Bioconjugate Chem. 1994, 5, 400-405
400
Fluorescent Derivatives of Diphenyl [1-(N-Peptidylamino)alkyl]phosphonateEsters: Synthesis and Use in the Inhibition and Cellular Localization of Serine Proteases Ahmed S. Abuelyaman, Dorothy Hudig,t Susan L. Woodard,t a n d J a m e s C. Powers* 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 April 5, 1994@
Three fluorescein- and one Texas Red-labeled derivatives of [ 1-(N-dipeptidylamino~alkyllphosphonate diphenyl esters were synthesized and evaluated as inhibitors of serine proteases. The two fluorophores, FITC and TXR,were attached to the peptide phosphonates via a n 6-aminocaproyl unit that acts as a spacer group and facilitates the binding of the phosphonate inhibitor to the targeted enzymes. These derivatives are potent and specific inhibitors of chymotrypsin, porcine pancreatic elastase (PPE), and human leukocyte elastase (HLE). FTC-Aca-Phe-Leu-PheP(OPh)z (3) inhibited chymotrypsin very potently (kobsd/[I] = 9500 M-l s-l) and 600-fold better than it did PPE (kobsd/[I] = 16 M-l s-l). FTCAca-Ala-Ala-MetP(OPh)z(1) was a more effective inhibitor of chymotrypsin (kobsd/[I] = 190 M-I s-l) than PPE and HLE (kobsd/[I] = 13 and 22 M-l s-l, respectively). Only HLE and PPE were inhibited ( 2 ) (kobsd/[I] = 41 and 22 M-l s-l, respectively). The specificity of by FTC-Aca-Ala-Ala-AlaP(OPh)z these inhibitors toward the targeted serine proteases depends on the sequence of the tripeptide portion and was not affected by the presence of the fluorescent label. Trypsin, for instance, was not inhibited by any of these compounds. In some cases, the inhibitory potency was increased by the fluorescent label. For example, chymotrypsin was inhibited by the fluorescent compounds, FTC-Aca-Ala-AlaMetP(OPh)z (1) and FTC-Aca-Phe-Leu-PheP(OPh)z (3), more potently than by the nonfluorescent compounds, Boc-Ala-Ala-MetP(0Ph)z ( 5 ) and Z-Phe-Leu-PheP(OPh)z (7). Initial experiments with cytotoxic lymphocytes indicate that FTC-Aca-Ala-Ala-MetP(OPh)z labels discrete granule-like regions where the serine proteases of the NK cell line, RNK-16, are stored.
INTRODUCTION
vestigations with inhibitors showed that the chymotrypsin-like proteases of the lymphocyte granules play an important role in perforin-dependent cytolysis (Ewoldt et al., 1992; Hudig et al., 1993). However, despite a considerable amount of research effort, the exact role of the serine proteases in cytolysis is unknown. In this paper, we report several fluorescent derivatives which should be useful in elucidating the biological role of granzymes in cytolysis. Diphenyl [1-(N-peptidylamino)alkyl]phosphonateesters are potent irreversible serine protease inhibitors (Oleksyszyn and Powers, 1991). In these inhibitors, the scissile peptide bond of a substrate is replaced by a diphenyl phosphonate functional group (Figure 1). Peptide phosphonates are highly specific inhibitors since proper interactions with the S1 pocket of the target serine protease are prerequisites for nucleophilic attack by the active site serine hydroxyl group on the phosphorus atom. + University of Nevada. This attack gives, via a pentacovalent intermediate, a * To whom correspondence should be addressed. very stable phosphonyl derivative (Powers et al., 1993). Abstract published in Advance ACS Abstracts, July 1, 1994. The selection of the right peptide sequence in these Abbreviations: Ac, acetyl; Aca, 6(or 6)-aminocaproyl; Boc, inhibitors produces a n important direct effect on the tert-butyloxycarbonyl; DCC, N,N '-dicyclohexylcarbodiimide;DCU, specificity and the rate of their interactions with serine dicyclohexylurea; DMF, N,N-dimethylformamide; FAB, fast proteases. For example, Z-PheP(OPh)zand Suc-Val-Proatom bombardment; FITC, 5-fluorescein isothiocyanate; FTC, PheP(OPh)z inhibited chymotrypsin potently with kohsd/ 5-fluoresceinyl(thiocarbamoyl); Hepes, N-(2-hydroxyethyl)pip[I] of 1200 and 44 000 M-l s-l, respectively, but they did erazine-N '-2-ethanesulfonic acid; HLE,human leukocyte elastase; HOBt, 1-hydroxybenzotriazole; Me, methyl; MeO, methoxy; Meznot inhibit porcine pancreatic or human leukocyte elastase SO, dimethyl sulfoxide; NK, natural killer lymphocyte; NMM, (designated PPE or HLE, respectively) (Oleksyszyn and N-methylmorpholine; Ph, phenyl; pNA, p-nitroanilide; PPE, Powers, 1991). The unaffected elastases cleave after Val, porcine pancreatic elastase; rt, room temperature; SUC, succinyl; Leu, or Ile. The inhibitor Boc-Val-Pro-ValP(OPh)z,on the T cell, thymus-derived lymphocyte; TEA, triethylamine; TXR, Texas Red, 9-[2(or 4)-(chlorosulfonyl)-4(or 2)sulfophenyl]-2,3,6,7,- other hand, inhibited PPE with a n inhibition rate of 12,13,16,17-0ctahydro-lH,5H, 1lH,15H-~antheno[2,3,4-&5,6,7- 11 000 M-l s-l and HLE with 27 000 M-' s-l, but did not inhibit chymotrypsin. Phosphonate inhibitors also i 3 ']diquinolizin-18-ium hydroxide, inner salt; TLC, thin layer chromatography; Z, benzyloxycarbonyl. have a number of other advantages. They are unreactive The lymphocyte serine proteases, which are also known as granzymes, are a group of proteolytic enzymes found in specialized granules of natural killer (NK)l cells and cytotoxic T lymphocytes (Tschopp et al., 1988). Studies with synthetic substrates revealed that at least five different substrate specificities are present among these proteases, namely, tryptase (trypsin-like, cleaving after arginine or lysine), Asp-ase (cleaving after aspartic acid), chymase (chymotrypsin-like, cleaving after phenylalanine, tryptophan, or tyrosine), Met-ase (cleaving after methionine), and Ser-ase (cleaving after serine) (Hudig et al., 1991). Upon killing, T and NK lymphocytes release the pore-forming protein, perforin, and several serine proteases which are believed to be essential for lymphocyte-mediated cytolysis. It is not clear if all the granzymes are involved in lymphocyte-mediated cytolysis, but in-
@
1043-1802/94/2905-0400$04.50/0 0 1994 American Chemical Society
Bioconjugate Chem., Vol. 5, No. 5, 1994 401
