Myeloperoxidase Inactivates TIMP-1 by Oxidizing Its N-terminal Cysteine Residue

An imbalance between the proteolytic activity of matrix metalloproteinases (MMPs) and the activity of tissue inhibitors of metalloproteinases (TIMPs) is implicated in tissue injury during inflammation. The N-terminal cysteine of TIMP-1 plays a key role in the inhibitory activity of the protein because it coordinates the essential catalytic Zn2+ of the MMP, preventing the metal ion from functioning. An important mechanism for controlling the interaction of TIMPs with MMPs might involve hypochlorous acid (HOCl), a potent oxidant produced by the myeloperoxidase (MPO) system of phagocytes. Here, we show that HOCl generated by the MPO-H2O2-chloride system inactivates TIMP-1 by oxidizing its N-terminal cysteine. The product is a novel 2-oxo acid. Liquid chromatography-mass spectrometry and tandem mass spectrometry analyses demonstrated that methionine and N-terminal cysteine residues were rapidly oxidized by MPO-derived HOCl but only oxidation of the N-terminal cysteine of TIMP-1 correlated well with loss of inhibitory activity. Importantly, we detected the signature 2-oxo-acid N-terminal peptide in tryptic digests of bronchoalveolar lavage fluid from patients with acute respiratory distress syndrome, demonstrating that TIMP-1 oxidation occurs in vivo. Loss of the N-terminal amino group and disulfide structure are crucial for preventing TIMP-1 from inhibiting MMPs. Our findings suggest that pericellular production of HOCl by phagocytes is a pathogenic mechanism for impairing TIMP-1 activity during inflammation.

The molecular masses of the four known human TIMPs (TIMP-1, -2, -3, and -4) range from 21 to 34 kDa (10). All contain 12 conserved cysteine residues whose six disulfide bonds fold the protein into N-and C-terminal domains. The N-terminal domain, which has three disulfide bonds, is highly conserved among TIMP family members and across species and is crucial for the inhibitory activity of TIMP-1 (3,(11)(12)(13).
TIMPs inhibit the proteolytic activity of MMPs by forming a stable, noncovalent 1:1 stoichiometric complex (3,11,14). The N-terminal residue of TIMP-1 (Cys 1 -Thr 2 -Cys 3 -Val 4 ) binds to the active site cleft of MMP-3 (15). Moreover, the N-terminal ␣-amino and carbonyl groups of the Cys 1 residue coordinate the essential catalytic Zn 2ϩ of the metalloproteinase. Importantly, all TIMPs have a conserved N-terminal sequence, CT/SCXPXHPQ, that contains Cys 1 (3). These findings indicate that Cys 1 plays a key role in inhibiting MMP activity.
Although the structure and inhibitory mechanism of TIMP-1 have been investigated intensively, less is known about the regulation of its inhibitory activity. Neutrophil elastase and Porphyromonas gingivalis (a bacterium involved in periodontal disease) inactivate TIMP-1 by proteolytic degradation in vitro (16,17). High concentrations of reagent hypochlorous acid (HOCl) and peroxynitrite also inactivate TIMP-1 in vitro (18,19) but by undefined molecular mechanisms.
At sites of inflammation, HOCl is generated by myeloperoxidase (MPO), an enzyme secreted by activated phagocytes. MPO converts hydrogen peroxide (H 2 O 2 ), another phagocyte product, to HOCl under physiological concentrations of Cl Ϫ (20) as shown in the equation Cl Ϫ ϩ H 2 O 2 ϩ H ϩ 3 HOCl ϩ H 2 O. HOCl is a potent antimicrobial oxidant that plays an important role in host defense (21), but it can also modify host proteins by oxidizing amino acid side chains, N-terminal amino groups, and disulfides (22)(23)(24).
In the current studies, we determined whether oxidants derived from phagocytes impair the ability of TIMP-1 to regulate MMP. We demonstrated that HOCl generated by MPO, but not H 2 O 2 alone, oxidizes Cys 1 and Met residues of TIMP-1. Loss of TIMP-1 inhibitory activity strongly correlated with Cys 1 oxidation and the formation of a novel 2-oxo acid product. Furthermore, we identified this novel oxidation product in bronchoalveolar lavage (BAL) fluid of patients with acute respiratory distress syndrome (ARDS). Our observations suggest an in vivo oxidative molecular mechanism by which MPO can prevent TIMP-1 from inhibiting MMPs.

