Purification and Characterization of Lymphocyte Chymase I, a Implicated in Perforin-Mediated Lysis

This information is current as Susan L. Woodard, Stephanie A. Fraser, Ulrike Winkler, Delwin of September 25, 2021. S. Jackson, Chih-Min Kam, James C. Powers and Dorothy Hudig J Immunol 1998; 160:4988-4993; ; http://www.jimmunol.org/content/160/10/4988 Downloaded from

References This article cites 39 articles, 12 of which you can access for free at: http://www.jimmunol.org/content/160/10/4988.full#ref-list-1

Why The JI? Submit online. http://www.jimmunol.org/

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

*average by guest on September 25, 2021

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 1998 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. Purification and Characterization of Lymphocyte Chymase I, a Granzyme Implicated in Perforin-Mediated Lysis1

Susan L. Woodard,* Stephanie A. Fraser,‡ Ulrike Winkler,* Delwin S. Jackson,§ Chih-Min Kam,§ James C. Powers,§ and Dorothy Hudig2*†

One mechanism of killing by cytotoxic lymphocytes involves the exocytosis of specialized granules. The released granules contain perforin, which assembles into pores in the membranes of cells targeted for death. Serine proteases termed are present in the cytotoxic granules and include several chymases (with -like specificity of cleavage). One chymase is selectively P reactive with an inhibitor, Biotinyl-Aca-Aca-Phe-Leu-Phe (OPh)2, that blocks perforin lysis. We report the purification and characterization of this chymase, lymphocyte chymase I, from rat natural killer cell (RNK)-16 granules. Lymphocyte chymase I is 30 kDa with a pH 7.5 to 9 optimum and primary substrate preference for tryptophan, a preference distinct from rat Downloaded from chymases. This chymase also reacts with other selective inhibitors that block perforin pore formation. It elutes by Cu2؉-immobilized metal affinity chromatography with other granzymes and has the N-terminal protein sequence conserved among granzymes. Chymase I reduces pore formation when preincubated with perforin at 37°C. In contrast, addition of the chymase without preincubation had little effect on lysis. It should be noted that the perforin preparation contained sufficient residual chymase activity to support lysis. Thus, the reduction of lysis may represent an effect of excess prolytic chymase I or a means to limit perforin lysis of bystander cells. In contrast, other chymases and granzyme K were without effect when added to http://www.jimmunol.org/ perforin during similar preincubation. Identification of the natural substrate of chymase I will help resolve how it regulates perforin-mediated pore formation. The Journal of Immunology, 1998, 160: 4988–4993.

here are multiple pathways that cytotoxic lymphocytes tions than (16), and, thus, recovery may be utilize in the killing of target cells, including the Fas- restricted by the initial low amounts. ligand-Fas pathway (1) and the granule exocytosis path- We recently identified a chymase-directed serine protease in- T P 3 way (2, 3; also reviewed in Ref. 4). The granule exocytosis path- hibitor, Bi-Aca-Aca-Phe-Leu-Phe (OPh)2, that preferentially re- way utilizes perforin and serine proteases of the granules acts with one granzyme of rat RNK-16 NK cells (17). This biotin-

(granzymes) to kill target cells. The granzymes A and K tagged reagent also prevented granule-mediated killing and by guest on September 25, 2021 and the Asp-ase granzyme B mediate cytotoxic activity through blocked NK killing (17). Here we describe the purification and their ability to induce DNA fragmentation after entering cells characterization of granzyme chymase I, the chymase that reacts through perforin pores (5–9). Another granzyme, a chymase (with preferentially with this tagged inhibitor. Our starting material was chymotrypsin-like specificity of cleavage), has a role in perforin granule extracts from 6 to 30 ϫ 109 rat RNK-16 lymphocytes per pore formation (10, 11). preparation of chymase I. Lymphocyte chymase granzymes have been difficult to purify to date. The enzyme activity has been detected in cytotoxic rat (10) Materials and Methods and human (M. Poe, unpublished results) lymphocytes. From com- Purification of chymase I puter models of granzyme serine protease active sites, mouse gran- zyme genes D, E, F and G, rat granzyme P7, and human granzyme Chymase I was purified from RNK-16 cell granules by size exclusion and cation exchange chromatography. H are predicted to be chymases (Refs. 12–14, respectively). To date, the only chymase purified from cytolytic granules has been a Isolation of RNK-16 granules. RNK-16 cells (18) were grown as ascites cells in F344 rats. Our RNK-16 subline has retained granule lytic activity carboxypeptidase, A-like protective protein (15). The while the live cells no longer bind to or kill YAC-1 target cells. To obtain abundance of the mRNA for granzymes D, E and F suggests that the granules, the cells were disrupted by nitrogen cavitation, and the lysate these granzymes may be expressed at ϳ100-fold lower concentra- was centrifuged over 54% Percoll (Sigma Chemical, St. Louis, MO). Gran- ules were isolated from the dense portion of the gradient as previously described (17). The granule pool was made1MinNaCl and then freeze