Fluorescent Phosphonate Protease Inhibitors
0
R
peptidyl,
/I
peptidyl,
OPh 6OPh
0 peptide substrate
peptide phosphonate
Serine Profease Active Site
CHz
7
0
FfC-Aca-y/lrNpyA H 0 CHz H
CHz r \ ,OPh OPh
.O
.! /
\
oxyanion hole
Serine Protease Active Sife
cell hybridoma 3A1 antibodies. Both fluors were then detected together on individual cells indentified a s simultaneous, two-color fluorescent events using dualparameter flow microfluorometry (Titus et al., 1982). In this paper we report the synthesis of several tripeptide phosphonate inhibitors labeled with FITC or TXR and the kinetics of their inhibition of several serine proteases. The fluorophores were coupled to the peptidyl phosphonates using an eaminocaproyl unit as a spacer group between the peptide and the fluorophore. We expected that the spacer would prevent unfavorable steric interactions between the fluorophore and the active site of the protease. We found that the four new inhibitors with the fluor-spacer-peptide-phosphonate structure are reactive and selective among serine proteases of different specificities. In addition, we report that FTCAca-Ala-Ala-MetP(OPh)2 irreversibly labels discrete granule-like regions of the NK cell line, RNK-16. The properties of these fluorophores indicate that these peptide phosphonates will be excellent tools for the study of the distribution of serine proteases in lymphocytes and their role during killing. Use of these inhibitors with different peptide sequences with varying specificities and different fluorophores will allow the simultaneous detection of different proteases in cytotoxic lymphocytes as well as in other biological systems. MATERIALS AND METHODS
Figure 1. Comparison of the structure of a peptide substrate for a serine protease and a peptidyl phosphonate inhibitor (top) and illustration of the inhibitor-enzyme complex formed with compound 3 (bottom). In the top illustration, the arrow indicates the scissile bond of the substrate. The phosphorus atom in the phosphonate is shown in bold throughout. The tetracovalent bond between the phosphonate and the enzyme active site serine in the enzyme-inhibitor complex is shown in bold. The interactions between the peptide amino acid residues and the enzyme sites SI, Sz, and Ss are numbered using the nomenclature of Schechter and Berger (Schechter and Berger, 1967).
toward most other nucleophiles, nontoxic to cells, stable under physiological conditions, and form stable phosphonylated enzyme derivatives upon reaction with serine proteases. Fluorescent derivatives of proteins, synthetic peptides, and inhibitors are important tools for the detection, localization, and quantification of numerous cellular constituents in biological systems. For example, fluorochrome-labeled gene probes are used for the rapid and quantitative detection of homologous RNA a t the single cell level (Pachmann et al., 1991). Proteins and antibodies labeled with fluorescein isothiocyanate (FITC) and the sulfonyl chloride of sulforhodamine 101 (commercially designated a s Texas Red) have been used to study the distribution of the receptors for IgE and IgG on basophilic leukemia cells (Titus et al., 1982). In many cases, proteins and synthetic peptides did not lose their biological activity after being labeled with FITC and Texas Red (TXR).These two fluorophores have the advantage that the excitation and emission spectra of their conjugates are widely separated from each other, a characteristic that permits them to be detected simultaneously. For example, to determine that T cells have Ig receptors, human peripheral blood leukocytes were treated with unlabeled normal rabbit IgG, washed, and then treated with Texas Red-conjugated goat anti-rabbit IgG antibodies together with FITC-conjugated mouse anti-human T
Materials. 5-Fluorescein isothiocyanate (FITC), sulforhodamine 101, 6-aminocaproic acid (Aca), and all common reagents and solvents were purchased from Aldrich Chemical Co., Milwaukee, WI. Porcine pancreatic elastase (PPE) was obtained from United States Biochemical Corp., Cleveland, OH. Human leukocyte elastase (HLE) was obtained from Athens Research and Technology, Inc., Athens, GA. Hepes was obtained from Research Organics, Inc., Cleveland, OH. Bovine trypsin was purchased from Sigma Chemical Co., St. Louis, MO. Preparative thin-layer chromatography was performed with plates precoated with 2 mm of silica gel G.F. and were obtained from EM Separations, Gibbstown, NJ 08027. NMR spectra were recorded on a Varian GEMINI 300. Elemental analyses were performed by the Atlantic Microlabs, Atlanta, GA. Diphenyl [ 1-(N-dipeptidylamino)alkyl]phosphonate esters were synthesized as previously described (Oleksyszyn and Powers, 1991; A. Abuelyaman, D. Jackson, D. Hudig, and J . Powers (unpublished results)). The sulfonyl chloride of sulforhodamine 101 (Texas Red) was prepared as previously described (Titus et al., 1982) and was used without further isolation in coupling reactions. 6-[6-Fluoresceinyl(thiocarbamoyl)aminolcaproic Acid [FTC-Aca-OH]. 5-Fluorescein isothiocyanate (0.20 g, 0.51 mmol) was dissolved in 5 mL of DMF. A solution of methyl 6-aminocaproate (0.15 g, 1.03 mmol) in 1 mL of DMF was added a t rt, and the mixture was stirred for 0.5 h. The solvent was removed in uucuo. The residue was purified on a silica gel column eluted with CHC13:MeOH (4:l). Fractions with Rf = 0.38 were collected and concentrated to give a dark orange oily residue which was triturated with HzO t o give FTC-AcaOMe as an orange sheetlike solid: yield 88%; one spot on TLC (Rf=0.70, CHC13:MeOH:HOAc, 16:3:1);‘H NMR (DMSO-d6)6 10.38-10.05 (m, 3H), 8.32 (bs, lH), 8.24 (s, lH), 7.73 (dd, lH), 7.16 (dd, lH), 6.70-6.52 (m, 6H), 3.59 (s, 3H), 3.47 (bs, 2H), 2.32 (t, 2H), 1.65-1.50 (m, 4H), 1.40-1.25 (m, 2H); high-resolution FAB-MS, mle (M H) calcd 535.1539, found 535.1560. Anal. Calcd for CZSHzsNz07S.HzO: C, 60.86; H, 5.11; N, 5.07; S, 5.80. Found: C, 60.53; H, 5.11; N, 5.05; S, 5.73.
