Characterization of the 46-kDa intermediates of matrix metalloproteinase 3 (stromelysin 1) obtained by site-directed mutation of phenylalanine 83.

The precursor of matrix metalloproteinase 3 (MMP-3/ stromelysin 1) is activated in vitro by proteinases or mercurial compounds by stepwise processes which include the initial formation of short-lived intermediates and the subsequent intermolecular cleavage of the His82-Phe83 bond to generate the fully activated mature MMP-3 (Nagase, H., Enghild, J. J., Suzuki, K., and Salvesen, G. (1990) Biochemistry 29, 5783-5789). To study the enzymatic properties of the intermediates we have mutated either His82 or Phe83 to Arg to obtain a stable MMP-3 intermediate. The mutant proteins were expressed in Chinese hamster ovary K-1 cells using a mammalian expression system. The proMMP-3(H82R) mutant was activated by chymotrypsin, elastase, and 4-aminophenylmercuric acetate to the 45-kDa MMP-3 with similar mechanism and kinetics as the wild-type. In contrast, the activation of the proMMP-3(F83R) mutant by proteinases or 4-aminophenylmercuric acetate resulted in 46-kDa forms, which retained 13, 14, or 15 amino acids of the pro-domain depending on the activators. The proteinase-activated MMP-3(F83R) intermediates exhibited little enzymatic activity, but they were partially active after treatment with SH-reacting reagents. These molecules could bind to the tissue inhibitor of metalloproteinases-1 and alpha 2-macroglobulin. However, the SH group of Cys75 of the intermediates was not modified by SH-reagents, indicating that the enzymatic activity generated by SH-reagents resulted from molecular perturbation of the enzyme rather than their interaction with Cys75. When gelatin and transferrin were digested with the 46-kDa intermediates the products were different from those generated by the wild-type MMP-3, suggesting an alteration in substrate specificity. The treatment of proMMP-3 with trypsin resulted in the formation of a 45-kDa MMP-3 with an NH2-terminal Thr85, whose activity and substrate specificity were similar to those of the 46-kDa MMp-3(F83R) obtained from the proMMP-3(F83R) mutant. These observations indicate that the correct processing at the His82-Phe83 bond is critical for expression of the full activity and the specificity of MMP-3.

Like other matrixins, MMP-3 is secreted from cells as an inactive zymogen (Okada et al., 1986(Okada et al., , 1988. ProMMP-3 comprises an NH 2 -terminal propeptide of 82 amino acids, a catalytic domain of 165 amino acids, and a COOH-terminal hemopexin/vitronectin-like domain of 213 amino acids (see Woessner (1991)). Latent proMMPs can be activated in vitro by proteolytic and non-proteolytic pathways (Okada et al., 1988;Nagase et al., 1990). The "cysteine-switch" model has been proposed to explain non-proteolytic activation of proMMPs by SH-reacting reagents (e.g. mercurial compounds, iodoacetamide, N-ethylmaleimide, oxidized glutathione) and chaotropic agents (Springman et al., 1990;VanWart and Birkedal-Hansen, 1990). This model suggests that activation occurs when the cysteinyl residue in the conserved propeptide sequence PRCG[V/N]PD transiently dissociates from the zinc atom at the active site and reacts with SH-reacting reagent, thereby preventing the reassociation of the cysteine-zinc complex. Studies on proMMP-3 activation with organomercurials and proteinases have indicated that the zymogen is processed in a stepwise manner : treatment with APMA results in the initial formation of a 46-kDa intermediate by cleavage of the Glu 68 -Val 69 bond; and proteinases also produce an intermediate by cleaving a peptide bond in a short segment (residues 34 -39), referred to as "bait" region, located near the middle of the pro-domain. The intermediates are then converted into the 45-kDa active form by a bimolecular reaction cleaving the His 82 -Phe 83 bond. Stepwise activation has been also shown for interstitial collagenase (MMP-1), gelatinase B (MMP-9), and matrilysin (MMP-7) Ogata et al., 1992;Crabbe et al., 1992), but the enzymatic properties of the transiently generated intermediates are not known. We have hypothesized that the intermediates generated during activation may bind to endogenous MMP inhibitors such as TIMP-1 and ␣ 2 M, so that this unique activation system provides an additional regulatory mechanism in matrixin activities.