† thawed three times to disrupt the granule membranes. The granule extract *Department of Microbiology, School of Medicine, School of Veterinary Sciences, Ϫ ϳ and ‡The Cell and Molecular Biology Graduate Program, University of Nevada, was stored at 20°C. The yield was 0.5 mg of granule extract per billion Reno, NV 89557; and §School of Chemistry and Biochemistry, Georgia Institute of RNK-16 cells. Technology, Atlanta, GA 30332 Superdex 200 size exclusion chromatography. Twenty milliliters of gran- ule extract were concentrated approximately fivefold by centrifugation Received for publication July 28, 1997. Accepted for publication January 26, 1998. with Centriprep 10 concentrators (Amicon, Beverly, MA) before loading 3 The costs of publication of this article were defrayed in part by the payment of page ml onto a 2.6 ϫ 70-cm Superdex 200 (Pharmacia Fine Chemicals, Pisca- charges. This article must therefore be hereby marked advertisement in accordance way, NJ) column. The running buffer was 20 mM HEPES, 1 M NaCl, 10% with 18 U.S.C. Section 1734 solely to indicate this fact. betaine, 0.1 mM EGTA, and 0.05% NaN3. Aprotinin (6.5 kDa), carbonic 1 This work was supported in part by National Institutes of Health Grants CA38942 (D.H.), GM42212 (D.H. and J.C.P.), GM 54401 (J.C.P. and D.H.), and T32CA09563 (S.A.F. and U.W.). 3 P Abbreviations used in this paper: (OPh)2, phosphonate phenylester; Aca, 6-amin- 2 Address correspondence and reprint requests to Dr. Dorothy Hudig, Mail Code 320, ocaproic acid; Bi, biotinyl; Boc, butyloxycarbonyl; IMAC, immobilized metal affinity School of Medicine, University of Nevada, Reno, NV 89557. E-mail address: chromatography; MES,2-morpholinoethanesulfonic acid; RNK, rat NK cell; SBzl, [email protected] thiobenzyl; Suc, succinyl; Z, benzyloxycarbonyl; DCI, 3,4-dichloroisocoumarin.

Copyright © 1998 by The American Association of Immunologists 0022-1767/98/$02.00 The Journal of Immunology 4989 anhydrase (29 kDa), BSA (66 kDa), and IgG (150 kDa) were used for size Table I. Purification of chymase I calibration. The void volume was determined using blue dextran (2000 kDa). Specific MonoS cation exchange chromatography. The leading one-third of the Protein Activity Activity Percent SD200 low m.w. chymase peak was omitted from the pool loaded onto an Step (mg) (mOD/min) (mOD/min/mg) Recovery Enrichment HR5/5 MonoS fast protein liquid chromatography (FPLC) column (Phar- macia) because these fractions contained granzyme B that would copurify Granules 13.7 55300 4033 100 1 with chymase I on MonoS. The SD200 chymase pool was diluted with an SD200 ND 59100 N/A 107 N/A equal volume of NaCl-free buffer (125 mM MES, 0.15 mM EGTA, 15% SD200 Pk 4 0.84 5650 6730 107a 17.6 betaine, pH 5.9). Pure ethylene glycol (enzyme grade, Fisher Biotech, Fair MonoS 0.008 1570 196000 29.9b 513 Lawn, NJ) was added to make the chymase solution 0.4 M NaCl, 20% a Initial chymase based on the 9.5% relative abundance observed after the SD200 ethylene glycol, 50 mM MES, 10% betaine, 0.05% NaN3, and 0.1 mM EGTA, pH 6.0. The diluted sample was then in MonoS running buffer A. column. b Recovery assumed to be equivalent for all four chymases. MonoS running buffer B was similar to buffer A but with 1 M NaCl. The sample was loaded and washed in 0.4 M NaCl (60% A, 40% B). A linear gradient was then run from 40% to 100% B. In additional experiments not P illustrated, chymase I was isolated using immobilized metal affinity chro- and FITC-Ala-Ala-Met (OPh)2 (25) were synthesized in the laboratory matography (IMAC) with Cu2ϩ bound to Poros IMAC beads (PerSeptive of J.C.P. Biosystems, Cambridge, MA) (19). Chymase effects on perforin lysis Chymase detection Purified chymase I (or other granzymes) and isolated perforin were incu-

bated together at 37°C for 30 min and then diluted before hemolytic assays. Downloaded from Substrate hydrolysis. Rates of hydrolysis of peptide thiobenzyl substrates 2ϩ were measured using Ellman’s reagent, dithiobis-(2-nitrobenzoic acid) (1.2 Perforin was prepared by Cu -IMAC (Poros resin; PerSeptive Biosys- mM in assay), to monitor the production of HSBzl. The extinction coef- tems) (19), followed by phenyl-Superose (Pharmacia) hydrophobic inter- ficient for the colored thiol product at 410 nm is 13,600 MϪ1 cmϪ1 (20). action chromatography (26). In detail, 50 ml of perforin solution was Rates were measured in esterase buffer (0.1 M HEPES, 0.5 M NaCl, pH mixed with 100 ml of chymase I, dilutions of chymase I in MonoS buffer, 7.5) at 25°C. For assaying activity from column runs, 10 to 100 ␮l of each or MonoS buffer as a control and incubated. Then 850 ml of salt-free 10 fraction were assayed in a total volume of 200 ␮l in 96-well microtiter mM HEPES pH 7.5 buffer was added (to bring the solution to physiologic salt), and the solution was diluted with HEPES-buffered saline (10 mM plates using a Molecular Devices kinetics microplate reader (Palo Alto, ␮ http://www.jimmunol.org/ CA). Rates were determined in duplicate and averaged. The substrate Suc- HEPES, pH 7.5, 0.15 M NaCl) with 10 g/ml BSA (crystalline; U.S.Bio- Phe-Leu-Phe-SBzl (Bachem Bioscience, King of Prussia, PA) was used for logic, Cleveland, OH, No. 10856). In some experiments, the preincubation the chymase assays, except where indicated in Table II. Boc-Lys-SBzl was omitted. Lysis was measured using five twofold serial dilutions of (Sigma), Boc-Ala-Ala-Met-SBzl (Enzyme System Products (ESP), Dublin, perforin (starting with a 1/20 final dilution) with rabbit RBC. The amount CA), and Boc-Ala-Ala-Asp-SBzl (ESP) were used to measure tryptase, of perforin needed to lyse 50% of the erythrocytes was defined as 1 LU and Met-ase, and Asp-ase activities, respectively. The Boc-Ala-Ala-[P1 ϭ Leu, calculated by linear regression of the log of the volume of perforin assayed Phe, Ser, Trp, or Tyr]-SBzl substrates were synthesized in the laboratory of vs lysis. Lysis of treated perforin was compared on the basis of LU per J.C.P. (21). Updated procedures for synthesis are available upon request. milliliter of perforin. Substrates were used at 90 ␮M, except for Suc-Phe-Leu-Phe-SBzl, for which solubility limited its concentration to 28 ␮M in the assay. Results

