Hydrolysis of a broad spectrum of extracellular matrix proteins by human macrophage elastase.

Macrophage elastase (ME) was originally named when metal-dependent elastolytic activity was detected in conditioned media of murine macrophages. Subsequent cDNA cloning of the mouse and human enzyme demonstrated that ME is a distinct member of the matrix metalloproteinase family. To date, the catalytic parameters that describe the hydrolysis of elastin by ME have not been quantified and its activity against other matrix proteins have not been described. In this report, we have examined the action of purified recombinant human ME (rHME), produced in Escherichia coli, on elastin and other extracellular matrix proteins. On a molar basis, rHME is approximately 30% as active as human leukocyte elastase in solubilizing elastin. rHME also efficiently degrades alpha1-antitrypsin (alpha1-AT), the primary physiological inhibitor of human leukocyte elastase. In addition, rHME efficiently degrades fibronectin, laminin, entactin, type IV collagen, chondroitan sulfate, and heparan sulfate. These results suggest that HME may be required for macrophages to penetrate basement membranes and remodel injured tissue during inflammation. Moreover, abnormal expression of HME may contribute to destructive processes such as pulmonary emphysema and vascular aneurysm formation. To further understand the specificity of HME, the initial cleavage sites in alpha1-AT have been determined. In addition, the hydrolysis of a series of synthetic peptides with different P'1 residues has been determined. rHME can accept large and small amino acids at the P'1 site, but has a preference for leucine.

macrophage elastase (HME) is most closely related to collagenase-1 (MMP-1) and stromelysin-1 (MMP-3), being 49% identical to each at the amino acid level (3). Moreover, the gene for macrophage elastase, composed of a common 10-exon, 9-intron structure, is on human chromosome 11q22.2/22.3 with at least six other MMPs (4). Despite these similarities, HME possesses certain distinct biological and biochemical properties. Expression appears to be largely restricted to tissue macrophages (4). Upon activation, it not only cleaves its 8-kDa N-terminal domain, but also has a unique propensity to autolytically release its 23-kDa C-terminal domain resulting in a mature active 22-kDa proteinase (3)(4)(5).
Macrophage elastase shares its elastolytic activity (6,7) with only a few MMPs, including the gelatinases (MMP-2 and MMP-9) and, to a lesser extent, matrilysin (8,9). However, despite this characteristic activity, the relative capacity of metalloelastase (human or mouse) to degrade elastin has never been quantified. In addition, the catalytic capacity of metalloelastase against other extracellular matrix components has never been described. We have recently generated mice that lack the capacity to produce macrophage elastase by genetargeting. Macrophages from these mice not only lost 95% of their elastolytic capacity, but were also unable to penetrate a synthetic basement membrane (Matrigel) (10). The purpose of this study was to define the capacity of HME to degrade extracellular matrix components and characterize its substrate specificity. Knowledge of the catalytic properties of macrophage elastase will help define potential roles for this enzyme in biologic processes associated with macrophage activation and metalloelastase expression in vivo. Current evidence suggests that these processes may include pathologic conditions such as atherosclerosis, tumor invasion/angiogenesis, cerebrovascular disease, and pulmonary emphysema, which currently represent the four leading causes of death in the United States.

MATERIALS AND METHODS
Reagents-4-Aminophenylmercuric acetate and heparin-agarose were obtained from Sigma. SP-Sepharose and Sephacryl S-200 were obtained from Pharmacia (Uppsala, Sweden). All other chemicals were reagent grade. Bovine ligament elastin and HLE were obtained from Elastin Products (Owensville, MO). Human ␣ 1 -AT was obtained from Athens Research Products (Athens, GA). Fibronectin, laminin, type IV collagen, chondroitan sulfate, and heparan sulfate were obtained from Collaborative Research Products (Bedford, MA). Entactin was a generous gift of Dr. Robert Senior and 92-kDa gelatinase and interstitial collagenase were kindly provided by Dr. Howard Welgus, both of the Washington University School of Medicine, St. Louis, MO.
