Mechanism-based inactivation of human leukocyte elastase via an enzyme-induced sulfonamide fragmentation process

ABB Archives of Biochemistry and Biophysics 429 (2004) 60–70 www.elsevier.com/locate/yabbi

Mechanism-based inactivation of human leukocyte elastase via an enzyme-induced sulfonamide fragmentation process Liuqing Wei, Zhong Lai, Xiangdong Gan, Kevin R. Alliston, Jiaying Zhong, Jeff B. Epp, Juan Tu, Asiri B. Perera, Michael Van Stipdonk, and William C. Groutas* Department of Chemistry, Wichita State University, Wichita, KS 67260, USA Received 12 April 2004, and in revised form 24 May 2004 Available online 2 July 2004

Abstract We describe herein the design and in vitro biochemical evaluation of a novel class of mechanism-based inhibitors of human leukocyte elastase (HLE) that inactivate the enzyme via an unprecedented enzyme-induced sulfonamide fragmentation cascade. The inhibitors incorporate in their structure an appropriately functionalized saccharin scaffold. Furthermore, the inactivation of the enzyme by these inhibitors was found to be time-dependent and to involve the active site. Biochemical, HPLC, and mass spectrometric studies show that the interaction of these inhibitors with HLE results in the formation of a stable acyl complex and is accompanied by the release of (L) phenylalanine methyl ester. The data are consistent with initial formation of a Michaelis–Menten complex and subsequent formation of a tetrahedral intermediate with the active site serine (Ser195 ). Collapse of the tetrahedral intermediate with tandem fragmentation results in the formation of a highly reactive conjugated sulfonyl imine which can either react with water to form a stable acyl enzyme and/or undergo a Michael addition reaction with an active site nucleophilic residue (His57 ). It is also demonstrated herein that this class of compounds can be used in the design of inhibitors of serine proteases having either a neutral or basic primary substrate specificity. Thus, the results suggest that these inhibitors constitute a potential general class of mechanism-based inhibitors of (chymo)trypsin-like serine proteases. Ó 2004 Elsevier Inc. All rights reserved. Keywords: Mechanism-based inhibitors; Enzyme-induced rearrangement; Sulfonamide; Fragmentation; Leucocyte elastase

The design and utility of novel mechanism-based (suicide) inhibitors in mechanistic enzymology and drug discovery are well documented [1]. A mechanism-based inhibitor is an inherently unreactive compound that acts as a substrate and is processed by the catalytic machinery of an enzyme, generating a highly reactive electrophilic species which, upon further reaction with an active site nucleophilic residue, leads to irreversible inactivation of the enzyme [2]. Inhibitors of this type offer many potential advantages, including high enzyme specificity, since the latent reactivity in the inhibitor is unmasked following catalytic processing of the inhibitor by the target enzyme only. In previous work, mechanistic and biochemical considerations were employed in the design of mecha-

nism-based inhibitors that lead to inactivation of a target protease by an enzyme-induced rearrangement such as, for example, the Lossen [3–5] and Gabriel– Colman rearrangements [6,7]. Mechanism-based inhibitors which, when processed by an enzyme, undergo a fragmentation reaction that results in the formation of a highly reactive electrophilic species while tethered to the active site of the enzyme constitute a particularly attractive way of inhibiting an enzyme. We describe herein the biochemical rationale underlying the design of the first example of a mechanism-based inhibitor (I) of the serine protease human leukocyte elastase (HLE)1 that appears to inactivate the enzyme via the formation

1

*

Corresponding author. Fax: 1-316-978-3431. E-mail address: [email protected] (W.C. Groutas). 0003-9861/$ - see front matter Ó 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.abb.2004.05.018

Abbreviations used: ESI, electrospray ionization; CID, collisioninduced dissociation; I, inhibitor; HLE, human leukocyte elastase; TOMI, Turkey ovomucoid inhibitor.

L. Wei et al. / Archives of Biochemistry and Biophysics 429 (2004) 60–70

of a highly reactive N-sulfonyl imine that arises by an unprecedented enzyme-induced sulfonamide fragmentation process.

Materials and methods Materials MeOSuc-AAPV-pNA, N-p-Tosyl-Gly-Pro-Lys p-nitroanilide and bovine trypsin were purchased from Sigma Chemical, St. Louis, MO. Human leukocyte elastase was purchased from Elastin Products, Owensville, MO. Dansyl chloride was purchased from Lancaster (Windham, NH). Thirty-three percent HBr in glacial acetic acid was obtained from Fisher Scientific. Substrate and inhibitor stock solutions were prepared in DMSO. The proton and carbon NMR spectra were recorded using a Varian XL-300 or 400 MHz NMR spectrometer. A Hewlett–Packard diode array UV/VIS spectrophotometer was used in the enzyme assays and inhibition studies.

61

N-Bromomethylsaccharin 1 [9] Saccharin (24.92 g; 0.136 mol) and paraformaldehyde (24.4 g; 0.271 mol) were dissolved in glacial acetic acid (80 mL), 33% HBr in glacial acetic acid (167 mL) was added, and the solution was heated at
80 °C for 3 h. The reaction mixture was allowed to cool to room temperature and then diluted with ice-cold water (100 mL). The precipitate was collected by suction filtration and rinsed with ice-cold water (30.74 g; 82% yield). 1 H NMR (CDCl3 ): d 5.50 (s, 2H), 7.82–8.10 (m, 4H).

Synthesis

Compound 2 N-Bromomethylsaccharin (44.73 g; 0.162 mol) was dissolved in dry THF (160 mL) and treated with thiolacetic acid (13.0 g; 0.170 mol) and triethylamine (17.0 g; 0.162 mol). The resulting solution was stirred overnight and the solvent was removed on the rotary evaporator. The residue was taken up in ethyl acetate (450 mL) and the solution was washed with 5% aqueous HCl (2 100 mL) and dried over anhydrous sodium sulfate. Removal of the solvent left a solid which was washed with hexane, yielding a light yellow solid (39.55 g; 90% yield). 1 H NMR (CDCl3 ): d 2.39 (s, 3H), 5.25 (s, 2H), 7.75–8.06 (m, 4H).