+
Abuelyaman et al.
402 Bioconjugate Chem., Vol. 5. No. 5, 1994
6-[4(or2)-[9-[2,3,6,7,12,13,16,17-octahydro-lH,5H,A solution of 1 N NaOH (3.0 mL) was added to FTC1lH, 15H-xantheno[2,3,4-i:5,6,74y 'I diquinolizinylAca-OMe followed by a minimum of MeOH to give a clear 18-ium]]3(or5)-sulfo-l-phenylsulfonamidolcaproic solution that was stirred a t rt for 1h. Most of the MeOH Acid, Hydroxide, Inner Salt [TXR-Aca-OH]. The was removed in vacuo, and the aqueous solution was intermediate TXR-Aca-OMe was synthesized from a placed in a n ice bath. Drops of concentrated HC1 were freshly prepared dry Texas Red solution in CHC13 (Titus added with stirring until the mixture became just acidic et al., 1982) and 1 equiv of 6-aminocaproic acid methyl (pH = 3-4). The orange suspension that formed was ester in the presence of 1 equiv of NMM. After being cooled for 3 additional h. The solid was isolated by stirred a t 0 "C for 0.5 h and a t rt overnight, the mixture vacuum filtration and dried to give FTC-Aca-OH as an was concentrated in vacuo to give a dark solid that was orange solid: yield 95%; one spot on TLC (Rf = 0.51, purified on a silica gel preparative plate using CHCl3: CHC13:MeOH:HOAc, 16:3:1);'H NMR (DMSO-&) 6 12.00 Me0H:HOAc (16:3:1) to give TXR-Aca-OMe: yield, 67%; (bs, lH), 10.15 (s, 2H), 9.90 (bs, lH), 8.30 (s, lH), 8.15 one spot on TLC (Rf= 0.71, CHC13:MeOH:HOAc,16:3: (bs, lH), 7.80 (d, lH), 7.20 (d, lH), 6.75-6.52 (m, 6H), 1);IH NMR (DMSO-&) 6 8.42 (d, lH), 7.98-7.85 (m, 2H), 3.60-3.40 (bs, 2H), 2.25 (t, 2H), 1.62-1.48 (m, 4H), 1.427.35 (d, lH), 6.52 (s, 2H), 3.60-3.40 (s and m, l l H ) , 3.051.27 (m, 2H); MS (FAB+)mle 521 (M 1). Anal. Calcd 2.95 (m, 4H), 2.92-2.80 (m, 2H), 2.67-2.55 (m, 4H), 2.27 for CZ~HZ~NZO~S.O.~HZO: C, 61.24; H, 4.76; N, 5.29; S, (t, 2H), 2.11-1.75 (m, 8H), 1.55-1.18 (m, 6H); high6.05. Found: C, 61.42; H, 4.58; N, 5.14; S, 6.00. Diphenyl [1-[[[[6-[5-Fluoresceinyl(thiocarbamo- resolution FAB-MS, mle (M H) calcd 734.2570; found 734.2568. Anal. Calcd for C38H43N308Sz-2.5H20: C, y1)aminolcaproyl]alanyl]alanyl]amino1-34methylthio)propyl]phosphonate [FTC-Aca-Ala-Ala-MetP- 58.60; H, 6.21; N, 5.39; S, 8.23. Found: C, 58.41; H, 5.83; N, 5.33; S, 8.09. (OPh)2, 11 (General Procedure). FTC-Aca-OH (0.13 An excess solution of 1 N NaOH was added to TXRg, 0.25 mmol) and the hydrochloride of H-Ala-Ala-MetPAca-OMe followed by a few drops of MeOH. The mixture (0Ph)z (0.13 g, 0.25 mmol) were dissolved in 25 mL of was stirred a t rt for 2 h and then cooled in an ice bath. DMF followed by addition of 1 equiv of TEA. The Concentrated HC1 was added carefully until a dark solid solution was stirred in a n ice bath for 15 min, and then completely precipitated. The solid was isolated by vacuum DCC (0.05 g, 0.25 mmol) was added and the mixture was filtration and then purified on a preparative plate to give stirred a t 0 "C for 4 h and a t rt for 48 h. The solvent TXR-Aca-OH: yield, 85%; one spot on TLC (Rf= 0.55, was removed in vacuo, and the residue was purified on CHC13:MeOH:HOAc, 16:3:1); lH NMR (DMSO-&) 6 8.40 a silica gel column eluted with CHC13:MeOH (9:l). (d, lH), 8.05-7.88 (m, 2H), 7.37 (d, l H ) , 6.50 (s, 2H), Fractions containing product were combined and con3.60-3.40 (m, 8H), 3.05-2.95 (m, 4H), 2.90-2.80 (m, 2H), centrated in vacuo to give a yellow oil that was triturated 2.68-2.58 (m, 4H), 2.10-1.95 (m, 6H), 1.90-1.78 (m, 4H), with water to give a bright yellow solid: yield 25%; 'H 1.50-1.15 (m, 6H); high-resolution FAB-MS mle (M NMR (DMSO-&) 6 10.15 (s, 2H), 9.95-9.82 (bs, lH), 8.45 H) calcd 720.2435; found 720.2413. (d, lH), 8.25 (s, lH), 8.18-7.98 (m, 3H), 7.73 (d, l H ) , Diphenyl [1-[[[[6-[4(or2)-[9-[2,3,6,7,12,13,16,17-0~7.49-7.35 (m, 4H), 7.36-7.10 (m, 7H), 6.73-6.52 (m, 6H), tahydro-lH,,SH,llH,lSH-xanthen0[2,3,4-@5,6,7-i 3 'I4.89-4.67 (m, l H ) , 4.42-4.20 (m, 2H), 3.58-3.42 (bs, diquinolizinyl-l8-iumll-3(or5)-sulfo-l-phenylsul2H), 2.72-2.32 (m, 2H), 2.22-1.92 (m and s, 7H), 1.63fonamidolcaproyl]phenylalanyl]leucyl]amino]-21.42 (m, 4H), 1.38-1.10 (m, 8H); MS (FAB+)mle 1004 C,Z . ~ .phenylethyllphosphonate, ~HZO: Hydroxide, Inner Salt (M Na). Anal. Calcd for C ~ ~ H ~ ~ N ~ O ~ ~ P S [TXR-Aca-Phe-Leu-Phep(OPh)2, 41. This compound 58.32; H, 5.49; N, 6.94; S, 6.35. Found: C, 58.43; H, 5.72; was prepared from TXR-Aca-OH(0.084 g, 0.12 mmol) and N, 6.68; S, 6.01. Diphenyl [ 1-[[ [[6-[5-Fluoresceinyl(thiocarbam- H-Phe-Leu-PheP(OPh)2hydrochloride (0.084 g, 0.13 "01) using the DCC/HOBt method in the presence of TEA. The oyl)amino]caproyl]alanyllalanyllaminolethyllreaction was carried out in CHC13 a t 0 "C for 2 h and a t phosphonate [FTC-Aca-Ala-Ala-AlaP(OPhh,21. The rt for 48 h. The crude product was purified on a silica general procedure for compound 1 was used, starting gel preparative TLC plate using CHC13:MeOH:AcOH(16: with H-Ala-Ala-AlaP(OPh)z. Crude product was purified 3:l) as the eluting solvent. The isolated product was on a silica gel preparative TLC plate using CHC13:MeOH dissolved in 20 mL of CHC13 and extracted twice with (85:15) as the eluant solvent to give a yellow solid: yield 10 mL of 5% aqueous NaHC03 and then with 10 mL of 35%; 'H NMR (DMSO-&) 6 10.30-10.05 (bs, 2H), 9.98HzO. The organic layer was dried (Na2S04) and then 9.85 (bs, lH), 8.55 (t, lH), 8.25 (s, lH), 8.18-7.95 (m, concentrated to give a dark purple solid: yield, 25-35%; 3H), 7.75 (d, lH), 7.48-7.32 (m, 4H), 7.28-7.08 (m, 7H), one spot on TLC, Rf = 0.79, CHC13:MeOH:AcOH (16:3: 6.75-6.52 (m, 6H), 4.78-4.55 (m, lH), 4.45-4.18 (m, 2H), 1); NMR spectrum was recorded and was consistent with 3.58-3.38 (m, 2H), 2.12 (t,2H), 1.62-1.05 (m, 15H); MS the proposed structure; high-resolution FAB-MS, mle (M (FAB') mle 921 (M). Anal. Calcd for C47H48N5011H) calcd 1315.501, found 1315.495. Anal. Calcd for PS.HzO: C, 60.06; H, 5.37; N, 7.50; S, 3.40. Found: C, C ~ ~ H ~ ~ N ~ O ~ Z PC,S63.98; Z . ~ HH,~6.18; O : N, 6.22; S, 4.74. 60.04; H, 5.35; N, 7.36; S, 3.39. Diphenyl [ 1-[[[[6-[5-Fluoresceinyl(thiocarbam- Found: C, 63.62; H, 6.03; N, 6.39; S, 4.87. Enzyme Assays. The hydrolysis of peptide p-nitrooyl)amino]caproyl]phenylalanyllleucyllaminol -2phenylethyllphosphonate [FTC-Aca-Phe-Leu-PheP- anilide substrates, catalyzed by chymotrypsin, PPE, HLE, and trypsin was measured in 0.1 M Hepes and 0.5 (OPh)2,31. The general procedure for compound 1 was M NaCl (0.01 M CaClz for trypsin), pH 7.5 buffer used, starting with H-Phe-Leu-PheP(OPh)z: yield 36%; containing 5-10% MezSO a t 25 "C. Stock solutions of 'H NMR (DMSO-dtj)6 10.15 (s, 2H), 9.85 (bs, lH), 8.87 substrates were prepared in MezSO (20 mM) and stored (d, lH), 8.77 (d, lH), 8.25 (s, lH), 8.10-7.90 (m, 3H), 7.75 a t -20 "C. Final substrate concentrations were 0.24 mM. (d, lH), 7.45-7.05 (m, 21H), 6.72-6.52 (m, 6H), 4.92Chymotrypsin activity was assayed with Suc-Val-Pro4.75 (m, lH), 4.62-4.32 (m and m, 2H), 3.50-2.55 (m, Phe-pNA (Tanaka et al., 1985). PPE was assayed with 6H), 2.09-1.92 (m, 2H), 1.60-0.90 (m, 9H), 0.85-0.65 Suc-Ala-Ala-Ala-pNA (Bieth et al., 1974). HLE was (m, 6H); MS (FAB') mle 389.9 (loo%, M' - fluoresceinylassayed with MeO-Suc-Ala-Ala-Pro-Val-pNA (Nakajima NHCS), 1116 (20%, M 1). Anal. Calcd for C62H62et al., 1979) and trypsin was assayed with Z-Arg-pNA Ns011PS.HzO: C, 65.66; H, 5.69; N, 6.17; S, 2.82. (Kanaoka et al., 1977). The initial rates of hydrolysis Found: C, 65.98; H, 5.68; N, 6.21; S, 2.80.
