Download - Native TIMP-free 70 kDa progelatinase (MMP-2) secreted at elevated levels by RSV transformed fibroblasts

JOURNAL OF CELLULAR PHYSIOLOGY 161:419428 (1994)

Native TIMP-Free 70 kDa Progelatinase (MMP-2) Secreted at Elevated levels by RSV

Transformed Fibroblasts STEINGRIMUR STEFANSSON, RONALD 1. AIMES, NANCY B. SEWARD,

DANIELA S . ALEXANDER, AND JAMES P. QUICLEY* Department of Pathology (S.S., N. B.S., D.S.A., 1. P.Q.) and Department of Biochemistry and

Cell Biology (R.T.A.), State University of New York at Stony Brook, Stony Brook, New York 1 1794

Rous sarcoma virus-transformed cultures of chicken embryo fibroblasts (RSVCEF) secrete elevated levels of a 70 kDa progelatinase, an avian form of the 72 kDa matrix metalloproteinase-2 (MMP-2). Affinity-purified preparations of secreted 70 kDa progelatinase are composed of two distinct populations of zymogen: a 70 kDa progelatinase tightly complexed with an avian form of TIMP-2 and a native 70 kDa progelatinase free of any detectable TIMP-2. These two forms of the progelatinase can be separated by Mono Q FPLC in the absence of denaturing agents. The homogeneity of the two separated forms i s demonstrated by both SDS-PACE and nondenaturing, native gel electrophoresis. The purified TIMP-free 70 kDa progelatinase i s stable in aqueous conditions and does not spontaneously autoactivate. Treatment of the TIMP-free progelatinase with the organomercurial, p-aminophenylmercuric acetate (APMA), results in rapid (5-60 minutes) autolytic conversion of the 70 kDa progelatinase to 67 kDa, 62 kDa and lower molecular weight forms of the enzyme. APMA treatment of the TIMP-free progelatinase yields a preparation that is enzymatically active with a high specific activity towards a peptide substrate. Identical treatment of TIMP-complexed progelatinase with APMA results in a significantly slower conversion process in which the 70 kDa progelatinase is only 50% converted after 6-24 hours and the specific enzyme activity of the preparation i s 8 to 18-fold lower. Purified avian TIMP-2 added to the TIMP-free progelatinase forms a complex with the progelatinase and prevents the rapid autolytic conversion induced by APMA. Comparative analysis of parallel cultures of transformed RSVCEF and normal CEF demonstrates that the transformed cultures contain threefold higher levels of the TIMP-free progelatinase than the normal CEF cultures which produce predominantly TIMP-complexed progelatinase. The presence in transformed cultures of elevated levels of a more readily activated TIMP-free progelatinase, the suppression of its rapid activation by TIMP-2, and the potential effect of the altered balance between TIMP-free and TIMP-complexed 70 kDa progelatinase on the invasive, malignant phenotype, are discussed. o 1994 Why-Liss, Inc

The matrix metalloproteinases (MMPs) are a family of zinc-dependent endoproteinases secreted by both normal and transformed cells and are capable of degrading both the collagenous and noncollagenous components of the extracellular matrix (ECM; Matrisian, 1990; Woessner, 1991; Liotta et al., 1991). These enzymes play an important role in both normal and pathological processes that involve remodeling of the ECM such as angiogenesis, wound healing, inflammation and tumor invasion (Liotta et al., 1991; Werb, 1989; Murphy and Docherty, 1992). Several members of the MMP family have been identified and include a number of stromel- ysins (Chin et al., 1985; Nicholson et al., 1989; Quantin et al., 1989), collagenases (Goldberg et al., 1986; Ma- cartney and Tschesche, 19831, and gelatinases (Collier et al., 1988; Wilhelm et al., 1989). All of these enzymes share several features: they are produced as zymogens, 0 1994 WILEY-LISS, INC.

are activated by organomercurials resulting in the autolytic cleavage of an &lo kDa amino-terminal frag- ment (Stetler-Stevenson et al., 1989b), function at neutral pH, and are inhibited by members of the tissue inhibitor of metalloproteinases (TIMP) family (Murphy et al., 1981). These shared properties are the result of a highly ordered and conserved domain structure found

Received April 12, 1994; accepted June 3,1994. *Address reprint requestsicorrespondence to James P. Quigley, Department of Pathology, State University of New York at Stony Brook, Stony Brook, NY 11794. The present address for Dr. Steingrimur Stefansson is Biochemis- try Division, The American Red Cross, 15601 Crabbs Branch Way, Rockville, MD 20855.

STEFANSSON ET AL 420

in the MMPs, including a propeptide domain, a zinc- binding active site domain and a carboxy terminal domain (Matrisian, 1990; Woessner, 1991). Two members of the MMP family, the 92 kDa and 72 kDa progelati- nases, also contain a gelatin-binding domain (Collier et al., 1988; Wilhelm et al., 1989), which may influence their substrate specificity and allows them to be purified by gelatin-Sepharose affinity chromatography (Moll et al., 1990).

A well-characterized member of the MMP family is the 72 kDa progelatinase (MMP-2). The production of this enzyme has been linked to the invasive behavior of a number of malignant cells (Liotta et al., 1991; Wilhelm et al., 1989; Moll et al., 1990; Chen et al., 1991; Garbisa et al., 1987; Albini et al., 1991; Levy et al., 1991). The 72 kDa progelatinase is secreted as a zymogen and can undergo proteolytic processing to a 62 kDa active enzyme mediated by intact cells or cell membranes (Brown et al., 1990; Ward et al., 1991a; Overall and Sodek, 1990). The activation involves the removal of 80 residues from the amino terminus which contains an unpaired cysteine (Stetler-Stevenson et al., 198913). This cysteine is thought (Van Wart and Birkedal-Hansen, 1990; Springman et al., 1990; Wind- sor et al., 1991) to coordinate with the active site zinc in the zymogen, maintaining it in a noncatalytic form.

The activity of the 72 kDa progelatinase is regulated by TIMP-2, a 21 kDa molecule with sequence similarity to the originally described metalloproteinase inhibitor, TIMP-1 (Stetler-Stevenson et al., 1989a; De Clerck et al., 1989) and to a newly described member of the TIMP family, ChIMP-3 (Pavloff et al., 1992). Like TIMP-1, TIMP-2 inhibits the activity of activated MMPs (Howard et al., 1991a). However, only TIMP-2 forms a high affinity complex with the latent 72 kDa progelatinase and copurifies with the zymogen under nondenaturing conditions, whereas TIMP-1 forms a complex and copurifies with the 92 kDa progelatinase (Collier et al., 1988; Wilhelm et al., 1989). It appears that TIMP-2 may have a dual role in the regulation of 72 kDa progelatinase in that it inhibits the enzymatic activity of the activated 62 kDa gelatinase (Howard et al., 1991b; Murphy et al., 1992) and also retards the membrane- mediated activation of the latent 72 kDa progelatinase (Ward et al., 1991b). The functional properties of the 72 kDa progelatinase and TIMP-2 have been difficult to assess due to the strong interaction between these two