In this study, we aimed to generate a stable intermediate of MMP-3 by mutating residues involved in the final activation site. Such a mutant was obtained by substitution of Phe 83 with Arg in proMMP-3. The intermediates generated from the proMMP-3(F83R) mutant have limited proteolytic activity but they interact with TIMP-1 and ␣ 2 M. In addition, these intermediates exhibit an altered substrate specificity when compared with the fully processed wild-type MMP-3. Our studies suggest that intermolecular processing of the His 82 -Arg 83 bond is critical for the expression of the full enzymatic activity and the specificity of MMP-3.
ProMMP-3 and its mutant cDNA constructs were stably transfected into CHO K-1 cells using Lipofectin according to the manufactures description. Transfected cells were initially grown in glutamine-free Dulbecco's modified Eagles's medium in the presence of dialyzed fetal bovine serum and 25 M methionine sulfoximine for 2 weeks. After colony formation, cells were harvested by treating with trypsin, split into 24-well plates, and grown for additional 2 weeks in the same medium in the presence of increasing concentration of methionine sulfoximine (100 -400 M). Expression of proMMP-3 and its mutants was confirmed by Western blot analysis.
Purification of Recombinant Proteins-Wild-type proMMP-3 and its mutants were purified from the medium conditioned by the relevant stably transfected CHO K-1 cell line by immunoadsorbent chromatography as described by Ito and Nagase (1988). In a typical preparation about 2 mg of proMMP-3 and 1-1.5 mg of proMMP-3 mutants were purified from 1 liter of the conditioned medium of a 4-day culture.
NH 2 -terminal Sequence Analyses-The activated proteins were separated by SDS-PAGE with 7.5% (w/v) total acrylamide under reducing conditions (Bury, 1981), and transferred to a poly(vinylidene difluoride)-Millipore Immobilon transfer membrane as described by Matsudaira (1987). The proteins transferred to the poly(vinylidene difluoride) membranes were located by staining with Coomassie Brilliant Blue R-250, and the bands of interest were excised, placed directly onto a Polybrene-treated glass filter, and sequenced by Applied Biosystem 447 A pulse liquid sequenator with "on-line" 120 A phenylthiohydantoin analysis.
Transverse Urea Gradient-Polyacrylamide Gel Electrophoreses (PAGE)-A three channel peristaltic pump was used to cast three 7.5% polyacrylamide gels simultaneously, containing a continuous 0 -8 M urea gradient in the buffer system (2-amino-2-methyl-1,3-propanediol/ glycine/HCl) as described by Mast et al. (1991). After acrylamide polymerization, the gel was rotated 90°, and a single sample containing approximately 20 g of protein in a total volume of 200 -300 l was loaded evenly across the top of the gel. Electrophoresis was performed at 10 mA (constant current) at 23°C for 3 h. Proteins were stained with Coomassie Brilliant Blue R-250.
Protein Determination-ProMMP-3 concentration was determined by using the bicinchoninic acid (Smith et al., 1985) with bovine serum albumin as standard. The amount of activated MMP-3 was determined by titration with TIMP-1 as described by Nagase (1995).
Purification of the 46-kDa Intermediates-The MMP-3 intermediates were generated by reacting of the proMMP-3(F83R) mutant with chymotrypsin or HNE. After inactivation of the serine proteinase with 2.5 mM DFP, the sample was incubated with ␣ 2 M and applied to a concanavalin A-Sepharose column. This step removed an active 45-kDa MMP-3, a minor component generated during the activation process. The unbound fractions were further purified by a column of GreenA dye matrix equilibrated with TNC buffer. The GreenA-bound fraction containing the 46-kDa protein was eluted by TNC buffer containing 0.5 M NaCl. The purified proteins were concentrated using a Centricon-10 concentrator. The 46-kDa intermediate generated by APMA was used directly.