P by guest on September 25, 2021 Biotinylation with Bi-Aca-Aca-Phe-Leu-Phe (OPh)2. Chymase-contain- Chymase I purification ing fractions were pretreated with either 0.1 mM inhibitor Bi-Aca-Aca- P Chymase I, the granzyme preferentially reactive with Bi-Aca-Aca- Phe-Leu-Phe (OPh)2 (22) or the solvent DMSO for 10 min at room tem- P perature. The treated fractions were assayed for loss of enzyme activity and Phe-Leu-Phe (OPh)2, was purified from RNK-16 granule extracts for biotinylation using protein blots (17). with approximately 30% recovery and 500-fold enrichment (Table I). The initial separation, by Superdex 200 size exclusion (Fig. Protein characterization 1A), was selected because the step was compatible with the high Protein determinations. Protein assays were done with bicinchoninic acid salt (Ͼ0.4 M NaCl) needed to stabilize the chymase activity. The (BCA) (23) (Pierce, Rockford, IL) with crystalline BSA as a standard. high salt was also needed to separate the chymase from the high M Gel electrophoresis. Samples were run on SDS-PAGE under reducing r conditions using 12% MiniProtean gels (Bio-Rad Laboratories, Richmond, proteoglycan (27) that complexes with granzymes (28) and inter- CA). Proteins were detected by silver staining. feres with their purification. Size exclusion also depleted most of Protein sequencing. Protein sequencing was performed on an Applied the (GrA and GrK) and the Asp-ase (GrB) from the chy- Biosystems (Foster City, CA) Procise 494 gas phase sequencer by Dr. mase peaks (Fig. 1B). The majority of granule protein (as indicated Matthew Williamson (University of California at San Diego, San Diego, by the OD ) eluted with the void volume where the high M CA). The protein was bound to ProBlott (Applied Biosystems) polyvinyli- 280 r dene difluoride (PVDF) membrane for sequencing. proteoglycan also eluted (not indicated). There were four chymase peaks with apparent molecular masses of 600, 190, 88, and 14

Determination of pH optimum kDa. The peak with the lowest apparent Mr was the most reactive Ͼ Hydrolysis rates of Suc-Phe-Leu-Phe-SBzl were measured in Good’s with the biotinylated inhibitor ( 99% inactivation after treatment buffer solutions containing 0.1 M MES (pH 5.58, 5.87, 6.12, and 6.43), 0.1 with 0.1 mM for 10 min) and remained inhibited after dialysis. The M PIPES (pH 6.44, 6.83, and 7.23), 0.1 M HEPES (pH 7.09, 7.50, 7.75, first peak (600 kDa) was 28% inhibited and the second (190 kDa) and 7.94) or 0.1 M TAPS (pH 7.94, 8.33, 8.69, and 9.10) and 0.75 M NaCl peak of chymase was ϳ68% inhibited when 50 ␮M inhibitor was in the presence of 1.2 mM Ellman’s reagent. The data were fit using a present in the enzyme assays. The third peak of chymase was nonlinear least squares algorithm to an equation of the form y ϭ ax2 ϩ bx ϩ c. unaffected. Neither the first or second chymase peak was inhibited after dialysis (or labeled with biotin detectable by SDS-PAGE). Chymase inhibition These data are consistent with unfavorable positioning of the in- Chymase isolated after SD200 was treated with 0.1-mM inhibitors for 20 hibitors’s reactive phosphonate with these chymases, despite abil- min at room temperature in esterase buffer and then immediately assayed. ity of the peptide inhibitor to compete with the peptide thioester The inhibitors were dissolved at 20 to 40 mM in anhydrous DMSO and substrate in the assays. stored at Ϫ20°C. Controls were treated with DMSO alone. The inhibitors The net recovery of total chymase after sizing was close to the came from the following sources: 3,4-dichloroisocoumarin (Boehringer Mannheim, Indianapolis, IN); PMSF and Tosyl-Phe-CH Cl (Sigma); initial activity. From this recovery, we assumed that the initial 2 amount of chymase I in the granules was proportional to the ϳ10% Z-Gly-Leu-Phe-CH2Cl (ESP). 2-(Z-NH(CH2)2CO-NH)C6H4SO2F (24), P P Bi-Aca-Aca-Phe-Leu-Phe (OPh)2 (22), Suc-Phe-Leu-Phe (OPh)2, (25) fraction appearing in the lowest Mr peak. We used the activity of 4990 CHYMASE I, A GRANZYME IMPLICATED IN PERFORIN-MEDIATED LYSIS Downloaded from

FIGURE 3. Characterization of the Mr and inhibitor-reactivity of chy- mase I by SDS-PAGE and protein blots. A, Purification of chymase I. Chymase I at different steps of purification was characterized by SDS- PAGE and silver stained in the gels. Lane 1,1␮g of unfractionated granule ␮ ␮ extract; lane 2,1 g of SD200 low Mr chymase; and lane 3, 0.25 gof

