Design and Characterization of a Fluorogenic Substrate Selectively

THEJOURNAL OF BIOLOGICAL CHEMISTRY 0 1994 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 269,No. 33, Issue of August 19,pp. 20952-20957, 1994 Printed in U.S.A.

Design and Characterization of a Fluorogenic Substrate Selectively Hydrolyzed by Stromelysin 1 (Matrix Metalloproteinase-3)* (Received for publication, May 19, 1994) Hideaki Nagasel, CynthiaG. FieldsBI, and Gregg B. Fields§nll** From the Wepartment of Biochemistry and Molecular Biology, University of Kansas Medical Center, Kansas City, Kansas 66103 and the Departments of §Laboratory Medicine and Pathology, 1€3iochemistry,and YThe Biomedical EngineeringCenter, University of Minnesota, Minneapolis, Minnesota55455

Members of the matrix metalloproteinase (MMP) fam- thetic libraries and consideration of secondary and terily have been implicated in disease states such as arthri- tiary structures of substrates and the enzyme. tis, periodontal disease, and tumor cell invasionand metastasis. Stromelysin 1 ("P-3) has a broad substrate specificity and participates in the activation of several The matrix metalloproteinase (MMP)l family has been imMMP zymogens. We examined known sequences of plicated in a variety of disease states, including arthritis, peMMP-3 cleavage sites in natural peptides and proteins riodontaldisease,andtumor cell invasionandmetastasis and compared sequence specificities of MMP-3 and in- (Woessner, 1991; Birkedal-Hansen et al., 1993; Stetler-Steventerstitial collagenase (MMP-1)in orderto design fluoro- son et al., 1993). Studies with tumorigenic cells have reported genic substrates that(i) would be hydrolyzed rapidly by of interstitia1 collagenase correlations betweenexpression " P - 3 , (ii) would discriminate between " P - 3 and (MMP-11, gelatinase A (MMP-21, stromelysin 1 (MMP-31, maMMP-1, and (iii) could be monitored continuouslywithB (MMP-g), and/or stromelysin 3 trilysin (MMP-7), gelatinase out interference from MMP amino acid residues. De(MMP-11) and invasive and metastatic behaviors (Goldberg et signed substrates were then screened for activity toward MMP-1, gelatinase A ("P-21, " P - 3 , and al., 1990; Stetler-Stevenson et al., 1993). It is not clear which The first of these substrates, MMPs are of primary importance for tumor cell invasion or if gelatinase B ("9). invasive behavior can be assigned to a single member of the NFF-1 (Mca-Pro-Lys-Pro-Gln-Gln-Phe-Phe-Gly-Leu-Lys(Drip)-Gly, where Mca is (7-methoxycoumarin-4-y1)acetyl MMP family(Furcht et al.,1994). The development of discrimiand Dnp is 2,4-dinitrophenyl), was hydrolyzed equally natory substrates may allow for the assignment of a specific MMP activity for pathogenesis. An assay specific for MMP-3 wellby " P - 3 and MMP-2 (kcaJKm-11,000 s-l M-'). MMP-1 had 25% of the activity of MMP-3 toward NFF-1. would be especially valuable, asMMP-3 can participate in the activation of zymogens of MMP-1, MMP-3, and MMP-9 (Vater The second substrate, NFF-2 (Mca-Arg-Pro-Lys-Pro--Ala-Nva-"p-Met-Lys(Dnp)-NH,,where Nva is norval- et al., 1983; Murphy etal., 1987; Brinckerhoff et al., 1990; ine), was hydrolyzed 60 times more rapidly by " P - 3 Nagase et al., 1990; Suzuki et al., 1990; Ogata et al., 1992) and (k,,JKm = 59,400 s-' "'1 than " P - 1 . Unfortunately, degrades a broad range of protein substrates. This variety of NFF-2showed little discrimination between " P - 3 , activities suggests that MMP-3 may be involved in pathogenMMP-2 (kcaJKm= 54,000 s-l M-I), and MMP-9 (kc,,&,,= esis of a number of connective tissue diseases. 55,300 s-' "I). The third substrate, NFF-3 (McaSynthetic peptide-based assays canbe developed t o discrimiArg-Pro-Lys-Pro-Val-Glu-Nva-"rp-Arg-Lys(Dnp)-NH,), nate between enzyme types. Oneadvantage for using synthetic was hydrolyzed rapidly by MMP-3 (k,,JKm= 218,000 s-' peptides is that moieties required for assay monitoring, such as and very slowly by " P - 9 (Kc,&,, = 10,100 s-l M-I), chromophores or fluorophores, can be incorporated easily. Flubut there was no significant hydrolysis by MMP-1 and orogenic substrates provide a particularly convenient enzyme MMP-2.NFF-3 is the first documented synthetic subassay method, as they can be monitored continuously and utistrate hydrolyzed by only certain members of the MMP lized at reasonably low concentration ranges.Although a numfamily and thus has important application for the discrimination of " P - 3 activity from that of other MMPs. ber of laboratories have reportedon the design and characterAlthough NFF-3 was designed by assuming that sub- ization of fluorogenic substrates for MMPs (Fields, 1988; Stack strate subsites were independent and hence free energy and Gray, 1989; Netzel-Arnett et al., 1991b; Knight et al., 1992; changes derived from singlemutation experiments were Niedzwiecki et al., 1992; Bickett et al., 1993; Geogheganet al., additive, we found discrepancies between predicted and 1993), none have identified a substrate that is hydrolyzed by only certain members of the MMP family. For example, the experimental k,,JK,,, values, one on the order of 2000substrate of Knight et al. 5000. Thus, the design of additional discriminatory MMP Mca-Pro-Leu-Gly-Leu-Dpa-Ala-Arg subsubstrates may require approaches other thanassuming (1992) and the Dnp-Pro-Leu-Gly-Leu-Trp-Ala-D-Arg-NH, additive free energy changes, such as screening syn- strate of Stack and Gray (1989) are hydrolyzed by MMP-1,