Inhibitors 8–12 were synthesized according to the scheme shown in Fig. 1 and are listed in Table 1. Detailed synthetic procedures are given below. Compound 13 was synthesized essentially as described in the literature [8] using the scheme shown in Fig. 2.

2-[N-(2,2-Diphenyl-ethyl)aminosulfanylmethyl]-1,1-1, 2-dihydro-1k6 -1,2-benzisothiazol-3-one 3 A solution of compound 2 (3.0 g; 11.1 mmol) in dry methylene chloride (9 mL) was cooled to 0 °C under a nitrogen atmosphere. Sulfuryl chloride (1.8 mL;

Fig. 1. Synthesis of inhibitors 8–12.

62

L. Wei et al. / Archives of Biochemistry and Biophysics 429 (2004) 60–70

Table 1 Inhibitory activity of compounds 8–13 toward human leukocyte elastase and bovine trypsin Compound

8 9 10 11 12 13

kobs /[I] (M1 s1 ) Elastase

Trypsin

270 870 870 90 60 Inactive

80 290 320 Inactive Inactive —

22.2 mmol) was added and the resulting solution was stirred for 20 min at 0 °C. The solvent and excess sulfuryl chloride were removed on the vacuum pump at 0 °C. The residue was dissolved in dry methylene chloride (10 mL) and cooled to )78 °C under nitrogen. A solution of 2,2diphenylethylamine (3.27 g; 16.6 mmol) and N-methylmorpholine (1.8 mL; 16.6 mmol) in dry methylene chloride (20 mL) was then added dropwise and the resulting mixture was stirred at 0 °C for 15 min and at room temperature for 45 min. The precipitate formed was filtered off and rinsed with ethyl acetate. The filtrate was evaporated off leaving a crude product which was purified using flash chromatography (1.0 g; 21% yield). 1 H NMR (CDCl3 ): d 2.85 (t, 1H), 3.65 (t, 2H), 4.15 (t, 1H), 4.85 (s, 2H), 7.05–7.40 (m, 10H), 7.80–8.10 (m, 4H). (S)-3-Phenyl-2(1,1,3-trioxo-1,3-dihydro-1k6 -1,2-benzisothiazol-2-yl-methylsulfanylamino)-propionic acid methyl ester 4 A solution of compound 2 (3.0 g; 11.1 mmol) in dry methylene chloride (9 mL) was cooled to 0 °C under a nitrogen atmosphere. Sulfuryl chloride (1.8 mL; 22.2 mmol) was added, and the resulting solution was stirred for 20 min at 0 °C. The solvent and excess sulfuryl chloride were removed on the vacuum pump at 0 °C. The residue was dissolved in dry methylene chloride

(10 mL) and cooled to )78 °C under nitrogen. A suspension of L -phenylalanine methyl ester hydrochloride (3.58 g; 16.6 mmol) and N-methylmorpholine (1.8 mL; 16.6 mmol) in dry methylene chloride (20 mL) was then added dropwise and the resulting mixture was stirred at 0 °C for 15 min and at room temperature for 45 min. The precipitate formed was filtered off and rinsed with ethyl acetate. The filtrate was evaporated off leaving a crude product which was purified by flash chromatography using silica gel and hexane/ethyl acetate as eluents g; 22% yield). 1 H NMR (CDCl3 ): d 3.02 (m, 2H), 3.32 (d, 1H), 3.65 (s, 3H), 3.96 (m, 1H), 4.82 (dd, 2H), 7.10–7.36 (m, 5H), 7.78–8.14 (m, 4H). (R)-3-Phenyl-2(1,1,3-trioxo-1,3-dihydro-1k6 -1,2-benzisothiazol-2-yl-methylsulfanylamino)-propionic acid methyl ester 5 Prepared using D -phenylalanine methyl ester hydrochloride and the same procedure as that used to prepare compound 4. Compound 5 (1.2 g; 27% yield).1 H NMR (CDCl3 ): d 3.02 (m, 2H), 3.35 (d, 1H), 3.65 (s, 3H), 3.96 (m, 1H), 4.82 (dd, 2H), 7.10–7.32 (m, 5H), 7.78–8.12 (m, 4H). 2-(N-Benzylaminosulfanylmethyl)-1,1-dioxo-1,2-dihydro1k6 -1,2-benzisothiazol-3-one 6 Prepared using benzylamine and the same procedure as that used to make compound 3. Compound 6 (1.2 g; 32% yield). 1 H NMR (CDCl3 ): d 3.20 (t, 1H), 4.18 (d, 2H), 4.87 (s, 2H), 7.118–7.40 (m, 5H), 7.80–8.17 (m, 4H). 2-(N-phenethylaminosulfanylmethyl)-1,1-dioxo-1 2-dihydro-1k6 -1,2-benzisothiazol-3-one 7 Synthesized using 2-phenethylamine and the same procedure as that used to make compound 6. Compound 7 (1.0 g; 26% yield). 1 H NMR (CDCl3 ): d 2.83 (t, 2H), 2.97 (t, 1H), 3.25 (q, 2H), 4.85 (s, 2H), 7.10–7.33 (m, 5H), 7.80–8.17 (m, 4H).

Fig. 2. Synthesis of inhibitor 13.