MonoS chymase I. Lanes 1 and 2 were overdeveloped in the silver stains http://www.jimmunol.org/ in an attempt to visualize the 31-kDa chymase I, which is invisible within the unfractionated granule extract and barely visible in the SD200 peak. B, Reactivity of chymase I and a chymase copurifying with perforin with P Bi-Aca-Aca-Phe-Leu-Phe (OPh)2. After separation by SDS-PAGE, the FIGURE 1. Separation of chymase I by Superdex 200 size exclusion proteins were blotted. The biotinylated protease inhibitor complexes were chromatography. Three milliliters of concentrated granule extract was detected with avidin-alkaline phosphatase. Lane 1, Inhibitor-treated SD200 loaded onto the Superdex 200 column. A, OD280 and chymase activities. perforin lytic peak; lane 2, DMSO-treated SD200 perforin lytic peak; lane B, Asp-ase (GrB), tryptase (GrA the larger, and GrK the smaller, tryptase), 3, Inhibitor-treated SD200 low m.w. chymase I; and lane 4, DMSO-con- trol-treated SD200 low m.w. chymase peak. The numbers indicate the mo-

and Met-ase activities. by guest on September 25, 2021 lecular mass of the markers in kDa.

this peak as 100% for our determination of the relative enrichment and recovery of chymase I (Table I). MonoS cation exchange chro- matography was used as the second step of purification. The ion GrB, which would copurify with chymase I by MonoS chroma- exchange depleted the contaminant tryptase GrK from the chy- tography (not illustrated). Chymase I eluted from MonoS at ϳ650 mase (Fig. 2). The leading SD200 fractions of chymase I were mM NaCl. The overall enzymatic yield of chymase I after purifi- excluded from further purification because they contained Asp-ase, cation was ϳ30%. The protein yield was 8 ␮g from 13.7 mg of starting protein. This recovery indicates that the maximal abun- dance of chymase I protein within the granule extract is less than 0.2% of the total protein, based on calculations assuming 30% recovery of the protein (rather than better than 30% recovery, which could be the case if some of the enzyme lost activity). Purified chymase I appears homogeneous after SDS PAGE. It was detected as a single silver-stained protein (Fig. 3A) and as a single biotinylated protein after reaction with the selective inhib- itor (Fig. 3B). Chymase I has a molecular mass of ϳ30 kDa by SDS PAGE (Fig. 3), which indicated that it was nonspecifically retained on the SD200 column (where it eluted with 15-kDa pro- teins, Fig. 3A, lane 2). Purified chymase from as many as 3 ϫ 1010 RNK-16 cells was sequenced. The limited N-terminal amino acid sequence that we obtained was Ile-(Ile/Leu)-Gly and is similar to the N-terminal Ile-Ile-Gly-Gly sequence common among the gran- zymes (29). In addition, the chymases eluted with GrB, GrA, GrK, and the other granzymes in a single sharp peak when a Cu2ϩ- FIGURE 2. Final purification of chymase I by MonoS cation exchange IMAC column with an imidazole gradient was used to separate chromatography. The low M chymase peak separated by SD200 chroma- r granzymes from other granule proteins (19). Chymase I could be tography was diluted to reduce the salt concentration and to adjust the buffer pH before ion exchange (see Materials and Methods). The leading separated from the other granzymes afterward by SD200 chroma- third of this chymase peak was omitted from the pool that was loaded onto tography. However, the chymase recovery was lower than with the the MonoS column to eliminate GrB (Asp-ase). procedure of Table I. The Journal of Immunology 4991

Table III. Inhibition of chymase Ia

Concentration Inhibitor (nM) % Inhibition

3,4-Dichloroisocoumarin 0.05 82.5 PMSF 1.0 98.9 b 2-(Z-NH(CH2)2CO-NH)C6H4SO2F 0.4 99.5 c Z-Gly-Leu-Phe-CH2Cl 0.01 97.4 Tosyl-Phe-CH2Cl 1.0 29.2 P d Bi-Aca-Aca-Phe-Leu-Phe (OPh)2 0.1 100.0 P e Suc-Phe-Leu-Phe (OPh)2 0.1 98.5 P f FITC-Ala-Ala-Met (OPh)2 0.1 91.3 a The chymase was treated for 20 min at room temperature with the concentrations of inhibitors indicated, and then the substrate Suc-Phe-Leu-Phe-SBzl was added di- rectly to assay the chymase in the microtiter wells. b 2-N-(3-(benzyloxycarbonylamino)propanoyl)-amino)benzenesulfonyl fluoride. c Benzyloxycarbonyl-Gly-Leu-Phe-chloromethyl ketone. d Biotinylaminocaproylaminocaproyl-Phe-Leu-Phe-phosphonate diphenyl ester. e Succinyl-Phe-Leu-Phe-phosphonate diphenyl ester. f 5-Fluoresceinylthiocarbamoyl-Ala-Ala-Ala-Met phosphonate diphenyl ester. Downloaded from

FIGURE 4. Determination of the pH optima for chymase I activity. Chymase I fractionated by size exclusion was assayed with a series of Phe-Leu-Phe-SBzl vs Boc-Ala-Ala-Tyr-SBzl substrates indicate Good’s buffers at the pH indicated. that it (chymase I) is distinct; the other chymases hydrolysed Boc- Ala-Ala-Tyr-SBzl (not indicated).