____~

RSV CEF FPLC APMA TIMP MMP SDS DMEM PAGE Me,SO EDTA CAB ECM T-F T-C

Abbreviations

Rous sarcoma virus chicken embryo fibroblasts fast protein liquid chromatography aminophenylmercuric acetate tissue inhibitor of metalloproteinases matrix metalloproteinases sodium dodecyl sulphate Dulbecco’s modified Eagle’s medium polyacrylamide gel electrophoresis dimethyl sulfoxide ethylenediaminetetraacetic acid calcium assay buffer extracellular matrix TIMP-free TIMP-complexed

proteins and because they copurify from most mammalian cell cultures. Several attempts have been made to separate and purify the 72 kDa progelatinase away from its complexed inhibitor using denaturants such as SDS (Chen et al., 1991; Goldberg et al., 1989) or by using reverse phase HPLC in the presence of acetoni- trile and trifluoroacetic acid (Howard et al., 1991b); however, the former render the gelatinase inactive and the latter cause the enzyme to become unstable and autoactivate upon dialysis into neutral buffer, yielding smaller active and inactive peptide fragments (Howard et al., 1991b). Recently, recombinant TIMP-free 72 kDa progelatinase has been produced (Murphy et al., 1992; Fridman et al., 1992) and TIMP-free 72 kDa progelatinase also has been isolated from the conditioned media of cytokine-treated human fibroblasts (Ward et al., 1991b) and rheumatoid synovial cells (Okada et al., 1990; Kolkenbrock et al., 1991). None of these proen- zymes isolated free of TIMP spontaneously activate. It is not clear whether the production of TIMP-free 72 kDa progelatinase is a normal physiological event and whether the absence of associated TIMP-2 in vivo would render the 72 kDa progelatinase more rapidly activated by cellular mechanisms as it appears to be with organomecurials in vitro.

Our laboratory has reported (Chen et al., 1991) the isolation and characterization of a 70 kDa progelatinase that is elevated in Rous sarcoma virus-transformed chicken embryo fibroblasts (RSVCEF). Bio- chemical properties and partial <amino terminal sequencing (Chen et al., 1991) along with the deduced protein sequence from the cloned DNA (Aimes et al., 1994) suggest that this is an avian counterpart to the mammalian 72 kDa progelatinase. The avian 70 kDa progelatinase also copurifies with a 21 kDa TIMP-2- like protein during gelatin-affinity (chromatography. However, gel filtration under nondenaturing conditions suggested that some of the proenzyme may exist free of inhibitor (Chen et al., 1991). We report here on the separation of a naturally produced, mixed population of 70 kDa progelatinase molecules, TIMP-complexed and TIMP-free, from the culture media of CEF and RSVCEF, using gelatin-affinity chromatography and Mono Q FPLC under nondenaturing conditions. This mixed population in RSVCEF cultures provides a system to study the properties of the native gelatinase in the presence or absence of its naturally occurring inhibitor and to determine whether malignant transformation alters the balance between the free enzyme and its inhibitor.

MATERIALS AND METHODS Gelatin-Sepharose and MonoQ HR 515 were pur-

chased from Pharmacia Fine Chemicals (Piscataway, NJ). Me,SO, Brij-35, Triton X-100, phosphate-buffered saline, 1,lO-phenanthroline, EDTA, CaCl,, p-aminophenylmercuric acetate (APMA) and gelatin were all purchased from Sigma Chemical Company (St. Louis, MO). SDS, SDS-PAGE protein molecular weight stan- dards and Coomassie Brilliant Blue R-250 were purchased from Bio-Rad (Rockville Center, NY). Bicincho- ninic acid (BCA) protein assay reagent was purchased from Pierce (Rockford, IL). Centriprep-10 concentration units were from Amicon (Danvers, MA). Dulbecco’s

42 1 TIMP-FREE PROGELATINASE IN RSV TRANSFORMED FIBROBLASTS

modified Eagle's medium, penicillin and streptomycin were purchased from Gibco (Grand Island, NY). Patho- gen-free COFAL negative fertilized eggs were obtained from SPAFAS (Norwich, CT). The gelatinase substrate, DNP-Pro-Leu-Gly-Leu-Trp-Ala-D-Arg-NH, was purchased from Peptides International (Louisville, KY). Fetal bovine serum was purchased from Hyclone (Lo- gan, UT). Ultragel AcA44 was purchased from IBF Bio- technics (Savage, MD).

Cell culture Primary cultures of chicken embryo fibroblasts

(CEF) were prepared from 10-day-old pathogen-free COFAL negative fertilized eggs as previously described (Fairbairn et al., 1985; Sullivan and Quigley, 1986). The cells were grown at 37"C, 5% CO, in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum, 100 U/ml penicillin, 100 U/ml streptomycin, 1 mM sodium pyruvate, and 4 mM glu- tamine.

Secondary cultures of CEF were infected with Rous sarcoma virus as described previously (Chen et al., 1991; Kawai and Hanafusa, 1971). After three days of growth, the cultures appeared morphologically transformed (Chen et al., 1991; Kawai and Hanafusa, 1971). Third and fourth passage cultures (CEF and RSVCEF) were used for obtaining serum-free conditioned medium. RSVCEF and CEF cultures were grown to 80- 90% confluence, washed twice with serum-free DMEM and placed in serum-free medium. Conditioned medium was collected after 24 hours at 37"C, centrifuged a t 500g to remove cells and cell debris and stored at 0°C until purification of the 70 kDa progelatinase.

Progelatinase purification Purification of the avian 70 kDa progelatinase from

RSVCEF conditioned media was performed essentially as described previously (Chen et al., 1991). Serum-free conditioned medium was concentrated 20 x by ultrafil- tration or was precipitated by 70% saturated ammo- nium sulphate and resuspended in DMEM at 1120 the original volume. The collected precipitate or the 20x conditioned medium concentrate was dialyzed overnight against gelatin-Sepharose column buffer (50 mM Tris, pH 7.5, 0.5 M NaC1, 5 mM CaCl,, 0.02% NaN,, 0.05% Brij-35) and applied to a gelatin-Sepharose column. After washing the column with column buffer, the progelatinase was eluted with 7.5% Me2S0 in 50 mM Tris pH 7.5, 0.5 M NaC1, 5 mM CaCl,, 0.02% NaN,, 0.05% Brij-35. The progelatinase preparation eluted from the gelatin-Sepharose column was generally 95- 100% pure as judged by SDS-PAGE. High molecular weight (> 150 kDa) gelatin-Sepharose binding proteins were present in some preparations. Gel filtration chromatography on AcA44 as described (Chen et al., 1991) removed the high molecular weight contaminating proteins in the void volume of the column and provided a pure preparation of RSVCEF 70 kDa progelatinase. The purified avian progelatinase migrates slightly faster than purified human 72 kDa progelatinase in SDS-PAGE and slightly slower than the 67 kDa standard protein (BSA) and has been assigned an apparent molecular weight of 70 kDa. The 70 kDa progelatinase from CEF cultures was purified in an identical manner

from serum-free conditioned medium harvested from CEF cultures.

TIMP purification Chicken TIMP was purified from preparations of 70

kDa progelatinase-TIMP complexes as described previously (Chen et al., 1991). Briefly, a preparation of 70 kDa progelatinase-TIMP complex, isolated by gelatin- Sepharose affinity and AcA44 gel filtration chromatography, was treated with 0.2% SDS for 15 minutes at 20°C to dissociate the progelatinase-TIMP complex. The sample was applied to an AcA44 column equili- brated in SDS column buffer (0.2 M NaC1,50 mM Tris pH 7.4, 0.2% SDS). The column was eluted at 20°C in SDS column buffer and the free TIMP fractionated in the 20-30 kDa region of the column, distinctly separated from the 70 kDa progelatinase. The eluted TIMP was dialyzed extensively against 40 mM Tris, 0.2 M NaC1,2.5% Triton X-100 and passed over an SM-2 (Bio- Rad) column to remove traces of detergent. The preparation was a homogenous 21 kDa molecule.