Characterization of Gelatin or Cm-Tf Digestion Products (Substrate Mapping)-Various forms of the wild-type MMP-3 and the mutant MMP-3 were incubated with either gelatin or Cm-Tf at 37°C for the indicated periods of time. The reaction was stopped by adding 20 mM EDTA and the products were analyzed on 10% SDS-PAGE under reducing conditions.
Binding of the MMP-3 Intermediates to TIMP-1-The ability of intermediates to bind TIMP-1 was examined by applying the samples to a column of Affi-Gel 10 coupled with TIMP-1. The proteins bound to the column were eluted with 5% formic acid (1-ml fractions), neutralized with 2 M Tris-HCl, pH 8.6, and precipitated with 3.3% (w/v) trichloroacetic acid. Precipitated proteins were solubilized in SDS-PAGE sample buffer and analyzed by Western blotting using anti-(human MMP-3) sheep serum.
Alkylation of Cys 75 of MMP-3 Intermediates --The 46-kDa intermediate (1 M) was incubated with 1 mM [ 14 C]iodoacetamide in the presence or absence of 20 mM EDTA at 22°C for 15 min. The reaction was stopped by the addition of 5 mM cysteine. In some cases the samples were dialyzed against TNC buffer at 4°C overnight with several changes of the dialysis buffer. The samples were then subjected to SDS-PAGE under reducing conditions, and stained with Commassie Brilliant Blue R-250. The radioactivity was visualized by the use of the Molecular Dynamics 400 A PhosphorImager.
Kinetic Analysis-The enzyme kinetic studies of MMP-3 and MMP-3 intermediates were carried out using the synthetic substrate NFF-3 as described by Nagase et al. (1994). Various concentrations of the sub-strate (1-75 M) were reacted with 10 nM activated MMP-3 or 20 nM 46-kDa intermediates at 37°C for 1 h. The reaction was terminated by the addition of 1 ml of acetic acid (3.3%, w/v). The K m and k cat values were determined according to Eisenthal and Cornish-Bowden (1974).

Expression of Recombinant
Proteins-ProMMP-3 and the mutant cDNAs were cloned into the expression vector pEE-14, and stably transfected into CHO K-1 cells. The expression levels of recombinant proteins from the selected CHO K-1 cells were 2-3 mg/liter of the conditioned medium after a 4-day culture. The proMMP-3 mutants isolated by immunoadsorbent affinity chromatography were homogenous on SDS-PAGE (see below). Transverse urea gradient-PAGE analyses of the wildtype and two mutant proteins showed a similar unfolding pattern by urea (data not shown), indicating that the mutation did not interfere with the folding and the stability of the protein significantly. The correct folding of the mutant protein is also deduced from the similar processing and the enzymatic activity observed between the wild-type and the mutant proMMP-3 when treated with trypsin (see below).
Activation of ProMMP-3(H82R) and ProMMP-3(F83R) Mutants by APMA-Incubation of the proMMP-3(H82R) mutant with 1.5 mM APMA at 37°C converted the precursor to the 45-kDa active species, and the rate of this reaction was similar to that of the wild-type proMMP-3 (Fig. 1). In contrast, the proMMP-3(F83R) mutant was rapidly converted to a stable 46-kDa intermediate form (Fig. 1B). After a prolonged incubation it was slowly converted to 45 and 28 kDa. While the APMA-activated proMMP-3(H82R) expressed a similar enzymatic activity as the wild-type, the 46-kDa MMP-3(F83R) exhibited only about 20% of the MMP-3 activity. NH 2 -terminal amino acid sequence analysis of the 46-and 45-kDa MMP-3(F83R) revealed that the initial cleavage occurred at the Glu 68 -Val 69 and then slowly at the Thr 85 -Phe 86 bonds (Fig. 2). The inability of the proMMP-3(F83R) mutant to undergo similar processing as the wild-type precursor suggests that the Phe 83 residue at P 1 Ј is a critical residue for the intermolecular processing to generate the fully active MMP-3.
Since the treatment of the proMMP-3(H82R) mutant with proteinases also processed this precursor in a similar manner and kinetics as the wild-type (data not shown), further studies were carried out with the F83R mutant.