Bacterial Expression and Purification of Recombinant HME-Fulllength HME cDNA was ligated as an NdeI/BamHI cassette into the pET 5b vector which permitted translation in the proper reading frame beginning with the HME initiation methionine. pET 5b alone (control plasmid) and pET 5b/HME were transformed into E. coli strain BL21(DE3) (Novagen Inc., Madison, WI). Single colonies of E. coli (Ϫ/ϩ rHME) from Lennox Broth (LB)/agar plates with 20 g/ml ampicillin were grown to log phase in 1 liter of LB media (with ampicillin) in a shaking incubator at 37°C. To induce T7 RNA polymerase and drive high level expression of rHME, isopropyl-␤-D-thiogalactopyranoside was added to a final concentration of 0.4 mM. Cells were maintained in culture for an additional 4 h. Cell pellets were resuspended in 10 ml of 50 mM Tris, pH 8.0, with 10 mM CaCl 2 and 150 mM NaCl and then lysed by sonication. After centrifugation, the rHME was localized in the pellet. It was solubilized with 40 ml of 8 M urea containing 50 mM Tris, pH 8.0, with 10 mM CaCl 2 and 30 mM NaCl continuously stirred for 2 h at 4°C. Following centrifugation, the supernatant containing soluble rHME was collected.
The extracts from several 1-liter preparations were dialyzed against 3 M urea in the Tris-containing solution and the lysates were applied sequentially to SP-Sepharose ion exchange chromatography, high resolution Sephacryl S-200 column gel filtration (column 2.5 ϫ 90 cm, bed volume 400 ml, flow rate 30 ml/h), and heparin-agarose affinity/ion exchange chromatography. Following these purification procedures, polyacrylamide gel electrophoresis (SDS-PAGE) and both Coomassie and silver staining demonstrated rHME without any visible contaminating proteins. The identity of this protein as HME was confirmed by Western blotting. N-terminal amino acid sequence analysis of the final product was determined by Edman degradation.
The bicinchoninic acid protein assay (Pierce Chemical Co.) and a TIMP inhibition assay were used to determine the concentration of rHME. The latter method involved preincubation of known concentrations of TIMP with fixed amounts of rHME followed by addition of Ac-Pro-Leu-Gly-S-Leu-Gly-OEt. Hydrolysis of this thiopeptolide substrate (Bachem Bioscience, King of Prussia, PA) by metalloproteinases was determined as described previously (11). The concentration of rHME was further confirmed by pulse liquid sequencing performed on an Applied Biosystems model 473A sequencer equipped with the model 610A analysis software. The initial yield of the protein was extrapolated from the repetitive yield calculations for the first 10 cycles of sequencing. The average repetitive yield during the run was 96.4%. Comparison of the TIMP inhibition assay and pulse liquid sequencing results demonstrated that greater than 90% of the purified rHME was catalytically active. Control plasmids subjected to the same purification scheme were catalytically inactive against all substrates tested.
The catalytic domain of matrilysin was expressed and purified to homogeneity in bacteria using the same techniques described for metalloelastase. Matrilysin expressed in bacteria had equal catalytic activity to matrilysin expressed in eukaryotic cells (12).