+
+
+
+
+
+
Bioconjugate Chem., Vol. 5, No. 5, 1994 403
Fluorescent Phosphonate Protease Inhibitors
were measured a t 410 nm (€410 = 8800 M-' cm-l (Erlanger et al., 1961))on a Beckman 35 spectrophotometer after 25-50 pL of a n enzyme stock solution was added to a cuvette containing 2.0 mL of buffer and 25 pL of substrate. Inhibition Kinetics-Incubation Method. Each inhibition reaction was initiated by adding a 50-pL aliquot of inhibitor (100-5000 pM in MeZSO) to 0.5 mL of a 0.1 M Hepes, 0.5 M NaCl(O.01 M CaClz for trypsin), pH 7.5 buffer containing 50 pL of a stock enzyme solution a t 25 "C. The enzyme stock solutions were 20 pM chymotrypsin, trypsin, and PPE in 1mM HCl (pH 3) and 0.4-4 pM HLE in 0.25 M NaAc and 1M NaCl a t pH 5.5. All the enzyme stock solutions were stored a t -20 "C prior to use. Aliquots (25 pL) were withdrawn a t various intervals, and the residual enzymatic activity was measured spectrophotometrically as described above. Pseudofirst-order inactivation rate constants ( k o b s d ) were obtained from plots of In uJu, vs time and had correlation coeficients greater than 0.98. Each Kobsd was calculated from 5-10 activity determinations which extended to 2-3 half-lives. Control experiments were carried out in the same way as described above except MeZSO was added in place of the inhibitor solution in MeZSO. The initial rates of substrate hydrolysis did not change during the first 60 min of incubation. These initial rates were used as u, in the calculation of the inhibition rate constants. Cell Labeling and Imaging. Cells of the NK line RNK-16 (Ward and Reynolds, 1983)were labeled as live cells, washed, treated with methanol as a fixative and permeabilizing agent, and then examined by confocal microscopy. The cells were treated a t 1 x lo7 cells/mL in RPMI 1640 culture media (Sigma Chemical Co., St. Louis, MO) containing 10 mM Hepes with 0.1 mM FTCAca-Ala-Ala-MetP(OPh)zfor 30 min a t 37 "C. Control cells were treated with the same volume of DMSO required to deliver the inhibitor (typically 1% final concentration). Cells were then washed several times with phosphate-buffered saline (PBS) to remove excess inhibitor. They were suspended overnight in cold 80% methanol to permeabilize the membrane. The cells were washed free of MeOH into PBS and fixed onto poly-Llysine (Sigma) coated microscope slides using 3% paraformaldehyde. Fixed slides were washed in PBS, coated with Vectashield (Vector Laboratories, Inc., Burlingame, CA) to prevent fluorescence fading, and topped with a coverslip. Fluorescent laser scanning confocal microscopy was done with a Bio-Rad MRC 600 confocal system utilizing a Zeiss Axiophot microscope and equipped with
N=C=S
&
0
HO
FITC
Texas red
(TXR).