Activation of ProMMP-3(F83R) by HNE and Chymotrypsin-To assess if a stable intermediate can be obtained during proteinase activation, proMMP-3(F83R) was treated with HNE or chymotrypsin. HNE initially converted the mutant to an intermediate of 49 kDa and then to a 46-kDa species (Fig. 3B) which was stable for 48 h at 37°C (data not shown). The proteolytic activity of the HNE-activated 46-kDa MMP-3(F83R) intermediate was less than 1% of that of the mature wild-type enzyme. NH 2 -terminal sequence analysis of these two forms indicated that the Val 58 -Thr 59 and the Val 69 -Met 70 bonds were directly cleaved by HNE. Treatment with chymotrypsin resulted in the formation of a 51-, 46-, and 45-kDa species (Fig. 4B). The proteolytic activity of the chymotrypsinactivated 46-kDa MMP-3(F83R) was also less than 1% of that of the mature MMP-3, while the chymotrypsin-activated 45-kDa species exhibited approximately 20% activity (Figs. 3 and 4, Table I). The NH 2 -terminal sequencing indicated that chymotrypsin cleaved at the Phe 53 -Leu 54 bond, and subsequently the Leu 67 -Glu 68 and the Phe 86 -Pro 87 bonds, generating the 46and 45-kDa forms, respectively (Fig. 2).
Activation of ProMMP-3 and ProMMP-3(F83R) Mutant by Trypsin-An activation time course of the wild-type proMMP-3 with 5 g/ml trypsin at 37°C indicated that maximal activity was reached after 5 min, and then the activity fell rapidly to the level of about 25% (Fig. 5A). SDS-PAGE analysis showed that proMMP-3 was rapidly processed from high molecular mass intermediates to 45 kDa (Fig. 5B). The treatment of proMMP-3(F83R) mutant also resulted in a 45-kDa species but with a faster kinetics than the wild-type. The specific activity generated from the mutant precursor was approximately 25% of the fully active MMP-3 (Fig. 5A). The NH 2 terminus of the trypsin activated wild-type and mutant was Thr 85 , indicating that trypsin cleaved the Arg 84 -Thr 85 bond in both cases.
The Role of the NH 2 -terminal Phe 83 on the Substrate Specificity of MMP-3-The proteolytic activities of the APMA-activated 46-kDa MMP-3(F83R) intermediate, the trypsin-activated 45-kDa [Thr 85 ]MMP-3, and the chymotrypsin-activated [Pro 87 ]MMP-3 (the residues in brackets denote the NH 2 -terminal amino acid) were examined for their abilities to digest type-I gelatin and Cm-Tf (Fig. 6, A and B). All three species gave similar digestion patterns against each substrate, but they were different from those of the mature MMP-3 with Phe 83 at the NH 2 terminus. The HNE-and chymotrypsinactivated MMP-3(F83R) intermediates in the presence of APMA gave similar digestion products to those generated by [Thr 85 ]MMP-3 (data not shown). The difference in digestion products was not due to the lower specific activities of these species as shown by the time course study (Fig. 6C). These results indicate that Phe 83 is not only essential to express the full enzymatic activity, but also influences the substrate specificity of MMP-3. A support for this was obtained by trypsin treatment of fully active [Phe 83 ]MMP-3, which resulted in a time-dependent decrease in activity and changes in substrate specificity that were similar to those obtained with [Thr 85 ]MMP-3 (Fig. 7). This suggests that the removal of a dipeptide Phe 83 -Arg 84 from MMP-3 results in alteration of the enzymatic activity and substrate specificity.  (Table II). The similar kinetic parameters for the APMAand the trypsin-activated forms indicates that the presence of a portion of the propeptide is not the determinant factor for the reduced enzymatic activity and the changed substrate specificity of the mutated protein.
Interaction ]MMP-3 could bind to TIMP-1-Affi-Gel 10 and was eluted from the column, the HNEgenerated [Met 70 ]MMP-3 was detected in the unbound fractions (Fig. 8). Like in the case of ␣ 2 M, its binding to TIMP-1 was attained after treatment with SH-reacting reagents.

FIG. 2. Sites cleaved in the propeptide of proMMP-3 and the proMMP-3(F83R) mutant after APMA or proteinases treatment.