L. Wei et al. / Archives of Biochemistry and Biophysics 429 (2004) 60–70

N-(2,2-Diphenyl-ethyl)-1-(1,1,3-trioxo-1,3-dihydro-1k6 1,2-benzisothiazol-2-yl)-methanesulfonamide 8 A solution of compound 3 (0.78 g; 1.8 mmol) in dry methylene chloride (4 mL) kept at 0 °C under nitrogen was treated with 70% m-chloroperbenzoic acid (1.8 g; 7.2 mmol) and the resulting mixture was allowed to warm up to room temperature and stirred overnight. The solvent was removed on the rotary evaporator, ethyl acetate (45 mL) was added to the residue, and the resulting solution was washed with saturated sodium bicarbonate (3 15 mL) and brine (15 mL). The organic phase was dried over anhydrous sodium sulfate and the solvent was evaporated off, leaving a crude product which was purified by flash chromatography (hexane/ ethyl ether) to yield compound 8 as a white solid (0.47 g; 50% yield), mp 141–142 °C. 1 H NMR (CDCl3 ): d 3.88 (t, 2H), 4.28 (t, 1H), 4.65 (t, 1H), 4.88 (s, 3H), 7.20–7.40 (m, 10H), 7.84–8.18 (m, 4H). Anal. Calcd. for C22 H20 N2 SO5 : C, 57.89; H, 4.39; N, 6.14. Found: C, 58.14; H, 4.27; N, 5.79. Compounds 9–12 These were obtained using the same procedure as that used in the preparation of compound 8. (S)-3-Phenyl-2-(1,1,3-trioxo-1,3-dihydro-1k6 -1,2-benzisothiazol-2-yl-methanesulfonylamino)-propionic acid methyl ester 9 White solid (0.33 g; 60% yield), mp 161–164 °C. 1 H NMR (CDCl3 ): d 3.16 (m, 2H), 3.75 (s, 3H), 4.58 (m, 1H), 4.92 (dd, 2H), 5.12 (d, 1H), 7.18 (7.36 (m, 5H)), 7.84–8.10 (m, 4H). Calcd. for C18 H18 N2 S2 O7 : C, 49.32; H, 4.11; N, 6.39. Found: C, 49.26; H, 3.97; N, 6.39. (R)-3-Phenyl-2-(1,1,3-trioxo-1,3-dihydro-1k6 -1,2-benzisothiazol-2-yl-methanesulfonylamino)-propionic acid methyl ester 10 White solid (0.78 g; 66% yield), mp 159–161 °C. 1 H NMR (CDCl3 ): d 3.16 (m, 2H), 3.75 (s, 3H), 4.58 (m, 1H), 4.92 (dd, 2H), 5.20 (d, 1H), 7.18–7.36 (m, 5H), 7.84–8.15 (m, 4H). Calcd. for C18 H18 N2 S2 O7 : C, 49.32; H, 4.11; N, 6.39. Found: C, 49.40; H, 4.18; N, 6.26. N-Benzyl-1-(1,1,3-trioxo-1,3-dihydro-1k6 -1,2-benzisothiazol-2-yl)-methanesulfonamide 11 White solid (0.38 g; 70% yield), mp 53–54 °C. 1 H NMR (CDCl3 ): d 4.33 (d, 2H), 4.88 (s, 2H), 5.35 (t, 1H), 7.20–7.40 (m, 5H), 7.80–8.18 (m, 4H). Calcd. for C15 H14 N2 S2 O5 : C, 49.10; H, 3.83; N, 7.65. Found: C, 49.13; H, 3.96; N, 7.63. N-Phenethyl-1-(1,1,3-trioxo-1,3-dihydro-1k6 -1,2-benzisothiazol-2-yl)-methanesulfonamide 12 White solid (0.60 g; 53% yield), mp 124–125 °C. 1 H NMR (CDCl3 ): d 2.90 (t, 2H), 3.45 (q, 2H), 4.78 (t, 1H),

63

4.90 (s, 2H), 7.10–7.38 (m, 5H), 7.84–8.18 (m, 4H). Calcd. for C16 H16 N2 S2 O5 : C, 50.53; H, 4.21; N, 7.37. Found: C, 50.84; H, 4.32; N, 7.24. Synthesis of L -phenylalanine methyl ester dansyl derivative 14. Dansyl chloride (0.54 g; 2 mmol) was added to a solution of (L) phenylalanine methyl ester hydrochloride (0.54 g; 2 mmol) in dry acetonitrile (10 mL) kept at 0 °C under nitrogen. Triethylamine (0.41 g; 4 mmol) was added, the ice bath was removed and the reaction mixture was stirred for 2 h. Evaporation of the solvent in vacuo left a crude product which was purified using a silica gel column (hexane/ethyl acetate) to yield 0.70 g (85% yield) of pure crystalline product, mp 120–121 °C. 1 H NMR (CDCl3 ): d 2.77 (s, 6H), 2.83 (d, 3H), 3.35 (s, 3H), 4.17 (q, 1H), 5.28 (d, 1H), 6.92 (m, 2H), 7.10 (m, 3H), 7.18 (d, 1H), 7.46 (t, 1H), 8.19 (t, 2H), 8.30 (d, 1H). Calcd. for C19 H18 N2 SO6 : C, 56.72; H, 4.78; N, 6.97. Found: C, 56.51; H, 4.59; N, 7.02. Reaction of inhibitor 9 with methoxide. Sodium metal (0.23 g; 10 mmol) was added to dry methanol (10 mL) under nitrogen and the resulting solution was added to a suspension of compound 9 (60 mg; 0.14 mmol) in dry methanol (5 mL) with stirring. After stirring for 10 min, TLC analysis showed complete disappearance of the starting material. The reaction mixture was acidified with 5% aqueous HCl (pH
2) and the solvent was removed in vacuo, leaving a solid residue that was analyzed by NMR. Biochemical studies Enzyme assays and inhibition studies: incubation method Human leukocyte elastase ([E]f ¼ 0.7 lM) was assayed by mixing 10 lL of a 70 lM enzyme solution in 0.05 M sodium acetate buffer, pH 5.5, 100 lL dimethyl sulfoxide, and 890 lL of 0.1 M Hepes buffer, pH 7.25, in a thermostatted test tube. A 100 lL aliquot was transferred to a thermostatted cuvette containing 880 lL of 0.1 M Hepes, pH 7.25, and 20 lL of a 7.0 mM solution of MeOSuc-Ala-Ala-Pro-Val-p-NA ([S]f ¼ 0.14 mM), and the change in absorbance was monitored at 410 nm for 1 min. In a typical inhibition run, 10 lL of a 21 mM solution of inhibitor 9 ([I]f ¼ 0.21 mM) in dimethyl sulfoxide and 90 lL dimethyl sulfoxide were mixed with 10 lL of a 70 lM enzyme solution ([E]f ¼ 0.7 lM) and 890 lL of M Hepes buffer, pH 7.25, and placed in a constant temperature bath, Aliquots (100 lL) were withdrawn at different time intervals and transferred to a cuvette containing 20 lL of a 7.0 mM substrate solution ([S]f ¼ 0.14 mM) and 880 lL of 0.1 M Hepes buffer, pH 7.25. The absorbance was monitored at 410 nm for 1 min. The kinetics data obtained by using the incubation method were analyzed by determining the slopes of the semilogarithmic plots of enzymatic activity remaining versus time using Eq. (1) below, where [E]t /[E]o is the amount of active enzyme remaining at time t [10].