Chymase I was reactive with selected irreversible inhibitors in- http://www.jimmunol.org/ Enzymatic properties of chymase I cluding the general irreversible serine protease inhibitors, 3,4-di- Chymase I has a broad pH optimum between pH 7.5 and 9 (Fig. 4), chloroisocoumarin (DCI) and PMSF (Table III). Chymase I was which is the pH optimum characteristic of serine-dependent pro- also inactivated by several irreversible chymase-selective inhibi- teases. When the optimal pH was determined for unfractionated tors representing classes of compounds that preferentially react chymases within the granule extract, the maximum activity toward with and inactivate serine proteases, including sulfonyl fluorides, the same substrate was at pH 7.5 (data not shown), indicating that peptide chloromethyl ketones, and peptide phosphonates chymase I (optimal at pH 8.7) is different from the predominant (Table III). chymases. Chymase I demonstrated a marked preference for tryp- tophan at the P1 amino acid of substrates (see Ref. 30 for substrate by guest on September 25, 2021 nomenclature) when used to hydrolyse several Boc-Ala-Ala- Biologic properties of chymase I ϭ [P1 X]-SBzl substrates (Table II). Comparisons were made at 90 A biotinylated serine protease inhibitor, Bi-Aca-Aca-Phe-Leu- mM substrate concentration. Ninety micromolar concentrations of PheP(OPh) , inactivates perforin-mediated lysis (17). A chymase substrate were employed to facilitate detection of the low hydro- 2 reactive with this inhibitor copurified with perforin on SD200. It lysis rates of Leu and Ser P1 substrates. Hydrolysis of the Trp P1 was detectable when the ϳ70-kDa perforin fraction was treated substrate was more than twofold faster than hydrolysis of the Phe with the inhibitor and then highly concentrated before SDS-PAGE P1 substrate. It is noteworthy that there was no detectable hydro- and protein blotting. (See the biotinylated 30-kDa protein of Fig. lysis of the Tyr P1 substrate by chymase I while this substrate was 3B, lane 1). Purified chymase I also reacted with the biotinylated concurrently hydrolysed by the unfractionated granule extract. Km values were not determined due to the low availability of chymase inhibitor (Fig. 3B, lane 3) and has the same Mr as the perforin- ϳ I. The turnover rate for the Suc-Phe-Leu-Phe-SBzl substrate at the associated chymase. An additional 45-kDa band visible in Figure solubility limit of 28 ␮M was 4.5 pmols/s whereas the rate for 3B, lanes 1 and 2, is an endogenous avidin-binding protein found Boc-Ala-Ala-Trp-SBzl at 90 ␮M was 7.5 pmols/s, both with 97 ng with the granule proteins. So far we have been unable to purify of chymase I, indicating that the commercial reagent was a rea- perforin to homogeneity and retain lytic activity. For this reason, sonably sensitive substrate despite its lack of an optimal P1 amino we supplemented highly enriched (and still lytic) perforin with acid. These substrates have also been used to indicate that chymase purified chymase I to determine its effects on perforin lysis. When I is unique. Ratios of enzyme activities of this peak toward Suc- perforin was preincubated with chymase I at 37°C (without calci- um), there was a dose-dependent loss of lytic activity (Fig. 5). Without the preincubation, there was no effect on perforin lysis Table II. Chymase I hydrolysis of peptide thiobenzyl ester substrates when identical amounts of chymase I were added in similar pro- tocols. Because calcium will inactivate perforin during preincuba- Ϫ Substrate Ratea (pmol s 1) tion, there are two other differences in these protocols: the pres- Boc-Ala-Ala-Asp-SBzl 0.2 ence or absence of calcium and the availability of a membrane to Boc-Ala-Ala-Leu-SBzl 0.3 accept the pore. Pretreatment of perforin with GrK or the two other Boc-Ala-Ala-Met-SBzl 3.1 chymase peaks separated by SD200 (at similar mOD minϪ1 mlϪ1 Boc-Ala-Ala-Phe-SBzl 3.3 Boc-Ala-Ala-Ser-SBzl 0.1 of enzyme activity) was without effect on lysis. Addition of human Boc-Ala-Ala-Tyr-SBzl 0 neutrophil , another “chymase,” was also without ef- Boc-Ala-Ala-Trp-SBzl 7.5 fect. It should be noted that these preparations of perforin contain Suc-Phe-Leu-Phe-SBzl 4.5 ϳ1 mOD minϪ1 of chymase per 1000 lytic units of perforin (ϳ100 Ϫ a Each rate was determined in duplicate with 97 ng of enzyme. mOD min 1 per mg of perforin) (31). 4992 CHYMASE I, A GRANZYME IMPLICATED IN PERFORIN-MEDIATED LYSIS

would have very little activity at the intragranular pH of 5.5 (28) and would need to be exocytosed to be fully active. Rat mast cell protease I, a chymase, also has optimal activity at pH 8.5 (34). The selective reactivity for tryptophan at P1 contrasts with the P1 spec- ificity of the other rat lymphocyte chymases (S.A.F., unpublished results) and with the specificity of rat mast cell chymases I and II, both of which have much greater activity toward Phe than Trp at

P1 as reflected in kcat/Km that are 209- and 90-fold better for Phe than Trp in Suc-Val-Pro-(Phe/Trp)-NA substrates (24). This reac- tivity suggests that lymphocyte chymase I-selective inhibitors could be designed that would incorporate the P1 amino acid Trp. The chymase-directed sulfonyl fluoride inhibitor 2-(Z-

