhibited by BIC 5 or 7 in pH 7.5 Hepes buffer with effective loss of all enzyme activity (step 1). Excess inhibitors were removed by gel filtration on a Sephadex G-25 column in a pH 5 acetate buffer (step 2). A low pH buffer was used to avoid reactivation of inhibited enzyme. After gel filtration, the inhibited PPE solution regained a small amount of enzyme activity (