a mixed-gas argonkrypton ion laser using 488-nm excitation for fluorescein. RESULTS AND DISCUSSION
Synthesis. Three fluorescein labeled and one Texas Red labeled diphenyl tripeptidylphosphonate esters were synthesized. They were made by joining two synthetic intermediates: a fluorophore-6-aminocaproic acid unit and a tripeptide phosphonate. The 6-aminocaproyl acts as a spacer group to facilitate the correct binding of the phosphonate inhibitors to the targeted serine proteases. The fluorophores, FITC and TXR (Figure 21, were incorporated into the .+caproyl units as follows. In the case of the FTC derivatives, commercially available fluorescein isothiocyanate, FITC, was coupled to methyl 6-aminocaproate, followed by saponification of the ester group to give FTC-Aca-OH. The TXR derivative, TXR-Aca-OH, was obtained from coupling the sulfonyl chloride functional group in Texas Red with methyl 6-aminocaproate in presence of 1equiv of TEA. Hydrolysis of the methyl ester resulted in TXR-Aca-OH. The procedure we used in preparing these two intermediates was very efficient and resulted in high yields (70-80%). Both FTC-AcaOH and TXR-Aca-OH were then coupled to the hydrochloride salts of tripeptidyl phosphonates using the DCCI HOBt method in the presence of TEA (Figure 3). The products were purified on silica gel preparative plates. The final products were characterized by NMR, mass spectroscopy and elemental analysis. The isolated Texas Red derivative compound (4) was initially isolated as a mixture of the zwitterionic form and as an acetate salt mixture, a result that was evident from the NMR and elemental analysis. Pure zwitterionic form was obtained after dissolving the crude product in CHC13 and then extracting with 5% aqueous NaHC03. Kinetic Studies. The specificity of these phosphonate inhibitors is dependent upon the amino acid sequence in
FTC-Aca-OH
1. H,N(CH,),COOMe
+ 2. OH, H30'
-
or
TXR- Aca-OH
Attachment of the E-Aminocaproic Acid Units to Peptide Phosphonates Fluor-Aca unit ipeptide phosphonate
FTC-Aca-OH
+
TXA-Aca-OH
+
Fluor-Aca-peptide phosphonate
Ala-Ala-MetP(OPh), Ala-Ala-AlaP(OPh), Phe-Leu-PheP(OPh),
1 FTC-Aca-Ala-Ala-MetP(OPh), 2 FTC-Aca-Ala-Ala-AlaP(OPh)2 3 FTC-A~a-Phe-Leu-Phe~(0Ph)~
Phe-Leu-Phe'(OPh),
4 TXR-Aca-Phe-Leu-PheP(OPh)2
or
TXR
Figure 2. Structures of fluorescein isothiocyanate (FITC) and
Synthesis of the Fluorescent e-Aminocaproic Acid Intermediates
FITC or TXR
SOPCl
Figure 3. Scheme for the synthesis of FITC and TXR labeled peptidyl phosphonates.
Abuelyaman et al.
404 Bioconjugafe Chem., Vol. 5, No. 5, 1994 Table 1. Inhibition of Serine Proteases by Fluorescent Peptide Phosphonatep
inhibitor
[IIOtM) FTC-Aca-Ala-Ala-MetP0Ph)z(1) 8.3 8.3 FI'C-Aca-Ala-Ala-AlaP(OPh)2 (2) FI'C-Aca-Phe-Leu-PheP(OPh)z (3) 8.3 TXRAca-Phe-Leu-PheP(OPh)2(4) 8.3 Boc-Ala-Ala-MetP(OPh)2 ( 5 ) 210 Z-Ala-Ala-AlaP(OPh)2 ( 6 ) 41.7 Z-Phe-Leu-PheP(OPh)2 (7) 10.4 a
chymetrypsin 190 NIb 9500 15 3 NI 110
PPE HLE 13 22 22 41 16 252 11 96 3 2 30 38 NI 27
Inhibition kinetics were measured in 0.1 M Hepes, 0.5 M NaCI,
pH 7.5 buffer, 5-1096 MezSO and a t 25 "C. No inhibition after 40 min of incubation of inhibitor and enzyme. No inhibition was
observed with trypsin after 1 h of incubation of the enzyme and the inhibitor in 0.1 M Hepes, 0.01 M CaC12, pH 7.5.
the tripeptide portion of the inhibitor. In each case, the amino acid sequence was chosen based on the specific sequence of a good substrate or inhibitor for the target enzyme. For example, previous studies showed that human Q31 chymase, cathepsin G, and related chymotrypsin-like enzymes have significant hydrolysis activity toward Suc-Phe-Leu-Phe-SBzl and were potently inhibited by the corresponding peptide chloromethyl ketone, Suc-Phe-Leu-Phe-CH2C1(Odake et al., 19911. Likewise, the p-nitroanilide substrates Suc-Ala-Ala-Ala-pNA and Boc-Ala-Ala-Ala-pNA were hydrolyzed effectively by PPE (Bieth et al., 1973, 1974) and peptide chloromethyl ketones such as Ac-Ala-Ala-Ala-AlaCHaCl are good inhibitors of HLE (Tuhy and Powers, 1975). The fluorescent compounds were evaluated as inhibitors of chymotrypsin, PPE, and HLE and were found to be potent and very specific with inactivation rate constants (k,hd[I]) as high as 9500 M-l s-l (Table 1). For instance, the Phe derivative, compound 3, inhibited chymotrypsin very potently (ko&IJ = 9,500 M-l s-l) and 600 times faster than it inhibited HLE (kob&I] = 16 M-' s-l). Chymotrypsin was also inhibited by FTC-Aca-AlaAla-MetP(OPh)2(1) more effectively (kobsd[I] = 190 M-' 6-l) than HLE and PPE (kob,&] = 22 and 13 M-' s-l, respectively). Compound 2, FTC-Aca-Ala-Ala-AlaP(OPh)2, inhibited PPE and HLE with &,&I] of 22 and 41 M-l s-l, respectively, but did not inhibit chymotrypsin. As expected, none of these phosphonates inhibited trypsin. The presence of the fluorescein fluorophore in these inhibitors resulted in either better or the same inhibitory potency compared to their nonfluorescent analogs. For example, with chymotrypsin, the presence of the fluorophores increased the potency of these inhibitors. The fluoresceinylated Phe analog 3 inhibited chymotrypsin very potently with kobsd[I] of 9500 M-l s-l, while the nonfluoresceinylated analog, Z-Phe-Leu-PheP(OPh)2(71, inhibited chymotrypsin but less potently with ko~sd[Ilof 110 M-l s-* (A. Abuelyaman, D. Jackson, D. Hudig, and J. Powers (unpublished results)). Compound 1,FTC-AcaAla-Ala-MetP(OPh)2,was found to be a better inhibitor of chymotrypsin and HLE, hobsd[I] = 190 and 22 M-' s-l respectively, than Boc-Ala-Ala-MetP(OPh)2(51, hobsd[Il= 3 and 2 M-l s-l, respectively. In contrast to FTC enhancement of the inhibitory potency of the chymotrypsin-directed inhibitors with chymotrypsin, the presence of FTC did not alter the efficacy of the elastasedirected inhibitor with the two elastases. FTC-Aca-AlaAla-AlaP(OPh)2(2) and its analog Z-Ala-Ala-AlaP(OPh)2 (6) inhibited HLE with almost equal rate constant values, kobdI1 = 38 and 41 M-' s-l, respectively. Compound 2 also showed similar efficacy with PPE as compound 6, kobsd[I] = 22 and 30 M-' s-l, respectively. Fluorescent
Figure 4. RNK-16 cytotoxic lymphocytes labeled with FTCAca-Ala-Ala-MetP(OPhh.