NH(CH2)2CO-NH)C6H4SO2F, the Z-Gly-Leu-Phe-CH2Cl chlo- romethyl ketone inhibitor, and the peptide phosphonate inhibitors that inactivated chymase I also blocked perforin-mediated lysis (see Refs. 10, 35, and 17, respectively.) Several observations are consistent with a physiologic role for chymase I although additional experiments will be needed to es- Downloaded from tablish this role. 1) Chymase I reacts preferentially with Bi-Aca- P Aca-Phe-Leu-Phe (OPh)2, an irreversible peptide phosphonate in- hibitor that biotinylated one granule protein (Fig. 7, Ref. 17), FIGURE 5. Supplementation of perforin with chymase I. Highly en- inactivated granule-mediated (perforin) lysis of RBCs, and riched perforin was incubated with the indicated concentrations of purified blocked rat NK activity (17). 2) A similar (if not identical) reactive chymase I for 30 min at 37°C, diluted, and then assayed using further chymase with an equivalent Mr was observed in lytic perforin twofold dilutions for lysis of RBC. Specifically, the indicated concentra- preparations (Fig. 3). 3) Furthermore, treatment of the perforin http://www.jimmunol.org/ Ϫ Ϫ tions (mOD min 1 ml 1) of chymase I (or buffer control) was added to a preparations with the biotinylated inhibitor also blocked lytic ac- final volume of 0.15 ml with purified perforin and incubated (see Materials tivity (our unpublished results). A simple interpretation is that this and Methods) before 20- to 80-fold dilutions in the lytic assays. The per- 30-kDa chymase participates in perforin-mediated cytotoxicity. forin preparation contained traces of other proteins, including perforin en- For this interpretation to be valid, chymase I would have to have hancing protein (PEPr) (26) and calreticulin (Ref. 40 and our unpublished results). a human species equivalent. Granzyme H (14, 36), so far found only in humans, is a potential species homologue for equivalent function because granzyme H contains a large, apparently un-

charged, substrate-binding pocket that is able to accommodate ar- by guest on September 25, 2021 Discussion omatic amino acids. The other human granzymes (A, B, K, and M) We have purified a new rat chymase. We believe that it is the first lack chymase activity (J.C.P., unpublished data). lymphocyte chymase granzyme purified and characterized enzy- We also found that preincubation of enriched perforin with pu- matically. Chymase I has many properties typical of a serine pro- rified chymase I depressed lytic activity. The experimental condi- tease of the chymotrypsinogen subfamily. The molecular mass of tions are compatible with premature proteolysis of a perforin-as- 30 kDa suggests that there are 1 or 2 glycosylation sites on a sociated granule protein by chymase I. Under calcium- and chymotrypsin-like structure that lacks additional domains. (The membrane-free conditions, the nascent product may have assumed kringle structures of many blood proteases are exam- a nonfunctional and irreversible conformation that will not en- ples of additional domains.) The pH optimum is also typical. Chy- hance lysis. Lysis was reduced as a consequence when calcium and mase I was inhibited by both DCI (32) and PMSF, “general” serine red cells are provided. Because the perforin contains only residual protease inhibitors that react with almost all serine proteases. The chymases, the enzyme supplementation (to physiologic ratios) may limited N-terminal amino acid sequence and its elution from be required to produce its effects during preincubation. Cu2ϩ-IMAC suggest that chymase I is likely to be a granzyme. The reduction of lysis could be interpreted as counter to a pos- We need additional protein sequence information to match chy- itive regulatory role for chymase I in lysis. A down-regulatory role mase I with the gene that encodes it. Amino acid sequences ex- for chymase I would place it as an intragranule protein to limit tending beyond amino acid 17 of chymase I will be needed. Ma- lysis of “by-stander” cells that are not targeted for lysis and add ture, fully processed granzymes conserve the first four N-terminal chymase I to the plasma proteins protein S (vitronectin) (37) and amino acids. The granzyme pairsD&EandF&Ghave amino apolipoprotein B (38, 39), which reduce perforin’s lytic activity. acids 5 to 8 in common so that this sequence could be ambiguous. However, we suggest that the issue of up- vs down-regulation of The granzymes retain a conserved motif at the next eight amino lysis remains open. Chymase I definitely reacts with a protein im- acids, 9 to 16. We are also developing an alternate approach for portant to lysis. This interaction appears specific because the other granzyme identification by making granzyme-specific Abs to pep- endogenous chymases lacked an effect on lysis under the same tide representing amino acids 17 to about 30 of the mature gran- conditions. zymes. Rat granzyme C (also termed P4; Ref. 13), rat granzyme The specificity of chymase I is also reflected by the enzymatic D/G1 (P7; Ref. 13), and rat granzyme J (P5; Ref. 33) genes have and the genetic variability among the chymase granzymes. Chy- been cloned and are not yet matched with the proteins they encode. mase I prefers Trp P1 while the two other rat lymphocyte chy- Based on computer models of the S1 substrate binding pockets, mases that are in separate SD200 peaks prefer Phe to Trp at P1 chymase I could correspond to GrD/G1 and is unlikely to be en- (S.A.F., unpublished results). The rodent granzymes predicted to coded by the GrC or GrJ genes. encode chymases differ in the genetic regions that specify their The enzymatic properties of chymase I have several implica- substrate binding sites (12). At this time, perforin and perforin tions. The pH optimum between 7 and 9 indicates that the chymase enhancing protein (PEPr; Ref. 26) are two proteins that participate The Journal of Immunology 4993 in lysis that are potential natural substrates to examine for selective single 30-kDa granzyme and blocks NK-mediated cytotoxicity. J. Immunol. 153: hydrolysis by lymphocyte chymase I. 5016. 18. Ward, J. M., and C. W. Reynolds. 1983. Large granular lymphocyte leukemia: a In summary, we have purified and characterized a native lym- heterogeneous lymphocytic leukemia in F344 rats. Am. J. Pathol. 111:1. phocyte chymase. Our data suggest that it (chymase I) may have a 19. Winkler, U., T. M. Pickett, and D. Hudig. 1996. Fractionation of perforin and granzymes by immobilized metal affinity chromatography (IMAC). J. Immunol. role in perforin lysis. At present, its absolute requirement for lysis Methods 191:11. is unresolved. The low abundance of this granzyme within gran- 20. Ellman, G. L. 1959. Tissue sulfhydryl groups. Arch. Biochem. Biophys. 82:70. ules indicates that recombinant expression of the rat chymase gran- 21. Harper, J. W., R. R. Cook, C. J. Roberts, B. J. McLaughlin, and J. C. Powers. 1984. mapping of the serine proteases human leukocyte , zymes would greatly benefit future studies. cathepsin G, porcine , rat mast cell proteases I and II: bovine chymotrypsin A alpha, and Staphylococcus aureus protease V-8 using tripeptide thiobenzyl ester substrates. Biochemistry 23:2995. Acknowledgments 22. Abuelyaman, A. S., D. S. Jackson, D. Hudig, S. L. Woodard, and J. C. Powers. We thank Drs. Kathleen Schegg and David Schooley of Biochemistry at 1997. Synthesis and kinetic studies of diphenyl 1-(N-peptidylamino)alkanephos- phonate esters and their biotinylated derivatives as inhibitors of serine proteases the University of Nevada (Reno, NV) for helpful advice on protein isola- and probes for lymphocyte granzymes. Arch. Biochem. Biophys. 344:271. tion and sequencing, Dr. Matthew Williamson at the University of Cali- 23. Smith, P. K., R. I. Krohn, G. T. Hermanson, A. K. Mallia, F. H. Gartner, fornia at San Diego (San Diego, CA) for protein sequencing, and Dr. M. D. Provenzano, E. K. Fujimoto, N. M. Goeke, B. J. Olson, and D. C. Klenk. Thomas F. Kozel at the University of Nevada (Reno, NV) for helpful 1985. Measurement of protein using bicinchoninic acid. Anal. Biochem. 150:76. 24. Powers, J. C., T. Tanaka, J. W. Harper, Y. Minematsu, L. Barker, D. Lincoln, suggestions with the manuscript. K. V. Crumley, J. E. Fraki, N. M. Schechter, G. G. Lazarus, K. Nakajima, K. Nakashino, H. Neurath, and R. G. Woodbury. 1985. Mammalian chymotryp- References sin-like : comparative reactivities of rat mast cell proteases, human and dog skin chymases, and human cathepsin G with peptide 4-nitroanilide substrates Downloaded from 1. Nagata, S., and P. Golstein. 1995. The Fas death factor. Science 267:1449. and with peptide chloromethyl ketone and sulfonyl fluoride inhibitors. Biochem- 2. Kagi, D., F. Vignaux, B. Ledermann, K. Burki, V. Depraetere, S. Nagata, istry 24:2048. H. Hengartner, and P. Golstein. 1994. Fas and perforin pathways as major mech- 25. Abuelyaman, A. S., D. Hudig, S. L. Woodard, and J. C. Powers. 1994. Fluores- anisms of T cell-mediated cytotoxicity. Science 265:528. cent derivatives of diphenyl 1-(N-peptidylamino)alkanephosphonate esters: syn- 3. Lowin, B., M. Hahne, C. Mattmann, and J. Tschopp. 1994. Cytolytic T-cell thesis and use in the inhibition and cellular localization of serine proteases. Bio- cytotoxicity is mediated through perforin and Fas lytic pathways. Nature 370: conjug. Chem. 5:400. 650. 26. Winkler, U., S. A. Fraser, and D. Hudig. 1997. Perforin enhancing protein (PEPr), a low molecular weight protein of cytotoxic lymphocyte granules, enhances per-