Enzyme assays All activation of the progelatinase was performed

with 2 mM APMA, diluted directly from a 100 mM APMA stock solution in Me,SO. When APMA-treated samples were applied to SDS-PAGE, the activation was stopped by adding SDS sample buffer to the reaction and storing the samples on ice until they were loaded onto the gels. The enzymatic activity of the gelatinase was measured using a synthetic substrate DNP-P-L-G- L-W-A-DR-NH, as described (Stack and Gray, 1989). The substrate was dissolved in Me2S0 and the concentration was determined b using an extinction coeffi- cient E~~~ = 16 mM-lcm- (Stack and Gray, 1989). As- says were carried out using 0.2 pg of enzyme in calcium assay buffer (CAB) which consists of 50 mM Tris pH 7.5, 10 mM CaC12, 0.2 M NaC1. The assays were performed at 25°C in a total volume of 0.25 ml using the substrate at a final concentration of 0.05 mM. The flu- orescence change was continuously monitored over a 60 minute time course in a Perkin-Elmer fluorimeter with the excitation wavelength at 280 nm and measuring the emission at 340 nm. In all cases the substrate hydrolysis by the enzyme was less than 10% of the total substrate added in the reaction. Substrate hydrolysis in the absence of added enzyme was less than 0.5% and was subtracted from all enzyme values.

Electrophoresis and zymography Electrophoresis was performed on Mini-Protean I1

electrophoresis system (Bio-Rad). All SDS-PAGE re- agents were prepared according to Laemmli (1970). Gelatin zymography SDS-PAGE was performedusing a final concentration of 0.28 mglml of copolymerized gelatin, as described previously (Chen et al., 1991). Native gel electrophoresis of the TIMP-free and TIMP-complexed 70 kDa progelatinase was performed using 5 1 5 % gradient gels made from a 30% wlv I 0.8% wlv acrylamidehisacrylamide stock solution and using a Trislglycine buffer system as described (Andrews, 1986) with the addition of a final concentration of 0.05% Brij-35 to the gels. Native and SDS-PAGE gels were stained with silver as described (Oakley et al.,

Y

STEFANSSON ET AL. 422

1980). Gelatin zymograms were washed with 2.5% Tri- ton X-100 and incubated overnight in CAB and stained with 0.2% Coomassie Brilliant Blue R-250 as described previously (Chen et al., 1991).

Protein determination The protein concentration of all samples was deter-

mined with the bicinchoninic acid (BCA) reagent (Pierce) and by direct comparison to purified protein standard on silver stained SDS-PAGE.

FPLC mono Q chromatography Preparations of 70 kDa progelatinase (10-50 kg total

protein in 1-5 ml) were dialyzed overnight at 4°C against 10 mM Tris, 0.05% Brij-35 pH 7.4. The samples were loaded onto a Mono Q HR 5/5 column linked to an FPLC apparatus a t a flow rate of 0.5 ml/min at 22°C. The proteins were eluted from the column at 1 ml/min with a 0.1 M to 1.0 M NaCl linear gradient in 10 mM Tris, 0.05% Brij-35 pH 7.4. Fractions (1 ml) were collected and absorbance at 280 nm was continuously re- corded.

RESULTS Separation by FPLC of two distinct populations of the MMPS 70 kDa progelatinase isolated from

cultures of RSVCEF Rous sarcoma virus-transformed chicken embryo fi-

broblasts produce an MMP-2-like 70 kDa progelatinase at elevated levels compared to normal chicken embryo fibroblasts (Chen et al., 1991). The purified 70 kDa gelatinase appears to be tightly complexed to an avian 21 kDa TIMP-2-like molecule (Chen et al., 1991). Gel filtration chromatography, however, indicated that purified preparations of the RSVCEF 70 kDa gelatinase may be composed of more than one form of the proenzyme. To address this possibility, gelatin-Sepharose- purified preparations of the avian 70 kDa progelatinase were subjected to FPLC Mono Q ion exchange chromatography and the bound proteins were eluted with a linear gradient of 0.1-1.0 M NaC1. Two distinct protein peaks are eluted a t 150 and 170 mM NaC1, respectively (Fig. 1A). Analysis of the two protein peaks by SDS-PAGE (Fig. 1B) indicates that the first peak contains a single protein band at 70 kDa while the second peak contains two proteins, the 70 kDa molecule and a 21 kDa protein which had been previously identified as an avian TIMP-2 (Chen et al., 1991). Gelatin substrate zymography demonstrates that both peaks contain the gelatinolytic proteinase migrating at 70 kDa (Fig. 1C).

Further characterization of the two protein peaks was carried out by native PAGE. Under nondenaturing conditions, a 21 kDa TIMP-70 kDa progelatinase complex with a combined molecular weight of 90-95 kDa would be expected to migrate slower than a TIMP-free 70 kDa proenzyme. Figure 2A illustrates that peak 1 from the Mono Q column migrates as a single band (Fig. 2A, lane 2) while peak 2 also migrates as a single band but at a slower electrophoretic rate (Fig. 2A, lane 3). The original gelatin-Sepharose-purified preparation )exhibits two distinct protein bands (Fig. 2A, lane 1) corresponding to the two species that were eluted in peaks 1 and 2.

A. .020 - -. -200 1

.OlO-

B. C.

0 0

67kDa-

0 0

7OkDa-

21 kDa - Fig. 1. Separation of TIMP-free and TIMP-complexed 70 kDa progelatinase by Mono Q FPLC. A A,,, tracing of a Mono Q fractionation of a preparation of the total gelatin-Sepharose-purified 70 kDa progelatinase from RSVCEF. The progelatinase preparation (40 Fg protein in 6 ml) was dialyzed overnight against column buffer, 10 mM Tris, 0.05% Brij, pH 7.4, loaded onto a Mono Q column and eluted with a linear gradient of the column buffer including 0.1 M to 1 M NaC1. Two peaks (1 and 2) were eluted a t 150 mM NaCl and 170 mM NaCI, respectively. B: Silver-stained 10% SDS-PAGE of TIMP-free 70 kDa progelatinase (peak 1, lane 1) and TIMP-complexed 70 kDa progelatinase (peak 2, lane 2). C: Coomassie-stained 8% SDS-PAGE gelatin zymogram of the TIMP-free 70 kDa progelatinase (peak 1, lane 1) and TIMP-complexed 70 kDa progelatinase (peak 2, lane 2).

A reconstitution experiment was carried out to con- firm that the isolated TIMP-free 70 kDa progelatinase is a native zymogen able to bind TIMP-2 and not a partially denatured species generated through the purification procedure. Avian TIMP-2 was purified as described previously (Chen et al., 1991) and mixed in equimolar amounts with the purified preparation of TIMP-free 70 kDa progelatinase (peak 1 from the Mono Q column) and analyzed by native PAGE (Fig. 2B). The reconstituted complex of 70 kDa progelatinase and TIMP-2 (Fig. 2B, lane 2) migrates slower than the TIMP-free 70 kDa progelatinase (Fig. 2B, lane 1) with a mobility equal to that of the native TIMP-complexed 70 kDa progelatinase which had been isolated from the Mono Q column (Fig. 2B, lane 3). Further demonstra- tion that the appearance of TIMP-free 70 kDa progelatinase was not the result of dissociation of 70 kDa progelatinase from the TIMP-70 kDa progelatinase complex during purification was obtained by resubject- ing the isolated TIMP-70 kDa progelalinase complex to gelatin-Sepharose and FPLC-Mono Q chromatography. No TIMP-free 70 kDa progelatinase was observed during multiple chromatographic procedures of the TIMP-70 kDa progelatinase complex (data not shown).