Materials
Sodium hypochlorite, hydrogen peroxide (H 2 O 2 ), trifluoroacetic acid, and HPLC-grade acetonitrile (CH 3 CN) were obtained from Fisher Scientific. Human matrilysin (MMP-7) and pro-gelatinase B (pro-MMP-9) were obtained from Calbiochem. Unless otherwise indicated, all other materials were purchased from Sigma-Aldrich. MPO was isolated by lectin affinity and size exclusion chromatographies from human neutrophils (25,26) and stored at Ϫ20°C. Purified enzyme had an A 430 /A 280 ratio of 0.8 and was apparently homogeneous on SDS-PAGE analysis. Its concentration was determined spectrophotometrically (⑀ 430 ϭ 0.17 M Ϫ1 cm Ϫ1 ) (27).
Proteolytic Digestion of Proteins-Native or oxidized TIMP-1 was reduced with 4 mM dithiothreitol at 70 -80°C in 50 mM ammonium bicarbonate and 10% acetonitrile for 20 min. The reduced protein was alkylated with 10 mM iodoacetamide at room temperature for 15 min. Alkylated protein was digested overnight at 37°C with sequencing grade modified trypsin (Promega, Madison, WI) at a ratio of 25:1 (w/w) protein/trypsin or endoproteinase Glu-C (Roche Applied Science) at a ratio of 20:1 (w/w) protein/Glu-C. Digestion was halted by acidification (pH 2-3) with 10% (v/v) trifluoroacetic acid.
BAL Fluid-All human studies were approved by the Human Studies Committee of the University of Washington. BAL was performed on patients with ARDS and normal control subjects as described (29) using a fiberoptic bronchoscope. The BAL fluid was stored at Ϫ70°C and thawed at 4°C before analysis. The six most abundant proteins (albumin, IgG, IgA, transferrin, haptoglobin, and antitrypsin) in the BAL sample were depleted using a multiple affinity removal spin cartridge (Agilent Technologies, Wilmington, DE). All depletion procedures were performed at room temperature according to the manufacturer's instructions. The flow-through fraction was separated on a macroporous reverse-phase C18 (mRP-C18) high recovery protein column (Agilent Technologies) using an Agilent 1100 HPLC system. The fraction containing TIMP-1 was reduced, alkylated, and digested overnight at 37°C with sequencing grade modified trypsin. Tryptic peptides were concentrated and desalted with 3 M Empore High Performance C18HD extraction disk cartridges according to the manufacturer's protocol, dried under vacuum, and resuspended in 0.1% formic acid and 5% acetonitrile.
LC-Electrospray Ionization-MS Analysis-LC-MS analyses of synthetic peptides were performed in the positive ion mode with a Finnigan Mat LCQ ion trap instrument (San Jose, CA) coupled to a Waters 2690 HPLC system (Milford, MA). Peptides were separated at a flow rate of 0.2 ml/min on a reversephase column (Vydac C18 MS column; 250 ϫ 2.1 mm), using solvent A (0.2% formic acid in water) and solvent B (0.2% formic acid in 90% CH 3 CN, 10% water). Peptides were eluted using a linear gradient of 0 -40% solvent B over 30 min. The spray voltage was 4.5 kV, and the temperature of the heated capillary was 220°C. The collision energy for MS/MS was 35%.
Micro-LC-MS analyses of proteolytic peptides of TIMP-1 were performed in the positive ion mode with a Finnigan LCQ Deca XP ion trap instrument (Thermo Electron Corp., San Jose, CA) coupled to a Surveyor HPLC system. Peptides were separated at a flow rate of 2 l/min on a BioBasic-18 MS column (100 ϫ 0.18 mm; Thermo Hypersil-Keystone, Bellefonte, PA) using solvent A (0.1% formic acid in water) and solvent B (0.1% formic acid in CH 3 CN). Peptides were eluted using a linear gradient of 5-35% solvent B over 90 min. The spray voltage was 3.5 kV, and the temperature of the heated capillary was 170°C. The collision energy for MS/MS was 35%.
Nano-LC-MS analyses of tryptic peptides of BAL samples were performed in the positive ion mode with a Finnigan LTQ linear ion trap mass spectrometer (Thermo Electron Corp.) coupled to a Paradigm MS4 LC system (Michrom BioResources, Inc.). Peptides were separated at a flow rate of 1.0 l/min on a Magic C18 AQ column (150 ϫ 0.15 mm, 5 m 200 Å; Michrom BioResources, Inc.) using solvent A (0.1% formic acid, 5% CH 3 CN in water) and solvent B (0.1% formic acid in 90% CH 3 CN). Peptides were eluted using a linear gradient of 5-35% solvent B over 180 min. The spray voltage was 1.8 kV, and the temperature of the heated capillary was 200°C. The collision energy for MS/MS was 35%.