64

lnð½Et =½Eo Þ ¼ kobs t

L. Wei et al. / Archives of Biochemistry and Biophysics 429 (2004) 60–70

ð1Þ

Hydroxylamine reactivation of inactivated HLE A solution containing 980 lL of 0.1 M Hepes buffer, pH 7.25, 10 lL of 21.9 mM inhibitor 9 ([I]f ¼ 219 lM) in DMSO, and 10 lL of 21.9 lM HLE ([E]f ¼ 219 nM) was incubated for 30 min. The enzyme was totally inactivated, as shown by withdrawing an aliquot and assaying for remaining enzyme activity. Excess hydroxylamine (100 lL of a 0.50 M solution in water) was then added to the fully inactivated enzyme. Aliquots (100 lL) were removed at various time intervals (from 1 min to 24 h) and assayed for remaining enzyme activity by mixing with 20 lL of 3.86 mM solution of MeOSuc-Ala-AlaPro-Val p-nitroanilide ([S]f ¼ 77.2 lM), and 880 lL of 0.1 M Hepes buffer, pH 7.25), and monitoring the absorbance at 410 nm. Enzyme activity was determined by comparing the activity of an enzyme solution containing no inhibitor (control) with the activity of an enzyme solution containing inhibitor at the same time point. Substrate protection In separate experiments, the kobs /[I] values were determined by incubating HLE with inhibitor 9 in the absence and presence of substrate. In the former case, 10 lL of HLE ([E]f ¼ 219 nM) was incubated with 10 lL of inhibitor 9 ([I]f ¼ 21.9 lM) dissolved in DMSO and 0.1 M Hepes buffer (980 lL), pH 7.25, in a thermostatted cuvette. Aliquots (100 lL) were withdrawn at different time intervals and added to a thermostatted cuvette containing 0.1 M Hepes buffer (880 lL), pH 7.25, and MeOSuc-Ala-Ala-Pro-Val pNA ([S]f ¼ 77.2 lM). The absorbance was monitored at 410 nm and the kobs /[I] M1 s1 value was then determined. The kobs /[I] value was also determined by repeating the experiment in the presence of substrate: HLE ([E]f ¼ 219 nM) was incubated with inhibitor 9 ([I]f ¼ 21.9 lM), and MeOSucAla-Ala-Pro-Val pNA ([S]f ¼ 231.6 lM) in 0.1 M Hepes buffer (974 lL, pH 7.25) in a thermostatted cuvette. Aliquots (100 lL) were withdrawn at different time intervals and added to a thermostatted cuvette containing MeOSuc-Ala-Ala-Pro-Val pNA ([S]f ¼ 77.2 lM) and 0.1 M Hepes buffer (880 lL), pH 7.25. Efficiency of inactivation (determination of partition ratio) Ten microliters of inhibitor 9 ([I]f ¼ 3.5–42 lM) in DMSO was incubated with 10 lL HLE ([E]f ¼ 0.70 lM) and 980 lL of 0.1 M Hepes buffer, pH 7.25, for 15 min. At the end of the 15-min incubation period, enzyme activity was assayed by transferring an aliquot (100 lL) to a cuvette containing 20 lL MeOSuc-Ala-Ala-Pro-Val p-NA ([S]f ¼ 0.14 lM) and 880 lL of 0.1 M Hepes buffer, pH 7.25, and monitoring the absorbance at 405 nm. The partition ratio was calculated as described by

Knight and Waley [11] by plotting the fraction of remaining enzyme activity ([E]t /[E]o ) versus the initial ratio of inhibitor to enzyme ([I]/[E]o ). Reactivation of the HLE–inhibitor 9 complex Forty microliters of a 70.0 lM solution of human leukocyte elastase was incubated with excess inhibitor 9 (10 lL of a 25.2 mM solution in dimethyl sulfoxide), 40 lL DMSO and 410 lL of 0.1 M Hepes buffer, pH 7.25, at 25.0 °C. After the solution was incubated for 30 min, a 25 lL was removed and assayed for enzymatic activity (the enzyme was found to be completely inhibited). Excess inhibitor was removed via Centricon-10 filtration by centrifuging at 14,000g for 45 min at 25.0 °C. Buffer (500 lL of 0.1 M Hepes buffer, pH 7.25) was added to the HLE–inhibitor complex and the centrifugation was repeated at 14,000g for 1 h at 25.0 °C. The HLE–inhibitor complex was dissolved in 2.0 mL buffer and aliquots (100 lL) were withdrawn at different time intervals and added to a cuvette containing 20 lL of 7.0 mM MeOSuc-Ala-Ala-Pro-Val p-nitroanilide, 880 lL of 0.1 M Hepes buffer, pH 7.25, and monitoring the absorbance at 410 nm. The amount of active enzyme was determined by comparing the activity of an enzyme solution containing no inhibitor (control) with the activity of an enzyme solution containing inhibitor at the same time point. The experiment was repeated using inhibitor 10. HPLC studies: product analysis and identification from the incubation of inhibitor 9 with HLE A solution of HLE (50 lL) containing 1 mg/mL enzyme was added to a test tube containing 300 lL of 0.1 M Hepes buffer, pH 7.25, containing 0.5 M NaCl and 0.2 mM inhibitor 9. A second sample containing 0.1 N Hepes buffer, pH 7.25, with 0.5 M NaCl and 0.2 mM inhibitor 9 but no enzyme was used as a control. Both samples were incubated in a water bath at 23 °C for 24 h. Fifty microliters of a saturated dansyl chloride solution (18 mg/mL) in methanol was added to each test tube and after 1 h the samples were analyzed by HPLC with no product detected. Both samples were stored at 4 °C for 72 h after which the clear solutions were re-injected sequentially along with compound 14 (used as a standard). Each sample (20 lL) and dansyl standard 14 (5 lL) were sequentially injected into a 250  4.6 mm C18 Luna [2] column (Phenomenex) with a flow rate of 1.5 mL/min. Mobile phase: A 10% acetonitrile, 25% methanol, 65% water; B 10% acetonitrile, 90% methanol using a gradient of 55% B to 100% B over 30 min using two 510 pumps, a 717+ refrigerated autosampler, and a 474 fluorescence detector (excitation k380 emission k470), all controlled by Millenium software (Waters). The fraction eluting from 10 to 11.5 min was collected and analyzed by electrospray ionization (ESI) and collision-induced dissociation (CID) mass spectrometry