TIMP-FREE PROGELATINASE IN RSV TRANSFORMED FIBROBLASTS 423

A.

1

8.

1 2

2

Fig. 2. Analysis of the TIMP-free and TIMP-complexed 70 kDa progelatinase by native PAGE. A Samples of gelatin-Sepharose-purified 70 kDa progelatinase (40 ng, lane 11, TIMP-free 70 kDa progelatinase (T-F) from peak 1, Figure 1 (30 ng, lane 2) and TIMP-complexed 70 kDa progelatinase (T-C) from peak 2, Figure 1 (15 ng, lane 3) were applied to a 5-15% native PAGE gel and stained with silver. B Proge- latinase-TIMP complex was reconstituted with purified chicken TIMP-2 (50 ng) and Mono Q-purified TIMP-free 70 kDa progelatinase

These results and the above described native PAGE experiments (Fig. 2) indicate that the conditioned medium of RSVCEF cultures contains a mixed population of TIMP-free 70 kDa progelatinase and TIMP-70 kDa progelatinase complex.

APMA treatment of TIMP-free and TIMP-complexed 70 kDa progelatinase: differential generation of active enzyme

A distinct difference in the susceptibility to and extent of activation of the two purified forms of the 70 kDa progelatinase was observed by monitoring the generation of active enzyme following treatment of the TIMP-free and TIMP-complexed 70 kDa progelatinase with the organomercurial, APMA. Enzymatic activity was measured in solution by monitoring the specific cleavage of a collagen-like peptide substrate as described in Materials and Methods. The results shown in Figure 3 demonstrate that the specific peptidolytic activity of the TIMP-free preparation after two hours of APMA treatment is 8-fold higher than equimolar amounts of the corresponding TIMP complex preparation subjected to the same APMA treatment. After 24 hours of treatment with APMA, the specific activities of both preparations are diminished but the specific activity of the TIMP-free preparation remains substantially higher (18-fold) than the corresponding TIMP complex preparation. The absence of any significant peptidolytic activity with no APMA treatment (0 treat-

3

3 4

(15 ng). The proteins were mixed together in CAB for 30 min at 25°C in a final volume of 45 pl before analyzing the products on a 5 1 5 % native PAGE gel. TIMP-free progelatinase (15 ng) is shown in lane 1. The reconstituted complex is shown in lane 2. Mono Q-purified TIMP- complexed 70 kDa progelatinase (25 ng) is shown in lane 3. A sample of the total gelatin-Sepharose-purified 70 kDa progelatinase preparation (30 ng) is shown in lane 4.

ment, Fig. 3) confirms that the isolated TIMP-free 70 kDa species is a stable zymogen requiring a distinct activation step to generate enzymatic activity.

APMA-induced conversion of TIMP-free and TIMP-complexed 70 kDa progelatinase to lower

molecular weight forms The originally isolated 70 kDa progelatinase prepa-

ration following a 2 hour APMA treatment underwent a limited autolytic conversion to a 62 kDa species (Chen et al., 1991). Since i t now appears that this preparation was composed of the two forms of the zymogen, TIMP- free and TIMP-complexed, i t was of interest to compare the rate and extent of autolytic conversion of the now separated forms of the zymogen. Equimolar amounts of the purified TIMP-free and TIMP-complexed 70 kDa progelatinase were treated with 2 mM APMA for vary- ing time periods and the appearance of lower molecular weight forms was monitored by SDS-PAGE and silver staining (Fig. 4). The TIMP-free 70 kDa progelatinase is more rapidly converted to lower molecular weight species than the TIMP-complexed form of the progelatinase (Fig. 4A). From the TIMP-free 70 kDa progelatinase, the 62 kDa form is generated as early as 1 hour after APMA treatment along with peptide fragments that migrate at lower apparent molecular weights of 3 0 4 3 kDa. Gelatin substrate zymography indicates that of the lower molecular weight fragments, the 43 kDa species is gelatinolytically active (data not shown).

424 STEFANSSON ET AL

T-F r j T-C

5 11 5 5 0 r 5.00

4.50

4.00

3.50

3.00

2.50

2.00

1 S O

1 .oo 0.50

0.00

4.13

0 2 24 APMA Treatment (Hours)

Fig. 3. Specific peptidolytic activities of TIMP-free and TIMP-complexed 70 kDa progelatinase following activation with APMA. Puri- fied preparations of TIMP-free IT-F, cross-hatched bars) and TIMP- complexed IT-C, stippled bars) 70 kDa progelatinase were activated for 2 and 24 hours a t 25°C with 2 mM APMA in 0.25 ml CAB. The

In comparison, APMA treatment of TIMP-complexed 70 kDa progelatinase results in the appearance of a distinct 62 kDa species only after 6 hours of treatment. Lesser amounts of the 3 0 4 3 kDa peptide fragments also appear with longer (6 to 24 hours) APMA treat- ments of the TIMP-complexed progelatinase. The ra- pidity of the APMA-induced conversion of the TIMP- free 70 kDa progelatinase is illustrated further in Figure 4B. Within 5 minutes of APMA treatment, approximately 50% of the TIMP-free 70 kDa progelatinase is converted to a 67 kDa intermediate form (Fig. 4B, lane 2). After 15 minutes of APMA treatment, the 67 kDa form is the dominant species while the 62 kDa form is just detectable (Fig. 4B, lane 3). At 60 minutes, approximately equal amounts of the 67 kDa and 62 kDa forms are apparent (Fig. 4B, lane 4) which may explain the appearance of a broad, diffuse band in the 1 hour time point of Figure 4A (lane 2) where the 62-67 kDa region was not distinctly resolved. After 60 minutes of APMA treatment there is no detectable 70 kDa progelatinase remaining, as it has been completely converted to the lower molecular weight forms. In contrast, a significant amount of the 70 kDa progelatinase remains throughout the entire 24 hour time course of APMA treatment of the TIMP-complexed 70 kDa progelatinase (Fig. 4A, lanes 7-10). The APMA-induced conversion of the 70 kDa progelatinase appears to be an autolytic process since addition of the zinc-metalloproteinase inhibitor, 1,lO-phenanthroline, prevents the generation of the 62 kDa species and maintains the 70 kDa progelatinase form (Fig. 4B, lane 6).