Reverse Zymography-Reverse zymography of TIMP-1 was performed by electrophoresis on 15% SDS-polyacrylamide gel copolymerized with 0.8 mg/ml gelatin and 160 ng/ml human recombinant MMP-9 (proenzyme) (30). Gels were washed with 2.5% Triton X-100 and then incubated in enzyme assay buffer (50 mM Tris, pH 7.5, 200 mM NaCl, 5 mM CaCl 2 , 0.02% brij-35 (w/v)) for 36 h. After the TIMP-1 band was stained with Simply Blue TM Safestain and destained with water, it was visualized against the clear background of the gel. The inhibitory activity of TIMP-1 was determined by scanning the bands with a densitometer.

RESULTS
MPO Inactivates TIMP-1-To determine whether HOCl can restrain the MMP inhibitory activity of TIMP-1, we exposed the protein to the MPO-H 2 O 2 -chloride system. Human TIMP-1 (1 M) was incubated in PBS (141 mM chloride, pH 7.4) with increasing concentrations of H 2 O 2 in the presence of the complete MPO system or with H 2 O 2 alone. The ability of TIMP-1 to inhibit gelatinase B (MMP-9) was then determined by reverse zymography (Fig. 1). The complete MPO-H 2 O 2 -Cl system dramatically decreased the inhibitory activity of TIMP-1. Band density analysis demonstrated that 78 or 94% of the inhibitory activity of TIMP-1 was lost when either 20 M or 50 M H 2 O 2 was included in the complete MPO system. In contrast, twice the concentration (100 M) of H 2 O 2 alone had no effect. We observed similar results when TIMP-1 was exposed to HOCl (data not shown). These observations strongly suggest that HOCl produced by MPO can inactivate TIMP-1.
MPO Oxidizes Specific Residues in TIMP-1-To investigate the molecular basis for the oxidative inactivation of TIMP-1, we first digested unmodified protein with trypsin and identified the resulting peptides through LC-MS and MS/MS analysis. We detected all the cysteine-containing peptides (Table 1, P1-P7), which account for the disulfide structure of TIMP-1 and also contain the most important residues for its inhibitory activity (33).
We used LC-MS/MS to determine which peptides in TIMP-1 were susceptible to oxidation by the complete MPO-H 2 O 2 -chloride system (1:20 molar ratio of protein to oxidant). Interestingly, two tryptic peptides, P1 (C 1 TC 3 VP-PHPQTAFC 13 NSDLVIR) and P2 (FVYTPAM 66 ESVC 70 GY-FHR), were selectively oxidized. Quantification revealed dramatic loss of both peptides ( Figs. 2A and 3A), with a decline of ion intensity of 50% for P1 and of almost 100% for P2. In contrast, H 2 O 2 alone oxidized neither P1 nor P2 (Figs. 2A and 3A). These findings suggest that the MPO-H 2 O 2chloride system selectively oxidizes P1 and P2. Importantly, these peptides contain Cys 1 and Cys 70 , which form the disulfide bond that is critical to the inhibitory activity of TIMP-1 (15).