L. Wei et al. / Archives of Biochemistry and Biophysics 429 (2004) 60–70

using a Finnigan LCQ-Decca ion-trap mass spectrometer (Thermoquest). Assay and inhibition of bovine trypsin Bovine trypsin was assayed spectrophotometrically using N-p-tosyl-Gly-Pro-Lys 4-nitroanilide. Ten microliters of bovine trypsin ([E]f ¼ 0.84 lM), 890 lL of 0.1 M Tris buffer, pH 7.51, containing 0.01 M CaCl2 and 100 lL DMSO were incubated in a thermostatted test tube. A 100 lL aliquot was transferred to a thermostatted cuvette containing 20 lL of a 5 mM solution of N-p- tosyl-GlyPro-Lys-pNA ([E]f ¼ 0.1 mM) and the change in absorbance was monitored at 410 nm for 1 min. In a typical inhibition run, 10 lL of inhibitor ([I]f ¼ 0.252 mM) in DMSO, 90 lL DMSO and 10 lL of an 84 lM enzyme solution in 890 lL of 0.025 M phosphate buffer, pH 7.51, containing 0.1 M NaCl were mixed and the solution was placed in a constant temperature bath. Aliquots (100 lL) were withdrawn at different time intervals and transferred to a cuvette containing 20 lL substrate (5 mM) and 880 lL of 0.025 M phosphate buffer, pH 7.51, containing 0.1 M NaCl. The absorbance was monitored at 410 nm for 1 min. The kinetics data were analyzed as described for HLE.

Results and discussion Rationale underlying the design of inhibitor (I) The biochemical rationale underlying the design of inhibitor (I) was based on the following observations:

65

(a) replacement of the carbonyl group in peptides with SO2 yields a-amido sulfonamides which are known to undergo a spontaneous fragmentation reaction, as illustrated in Fig. 3A [12,13]; (b) in contrast to acyclic a-amido sulfonamides, N-(phthalimidosulfonyl)-L phenylalanine methyl ester 13 (Fig. 3B, X ¼ CO) and related compounds have been synthesized and shown to be stable [8]; (c) inhibitors based on the saccharin scaffold are known to dock to the active site of (chymo)trypsin-like proteases, an event that is followed by acylation of the active site serine (Ser195 ) [14–18]; and, (d) the imidazole ring of histidine residues located at the active site of enzymes is known to undergo facile Michael addition reactions with conjugated systems [19,20]. Based on these considerations, we reasoned that an entity such as (I) might inactivate a target serine protease via a sequence of steps involving the initial formation of a Michaelis–Menten complex, followed by enzyme-induced ring opening and tandem fragmentation, leading to the release of an amine or aminoacid ester, sulfur dioxide, and the formation of a Michael acceptor (in this instance, an N-sulfonyl imine) [21] capable of reacting with an active site nearby nucleophilic residue (His57 ), ultimately leading to inactivation of the enzyme. The postulated mechanism of action of (I) is illustrated in Fig. 4. Our initial efforts focused on (I), an entity that incorporates a saccharin scaffold with appended recognition and reactivity elements. The initial selection of Phe as the P02 residue [22] was based on the fact that the P02 residue in the Turkey Ovomucoid Inhibitor (TOMI), a macromolecular inhibitor of HLE, is an aromatic residue (Tyr) [23,24] that is likely involved

Fig. 3. (A) Spontaneous fragmentation of a-amido sulfonopeptides to yield a Michael acceptor; (B) a-amido sulfonamide motif in phthalimide and saccharin derivatives.

66

L. Wei et al. / Archives of Biochemistry and Biophysics 429 (2004) 60–70

Fig. 4. Postulated mechanism of action of inhibitor (I).

in a p-stacking interaction with Phe-41 (located close to the S02 subsite) and, second, the results of previous studies [14,18] which have shown that the presence of an aromatic residue at the P02 position leads to enhanced inhibitory potency. The effect of stereochemistry and chain length on inhibitory activity was also investigated. Inactivation of HLE by derivatives of (I): relationship of structure to inhibitory potency and specificity Synthesis Inhibitors 8–12 were synthesized using (Fig. 1) starting with saccharin and are listed in Table 1. The most troublesome aspect of the synthesis was the formation of sulfenamides 3–7. The reaction leading to the formation of the sulfenamides was highly capricious, with slight changes in the reaction conditions and the nature of the amine component having a major effect on the yields of the desired products. Extensive experimentation was carried out to optimize the formation of the desired compounds and to minimize the formation of the major side product (identified by NMR as Nchloromethylsaccharin). The best yields realized were in the 20–30% range. It is possible that the low yields are due to the high reactivity of the intermediate sulfenyl chloride and its sensitivity to adventitious water. The S0