Cleavage sites identified by NH 2 -terminal sequence analysis of the products generated by APMA and proteinases are shown by arrows. The His 82 -Phe 83 bond is cleaved by MMP-3 . X indicates that the cleavage did not occur due to the substitution of Phe 83 with Arg. The residue in brackets indicates the NH 2 terminus of the individual MMP-3 species generated after a proteinase or APMA treatment. Molecular masses of individual enzyme species were estimated by SDS-PAGE. The cleavage sites in the boxed bait region are taken form Nagase et al. (1990). The conserved cysteine-switch sequence is boxed by dotted lines. CT, chymotrypsin; T, trypsin. Dotted arrows indicate sites cleaved by autolysis in the presence of APMA.
with [ 14 C]iodoacetamide in the absence and presence of 20 mM EDTA. As shown in Fig. 9, [ 14 C]iodoacetamide was incorporated into the 46-kDa MMP-3(F83R) intermediates only in the presence of EDTA, indicating that the SH-group of Cys 75 in the proteinase-activated intermediates is probably bound to the zinc atom at the active site. However, the treatment of the 46-kDa intermediate with SH-reacting reagents, such as APMA, iodoacetamide, and DTNB resulted in an increase in specific activity (Tables I and II). The increased activity was not due to the modification of Cys 75 by SH-reagents. This was concluded from the following observations. First, incubation of the proteinase-generated 46-kDa intermediates did not react with [ 14 C]iodoacetamide unless the reaction was carried out in the presence of EDTA (Fig. 9A). Second, even after the HNEactivated 46-kDa species were reacted with 1 mM iodoacet-amide (Fig. 9B), or 1 mM APMA and 1 mM iodoacetamide (data not shown) for 2 h at 37°C, the samples, after dialysis of these reagents, incorporated [ 14 C]iodoacetamide only in the presence of EDTA (Fig. 9B), indicating that the SH group of Cys 75 was not modified by the SH reagents. The APMA-activated [Val 69 ]MMP-3 did not react with [ 14 C]iodoacetamide, but it incorporated [ 14 C]iodoacetamide in the presence of EDTA (Fig.  9B), indicating that the SH group of the [Val 69 ]MMP-3 was also not modified by APMA. These results suggest that an increase in activity of the 46-kDa intermediate by APMA or other SH reagent treatment is not due to the reaction with the SH group of Cys 75 , but most likely due to molecular perturbation of the enzyme induced by these reagents.

DISCUSSION
The involvement of matrixins in extracellular matrix degradation is controlled, in part, by the activation of their zymogens and the inhibition of the activated enzymes by their endogenous inhibitors. Promatrixins are activated in vitro by several proteinases, SDS, HOCl, chaotropic agents (see Woessner, 1991), low pH (Davis and Martin, 1990), and elevated temperature (Koklitis et al., 1991). Our previous work, demonstrating that the activation of proMMP-3 occurs in a stepwise manner , led us to investigate enzymatic properties of the intermediates and their ability to interact with endogenous inhibitors, TIMP-1 and ␣ 2 M. A stable intermediate was generated by mutating Phe 83 to Arg, whereas the proMMP-(H82R) mutant was converted to the active 45-kDa species. These results are in agreement with the substrate specificity of MMP-3 reported by Niedzwiecki et al. (1992); i.e. the activity of MMP-3 decreased more then a 100-fold when phenylalanine at the P 1 Ј site was replaced by arginine.

FIG. 5. Activation of proMMP-3 and proMMP-3(F83R) mutant by trypsin.
A, the wild-type proMMP-3 (Ⅺ) and mutant proMMP-3(F83R) (ࡗ) (4 g/ml) were treated with 5 g/ml trypsin at 22°C for the indicated period of time, and the proteolytic activity of the samples was assayed against [ 3 H]Cm-Tf after termination of trypsin activity with 2.5 mM DFP. B, SDS-PAGE analysis of the trypsin-treated proMMP-3 and its mutant. The wild-type and the mutant proMMP-3 were treated with trypsin as above and the samples were processed as described in the legend to Fig. 3B. Lanes 1, 3, 5, and 7 wild-type proMMP-3; lanes 2, 4, 6, and 8, proMMP-3(F83R). low, their binding constants with TIMP-1 could not be determined. Nonetheless, our observations are in good agreement with those by Ward et al. (1991) reporting that TIMP-1 was able to bind to MMP-3 intermediates during activation by APMA. TIMP-2 also interferes with the proMMP-1 processing during activation (DeClerck et al., 1991).