Rapid activation of TIMP-free progelatinase is prevented by TIMP-2

The rapid activation (Fig. 3) and autolytic conversion (Fig. 4) of TIMP-free 70 kDa progelatinase by APMA

peptidolytic activity was measured fluorometrically using the synthetic substrate DNP-Pro-Leu-Gly-Leu-Trp-Ala-D-Arg-NH, as described in Materials and Methods. The specific activity is expressed as pmoles peptide hydrolyzed per hour per pg of progelatinase.

compared to the slow and incomplete conversion and activation of equal amounts of TIMP-complexed 70 kDa progelatinase suggests a negative regulatory role of the complexed TIMP. To examine this palssibility, TIMP- free 70 kDa progelatinase was treated with APMA before and after purified TIMP-2 was added to form a complex with the 70 kDa progelatinase. The results of SDS-PAGE analysis of a 10 minute APMA activation are shown in Figure 5. TIMP-free 70 kDa progelatinase (Fig. 5, lane 1) when treated with 2 mM APMA, is converted by approximately 50% to the 67 kDa intermediate form (Fig. 5, lane 2). When APMA treatment is carried out with TIMP-free 70 kDa progelatinase that had been preincubated with purified TIMP, the rapid conversion is prevented and the progelatinase is maintained as a 70 kDa species (Fig. 5, lane 3 ) . Treatment of the naturally produced, TIMP-complexed progelatinase with APMA is similarly maintained as a 70 kDa species (Fig. 5, lane 4).

Transformed RSVCEF cultures aire enriched in TIMP-free 70 kDa progelatinase

In order to determine if the occurrence of the TIMP- free form is linked to transformation by RSV, parallel cultures of normal CEF and RSVCEF were analyzed for their levels of TIMP-free proenzyme and TIMP-complexed proenzyme. The total 70 kDa ,progelatinase in the conditioned media from both cultures was purified by gelatin-Sepharaose affinity chromatography and equal amounts of protein from each preparation was analyzed by Mono Q FPLC. Representative tracings of the Mono Q elutions from the RSVCEF and CEF preparations are illustrated in Figure 6. Both normal CEF cultures and transformed RSVCEF cultures yield TIMP-free and TIMP-complexed 70 kDa progelatinase. However, the ratio of TIMP-free proenzyme to TIMP-

TIMP-FREE PROGELATINASE IN RSV TRANSFORMED FIBROBLASTS 425

A. APMA

hcubation T-F T-C

B.

70kDa, 67kDa - 62kDa'

Time(h1 0 1 2 6 24 0 1 2 6 24

kDa 97-

6 7 e

45-

31 -

t 7 0 k D a -62kDa

2+ - c T I M P - 2

1 2 3 4 5 6 7 8 9 1 0

1 2 3

Fig. 4. APMA activation of TIMP-free and TIMP-complexed 70 kDa progelatinase. A 80 ng of TIMP-free 70 kDa progelatinase (T-F, lanes 1-5) and 80 ng of TIMP-complexed 70 kDa progelatinase (T-C, lanes 6-10) in 0.05 ml CAB were left untreated (lanes 1 and 6) or incubated with 2 mM APMA a t 25°C for 1 hour (lanes 2 and 7),2 hours (lanes 3 and 8), 6 hours (lanes 4 and 9) or 24 hours (lanes 5 and 10). All samples were electrophoresed in SDS-10% polyacrylamide gels and silver- stained. Standard proteins were electrophoresed in a parallel lane and

complexed proenzyme is greater in the transformed cultures (Fig. 6). In a series of four separate experiments, parallel cultures of CEF and RSVCEF were analyzed from the relative areas encompassed by peaks 1 and 2 from Mono Q-FPLC chromatograms. RSVCEF cultures yielded a T-F/T-C ratio of 0.96 while CEF cultures yielded a T-F/T-C ratio of 0.35. RSVCEF cultures thus contained approximately equal amounts of TIMP-free and TIMP-complexed proenzyme while normal CEF cultures contained almost three times more TIMP-complexed proenzyme than TIMP-free proenzyme.

DISCUSSION Most mammalian cells that express the MMP-2 72

kDa progelatinase produce the zymogen in cell culture tightly complexed with TIMP-2 (Wilhelm et al., 1989; Stetler-Stevenson et al., 1989a; De Clerck et al., 1989; Goldberg et al., 1989; Howard et al., 1991b). Highly transformed cultures of RSVCEF, however, produce increased amounts of a distinct TIMP-free form of an avian MMP-2 70 kDa progelatinase in addition to the TIMP-progelatinase complex. The TIMP-free 70 kDa progelatinase appears to be a naturally produced, na-

4 5 6

their apparent molecular weight and position in the gel are indicated on the left in kilodaltons (kDa). B Eighty ng of TIMP-free 70 kDa progelatinase in 0.05 ml CAB was left untreated (lane 1) or activated with 2 mM APMA at 25°C for 5 minutes (lane 2), 15 minutes (lane 3), 60 minutes (lane 4), 6 hours (lane 5), 6 hours in the presence of 5 mM 1,lO-phenanthroline (lane 6). All samples were electrophoresed in SDS-8% polyacrylamide gels and silver-stained.

tive form of the progelatinase and not a species generated from TIMP-complexed 70 kDa progelatinase during the purification procedure. The conditions used during gelatin-Sepharose, gel filtration and Mono Q chromatography would not be expected to dissociate the 70 kDa molecule from the tightly associated TIMP-2. Furthermore, the isolated TIMP-complexed 70 kDa progelatinase was resubmitted to these same chromatographic procedures and TIMP-free progelatinase was never observed. The purified TIMP-free progelatinase also does not appear to be an altered form of the zymogen that is deficient in its ability to bind TIMP-2. The isolated TIMP-free progelatinase can bind purified avian TIMP-2 forming a complex indistinguishable from the naturally produced TIMP-complexed 70 kDa progelatinase both electrophoretically (Fig. 2) and functionally (Fig. 5).

The purified TIMP-free progelatinase does not autoactivate upon prolonged incubation in calcium-containing medium and appears to be a stable zymogen. Previ- ously, it had been reported that an isolated human MMP-2 72 kDa progelatinase undergoes spontaneous activation and autolytic conversion when dialyzed into


97kDa-

67kDa- c 7OkDa

45kDa -

31 kDa -

+ TIMP-2 21 kDa-

1 2 3 4

Fig. 5. Treatment of TIMP-free 70 kDa progelatinase with purified TIMP-2 prevents APMA-induced rapid activation. Purified TIMP-free 70 kDa progelatinase (30 ng) was left untreated (lane l), treated with 2 mM APMA at 25°C for 10 minutes (lane 2), or pretreated with purified TIMP (20 ng) for 1 hour at 25°C and then treated with 2 mM APMA at 25°C for 10 minutes (lane 3). Purified TIMP-complexed 70 kDa progelatinase (30 ng) was treated with 2 mM APMA at 25°C for 10 minutes (lane 4). All samples were electrophoresed in SDS-10% polyacrylamide gels and silver-stained.

0 0

I

x 0 - 200 .02 - RSVCEF , G\, y 0 +: 0

- 0 T co 4" .02- -200 % z u 0 0

G',I C E F

0

- 150

I . .01- ,-' - 100

T-F T-C

Fig. 6. Appearance of TIMP-free and TIMP-complexed 70 kDa progelatinase in normal and transformed cultures. 70 kDa progelatinase preparations from parallel cultures of CEF and RSVCEF were purified by gelatin-Sepharose (see Materials and Methods) and equal amounts of protein (50 kg) from each preparation were applied to Mono Q FPLC as described in Figure 1. An A,,, tracing is shown for each preparation. TIMP-free 70 kDa progelatinase (T-F) and TIMP- complexed 70 kDa progelatinase (T-C) are eluted from the column at 150 mM and 170 mM NaCl, respectively.

calcium-containing, neutral buffer (Howard et al., 1991b). However, the strong acidic conditions that were used to isolate the human TIMP-free MMP-2 may have altered the structural conformation of the zymogen al- lowing for the observed autolytic activation. Other studies on isolated and recombinant human TIMP-free MMP-2, in which the TIMP-free zymogen was not subjected to such acidic conditions, demonstrated that pu-

rified TIMP-free 72 kDa progelatinase IS indeed a relatively stable zymogen (Murphy et al., 1992; Okada et al., 1990; Kolenbrock et al., 1991; Fridman et al., 1993; Kleiner et al., 1993).