4. Podack, E. R. 1995. Execution and suicide: cytotoxic lymphocytes enforce Dra- http://www.jimmunol.org/ conian laws through separate molecular pathways. Curr. Opin. Immunol. 7:11. forin lysis. Biochem. Biophys. Res. Commun. 236:34. 5. Shi, L., R. P. Kraut, R. Aebersold, and A. H. Greenberg. 1992. A natural killer 27. MacDermott, R. P., R. E. Schmidt, J. P. Caulfield, A. Hein, G. T. Bartley, J. Ritz, cell granule protein that induces DNA fragmentation and apoptosis. J. Exp. Med. S. F. Schlossman, K. F. Austen, and R. L. Stevens. 1985. Proteoglycans in cell- 175:553. mediated cytotoxicity: identification, localization, and exocytosis of a chondroitin 6. Shiver, J. W., L. Su, and P. A. Henkart. 1992. Cytotoxicity with target DNA sulfate proteoglycan from human cloned natural killer cells during target cell breakdown by rat basophilic leukemia cells expressing both cytolysin and gran- lysis. J. Exp. Med. 162:1771. zyme A. Cell 71:315. 28. Masson, D., P. J. Peters, H. J. Geuze, J. Borst, and J. Tschopp. 1990. Interaction 7. Shi, L., C. Kam, J. C. Powers, R. Aebersold, and A. H. Greenberg. 1992. Puri- of chondroitin sulfate with perforin and granzymes of cytolytic T-cells is depen- fication of three lymphocyte granule serine proteases that induce apoptosis dent on pH. Biochemistry 29:11229. through distinct substrate and target interactions. J. Exp. Med. 176:1521. 29. Green, G. D. J., and E. Shaw. 1979. Thiobenzyl benzyloxycarbonyl-L-lysinate, 8. Heusel, J. W., R. L. Wesselschmidt, S. Shresta, J. H. Russell, and T. J. Ley. 1994. substrate for the sensitive colorimetric assay for -like enzymes. Anal. Bio-