Degradation of Basement Membrane Components by MMPs-To qualitatively compare the degradative capacity of rHME to either matrilysin or 92-kDa gelatinase, the basement membrane components fibronectin, laminin, entactin, chondroitan sulfate, heparan sulfate, and type IV collagen were incubated with each enzyme and cleavage products resolved with polyacrylamide gel electrophoresis. Specifically, fibronectin and laminin (5 g) were each incubated at 37°C for 18 h with rHME or matrilysin in a final reaction mixture volume of 30 l. Entactin was incubated under similar conditions for 2 h. Native and pepsinized type IV collagen (20 g) were incubated with rHME or 92-kDa gelatinase at 25°C for 18 h. Two different concentrations (46 and 460 nM for type IV collagen; 23 and 230 nM for all other substrates) of each enzyme were utilized. Chondroitan sulfate and heparan sulfate (10 g) were incubated with rHME or matrilysin at 37°C for 18 h and two different concentrations (0.23 and 2.3 M) of each enzyme were utilized. The reaction mixtures were stopped with SDS sample buffer containing dithiothreitol, boiled, and then subjected to polyacrylamide gel electrophoresis. 10% chondroitan sulfate and type IV collagen, 8% fibronectin and entactin, and 6% laminin slab gels were stained with 1% Coomassie Brilliant Blue.
To further estimate the concentration of rHME or matrilysin required to produce 50% substrate cleavage, varying concentrations of both metalloproteinases were incubated with substrates as otherwise described above. For entactin, duration of incubation was reduced to 15 min and reaction temperature was reduced to 25°C. Samples were resolved by SDS-polyacrylamide gel electrophoresis and stained with 1% Coomassie Brilliant Blue. Subsequently, bands representing substrate alone as well as residual non-cleaved substrate were scanned with an Apple Color One Scanner to permit quantitative comparison. Degradation was interpolated linearly to estimate the concentration of metalloproteinase required to produce 50% substrate cleavage.
Elastin degradation was quantified by measuring solubilization of insoluble 3 H-elastin at 37°C and pH 7.5. Data are expressed as micrograms of elastin degraded from at least three separate experiments each performed in duplicate. These calculations are based on measurements of 3 H-elastin counts/min, corrected for buffer blanks. The radiolabeled elastin used for these studies had a specific activity of 1900 cpm/g.
Degradation of ␣ 1 -AT by MMPs-␣ 1 -AT (10 g) was incubated at 37°C for 18 h with rHME or interstitial collagenase in a final reaction mixture volume of 60 l with enzyme concentrations of 10 and 100 nM. The reactions were quenched by addition of SDS sample buffer containing dithiothreitol and the mixtures were boiled and subjected to polyacrylamide gel electrophoresis. 8% Slab gels were stained with 1% Coomassie Brilliant Blue. To quantify the degradation of ␣ 1 -AT, gels were scanned using a Gilford spectrophotometer set at 560 nm equipped with a linear gel scanning device (12).
The effect of the cleavage of ␣ 1 -AT, by rHME, on its ability to inhibit the elastase activity of HLE was examined. Specifically, 3 H-labeled insoluble elastin was incubated for 24 h at 37°C with 0.4 g of HLE alone, 0.4 g of HLE preincubated with molar excess (40 g) of intact ␣ 1 -AT or ␣ 1 -AT previously incubated with rHME for 18 h at 37°C.
N-terminal Amino Acid Sequence Analysis of ␣ 1 -AT Cleavage Products-Amino acid sequence analysis was performed on ␣ 1 -AT degraded by purified rHME. 5 g of ␣ 1 -AT was incubated with 44 ng of rHME for 10 min at 37°C and subsequently resolved by 12% SDS-polyacrylamide gel electrophoresis. To resolve lower molecular weight degradation products, incubation time was extended to 30 min and a 4 -20% gradient gel (Bio-Rad) was utilized. Proteins were transferred to a Problott membrane (Applied Biosystems) and visualized with 0.1% Coomassie Blue, excised, and sequenced by Edman degradation using an Applied Biosystems 473 sequenator.