* This work was supported by National Institutes of Health Grants

The abbreviations used are: MMP, matrix metalloproteinase; Cha,

AR 39189, AR 40994 (to H. N.), KD 44494, and AR 01929 (to G. B. F.) cyclohexylalanine;DCM, dichloromethane; DIPCDI, N,N"diisopropyland by the American Cancer Society and Millipore Carp. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be herebymarked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. ** Recipient of a McKnight-Land grant professorship anda National Institutes of Health research careerdevelopment award. Towhom correspondence should be addressed: Dept. of Laboratory Medicine and Pathology, Box 107, 420 Delaware St. S.E., University of Minnesota, Minneapolis, MN 55455.Tel.: 612-626-2446; Fax: 612-625-1121.

carbodiimide; DMF, N,N-dimethylfonnarnide; DMPAMP, 4-(2',4'-di-

methoxyphenylaminomethy1)phenoxy; Dnp, 2,4-dinitrophenyl; Dpa, N-3-(2,4-dinitrophenyl)-~-2,3-diaminopropionyl; FABMS, fast atom bombardment mass spectrometry; Fmoc, 9-fluorenylmethoxycarbonyl; HMP,4-hydroxymethylphenoxy;HOBt,l-hydroxybenzotriazole; Mca, (7-methoxycoumarin-4-y1)acetyl;Nma, N-methylanthranilyl;Nva, norvaline; resin, copoly(styrene-1%-divinylbenzene);RP-HPLC, reversedphase high performance liquid chromatography; TIMP,tissue inhibitor of metalloproteinase.