subsites were probed by varying the nature of the amine component and chirality. Lastly, in order to make a direct comparison between the saccharin and phthalimide templates, phthalimide analog 13 was also synthesized (Fig. 2). The reaction sequence used to make 13 was not applicable to the synthesis of the saccharin derivatives. Inactivation kinetics The inhibitory activity of compounds 8–13 toward HLE was evaluated using the incubation method. Thus, incubation of HLE with inhibitor 9 led to rapid time-dependent loss of enzymatic activity (Fig. 5). The enzyme slowly regained virtually all of its activity after 24 h. Kitz and Wilson analysis [10] of the data (Fig. 6) yielded a bimolecular rate constant kobs /[I] of 870 M1 s1 . Interestingly, the potency of the corresponding D isomer 10 was comparable to that of the Lisomer (kobs /[I] 870 M1 s1 ). As shown in Table 1, with the exception of phthalimide derivative 13, the rest of the compounds were also found to inhibit HLE. It is evident from Table 1 that inhibitory activity is dependent on the nature of R. Assuming that R is oriented toward the S0 subsites, the spatial requirements observed for R may simply reflect the inability of the saccharin template to bind to the active site of the enzyme in a

L. Wei et al. / Archives of Biochemistry and Biophysics 429 (2004) 60–70

Fig. 5. Time-dependent loss of enzymatic activity. (A) Excess inhibitor 9 ([I]f ¼ 0.21 mM) was incubated with human leukocyte elastase ([E]f ¼ 0.70 lM) in 0.1 M Hepes buffer, pH 7.25, 25 °C, aliquots were withdrawn at different time intervals and assayed for enzymatic activity using MeOSuc-Ala-Ala-Pro-Val p-nitroanilide ([S]f ¼ 0.14 mM) (solid circles); (B) A 300-fold excess of inhibitor 9 was incubated with bovine trypsin ([E]f ¼ 0.84 lM) in 0.1 M Tris buffer, pH 7.51, containing 0.021 M CaCl2 , aliquots were withdrawn at different time intervals and assayed for enzymatic activity using N-p-Tosyl-Gly-ProLys p-nitroanilide ([S]f ¼ 0.1 mM) (open circles).

Fig. 6. Kinetics of inactivation of human leukocyte elastase ([E]f ¼ 0.70 lM) by compound 9. Inhibitor 9 was incubated with human leukocyte elastase (the [I]/[E] ratio varied between 50 and 600) in 0.1 M Hepes buffer, pH 7.25, at 0 °C. Aliquots were withdrawn periodically and assayed for remaining enzyme activity using MeOSucAla-Ala-Pro-Val p-nitroanilide.

strictly substrate-like fashion. Furthermore, in the case of compound 12, for example, the short alkyl chain (one methylene group) may not be sufficient to place the phenyl ring close enough to Phe-41 of HLE for an optimal hydrophobic p-stacking interaction. The fact that phthalimide derivative 13 is devoid of any inhibitory activity suggests that the interaction of the saccharin and phthalimide templates with the active site involves a

67

Fig. 7. Kinetics of inactivation of bovine trypsin by compound 9. Excess inhibitor (the [I]/[E] ratio varied between 300 and 1000) was incubated with bovine trypsin ([E]f ¼ 0.84 lM) in 0.1 M Tris buffer, pH 7.51, containing 0.021 M CaCl2 . Aliquots were withdrawn at different time intervals and assayed for remaining enzyme activity using N-pTosyl-Gly-Pro-Lys p-nitroanilide ([S]f ¼ 0.1 mM).

poorly understood delicate interplay of binding and electronic interactions that affect potency. The specificity of compounds 8–12 was briefly investigated using bovine trypsin. Thus, incubation of compound 9 with the enzyme led to rapid inactivation of the enzyme (Table 1), followed by gradual and full regain of enzyme activity after 24 h. The regain in enzymatic activity was somewhat faster when compared to the analogous interaction with HLE. Replotting of the data yielded a kobs /[I] value of 290 M1 s1 (Fig. 7). These observations are in agreement with previous studies related to the inhibition of human mast cell tryptase, a tetrameric trypsin-like serine protease, by saccharin derivatives [25]. The results of those studies, and those of the present studies (Table 1), clearly indicate that (a) inhibitor recognition by these enzymes is critically dependent on interactions with the S0 subsites of the enzyme, and (b) potent inhibitors of trypsin-like enzymes can, in principle, be designed that primarily exploit S0 interactions only and do not incorporate in their structure a Lys or Arg side chain. Since the presence of the latter is associated with poor pharmacokinetics [26] and low selectivity, optimized derivatives of (I) may offer several distinct advantages. Inactivation of HLE by derivatives of (I): mechanism of inactivation The proposed tentative mechanism of action of (I) was probed as follows: a substrate protection experiment was carried out in order to demonstrate that the interaction of (I) with HLE involves the active site. This is clearly evident in Fig. 8 where the presence of substrate in the incubation mix led to a decrease in kobs /[I]

68

L. Wei et al. / Archives of Biochemistry and Biophysics 429 (2004) 60–70

Fig. 8. Substrate protection. HLE (219 nM) was incubated with inhibitor 9 (21.9 lM) in 0.1 M Hepes buffer, pH 7.25, aliquots were taken at different time intervals and assayed for remaining enzyme activity using MeOSuc-Ala-Ala-Pro-Val p-nitroanilide (77.2 lM). The experiment was repeated by incubating HLE (219 nM), inhibitor 9 (21.9 lM) and MeOSuc-Ala-Ala-Pro-Val p-nitroanilide (0.232 mM) in 0.1 M Hepes buffer, pH 7.25. Aliquots were taken at different time intervals and assayed for remaining enzyme activity using MeOSuc-Ala-AlaPro-Val p-nitroanilide (77.2 lM). The data were then analyzed using the method of Kitz and Wilson [10].