PЈ 1 Substrate Specificity of Cleavage by HME-All peptides were synthesized as described in Ref. 13. The initial rates of hydrolysis (v i ) of the peptides were measured fluorometrically at 37°C using excitation at 278 nm and emission at 358 nm (14) in a Perkin-Elmer LS-50B Luminescence Spectrometer. The assays were conducted in 50 mM Tris-Cl, pH 7.5, 200 mM NaCl, 10 mM CaCl 2 , 2.5% Me 2 SO, and 0.005% Brij-35. Measurements of v i were made at 50, 100, 150, 200, and 250 M substrate. The data were then fit to the equation . For some substrates K m values could not be obtained due to limited substrate solubility. In these cases, only k cat /K m was determined by measuring v i at % Ͻ Ͻ K m . Verification of the proper cleavage products was made by reversed-phase high performance liquid chromatography.

RESULTS
Expression and Purification of Recombinant HME-rHME was expressed in Escherichia coli and, after cell lysis, solubilized in urea, dialyzed, and sequentially subjected to SP-Sepharose ion exchange chromatography, Sephacryl S-200 column gel filtration, and heparin-agarose affinity/ion exchange chromatography (Fig. 1). After the first chromatographic step, all of the rHME migrated with the apparent molecular mass of 22 kDa corresponding to the mature processed form. At each purification step, equal amounts of protein were incubated with insoluble 3 H-elastin. The specific activity increased greater than 250-fold after the final purification step (Table I). Nterminal sequence analysis was performed on the final material to confirm its purity and identify the N terminus of the active form of recombinant HME. A single Phe-Arg-Glu sequence was identified that corresponds to cleavage at the His 99 -Phe 100 bond, just C-terminal to the conserved cysteine switch motif. It should be noted, however, that the N terminus of mature native HME has not been identified to date. Finally, this expression and purification procedure gave ϳ500 g of purified rHME from 1 liter of E. coli.
Degradation of Basement Membrane Components by HME-The basement membrane proteins fibronectin, entactin, laminin, chondroitan sulfate, and heparan sulfate were incubated with rHME or matrilysin at 37°C as described under "Materials and Methods." Native and pepsinized type IV collagen was also incubated with rHME or 92-kDa gelatinase under similar conditions at 25°C. A lower incubation temperature was se-lected to avoid conditions that would favor denaturation of the collagen molecule in solution. Cleaved products were resolved by SDS-polyacrylamide gel electrophoresis and stained with 1% Coomassie Brilliant Blue (Fig. 2). The amounts of rHME or matrilysin required to produce 50% cleavage of fibronectin, entactin, laminin, and chondroitan sulfate are summarized in Table II. rHME and matrilysin both effectively cleave entactin. rHME, however, is more potent than matrilysin in cleaving fibronectin, laminin, and chondroitan sulfate. rHME and matrilysin also efficiently degrade the proteoglycan heparan sulfate (data not shown). Finally, rHME is comparable to 92-kDa gelatinase in its ability to degrade pepsinized type IV collagen. Neither enzyme, however, effectively cleaved native type IV collagen (data not shown).
Elastolytic Activity of rHME-The elastolytic activity of rHME was measured by quantifying solubilization of 3 H-elastin as measured by the release of 3 H into the supernatant. Activity is expressed as micrograms of elastin degraded. Because elastin is an insoluble substrate, classic Michaelis-Menten kinetics could not be applied in comparing rHME and HLE. Under conditions of substrate excess with small amounts of rHME or HLE, the amount of elastin degraded was linearly related to enzyme concentration (Fig. 3). The specific activity of HME as calculated from the first three data points is 33 g of elastin degraded/mg of enzyme/min. The specific activity of rHME against elastin was one-third that of HLE (Fig. 4). Incubation of rHME with insoluble elastin was extended for up to 96 h and catalysis continued to proceed in a linear manner (data not shown), demonstrating that the 22 kDa species is stable under the conditions tested.