20952

Fluorogenic Substrate Stromelysin

20953

(0.548 ml; 3.5 mmol) was added and coupling proceeded for 4.5 h. The resultant Fmoc-Lys(Dnp)-DMPAMPresin was washed with DMF and DCM and stored under vacuum. Peptide Synthesis and Purification-Incorporation of individual amino acids was by Fmoc solid-phase methodology on an Applied Biosystems 431A peptide synthesizer using cyclesdescribedpreviously (Fields et al., 1991; Fields et al., 1992a). Peptide-resins were analyzed prior to Mca incorporation by Edman degradation analysis to evaluate the efficiency of assembly (Fields and Fields, 1993; Fields et al., 1993~). The N termini of peptide-resins were acylated with 7-methoxycoumarin-4-acetic acid using standard synthesis cycles (Fields et al., 1991; Fields et al., 1992b) and 2-4 h coupling times. Mca-peptideswere cleaved from the resin and side chain deprotected by treatment with trifluoroacetic acid plus appropriate scavengers (Kinget al., 1990).Peptides were purified by preparative RP-HPLCon a Beckman System Gold with a Regis Chemical ODS C,, column (10-pmparticle size, 60-8, pore size, 250 x 21.1 mm). The elution gradient was 20-100% acetonitrile containing 0.1% trifluoroacetic acid in 35 min at a flow rate of 5.0 mumin. Detection was at 235 nm. Analytical RP-HPLCwas performed on a Hewlett-Packard 1090 liquid chromato aph equipped with a Dynamax C,, column (12-pmparticle size, 300-rpore size, 250 x 4.6 mm). The elution gradient was 0-100% acetonitrile containing 0.1% trifluoroacetic acid in 60 min at a flow rate of 1.0 mumin.Diode array detection was at 220, 254, and 280 nm. Peptide Analyses-Ninhydrin analysis (Fields et al., 1992b) was used to monitor all manual coupling and deprotection steps. Edman degradation sequence analysis of peptide-resins was performedon an Applied Biosystems 477A protein sequencer/l20A analyzer as described previously (Fields et al., 1993~). Fulvene-piperidine concentrations (301 nm) and scanning UV spectra (200-320 nm) were determined with a Beckman DU-70 spectrophotometer.Peptides were characterized by either electrospray mass spectrometry as described (Fields et al., 1993~)or EXPERIMENTALPROCEDURES FABMS on a VG 7070E-HF with a glycerol matrix. Matrix Metalloproteinases-AllMMPswerepurified in zymogen Materials-All standard peptide synthesis chemicals wereanalytical reagent grade or better and purchased from Applied Biosystems, Inc. form. Pro-MMP-1(Suzuki et al., 1990), pro-MMP-2(Okada et al., 1990), (Foster City,CA) or Fisher. 1,8-Diazabicyclo[5.4.0lundec-7-ene,1,2- and pro-MMP-3(Ito and Nagase, 1988)were purified from the culture ethanedithiol, and thioanisole were fromAldrich, HMP resin (substitu- medium of human rheumatoid synovial cells stimulated with rabbit the tion level = 0.97 mmoug) from Bachem Biosciences(Philadelphia, PA), macrophage-conditioned medium. Pro-MMP-9 was purified from Fmoc-DMPAMPresin (substitution level = 0.43 mmoug) from Novabio- culture medium of HT-1080 cells as described (Morodomi etal., 1992). chem (La Jolla, CA), 2-(lH-benzotriazole-l-yl)-l,1,3,3-tetramethyluro-Pro-MMP-2 and pro-MMP-9were activated by reacting with 1 mM nium hexafluorophosphate fromRichelieuBiotechnologies (St.-Hya- 4-aminophenylmercuricacetate at 37 "C for 45 min and 32 h, respectively. Pro-MMP-1 was activated by reacting with 1 mM 4-aminophencinthe, Quebec), and Lys(Dnp) from Sigma. Fmoc-amino acid ylmercuric acetate and an equimolar amount of MMP-3 at 37 "C for6 h. MA) or Novaderivatives were obtained from Millipore Corp. (Bedford, After activation, MMP-3 was completely removed from MMP-1 by afbiochem. Amino acids are of the L-configuration (except for Gly). Preparation of Fmoc-Lys(Dnp)-Fmoc-Lys(Dnp) was prepared from finity chromatography using an anti-MMP-3 IgG Affi-Gel 10 column. Pro-MMP-3 was activated by reacting with 5 pg/ml chymotrypsin at Lys(Dnp) as follows. Fmoc N-hydroxysuccinimide ester (1.89 g; 5.60 