Little activity was detected with the proteinase-activated MMP-3(F83R) intermediates, whereas the APMA-activated intermediate expressed about 15-20% of the full MMP-3 activity against [ 3 H]Cm-Tf and the synthetic substrate. The lack of activity with the former is likely to be due to the retention of the Cys 75 -Zn 2ϩ interaction since Cys 75 did not react with [ 14 C]-iodoacetamide unless the sample was treated with EDTA. This contrasts with our previous studies with the wild-type proMMP-3 whose intermediates underwent autoprocessing, indicating that those intermediates have proteolytic activity . This discrepancy may be explained by the different length of the remaining propeptide. The molecular mass of the major intermediate generated from the wild-type proMMP-3 by proteinases is 53 kDa, whereas that of the stable intermediate form the proMMP-3(F83R) is 46 kDa. It is, therefore, speculated that the longer propeptide moiety of the 53- kDa species may be more favorable than a shorter propeptide to create an open active site, which allows a rapid intermolecular cleavage of the His 82 -Phe 83 bond. A similar open structure may be generated with the initial proteolytic cleavage in the bait region of the proMMP-3(F83R) mutant, but the cleavage of the His 82 -Phe 83 does not take place in this case due to the mutation at the P 1 Ј site. Instead, the action of activator proteinases further pruned the propeptide to the 13-15 amino acid length, which apparently interact with the catalytic site of MMP-3 more tightly, possibly with the aid of the Cys 75 -Zn 2ϩ interaction. However, the 46-kDa MMP-3(F83R) intermediate was converted to a 45-kDa form by trypsin and the trypsintreated form expressed about 20% of the MMP-3 activity (data not shown), suggesting that the removal of a small stretch of the propeptide by cleaving the Arg 84 -Thr 85 bond generates an active MMP-3.
The [Val 69 ]MMP-3(F83R) intermediate generated by APMA expressed partial activity. It is notable, however, that the Cys 75 of this intermediate was not modified by APMA (Fig. 9B). This result was unexpected, because the activation of proMMPs is considered to require the disruption of the Cys-Zn interaction (VanWart and Birkedal-Hansen, 1990;Springman et al., 1990;Chen et al., 1993) and the resulting free SH group would react with APMA. Our observation, however, re-emphasizes the notion that the activation of proMMP-3 by APMA and other SH-reagents occurs not through their binding with Cys 75 but rather by induction of conformational changes in the enzyme molecule (Chen et al., 1993). The proteinase-activated intermediate exhibited little enzymatic activity, but an increase in activity was observed in the presence of APMA, iodoacetamide, or DTNB. Again, under these conditions the SH group of Cys 75 did not react with these agents, but it did in the presence of EDTA. This suggests that the zinc atom at the active site still remains bound to the Cys 75 in the partially activated intermediates, possibly forming a pentadentate coordination. Alternatively, the side chain of Cys 75 may be interacting with another moiety of the enzyme molecule. Although we cannot exclude the possibility that this unexpected finding is due to a subtle conformational change in the enzyme molecule introduced by mutagenesis, this seems unlikely because the structure and the stability of the proMMP-3(F83R) appear to be indistinguishable from those of the wild-type proMMP-3 as demonstrated by a urea denaturation curve and by activation by trypsin (Fig. 5). Furthermore, recent studies by Shapiro et al. (1995) showed that the activation of progelatinase B (proMMP-9) by APMA results in a 83-and 67-kDa species, both of which retained 13 residues of the propeptide including the conserved cysteine. These forms of MMP-9 exhibited enzymatic activity only in the presence of APMA. No activity could be detected after removal of APMA by dialysis, suggesting that APMA does not covalently modify the cysteinyl residue in the propeptide of proMMP-9.