Cleavage of ␣ 1 -AT by rHME-Human ␣ 1 -AT was incubated with rHME or interstitial collagenase at concentrations of 10 and 100 nM for 18 h at 37°C. The reaction products were resolved by SDS-polyacrylamide gel electrophoresis. As shown in Fig. 5, rHME completely converted 54-kDa ␣ 1 -AT into a stable 50-kDa species at both concentrations used, whereas interstitial collagenase demonstrated incomplete conversion even at 100 nM. Incubation of 10 g of ␣ 1 -AT with varying concentrations of rHME at 37°C for 30 min yielded 50% degradation of substrate at a concentration of 2 nM. Under similar conditions, interstitial collagenase and matrilysin yielded 50% degradation of substrate at concentrations of 75 and 47 nM, respectively. Preincubation of ␣ 1 -AT with rHME abolished its ability to inhibit the HLE-mediated degradation of elastin (data not shown). Thus, the stable 50-kDa form of ␣ 1 -AT is functionally inactive.
N-terminal Amino Acid Sequence Analysis of ␣ 1 -AT Cleavage Products-Previous studies of the hydrolysis of ␣ 1 -AT by MMPs have shown cleavage sites within the reactive loop (12,15,16). To determine initial HME cleavage sites in ␣ 1 -AT, the reaction was carried out for 10 min at 37°C and the products FIG. 1. Expression and purification of rHME. pET 5b vector (control) and pET 5b/rHME were expressed in E. coli, solubilized in urea, and purified using the chromatographic techniques outlined under "Materials and Methods." Top panel, 8% SDS-polyacrylamide gel, Coomassie stained, demonstrates the sequential purification to a single band. (Note the prominent band in the vector control migrating ϳ30 kDa representing translation of pET5 in the absence of inset.) In addition, no other proteins were identified when silver staining was utilized (data not shown). Bottom panel, Western analysis, with a peptide antibody specific for rHME, demonstrates the 54-kDa pro-enzyme, 45-kDa N-terminal active form and several intermediate forms in urea. After the first chromatographic step, all rHME migrated at 22 kDa, the mature processed form.  analyzed by SDS-PAGE (Fig. 6). Cleavage products with molecular masses of 50, 29, 25, and 4 kDa were identified, corresponding to two scissions of intact ␣ 1 -AT to yield 50 and 4 kDa or 29 and 25 kDa products, respectively. N-terminal amino acid sequence analysis identified the two major cleavage sites: Phe 352 -Leu 353 , generating the 50-and 4-kDa degradation products and Glu 199 -Val 200 , generating the 29-and 25-kDa fragments. As noted above, however, upon prolonged incubation the only stable degradation product was the 50-kDa species, corresponding to cleavage at Phe 352 -Leu 353 within the reactive loop.
Peptide Substrate Specificity of Cleavage by HME-The most important subsite in determining MMP substrate specificity is PЈ 1 . Thus, the relative rates of hydrolysis of nine octapeptides differing only in the residues in subsite PЈ 1 have been quanti-fied for HME (Table III). The preference follows the order: Leu Ͼ Ͼ Ala Ͼ Lys Ͼ Phe Ͼ Tyr Ͼ Trp Ͼ Arg Ͼ Ser Ͼ Glu. In general, HME tolerates a variety of large and small residues at the PЈ 1 position. This is consistent with the expectation that HME should have a deep SЈ 1 pocket based on its predicted amino acid sequence (see "Discussion"). DISCUSSION Elastin is a highly cross-linked and hydrophobic insoluble extracellular matrix protein that imparts elastic recoil to a FIG. 3. Degradation of 3 H-elastin by rHME. rHME was incubated with 10 l of 3 H-elastin for 4 or 8 h at 37°C at varying enzyme concentrations. Solubilized products were counted in a liquid scintillation counter. Elastolytic capacity was initially quantified as amount of 3 H released into the supernatant (in cpm). As 1 g of elastin corresponds to 1900 cpm released, elastolytic activity was expressed as micrograms of elastin degraded. Calculation of specific activity from the first three data points reveals 33 g of elastin degraded per mg of enzyme/min.