mmol) was dissolved in 30 ml of dimethoxyethane and stored at 4 "C. 37"C for 2 h. Chymotrypsin was inactivated with 2 mM diisopropyl Lys(Dnp) (1.63 g; 4.67 mmol) was dissolved in 10 ml of 10% aqueous fluorophosphate. The amounts of active MMP-1, MMP-2, and MMP-3 Na,CO, and added slowly to the dimethoxyethane solution. The reac- were determined by titration with TIMP-1over a concentration range of tion proceeded for2 h at 4 "C and overnight at room temperature. The 0.1-3 pg/ml, while the amount of active MMP-9 was determined by solution was filtered and the filtrate acidified to pH -3 with concen- titration with TIMP-2 ( 0 . 1 3 pglml). TIMP-1 was isolated from the trated HCl. Dimethoxyethane was removedby heating the solution medium of HT-1080 cells (Morodomi etal., 1992) and TIMP-2 fromthe under reduced pressure. The solution was extracted with ethyl acetate, medium of human uterine cervical fibroblasts.2 Assays-Substrates were prepared as 10 mM stock solutions in diand the ethyl acetate layer was reduced by heating under reduced methyl sulfoxide.Fluorescent assays were performedat A,, = 325 nmand pressure to an oil. The oil (Fmoc-Lys(Dnp))was used without further purification. A,, = 393 nm, which should encounter no interference from Trp residues, using a Hitachi fluorescence spectrophotometerF-3010. Initial total hyPreparation of Fmoc-Lys(Dnp) Resins-Fmoc-Lys(Dnp)-Gly-HMP drolysis assays were run at a substrate concentration of 5 p~ to avoid resin was prepared as follows. Fmoc-Gly (5.77 g; 19.4 mmol), HOBt (2.97 g;19.4 mmol), and 4-(dimethylamino)pyridine (0.237 g; 1.94 filtering effects.The change in fluorescenceat this concentrationwas 110 mmol) were dissolved in 100 ml of DCM-DMF (1:l)and added to 5.0 g based on total hydrolysis of 1p~ substrate. Atypical assay was carried of HMP resin (4.85 mmol). DIPCDI(3.04 ml; 19.4 mmol) was added, and out by incubating 100pl of various concentrations of a substratewith 10 esterification proceeded for 3 h. An additional 200ml of DMF was pl of an enzyme solution (2-20 n ~in) 50 mM Tris-HC1, pH 7.5, 0.15 M added, and the reaction continued for 3.5h. The resin was washed with NaCl, 10 nm CaCl,, 0.05% Brij 35,0.02%NaN,. The reaction was stopped DMF and DCM and stored under vacuum overnight. Fulvene-piperi- by addition of 900 plof 3% (v/v) glacial acetic acid. MMP-3activity at pH dine analysis (Fields et al., 1992b) gave a substitution level of 0.45 6.0 was measured in 50 mM sodium acetate buffer, pH 6.0, instead of mmoVg for Fmoc-Gly-HMP resin. 3.0 g of Fmoc-Gly-HMP resin (1.36 Tris-HC1 buffer. The amount of substrate hydrolysis was calculated mmol) was deprotected by treatment with of ml 50 1,8- based on the fluorescencevalues of the Mca-Arg-Pro-Lys-Pro-Glnstanddiazabicyclo[5.4.0lundec-7-ene-piperidine-DMF(1:1:48)for0.5 h and ard solution after subtraction of the reaction blank value (stopping sowashed 3 times with DMF. Fmoc-Lys(Dnp) (0.232g; 4.07 mmol) and lution added before the enzyme).Individual kinetic parameters (k,,,and HOBt (0.623 g; 4.07 mmol) were dissolvedin 50 ml in DCM-DMF (1:l) K,) were determined over a substrateconcentration range of 2.5-75 PM and added to the resin. DIPCDI (0.637 ml; 4.07 mmol)was added, and and calculated by double-reciprocal plots. coupling proceeded for 3.5 h. The resultant Fmoc-Lys(Dnp)-Gly-HMP resin was washed with DMF and DCM and stored under vacuum. RESULTSANDDISCUSSION Fmoc-Lys(Dnp)-DMPAMPresin was prepared as follows. 2.0 g of MMP-3 has a broad substrate specificity that includes extraFmoc-DMPAMP resin (0.86 mmol) wasdeprotected by treatment with 20 ml of piperidine-DMF (1:l) for 0.5h and washed 3 times with DMF. cellular matrix proteins and protease inhibitors (Table I). In Fmoc-Lys(Dnp)(0.2 g; 3.5 mmol) and HOBt (0.536 g; 3.5 mmol) were dissolved in 20 ml in DCM-DMF (1:l)and added to the resin. DIPCDI Y. Itoh, S. Binner, and H. Nagase, manuscript in preparation.