The 46-kDa MMP-3(F83R) intermediates and the [Thr 85 ]-MMP-3 showed similar specific activity on [ 3 H]Cm-Tf and NFF-3, which was considerably lower than that of the mature wild-type MMP-3. Kinetic parameters on NFF-3 and the cleavage patterns of Cm-Tf and gelatin indicate that these enzyme species have similarly altered substrate specificity. Thus, the reduced enzymatic activity associated with the intermediates is not simply due to the presence of a part of the propeptide at the NH 2 -terminal end of the enzyme. Although the altered activity and specificity cannot be readily explained, our observations are analogous to those for interstitial collagenase (MMP-1) and neutrophil collagenase (MMP-8). Both collagenases express the full collagenolytic activities when they possess Phe 81 and Phe 79 at their NH 2 termini, respectively, but [Val 82 ]MMP-1 and [Met 80 ]MMP-8 have only 40 to 20% of the full activities (Suzuki et al., , 1995Knä uper et al., 1993 Reinemer et al., 1994). It is postulated that a small structural and rotational difference may result in different stabilization of the active site or of the transition state, or that the mobile NH 2 -terminal peptide may interfere with the substrate . The changes in both k cat and K m with the 46-kDa intermediate, and the 45-kDa [Thr 85 ]MMP-3 suggest that both the enzymatic efficiency and the interaction with a substrate are altered without a correct positioning of the NH 2 -terminal Phe 83 .
In summary, our studies have shown that intermolecular processing of the His 82 -Phe 83 bond is critical for the expression of full enzymatic activity and specificity of MMP-3. The importance of the correct NH 2 terminus may be related to other members of the matrixin family. All matrixins identified to date have either Phe or Tyr at this position. The activation of proMMP-1 by human and rat mast cell chymases generates [Thr 84 ]MMP-1 and [Val 82 ]MMP-1, respectively (Saarinen et al., 1994;Suzuki et al., 1995), and reduced collagenolytic activity was detected (Suzuki et al., 1995). In the case of MMP-3, not only the reduction in activity, but also the changes in substrate specificity occur when the His 82 -Phe 83 bond is not correctly processed or the Phe 83 is removed from the NH 2 terminus of the mature enzyme. Although it is not known whether altered FIG. 9. Incorporation of [ 14 C]iodoacetamide into Cys 75 of the 46-kDa intermediate. A, proMMP-3(F83R) (1 M) was treated with either HNE (10 g/ml) or chymotrypsin (10 g/ml) at 37°C. After inactivation of the activating proteinases with 2.5 mM DFP the samples were incubated with 1 mM [ 14 C]iodoacetamide in the presence or absence of 20 mM EDTA at 23°C for 15 min. The reactions were terminated by the addition of 5 mM cysteine and the samples were subjected to SDS-PAGE (7.5% acrylamide) under reducing conditions, and the incorporation of [ 14 C]iodoacetamide to the proteins was visualized by fluorography according to Laskey and Mills (1975). Lanes 1 and 2, proMMP-3(F83R); lanes 3 and 4, the mature 45-kDa [Phe 83 ]MMP-3 that lack propeptide; lanes 5 and 6, proMMP-3(F83R) treated with HNE for 2 h; lanes 7 and 8, proMMP-3(F83R) treated with chymotrypsin for 30 min. B, proMMP-3(F83R) was activated either by 1 mM APMA (lanes 1 and 2) or HNE (10 g/ml) at 37°C for 1 h (lanes 3-6), and the samples were then reacted with either TNC buffer (lanes 3 and 4) or 1 mM iodoacetamide (lanes 5 and 6) at 37°C for 2 h. Under the conditions used HNE partially converted proMMP-3(F83R) to the 46-kDa form. After dialysis against TNC buffer the samples were reacted with [ 14 C]iodoacetamide in the presence or absence of 20 mM EDTA. The radioactivity incorporated was visualized using a PhosphorImager. processing of proMMP-3 occurs in vivo, a number of peptide bonds near the activation site can become a target of various proteinases. Such proteinases may arise not only from resident connective tissue cells but also inflammatory cells, plasma, and opportunistic microorganisms under certain pathological conditions.