Met and Phe derivatives 1 and 3 are better inhibitors of PPE (kO&I] = 13and 16 M-l s-l, respectively) compared to their nonfluorescent analogs, compounds 5 ( k o ~ s ~=1 3 3 M-' s-l) and 7 (no inhibition). The fluoresceinylated Ala analog, compound 2, on the other hand, showed almost the same inhibitory effect with PPE as did its nonfluorescent analog, Z-Ala-Ala-AlaP(OPh)2(61,kobsd[I] = 22 and 30 M-l s-l, respectively. Addition of the Texas Red fluorophore resulted in a less effective compound in comparison with an analogous compound with the FTC-fluorophore. The TXR-labeled derivative, TXR-Aca-Phe-Leu-PheP( OPh)2(4), showed less inhibition with all three enzymes compared to FTC-AcaPhe-Leu-PheP(OPh)2 (3).However, when TXR-Aca-PheLeu-PheP(OPh)2(4) was compared with nonfluorescent Z-Phe-Leu-PheP(OPh)2(7),the effects varied with the different enzymes. Inhibition of chymotrypsin was less with TXR-Aca-Phe-Leu-PheP(OPh)2 (4) while inhibition of HLE and PPE was greater (Table 1). From the results discussed above, i t can be concluded that attaching the fluorophores in these phosphonate inhibitors resulted in either a similar or equal inhibitory potency with HLE and PPE. In the case of chymotrypsin, the presence of the FTC fluorophore increased the potency of these inhibitors substantially, a n observation that implies interactions between this fluorophore and chymotrypsin. I t is also clear that the specificities of these fluorescent compounds are parallel to those of nonfluorescent analogs and are dependent on the sequences of the peptide portions. Cell Labeling. Fluorescent peptide phosphonates inactivated and labeled intracellular granzymes. They traversed the membranes of living cells and of their granules and irreversibly labeled the cells for subsequent analyses. The inhibitor FTC-Aca-Ala-Ala-MetP(OPh)2 reacted with distinct, granule-like regions of the cytotoxic lymphocytes cell line RNK-16 (Figure 4), similar to granules as observed in living cells by diffential interference contrast microscopy (Yannelli e t al., 1986). Examination of individual sections through the cells (not illustrated) indicated that these fluorescent regions are interspersed in the cytoplasm. The fluorescent regions had a granule-like morphology similar to that detected by granzyme reactivity with the reversible serine protease inhibitor soybean trypsin inhibitor (Burkhardt et al., 1989). In the latter experiments, the cells were fixed and permeabilized before the macromolecular inhibitor was used. Thus, the fluorescent phosphonate compounds
Fluorescent Phosphonate Protease Inhibitors
offer technical advantages over previous approaches. The new reagents also have far greater specificity: soybean trypsin inhibitor inactivated both tryptases and chymases of cytotoxic lymphocytes (Hudig et al., 1987) whereas the FTC-Aca-Ala-Ala-MetP(OPh)2 inactivated RNK-16 chymase and little Asp-ase or tryptase granule protease activities (Woodard et al., unpublished results). SUMMARY
The synthesis of four different peptidyl phosphonates labeled with two fluorophores, TXR and FITC, was accomplished from two synthetic intermediates. The intermediates TXR-Aca-OH and FTC-Aca-OH were prepared and then coupled to tripeptidyl phosphonates using the DCC method. The inhibitory potency of these compounds, as serine protease inhibitors, was evaluated using inactivation rate constants (kobsd/[II) with the representative enzymes chymotrypsin, PPE, and HLE. These inhibitors were found to be very potent and specific irreversible inhibitors. The most potent inhibitorenzyme interaction was between FTC-Aca-Phe-Leu-PheP(0Ph)z (3)and chymotrypsin (kobsd/[I] = 9500 M-' s-l). Specificity was indicated by the observation that chymotrypsin was inhibited 600 times faster than HLE by this inhibitor (3). FTC-Aca-Ala-Ala-AlaP(OPh)z, on the other hand, inhibited only HLE and PPE and did not inhibit chymotrypsin. None of these compounds inhibited trypsin. It is also noteworthy that the specificity of these inhibitors was unaffected by the presence of the fluorescent tags. Initial experiments with the inhibitor, FTCAca-Ala-Ala-MetP(OPh)2,and rat RNK-16 cytotoxic lymphocytes indicate that granzymes can be labeled within living cells. Thus, these compounds are expected to be excellent reagents for the identification of fully mature cytotoxic lymphocytes within tissues and for the comparison of the granzyme contents of individual cells. ACKNOWLEDGMENT
This research was supported by a grant, R01 GM42212 (to D.H. and J.C.P.), from the National Institutes of Health. We thank the Texaco Corp. for a graduate fellowship awarded to A.S.A. We also thank Dr. ChihMin Kam for assistance with the kinetic experiments. LITERATURE CITED Bieth, J., and Wermuth, C. G. (1973) The Action of Elastase on p-Nitroanilide Substrates. Biochem. Biophys. Res. Commun. 53 (2), 383-390. Bieth, J.,Siess, B., and Wermuth, C. G. (1974) Synthesis and Analytical Use of a Highly Sensitive and Convenient Substrate of Elastase. Biochem. Med. 11, 350-357. Burkhardt, J . K., Hester, S., and Argon, Y. (1989) Two Proteins Targeted to The Same Lytic Granule Compartment Undergo Very Different Posttranslational Processing. Proc. Natl. h a d . Sci. U.S.A. 86, 7128-7132. Erlanger, B. F.,Kokowsky, N., and Cohen, W. (1961) Preparation and Properties of Two Chromogenic Substrates of Trypsin. Arch. Biochem. Biophys. 95, 271-278. Ewoldt, G. R., Winkler, U. W., Powers, J. C., and Hudig, D. (1992) Sulfonyl Fluoride Serine Srotease Inhibitors inactivate RNK-16 Lymphocyte Granule Proteases and Reduce Lysis by Granule Extracts and Perforin. Mol. Immunol. 29, 713-721.