possible presence of a labile acyl linkage that leads to active enzyme upon treatment with hydroxylamine. Thus, it can be reasonably assumed that the interaction of the inhibitor with HLE leads to acylation of the active site serine, however, further structural studies are needed to arrive at a definitive conclusion regarding the precise structure of the acyl enzyme complex. The partition ratio, a parameter that corresponds to the number of molecules of inhibitor necessary to inactivate a single molecule of enzyme, and thus describes how efficiently a mechanism-based inhibitor inactivates an enzyme [2], was determined by plotting the fraction of remaining enzyme activity after a 15-min incubation period versus the initial ratio of inhibitor to enzyme (Fig. 10). The extent of inactivation was found to be linearly dependent on the inhibitor to enzyme molar ratio. Extrapolation of the linear part of the curve to the line of complete inactivation yielded a partition ratio of 40 for inhibitor 9, attesting to a moderately efficient inhibitor. The reactivation of the complex formed between HLE and inhibitor 9 was also investigated. Thus, HLE was totally inactivated using excess inhibitor 9. The excess inhibitor was then removed by Centricon-10 filtration, and the regain in enzymatic activity was monitored. The experiment was repeated using inhibitor 10. In both cases total regain of enzymatic activity was observed, suggesting the likely formation of one enzyme–inhibitor complex. The deacylation rate constants (kdeacyl ) for the HLE–inhibitor complexes derived from inhibitors 9 and 10 were determined by replotting the data according to Eq. (1). These were found to be 0.0028 and 0.0041 s1 , respectively. To establish that the interaction of 9 with HLE produces an N-sulfonyl imine that arises from an enzyme-induced fragmentation process, the products

Fig. 9. Effect of hydroxylamine on enzyme reactivation. Human leukocyte elastase (219 nM) was totally inactivated by incubating with a 100-fold excess of inhibitor 9 for 30 min in 0.1 M Hepes buffer, pH 7.25. Excess hydroxylamine was added (0.045 mM final concentration, closed circles), and aliquots were removed at different time intervals and assayed for remaining enzyme activity using MeOSuc-AlaAla-Pro-Val p-nitroanilide (77.2 lM).

from 320 to 220 M1 s1 . Next, the possible formation of an HLE–inhibitor acyl enzyme complex (or complexes) was investigated by adding excess hydroxylamine ((0.50 M) in 0.1 M Hepes buffer, pH 7.25) to HLE that had been fully inactivated with a 100-fold excess of inhibitor 9, and the regain in enzymatic activity was monitored over a 24 h period (Fig. 9). The data suggest that the regain in enzymatic activity arises from the

Fig. 10. Inactivation of human leukocyte elastase as a function of the molar ratio of inhibitor 9 to enzyme. HLE (70 lM) and various amounts of inhibitor 9 (0.35–4.20 lM in 0.1 M Hepes buffer, pH 7.25, were incubated for 15 min, aliquots were withdrawn at the end of the incubation period and assayed for remaining enzyme activity using MeOSuc-Ala-Ala-Pro-Val p-nitroanilide. The fractional activity remaining is proportional to the molar ratio of inhibitor to enzyme.

L. Wei et al. / Archives of Biochemistry and Biophysics 429 (2004) 60–70

formed by incubating inhibitor 9 with HLE were identified using HPLC and mass spectrometry. Initial HPLC analysis of the assay solution showed that a trace of (L) Phe–OCH3 was observed in this assay, possibly arising from the slow hydrolysis of inhibitor 9, which was labeled by dansyl chloride resulting in a small background. Therefore, a control given identical treatment in each step of the experiment, except for the addition of HLE, was used in order to subtract out this background. HPLC analysis of the two samples, after performing the product identification procedure described under Materials and methods, showed that a peak with an identical retention time to compound 14 standard’s peak (10.3 min) was observed in both samples. The sample including HLE had an area 2.7 times greater than the control, demonstrating that this compound is a product arising from the HLE-inhibitor 9 reaction (Fig. 11). The fractions eluting from 10 to 11.5 min of both the standard 14 (Fig. 11A) and the HLE–inhibitor reaction (Fig. 11C) were collected and analyzed by mass spectrometry. ESI revealed identical protonated molecule peaks of 413 amu, and CID of the 413 peak of each sample gave identical gave identical fragmentation patterns, with a principal peak at 301 amu. Therefore, the product arising from the inhibition of HLE by inhibitor 9 is confirmed to be (L) Phe–OCH3 . The chemical competence of the cascade steps outlined

69

in Fig. 4 was readily established by stirring inhibitor 9 with excess sodium methoxide in methanol at room temperature which led to rapid disappearance of the inhibitor. NMR product analysis revealed the presence of a roughly 1:1 mixture of (o-carboxymethyl)benzenesulfonamide and Phe–OCH3 . Taken together, these results suggest that the binding of (I) to the active site of HLE leads to acylation of the enzyme with concomitant sulfonamide fragmentation resulting in the release of L Phe–OCH3 , SO2 and the formation of N-sulfonyl imine (II) (Fig. 4). The available evidence suggests that (II) ultimately leads to the formation of a stable acyl enzyme whose structure can be tentatively represented by structure (IV). The slow deacylation rate observed with (IV) may be the result of a conformational change in the enzyme that perturbs the positions of the catalytic residues and/or H-bonding between the SONH2 group of the tethered inhibitor and the imidazole ring of His-57, thereby impairing the ability of His-57 to function in general base catalysis. Lastly, while the initial design of inhibitor (I) and postulated mechanism of action invoked the likely involvement of a ‘‘double hit’’ mechanism leading to the formation of species (III), the available data can be adequately explained by the formation of a stable acyl enzyme species, particularly in light of earlier high-field NMR studies [16] which demonstrated the formation of formaldehyde from species

Fig. 11. HPLC analysis of products formed by incubating human leukocyte elastase with inhibitor 9. (A) Standard: compound 14 (N-dansyl-L -Phe– OCH3 ) (synthesis described under Materials and methods); (B) Control: 20 lL injection of assay composed of 300 lL of 0.1 M Hepes buffer, pH 7.25, containing 0.5 M NaCl, and 0.2 mM compound 9 after incubation for 24 h at 23 °C, addition of 50 lL saturated dansyl chloride in methanol, and 72 h incubation at 4 °C; (C) Compound 9/HLE reaction: 20 lL injection of assay composed of 300 lL of 0.1 M Hepes buffer, pH 7.25, with 0.5 M NaCl, 0.2 mM compound 9, and 50lL HLE (1 mg/mL) after incubation for 24 h at 23 °C, addition of 50 lL saturated dansyl chloride in methanol and 72 h incubation at 4 °C; and (d) Overlay of (B and C): the peaks with 10.3 min retention times were one compound and had identical molecular ion peaks and fragmentation patterns, corresponding to standard 14.