FIG. 4. Elastolytic activity of rHME relative to HLE. Varying concentrations of rHME or HLE were incubated with 3 H-insoluble elastin for 8 h. Elastolytic capacity was expressed as micrograms of elastin degraded. Comparisons are best made using small concentrations of enzymes with linear degradation and substrate excess. Note that rHME is approximately one-third as effective in solubilizing elastin as HLE at any given concentration; the specific activity of HLE was 88 g of elastin degraded per mg of enzyme/min compared with 33 g of elastin degraded per mg of enzyme/min for HME in these experiments.
FIG. 5. Degradation of ␣ 1 -AT by rHME and interstitial collagenase. 10 g of ␣ 1 -AT was incubated with 10 and 100 nM rHME or interstitial collagenase at 37°C for 18 h. Samples were applied to an 8% SDS-polyacrylamide gel and Coomassie stained. Results indicate complete conversion of ␣ 1 -AT to a stable 50-kDa species by rHME at both concentrations whereas interstitial collagenase led to incomplete degradation. Arrows indicate undigested ␣ 1 -AT at 54 kDa and the major stable degradation product at 50 kDa. Incubation of 10 g of ␣ 1 -AT with varying concentrations of rHME at 37°C for 30 min yields 50% degradation of substrate at a concentration of 2 nM whereas interstitial collagenase or matrilysin, under similar conditions, produces 50% degradation of substrate at 75 and 47 nM, respectively (not shown).
FIG. 6. Sites of cleavage of ␣ 1 -AT by rHME. 5 g of ␣ 1 -AT was incubated with 44 ng of rHME at 37°C for 30 min and initial degradation products were resolved with 12% SDS-polyacrylamide gel electrophoresis. Proteins were transferred to a Problott membrane, visualized with 0.1% Coomassie Blue, excised and sequenced utilizing Edman degradation as described under "Materials and Methods." Above, the first three protein bands share the same N-terminal sequence. The lower two molecular weight products reveal cleavage sites at Glu 199 -Val 200 and Phe 352 -Leu 353 , which are indicated below, in the schematic representation of the ␣ 1 -AT protein.
variety of tissues including musculoskeletal ligaments, arterial vessels, and lung parenchyma (18). These properties contribute to its extreme stability and resistance to proteolysis by all but a limited number of proteinases. When compared in parallel assays, rHME has approximately one-third the elastolytic capacity of HLE. A similar difference in elastolytic capability between recombinant 92-kDa gelatinase and HLE has been observed (9). It is probable, however, that under most circumstances HLE remains within the neutrophil, degrading internalized foreign material, and that only inadvertent release of HLE from the cell, during "frustrated phagocytosis" or death of the short-lived neutrophil, may cause tissue destruction. On the other hand, in response to mediators of inflammation, MMPs, including HME, are characteristically secreted into the extracellular space where they modulate matrix remodeling. With sustained accumulation of macrophages and neutrophils, e.g. in chronic inflammation of the lung induced by cigarette smoking, HME and HLE expression may be poorly regulated with consequent elastin destruction, ultimately producing the distinctive changes of pulmonary emphysema.