A distinct difference between the TIMP-free 70 kDa progelatinase and the TIMP-complexed progelatinase is observed upon activation of the two avian zymogen forms by the organomercurial APMA. The TIMP-free progelatinase undergoes rapid and extensive conversion to lower molecular weight forms. A unique 67 kDa intermediate form is observed within 5 minutes while the well established 62 kDa form of MMP-2 appears within 15 minutes after APMA treatment (Fig. 4B). Complete conversion of the TIMP-free 70 kDa species occurs within 30-60 minutes, while the TIMP-complexed 70 kD progelatinase undergoes conversion more slowly and is not fully processed to lower molecular weight forms even after 24 hours of AlPMA treatment (Fig. 4A). Concomitant with the generation of lower molecular forms, an enzyme preparation with high specific activity is obtained from the T1MF'-free form upon APMA treatment, while the TIMP-cornplexed form of progelatinase yields an enzyme preparation with a substantially lower specific activity (Fig. 3). A number of other laboratories studying mammalian 72 kDa progelatinase also have reported that when MMP-2, free of any complexed TIMP-2, is subjected to APMA treatment, a 10 to 20-fold increase in specific gelatinase activity is observed over that of identically treated TIMP-complexed 72 kDa progelatinase (Kolkenbrock et al., 1991; Fridman et al., 1993; Kleirier et al., 1993). Kleiner et al. (1992), employing cross-linking studies where the TIMP-2 is maintained covalently linked to 72 kDa progelatinase, observed only a 50% decrease in specific activity upon APMA treatment compared to a 93% decrease in specific activity upon APMA treatment of the non-cross-linked TIMP-72 kDa progelatinase. These and other studies (Fridman et al., 1993; Kleiner et al., 1993) suggest that when native TIMP- complexed 72 kDa progelatinase is activated, the TIMP-2 is free to interact with the active site of the converted gelatinase, substantially reducing the specific activity of the generated enzyme. It is not yet clear whether the resulting interaction of TIMP-2 with the active site of the gelatinase is via an intermolecular or intramolecular reaction. Whatever the exact mechanism, the final result appears to be that TIMP-2, complexed to its stabilization site on the 72 kDa progelatinase, subsequently reduces the enzymatic activity of the activated enzyme and retards the progressive conversion to lower molecular weight active and inactive forms (Fridman et al., 1992, 1993; Kleiner et al., 1993). The implication of all these studies is that if a TIMP- free form of 72 kDa progelatinase is present when nat- ural activation of the zymogen occurs ait the cell membrane or within the tissue stroma, a rapid generation of a high specific activity gelatinase would occur locally, relatively unimpeded by TIMP-2 inhibition.

The observation in the present study that transformed RSVCEF cultures naturally produce such a TIMP-free progelatinase at higher levels than non- transformed, normal CEF (Fig. 6) may be related to the highly invasive and degradative phlenotype of the RSVCEF cultures. Our laboratory has reported that

427 TIMP-FREE PROGELATINASE IN RSV TRANSFORMED FIBROBLASTS

RSVCEF extensively and rapidly degrade the connective tissue matrix laid down by normal fibroblasts (Fairbairn et al., 1985). Matrix degradation could be inhibited 50-60% by an anti-catalytic antibody to the serine protease, urokinase-type plasminogen activator (uPA; Sullivan and Quigley, 1986). Nearly complete inhibition of matrix degradation was obtained, however, only when an inhibitor of metalloproteinases, 1,lO-phenanthroline, was used in combination with the anti-uPA monoclonal antibody (Sullivan and Quigley, 19861, implying that an active metalloproteinase(s) also is involved catalytically in RSVCEF-mediated matrix degradation. A 70 kDa MMP-2-like metalloproteinase has been shown to be produced by RSVCEF cultures at levels 3 to 5-fold greater than normal CEF cultures (Chen et al., 1991). This enzyme is the only major metalloprotease detected in the RSVCEF cultures and was suggested to be the candidate enzyme responsible for the observed 1,lO-phenanthroline-sensitive, matrix degrading activity (Chen et al., 1991). The present observation that these same RSVCEF cultures produce elevated levels of the more readily activated TIMP-free 70 kDa progelatinase is consistent with the proposed catalytic role of this MMP-2 in the matrix degradation. Also consistent with the observed elevated levels of TIMP-free 70 kDa progelatinase and its proposed catalytic involvement in matrix degradation is the observation that RSVCEF cultures, in contrast to CEF cultures, produce increased levels of the 62 kDa and 43 kDa activated forms of the progelatinase (Chen et al., 1991). It is these activated forms that would be the target of the added 1,lO-phenanthroline which along with the added anti-uPA antibody prevents matrix degradation (Sullivan and Quigley, 1986).

Elevated levels of a TIMP-free MMP-2 in RSVCEF cultures over those of their normal counterpart cells, CEF, indicate that malignant transformation can alter the balance between proteolytic enzymes and their nat- ural inhibitors. A number of reports have suggested that in both the cancerous and normal tissue milieu, it is the stoichiometric balance between proteases and their inhibitors that dictates an invasive, migratory phenotype or a noninvasive, stationary cellular behavior (Matrisian, 1990; Liotta et al., 1991; Mignatti and Rifkin, 1993; Testa and Quigley, 1990). Experimentally altering that balance in favor of inhibitors by the addition of the MMP inhibitors TIMP-1 (Schultz et al., 1988; Alvarez et al., 1990) or TIMP-2 (Albini et al., 1991) substantially reduces tumor cell invasion and metastasis. Overexpression of TIMP-1 (Tsuchiya et al., 1993) and TIMP-2 (De Clerck et al., 1992) in tumor cells by gene transfection also alters the balance in favor of inhibitor and reduces invasive and metastatic behavior. Conversely, altering the balance in favor of proteases by transfection of an MMP cDNA into nonmeta- static cells (Bernhard et al., 1994) or transfection of an anti-sense construct to TIMP-1 cDNA into Swiss 3T3 cells (Khokha et al., 1989) confers metastatic capacity to the transfected cells. In the present study the local proteaselprotease inhibitor balance has been altered by RSV transformation which allows for the elevated production of a TIMP-free MMP-2. How the enhanced expression of the TIMP-free 70 kDa progelatinase is brought about bv RSV transformation is at present un-

resolved. Malignant transformation by the src oncogene clearly enhances the production of the 70 kDa progelatinase at the mRNA level (Aimes et al., 1994) and at the protein level as demonstrated by the use of temperature-sensitive transformation mutants of RSV (Chen et al., 1991). This enhanced production of the progelatinase may titer out the available TIMP-2, or TIMP-2 itself may be downregulated by RSV transformation. Alternatively, the production andlor secretion of TIMP-free 70 kDa progelatinase may proceed through a different cellular pathway than that of the normally produced TIMP-complexed 70 kDa progelatinase and the aberrant pathway may be dominant in the RSV transformed cells. With either mechanism, the overall balance between proteases and protease inhibitors has shifted in favor of the former with the enhanced production in src-transformed cells of a readily activatable, inhibitor-free MMP-2.