Cytotoxic lymphocytes require granzyme B for the rapid induction of DNA frag- chem. 93:223. by guest on September 25, 2021 mentation and apoptosis in allogeneic target cells. Cell 76:977. 30. Schechter, I., and A. Berger. 1967. On the size of the active site in proteases. 9. Smyth, M. J., and J. A. Trapani. 1995. Granzymes: exogenous proteinases that I. Papain. Biochem. Biophys. Res. Commun. 27:157. induce target cell apoptosis. Immunol. Today 16:202. 31. Winkler, U. 1995. The role of granzymes in perforin-dependent cytotoxicity. 10. Ewoldt, G. R., U. Winkler, J. C. Powers, and D. Hudig. 1992. Sulfonyl fluoride Doctoral dissertation, University of Nevada, Reno, NV. serine protease inhibitors inactivate RNK-16 lymphocyte granule proteases and 32. Harper, J. W., K. Hemmi, and J. C. Powers 1985:. Reaction of serine proteases reduce lysis by granule extracts and perforin. Mol. Immunol. 29:713. with substituted isocoumarins: discovery of 3,4-dichloroisocoumarin, a new gen- 11. Talento, A., M. Nguyen, S. Law, J. K. Wu, M. Poe, J. T. Blake, M. Patel, T. Wu, eral mechanism based serine protease inhibitor. Biochemistry 24:1831. C. Manyak, M. Silberklang, G. Mark, M. Springer, N. H. Sigal, I. L. Weissman, 33. Ewoldt, G. R., D. Hudig, P. K. Darcy, J. L. Harris, C. S. Craik, and M. J. Smyth. R. C. Bleackley, E. R. Podack, M. L. Tykocinski, and G. C. Koo. 1993. Trans- 1997. cDNA cloning of a new granzyme, Granzyme J. Immunogenetics 45:452. fection of mouse CTL with an antisense vector reduces lytic activity. 34. Yoshida, N., M. T. Everitt, H. Neurath, R. G. Woodbury, and J. C. Powers. 1980. J. Immunol. 149:4009. Substrate specificity of two chymotrypsin-like proteases from rat mast cells: stud- 12. Odake, S., C. Kam, L. Narasimhan, M. Poe, J. T. Blake, O. Krahenbuhl, ies with peptide 4-nitroanilides and comparison with cathepsin G. Biochemistry J. Tschopp, and J. C. Powers. 1991. Human and murine cytotoxic T lymphocyte 19:5799. serine proteases: subsite mapping with peptide thioester substrates and inhibition 35. Zunino, S. J., N. J. Allison, C. Kam, J. C. Powers, and D. Hudig. 1988. Local- of enzyme activity and cytolysis by isocoumarins. Biochemistry 30:2217. ization, implications for function, and gene expression of chymotrypsin-like pro- 13. Ewoldt, G. R., M. J. Smyth, P. K. Darcy, J. L. Harris, C. S. Craik, B. Horowitz, teinases of cytotoxic RNK-16 lymphocytes. Biochim. Biophys. Acta 967:331. S. L. Woodard, J. C. Powers, and D. Hudig. 1997. RNKP-4 and RNKP-7, new 36. Haddad, P., D. Jenne, J. Tschopp, M. V. Clement, D. Mathieu-Mahul, and granzyme serine proteases expressed in activated rat lymphocytes. J. Immunol. M. Sasportes. 1991. Structure and evolutionary origin of the human granzyme H 158:4574. gene. Int. Immunol. 3:57. 14. Meier, M., P. C. Kwong, C. J. Fregeau, E. A. Atkinson, M. Burrington, 37. Tschopp, J., D. Masson, S. Schafer, M. Peitsch, and K. T. Preissner. 1988. The N. Ehrman, O. Sorensen, C. C. Lin, J. Wilkins, and R. C. Bleackley. 1990. heparin binding domain of S-protein/vitronectin binds to complement compo- Cloning of a gene that encodes a new member of the human cytotoxic cell pro- nents C7, C8, and C9 and perforin from cytolytic T-cells and inhibits their lytic tease family. Biochemistry 29:4042. activities. Biochemistry 27:4103. 15. Hanna, W. L., J. M. Turbov, H. L. Jackman, F. Tan, and C. J. Froelich. 1994. 38. Jadot, M., C. Colmant, S. Wattiaux-De Coninck, and R. Wattiaux. 1984. Intraly- Dominant chymotrypsin-like esterase activity in human lymphocyte granules is sosomal hydrolysis of glycyl-L-phenylalanine 2-naphthylamide. Biochem. J. 219: mediated by the serine carboxypeptidase called -like protective pro- 965. tein. J. Immunol. 153:4663. 39. Matsuda, K., S. Ishiura, H. Koizumi, H. Kawasaki, K. Arahata, and H. Sugita. 16. Prendergast, J. A., C. D. Helgason, and R. C. Bleackley. 1992. Quantitative 1991. Apolipoprotein B is a major perforin inhibitor protein in human serum. polymerase chain reaction analysis of cytotoxic cell proteinase gene transcripts in Mol. Immunol. 28:1211. T cells. J. Biol. Chem. 267:5090. 40. Dupuis, M., E. Schaerer, K. Krause, and J. Tschopp. 1993. The calcium-binding 17. Woodard, S. L., D. S. Jackson, A. S. Abuelyaman, J. C. Powers, U. Winkler, and protein calreticulin is a major constituent of lytic granules in cytolytic T lym- D. Hudig. 1994. Chymase-directed serine protease inhibitor that reacts with a phocytes. J. Exp. Med. 177:1.