Several circulating and locally secreted inhibitors of HLE have been identified. The primary physiological inhibitor is ␣ 1 -AT which forms a stable complex with the enzyme (19). A variety of MMPs are capable of degrading ␣ 1 -AT (12,15,16) and abolishing its ability to inhibit HLE. MMPs, therefore, may significantly modulate the physiological role of ␣ 1 -AT. We now show that HME, in particular, is an order of magnitude more active than matrilysin, previously shown to be the most potent MMP capable of degrading ␣ 1 -AT (12). In addition, two initial cleavages were identified, rather than the single degradation product reported for other MMPs. The first is a cleavage at the Glu 199 -Val 200 bond. Stromelysin is the only MMP which makes Glu-Val cleavages (2). The second site is within the active site loop at the Phe 352 -Leu 353 bond. This is the same site at which interstitial collagenase and the 92-kDa gelatinase cleave ␣ 1 -AT (20,21). Surprisingly, however, MME, the murine orthologue to HME, cleaves at the Pro 357 -Met 358 bond instead. Of note, when ␣ 1 -AT is inactivated by MME, a potent chemotactic factor for neutrophils is generated (15). It is therefore possible that, in addition to HME's inherent elastolytic capacity, elastin degradation may be further augmented by inactivating the primary inhibitor of HLE and by attracting additional neutrophils to sites of inflammation. HLE, conversely, can cleave and inactivate TIMP (19), releasing HME from inhibition. HME is able to mediate degradation of several extracellular matrix components. HME was unable to cleave interstitial collagens and was minimally active against denatured collagen or gelatin (data not shown). It is, however, at least as active as any other MMP against the matrix proteins that comprise basement membranes and readily degrades fibronectin, laminin, entactin, chondroitan sulfate, and heparan sulfate. HME also degrades pepsinized type IV collagen as efficiently as the 92-kDa gelatinase (17). These cleavages are likely limited to the non-helical collagen domains. It is interesting to note that type IV collagen is relatively resistant to all inflammatory cell proteinases when compared with other basement membrane proteins. An attractive hypothesis would suggest that inflammatory cells release proteinases which dissolve the more susceptible basement membrane substrates leaving the type IV collagen backbone intact. Cells may then migrate through the disrupted basement membrane while the structural architecture is maintained by type IV collagen which would then provide a lattice for renewed biosynthesis of matrix components.
To further probe the substrate specificity of HME, its action on a series of synthetic peptides with different PЈ 1 residues has been examined. These peptides were chosen based on knowledge gained from structure-function studies performed on other MMPs. The structures for the catalytic domains of several members of the MMP family have recently become available including human fibroblast collagenase (MMP-1) (22-24), human stromelysin-1 (MMP-3) (25,26), human matrilysin (MMP-7) (27), and human neutrophil collagenase (MMP-8) (28 -30). These structures revealed that the SЈ 1 subsite is the most well defined pocket in these MMPs and consists of a hydrophobic pocket which varies greatly in its depth. Mutational analyses of the SЈ 1 pocket (29) revealed that residue 214 (numbering according to Browner (27)) which lies at the bottom of the SЈ 1 pocket is critical in determining its shape. For MMP-1 and MMP-7, which have an Arg 214 and a Tyr 214 , respectively, these residues point into the SЈ 1 pocket, thus forming a shallow pocket. MMP-3 and MMP-8 each have a Leu 214 residue which points away from the pocket, thus allowing a deep SЈ 1 pocket to exist. These SЈ 1 pocket "types" have been corroborated by substrate specificity studies using libraries of synthetic peptides (30 -32). MMP-1 and MMP-7, which have shallow SЈ 1 pockets, prefer small hydrophobic amino acids at the PЈ 1 position. In contrast, MMP-3 and MMP-8, which have deep SЈ 1 pockets, can accommodate large and small PЈ 1 amino acids with similar efficiency.
Although the structure for macrophage elastase has not yet been determined, the prediction based on the Leu 214 which this enzyme possesses suggests that there should be a deep SЈ 1 pocket which can accept large and small PЈ 1 substrates. The data in Table III support this hypothesis. HME is able to tolerate large hydrophobic PЈ 1 residues, such as Trp and Phe, much better than MMP-1, which contains a shallow SЈ 1 pocket. Overall, HME had a preference for Leu at the PЈ 1 position and, in fact, rHME cleaved the reactive loop of ␣ 1 -AT at Phe 352 -Leu 353 . Of interest, HME is the only MMP known that can accommodate Arg at PЈ 1 . This finding could help in the design and synthesis of a selective HME inhibitor. This may be useful in conditions with aberrant inflammatory responses which may lead to macrophage-mediated tissue destruction. Dnp-Arg-Pro-Leu-Ala-Leu-Trp-Arg-Ser-NH 2