Bioconjugate Chem., Vol. 5, No. 5, 1994 405 Hudig, D., Gregg, N. J . , Kam, C.-M., and Powers, J. C. (1987) Lymphocyte Granule-Mediated Cytolysis Requires Serine Protease Activity. Biochem. Biophys. Res. Commun. 149, 882-888. Hudig, D., Allison, N. J., Pickett, T. M., Winkler, U., Kam, C.M., and Powers, J. C. (1991) The Function of Lymphocyte Proteases: Inhibition and Restoration of Granule Mediated Lysis with Isocoumarin Serine Protease Inhibitors. J . Zmmunol. 147, 1360-1368. Hudig, D., Ewoldt, G. R., and Woodard, S. L. (1993) Proteases and Lymphocyte Cytotoxic Killing Mechanism. Curr. Opin. Immunol. 5,90-96. Kanaoka, Y., Takahashi, T., Nakayama, H., Takada, K., Kimura, T., and Sakakibara, S. (1977) Synthesis of a Key Fluorogenic Amide, ~-Arginine-4-methylcoumaryl-7-amide (L-ArgMCA) and Its Derivatives. Fluorescence Assays for Trypsin and Papain. Chem. Pharm. Bull. 25, 3126-3128. Nakajima, K., Powers, J . C., Ashe, B. M., and Zimmerman, M. (1979) Mapping the Extended Substrate Binding Site of Cathepsin G and Human Leukocyte Elastases. Studies with Peptide Substrates Related to al-Protease Inhibitor Reactive Site. J . Biol. Chem. 254, 4027-4032. Odake, S.,Kam, C.-M., Narasimhan, L., Poe, M., Blake, J. T., Krahenbuhl, O., Tschopp, J.,and Powers, J. C. (1991) Human and Murine Cytotoxic T Lymphocyte Serine Proteases: Subsite Mapping with Peptide Thioester Substrates and Inhibition of Enzyme Activity and Cytolysis by Isocoumarins. Biochemistry 30, 2217-2227. Oleksyszyn, J., and Powers, J . C. (1991) Irreversible Inhibition of Serine Proteases by Peptide Derivatives of (a-Aminoalky1)phosphonate Diphenyl Esters. Biochemistry 30, 485-493. Pachmann, K., Reinecke, K., Emmerich, B., and Thiel, E. (1991) Highly Fluorochrome Labeled Gene Probes for Quantitative Tracing of RNA in Individual Cells by in situ Hybridization. Bioconjugate Chem. 2 (11, 19-25. Powers, J. C., Odake, S., Oleksyszyn, J., Hori, H., Ueda, T., Boduszek, B., and Kam, C.-M.(1993) Proteases-Structures, Mechanism, and Inhibitors. In: Proteases, Protease Znhibitors and Protease-Derived Peptides (J. C. Cheronis, and J . E. Repine, Eds.) pp 3-19, Birkhauser Verlag, Basel, Boston, Berlin. Schechter, I., and Berger, A. (1967) On the Size of the Active Site in Protease. 1Papain. Biochem. Biophys. Res. Commun. 27, 157-162. Tanaka, T.,Minematsu, Y., Reilly, C. F., Travis, J., and Powers, J . C. (1985) Human Leukocyte Cathepsin G. Subsite Mapping with 4-Nitroanilide, Chemical Modification and Effect of Possible Cofactors. Biochemistry 24, 2040-2047. Titus, J . A., Haugland, R., Sharrow, S. O., and Segal, D. M. (1982) Texas Red, A Hydrophilic, Red-emitting Fluorophore for Use with Fluorescein in Dual Parameter Flow Microfluorometric and Fluorescence Microscopic Studies. J . Immunol. Methods 50 (21, 193-204. Tschopp, J.,and Jongeneel, C. V. (1988) Cytotoxic T Lymphocyte Mediated Cytolysis. Biochemistry 27, 2641-2646. Tuhy, P. M., and Powers, J . C. (1975) Inhibition of Human Leukocyte Elastase by Peptide Chloromethyl Ketones. FEBS Lett. 50 (31, 359-361. Ward, J. M., and Reynolds, C. W. (1983) Large Granular Lymphocyte Leukemia.A Heterogeneous Lymphocytic Leukemia in F344 Rats. A m . J . Path. 111, 1-11. Yannelli, J. R., Sullivan, J. A,, Mandell, G. L., and Engelhard, V. H. (1986) Reorientation and Fusion of Cytotoxic T Lymphocyte Granules after Interaction With Target Cells as Determined by High Resolution Cinemicrography. J . Immunol. 136, 377-382.