MMP-2, MMP-3, and MMP-7 (Knight et al., 19921, while the Dnp-Pro-Cha-Gly-Cys(CH,)-His-Ala-Lys(Nma)-NHz substrate of Bickett et al. (1993) is hydrolyzed by at least MMP-1 and MMP-9. We have examined known sequences of MMP-3 cleavage sites in naturalpeptides (insulin (Azzo and Woessner, 1986; Wilhelm et al., 1993) and substance P (Teahan et al., 198911, protease inhibitors (ovostatin (Enghild et al., 19891, qmacroglobulin (Enghild et al., 1989), antithrombin I11 (Mast et al., 1991),cy,-antichymotrypsin(Mast et al., 1991),and a,-protease inhibitor (Mast et al., 1991)),MMP zymogens (pro-MMP-1 (SUzuki et al., 1990),pro-MMP-3 (Nagase et al., 19901, pro-MMP-8 (Knauper et al., 1993), and pro-MMP-9 (Ogata et al., 199211, and extracellular matrix proteins (types 11, M, and XI collagen (Wu et al., 1991), aggrecan (Flannery et al., 1992; Poe et al., 1992), fibronectin (Wilhelm et al., 19931, and cartilage link protein (Nguyen et al., 1989))and compared sequence specificity studies of MMP-3 (Fields, 1988; Niedzwiecki et al., 19921 and MMP-1 (Weingarten et al., 1985; Weingarten and Feder, 1986; Fields et al., 1987; Fields, 1988; Netzel-Amett et al., 1991a) in order to design fluorogenic substrates that (i)would be hydrolyzed rapidly by MMP-3, (ii) would discriminate between MMP-3 and MMP-l, and (iii) could be monitored continuously without interference from MMP amino acid residues. These designed substrates were also compared forMMP-2 and MMP-9 activity, and individual kinetic parameters for all four MMPs were determined.

Fluorogenic Substrate Stromelysin

20954

TABLEI Sequences of protein cleavage sitesof MMP-3 Protein

k,,JKmb

Sequence"

Ref.

s-l "1

Human a,-macroglobulin 56,000' 35,000 5,400 4,000d

Human a,-antichymotrypsin a,-Protease inhibitor Human aggrecan Substance P Antithrombin I11 Chicken ovostatin Human pro-MMP-1 Human pro-MMP-3 Human pro-MMP-3 Human pro-MMP-8 Human pro-MMP-9 Human pro-MMP-9 Bovine d ( I 1 ) collagen, N-telopeptide Bovine al(I1) collagen, N-telopeptide Bovine al(IX) collagen, NC2 Bovine a2(M) collagen, NC2 Bovine a3(M) collagen, NC2 Bovine al(XI) collagen, N telopeptide Human cartilage link Human fibronectin Bovine insulin, B chain

1,790