70

L. Wei et al. / Archives of Biochemistry and Biophysics 429 (2004) 60–70

(II). In previous studies using saccharin derivatives of the type saccharin–CH2 X, where X is a good leaving group (halide, carboxylate, etc.), a similar mechanism involving the formation an N-sulfonyl imine (structure (II), Fig. 4) was proposed for the inactivation of HLE [14,15]. It should be noted, however, that compounds represented by (I) are intrinsically more stable chemically and employ a sulfonamide fragmentation reaction as the driving force for the formation of the N-sulfonyl imine. In summary, the findings cited herein describe the first example of a heterocyclic mechanism-based inhibitor (I) of HLE that generates a highly reactive electrophilic species via an enzyme-induced sulfonamide fragmentation process, leading to inactivation of the enzyme. Acknowledgments Financial support of this work by the National Institutes of Health (HL 57788) and Supergen, Inc. is gratefully acknowledged. References [1] (a) R.B. Silverman, The Organic Chemistry of Enzyme-Catalyzed Reactions (revised edition), Academic Press, San Diego, 2002; (b) R.B. Silverman, The Organic Chemistry of Drug Action and Drug Design, Academic Press, San Diego, 1992. [2] R.B. Silverman, Methods Enzymol. 249 (1995) 240–283. [3] W.C. Groutas, P.K. Giri, J.P. Crowley, J.C. Castrisos, M.J. Brubaker, Biochem. Biophys. Res. Comm. 141 (1986) 741–748. [4] W.C. Groutas, M.A. Stanga, M.J. Brubaker, J. Am. Chem. Soc. 111 (1989) 1931–1932. [5] W.C. Groutas, R. Venkataraman, M.J. Brubaker, M.A. Stanga, Biochemistry 30 (1991) 4132–4136. [6] W.C. Groutas, L.S. Chong, R. Venkataraman, J.B. Epp, R. Kuang, M.J. Brubaker, N. Houser-Archield, H. Huang, J.J. McClenahan, Biochem. Biophys. Res. Commun. 194 (1993) 1491–1499.

[7] W.C. Groutas, J.R. Hoidal, L.S. Chong, R. Venkataraman, J.B. Epp, R. Kuang, N. Houser-Archield, Bioorg. Med. Chem. 3 (1995) 187–193. [8] D. Merricks, P.G. Sammes, E.R.H. Walker, K. Henrick, M.M. McPartin, J. Chem. Soc. Perkin Trans. I (1991) 2169– 2176. [9] R.C. Desai, R.P. Farrell, G.-H. Kuo, D.J. Hlasta, Synlett (1994) 933–934. [10] R. Kitz, I.B. Wilson, J. Biol. Chem. 12 (1962) 2940–2945. [11] G.C. Knight, S.G. Waley, Biochem. J. 225 (1985) 435–439. [12] T. Sommerfeld, D. Seebach, Angew. Chem. Int. Ed. 34 (1995) 553–554. [13] S. Paik, E.H. White, Tetrahedron 52 (1996) 5303–5318. [14] W.C. Groutas, N. Houser-Archield, L.S. Chong, R. Venkataraman, J.B. Epp, H. Huang, J.J. McClenahan, J. Med. Chem. 36 (1993) 3178–3181. [15] D.J. Hlasta, C. Subramanyam, M.R. Bell, P.M. Carabateas, J.J. Court, R.C. Desai, M.L. Drozd, W.M. Eickhoff, E.W. Ferguson, R.J. Gordon, J.A. Johnson, V. Kumar, A.L. Maycock, K.R. Mueller, E.D. Pagani, D.T. Robinson, M.T. Saindane, P.J. Silver, S. Subramanian, J. Med. Chem. 38 (1995) 739–744. [16] W.C. Groutas, R. Kuang, R. Venkataraman, J.B. Epp, S. Ruan, O. Prakash, Biochemistry 36 (1997) 4739–4750. [17] R. Kuang, J.B. Epp, S. Ruan, H. Yu, P. Huang, S. He, J. Tu, N.M. Schechter, J. Turbov, C.J. Froelich, W.C. Groutas, J. Am. Chem. Soc. 121 (1999) 8128–8129. [18] W.C. Groutas, J.B. Epp, R. Kuang, S. Ruan, L.S. Chong, R. Venkataraman, J. Tu, S. He, H. Yu, Q. Fu, Y. Li, T.M. Truong, N.T. Vu, Arch. Biochem. Biophys. 385 (2001) 162–169. [19] J.A. Banzon, J.M. Kuo, B.W. Miles, D.R. Fischer, P.J. Stang, F.M. Raushel, Biochemistry 34 (1995) 743. [20] J.M. Schwab, W. Li, C. Ho, J. Am. Chem. Soc. 106 (1984) 7293– 7294. [21] G.I. Georg, G.C.B. Harriman, S.A. Peterson, J. Org. Chem. 60 (1995) 7366–7368. [22] I. Schechter, A. Berger, Biochem. Biophys. Res. Commun. 27 (1967) 157–162. [23] W. Bode, E.F. Meyer, J.C. Powers, Biochemistry 28 (1989) 1951– 1963. [24] W. Bode, A.-Z. Wei, R. Huber, E.F. Meyer, J. Travis, S. Neumann, EMBO J. 5 (1986) 2453–2458. [25] K.D. Combrink, H.B. Gulgeze, N.A. Meanwell, B.C. Pearce, P. Zulan, G.S. Bisacchi, D.G.M. Roberts, P. Stanley, S.M. Seiler, J. Med. Chem. 41 (1998) 4854–4860. [26] P.E.J. Sanderson, Ann. Rep. Med. Chem. 36 (2001) 79–88.