ACKNOWLEDGMENTS The authors thank Ms. Betty Draskin for her assis-

tance in preparing the manuscript. The work was supported by a USPHS grant, CA55852, from the National Institutes of Health awarded to J.P.Q. RTA is an Insti- tute for Cell and Developmental Biology Predoctoral Scholar supported by Merck and Company.

LITERATURE CITED Aimes, R.T., French, D.L., and Quigley, J.P. (1994) Cloning of a 72

kDa matrix metalloproteinase (gelatinase) from chicken embryo fibroblasts using gene family PCR Expression of the gelatinase increases upon malignant transformation. Biochem. J., 300t729.

Albini, A., Melchiori, A,, Santi, L., Liotta, L.A., Brown, P.D., and Stetler-Stevenson, W.G. (1991) Tumor cell invasion inhibited by TIMP-2. J . Natl. Cancer Inst., 83t775-779.

Alvarez, D.A., Carmichael, D.F., and De Clerck, Y.A. (1990) Inhibi- tion of collagenolytic activity and metastasis of tumor cells by a recombinant human tissue inhibitor of metalloproteinases. J. Natl. Cancer Inst., 82:589-595.

Andrews, A.T. (1986) Electrophoresis: Theory, Techniques, and Bio- chemical and Clinical Applications. Oxford University Press, New York, p. 79.

Bernhard, E.J., Gruber, S.B., and Muschel, R.J. (1994) Direct evidence linking matrix metalloproteinase 9 (92 kDa gelatinaseicollagenase) to the metastatic phenotype in transformed rat embryo cells. Proc. Natl. Acad. Sci. USA, 91:42934297.

Brown, P.D., Levy, A.T., Margulies, I.M.K., Liotta, L.A., and Stetler- Stevenson, W.G. (1990) Independent expression and cellular processing of M, 72,000 type IV collagenase and interstitial collagenase in human tumorigenic cell lines. Cancer Res., 50:618&6191.

Chen, J.-M., Aimes, R.T., Ward, G.R., Youngleib, G.L., and Quigley, J.P. (1991) Isolation and characterization of a 70-kDa metalloprotease (gelatinase) that is elevated in Rous sarcoma virus-transformed chicken embryo fibroblasts. J . Biol. Chem., 266t5113-5121.

Chin, J.R., Murphy, G., and Werb, 2. (1985) Stromelysin, and connective tissue-degrading metalloendopeptidase secreted by stimulated rabbit synovial fibroblasts in parallel with collagenase. J . Biol. Chem., 260r12367-12376.

Collier, I.E., Wilhelm, S.M., Eisen, A.Z., Marmer, B.L., Grant, G.A., Seltzer, J.L., Kronberger, A., He, C., Bauer, E.A., and Goldberg, G.I. (1988) H-ras oncogene-transformed human bronchial epithelial cells (TBE-1) secrete a single metalloprotease capable of degrading basement membrane collagen. J . Biol. Chem., 263t6579-6587.

De Clerck, Y.A., Yean, T.-D., Ratzkin, B.J., Lu, H.S., and Langley, K.E. (1989) Purification and characterization of two related but distinct metalloproteinase inhibitors secreted by bovine aortic en- dothelial cells. J . Biol. Chem., 264t17445-17453.

De Clerck, Y.A., Perez, N., Shimada, H., Boone, T.C., Langley, K.E., and Taylor, S.M. (1992) Inhibition of invasion and metastasis in cells transfected with an inhibitor of metalloproteinases. Cancer Res., 52t701-708.

Fairbairn, S., Gilbert, R., Ojakian, G., Schwimmer, R., and Quigley, J.P. (1985) The extracellular matrix of normal chick embryo fibro-


blasts: Its effect on transformed chick fibroblasts and its proteolytic degradation by the transformants. J. Cell Biol., lOlt1790-1798.

Fridman, R., Fuerst, T.R., Bird, R.E., Hoyhtya, M., Oelkuct, M., Kraus, S., Komarek, D., Liotta, L.A., Berman, M.L., and Stetler- Stevenson, W.G. (1992) Domain structure of human 72-kDa gelatinaseitype IV collagenase. J. Biol. Chem., 267t15398-15405.

Fridman, R., Bird, R.E., Hoyhtya, M., Oelkuct, M., Komarek, D., Liang, C.-M., Berman, M.L., Liotta, L.A., Stetler-Stevenson, W.G., and Fu- erst, T.R. (1993) Expression of human recombinant 72 kDa gelatinase and tissue inhibitor of metalloproteinase-2 (TIMP-2): Characterization of complex and free enzyme. Biochem. J., 289t411416.

Garbisa, S., Pozzatti, R., Muschel, R.J., Saffiotti, U., Ballin, M., Gold- farb, R.H., Khoury, G., and Liotta, L.A. (1987) Secretion of type IV collagenolytic protease and metastatic phenotype: Induction by transfection with c-Ha-ras hut not c-Ha-ras plus Ad2-Ela. Cancer Res., 47t1523-1528.

Goldbere. G.I., Wilhelm. S.M., Kronberger, A,, Bauer, E.A., Grant, G.A., and Eisen, A.Z. (1986) Human filbroblast collagenase. J. Biol. Chem., 261 t6600-6605.

Goldberg, G.I., Marmer, B.L., Grant, G.A., Eisen, A.Z., Wilhelm, S., and He, C. (1989) Human 72-kilodalton type IV collagenase forms a complex with a tissue inhibitor of metalloproteases designated TIMP-2. Proc. Natl. Acad. Sci. USA, 86t8207-8211.

Howard, E.W., Bullen, E.C., and Banda, M.J. (1991a) Preferential inhibition of 72- and 92-kDa gelatinases by tissue inhibitor of metalloproteinases-2. J. Biol. Chem., 266t13070-13075.

Howard, E.W., Bullen, E.C., and Banda, M.J. (1991b) Regulation of the autoactivation of human 72-kDa progelatinase by tissue inhibitor of metalloproteinases-2. J . Biol. Chem., 266t13064-13069.

Kawai, S., and Hanafusa, H. (1971) The effects ofreciprocal changes in temperature on the transformed state of cells infected with a Rous sarcoma virus mutant. Virology, 46t470479.

Khokha, R., Waterhouse, P., Yagel, S., Lala, P.K., Overall, C.M., Norton, G., and Denhardt, D.T. (1989) Antisense RNA-induced re- duction in murine TlMP levels confers oncogenicity on Swiss 3T3 cells. Science, 243t947-950.

Kleiner, D.E., Jr., Unsworth, E.J., Krutzsch, H.C., and Stetler-Steven- son, W.G. (1992) Higher-order complex formation between the 72- kilodalton type IV collagenase and tissue inhibitor of metalloproteinases-2. Biochemistry, 31t1665-1672.

Kleiner, D.E., Tuuttila, A., Tryggvason, K., and Stetler-Stevenson, W.G. (1993) Stability analysis of latent and active 72-kDa type IV collagenase: The role of tissue inhibitor of metalloproteinases-2 ITIMP-2). Biochemistry, 32t1583-1592.

Kolkenbrock, H., Orgel, D., Hecker-Kia, A., Noack, W., and Ulbrich, N. (1991) The complex between a tissue inhibitor of metalloproteinases (TIMP-2) and 72-kDa progelatinase is a metalloproteinase inhibitor. Eur. J . Biochem., 198t775-781.

Laemmli, U.K. (1970) Cleavage of structural proteins during the as- sembly of the head of bacteriophage T4. Nature, 227t680-685.

Levy, A.T., Cioce, V., Sobel, M.E., Garbisa, S., Grigini, W.F., Liotta, L.A., and Stetler-Stevenson, W.G. (1991) Increased expression ofthe Mr 72,000 type IV collagenase in human colonic adenocarcinoma. Cancer Res., 51t439-444.

Liotta, L.A., Steeg, P.S., and Stetler-Stevenson, W.G. (1991) Cancer metastasis and angiogenesis: An imbalance of positive and negative regulation. Cell, 64t327-336.

Macartney, H.W., and Tschesche, H. (1983) Latent and active human polymorphonuclear leukocyte collagenase. Eur. J. Biochem., 130; 71-78.

Matrisian, L.M. (1990) Metalloproteinases and their inhibitors in matrix remodeling. Trends Genet, 6t121-125.

Mignatti, P., and Rikin, D.B. (1993) Biology and biochemistry of proteinases in tumor invasion. Physiol. Rev., 73t161-195.

Moll, U., Youngleib, G., Rosinski, K., and Quigley, J.P. (1990) Tumor promoter-stimulated Mr 92,000 gelatinase secreted by normal and malignant human cells: Isolation and characterization of the enzyme from HT1080 tumor cells. Cancer Res., 5Ot6162-6170.

Murphy, G., and Docherty, A.J.P. (1992) The matrix metalloproteinases and their inhibitors. Am. J. Respir. Cell. Mol. Biol., 7t120-125.

Murphy, G., Cawston, T.E., and Reynolds, J.J. (1981) An inhibitor of collagenase from human amniotic fluid, Purification, characterization and action on metalloproteinases. Biochem. J. , 195t167-170.

Murphy, G., Willenbrock, F., Ward, R.V., Cockett, M.I., Eaton, D., and Docherty, A.J.P. (1992) The C-terminal domain of 72 kDa gelatinase A is not required for catalysis, hut is essential for membrane

activation and modulates interactions with tissue inhibitors of metalloproteinases. Biochem. J. , 283t637-641.

Nicholson, R., Murphy, G., and Breathnach, R. (1989) Human and rat malignant-tumor-associated mRNAs encode stromelysin-like metalloproteinases. Biochemistry, 28t5195-5203.

Oakley, B.R., Kirsch, D.R., and Morris, R. (1980) A simplified ultra- sensitive sliver stain for detecting proteins in polyacrylamide gels. Anal. Biochem., 105t361-363.

Okada, Y., Morodomi, T., Enghild, J.J., Suzuki, K., Nakanishi, I., Salvesen, G., and Nagase, H. (1990) Matrix metalloproteinase 2 from human rheumatoid synovial fibroblasts. Purification and activation of the precursor and enzymic properties. Eur. J. Biochem., 194t721-730.

Overall, C.M., and Sodek, J. (1990) Concanavalin A produces a matrix-degradative phenotype in human fibroblasts. J. Biol. Chem., 265t21141-21151.

Pavloff, N., Staskus, P.W., Kishnani, N.S., and Hawkes, S.P. (1992) A new inhibitor of metalloproteinases from chicken: ChIMP-3. J. Biol. Chem., 267,17321-17326.

Quantin, B., Murphy, G., and Breathnach, R. (1989) Pump-1 cDNA codes for a protein with characteristics similar to those of classical collagenase family members. Biochemistry, 285327-5334.

Schultz, R.M., Silberman, S., Persky, B., Bajkowski, A.S., and Car- michael, D.F. (1988) Inhibition by human recombinant tissue inhibitor of metalloproteinases of human amnion invasion and lung colo- nization by murine B16-Fl0 melanoma cells. Cancer Res., 48t5539- 5545.

Springman, E.B., Angleton, E.L., Birkedal-Hainsen, H., and Van Wart, H.E. (1990) Multiple modes of activation of latent human fibroblast collagenase: Evidence for the role of C>ys7" active-site zinc complex in latency and a "cysteine switch" mechanism for activation. Proc. Natl. Acad. Sci. USA, 87t36P368.

Stack, M.S., and Gray, R.D. (1989) Comparison of vertebrate collagenase and gelatinase using a new fluorogenic substrate peptide. J. Biol. Chem., 264t4277-4281.

Stetler-Stevenson, W.G., Krutzsch, H.C., and Liotta, L.A. (1989a) Tis- sue inhibitor of metalloproteinase (TIMP-2). J . Biol. Chem., 264: 17374-17378.

Stetler-Stevenson, W.G., Krutzsch, H.C., Wacher, M.P., Margulies, I.M.K., and Liotta, L.A. 11989b) The activation of human type IV collagenase proenzyme. J. Biol. Chem., 264t1353-1356.

Sullivan, L.M., and Quigley, J.P. (1986) An anticatalytic monoclonal antibody to avian plasminogen activator: Its effect on behavior of RSV-transformed chick fibroblasts. Cell, 45:90E~915.

Testa, J.E., and Quigley, J.P. (1990) The role of urokinase-type plasminogen activator in aggressive tumor cell behavior. Cancer Me- tastasis Rev., 9t353-367.

Tsuchiya, Y., Sato, H., Endo, Y., Okada, Y., Mai, M., Sasaki, T., and Seiki, M. (1993) Tissue inhibitor of metalloproteinase 1 is a negative regulator of the metastatic ability of a human gastric cancer cell line, KKLS, in the chick embryo. Cancer Res., 53t1397-1402.

Van Wart, H.E., and Birkedal-Hansen, H. (1990) The cysteine switch A principle of regulation of metalloproteinase activity with potential applicability to the entire matrix metallopiroteinase gene family. Proc. Natl. Acad. Sci. USA, 87.5578-5582.

Ward, R.V., Atkinson, S.J., Slocombe, P.M., Docherty, A.J.P., Rey- nolds, J.J., and Murphy, G. (1991a) Tissue inhibitor of metalloproteinases-2 inhibits the activation of 72 kDa progelatinase by fibroblast membranes. Biochim. Biophys. Acta, 1079t242-246.

Ward, R.V., Hembry, R.M., Reynolds, J .J . , and ]Murphy, G. (1991b) The purification of tissue inhibitor of metalloproteinases-2 from its 72 kDa progelatinase complex. Biochem. J. , 278t179-187.

Werb, Z. (1989) Proteinases and matrix degradation. In: Textbook of Rheumatology. W.N. Kelly, E.D. Harris, Jr . , 13. Ruddy, and C.B. Sledge eds. Saunders, Philadelphia, pp. 300-321..

Wilhelm, S.M., Collier, I.E., Marmer, B.L., Eisen, A.Z., Grant, G.A., and Goldberg, G.I. (1989) SV40-transformed human lung fibroblasts secrete a 92-kDa type IV collagenase which is identical to that secreted by normal human macrophages. J. Biol. Chem., 264: 17213-17221.

Windsor, L.J., Birkedal-Hansen, H., Birkedal-Hansen, B., and Engler, J.A. (1991) An internal cysteine plays a role in the maintenance of the latency of human fibroblast collagenase. Biochemistry, 30.641- 647.

Woessner, J.F., Jr. (1991) Matrix metalloproteinases and their inhibitors in connective tissue remodeling. FASEB J, ,5:2145-2154.