MPO Oxidizes Cys 1 of P1 to an Unusual Product-P1 contains 3 disulfide-bonded cysteine residues that might be targets for HOCl. Therefore, we used LC-MS and MS/MS to analyze P1 and its oxidation product(s) in tryptic digests of native or oxidized TIMP-1. Unmodified P1 showed major ions of massto-charge ratio (m/z) 791.8 and 1186.9, which is consistent with the anticipated m/z of the triply and doubly charged peptide ions [M ϩ 3H] 3ϩ and [M ϩ 2H] 2ϩ . MS/MS analysis of the peptide ion of m/z 791.8 confirmed the sequence of P1 from the digest of unmodified TIMP-1 (Fig. 2, A2). Reconstructed ion chromatograms revealed that the relative abundance of this peptide decreased significantly after TIMP-1 was exposed to the complete MPO-H 2 O 2 -chloride system ( Fig. 2A). Unexpectedly, we failed to detect any known product(s) (e.g. Cys ϩ 16, Cys ϩ 32, Cys ϩ 48) (32, 34) that would be anticipated if cysteine residues or disulfide bonds were oxygenated. These observations suggested that HOCl might convert P1 to unusual products. Indeed, LC-MS/ MS detected a new peak of material with m/z 1142.7 (Fig. 2B). MS/MS analysis of this ion revealed a series of fragment ions with the same m/z as y-type ions from MS/MS analysis of P1 (y 16 ϩ , y 16 ϩ2 , y 14 ϩ2 , y 13 , y 12 , y 14 , y 5 , y 4 ,  etc.). Thus, the new peak of material was derived from P1. The isotopic pattern demonstrated that the ion was doubly charged, suggesting that the precursor peptide had lost 88 atomic mass units when it formed the product peptide (Fig. 2, B1). We therefore named this product P1-88. MS/MS analysis of P1-88 revealed a series of y-type ions and a prominent ion of m/z 1120.9, which is consistent with loss of 44 atomic mass units from the doubly charged precursor ion ([M Ϫ 44 ϩ 2H] 2ϩ ) (Fig. 2, B2). Detection of the fragment ion y 19 ϩ2 (Fig. 2, B2, m/z 1106.4) located the modification on the N-terminal cysteine residue. This observation indicates that Cys 1 is modified to Cys 1 -88 in P1-88 and raises the possibility that generation of P1-88 contributes to TIMP-1 inactivation. MPO Oxidizes the Met Residue in P2 to Methionine Sulfoxide-The methionine residue, disulfide bond, and 2 tyrosine residues in P2 are potential targets of oxidation. To identify the site at which HOCl modifies P2, we exposed TIMP-1 in PBS to H 2 O 2 (1:20, molar/molar, protein/oxidant) and MPO and used reconstructed ion chromatograms to detect the product. The results revealed that P2 precursor disappeared in the tryptic digest of oxidized TIMP-1, and a new peak of material was observed (Fig. 3, A and  B). LC-MS analysis of this material demonstrated a major ion at m/z 661.3 (Fig. 3, B1), which is consistent with the addition of 16 atomic mass units to triply charged P2 ([M ϩ 16 ϩ 3H] 3ϩ ). MS/MS analysis of the modified peptide indicated that the methionine residue (Met 66 ) of P2 had gained 16 atomic mass units (Fig. 3, B2), whereas the 2 tyrosine residues and 1 cysteine residue remained unmodified. These observations demonstrate that HOCl generated by the MPO-H 2 O 2 -chloride system readily converts the thioether group of Met 66 in P2 to the sulfoxide.
To determine whether the MPO-H 2 O 2 -chloride system also oxidizes other methionine residues in TIMP-1, we digested TIMP-1 with endoproteinase Glu-C, which cleaves peptide bonds C-terminal of glutamic acid when the reaction is performed in ammonium bicarbonate buffer (35). We detected a peptide containing all 3 methionine residues (IKM 42 TKM 45 YKG-FQALGDAADIRFVYTPAM 66 E). LC-MS and MS/MS analysis of oxidized TIMP-1 revealed that all 3 methionine residues (Met 42 , Met 45 , and Met 66 ) had gained 16 atomic mass units when the protein was exposed to a 10:1 mole ratio of oxidant: TIMP-1 (data not shown). These results suggest that HOCl generated by the MPO-H 2 O 2 -chloride system readily con- verts all 3 methionine residues in TIMP-1 to methionine sulfoxide.
Loss of TIMP-1 Activity Strongly Associates with Loss of N-Terminal Peptide and Formation of Cys 1 -88-To determine quantitatively whether oxidation of specific amino acid residues in P1 or P2 correlates with loss of TIMP-1 activity, we first incubated 1 M human TIMP-1 with increasing concentrations of H 2 O 2 in the complete MPO system or with H 2 O 2 alone in PBS. We then monitored the influence of control and oxidized TIMP-1 on MMP-7 activity using the fluorescent Mca peptide substrate. P2 disappeared completely from the tryptic digest of TIMP-1 that had been exposed to the MPO-H 2 O 2 -chloride system at low molar ratios of oxidant to protein (Fig. 4A). However, there was little correlation between P2 loss and the inhibitory activity of the protein (Fig. 4B). In contrast, as the mole ratio of oxidant to TIMP-1 increased from 0 to 50, both the relative intensity of the precursor peptide P1 and the inhibitory ability of TIMP-1 decreased to a similar extent (Fig. 4B). Thus, loss of P1 associated with loss of inhibitory activity.
P1 contains the Cys 1 residue that is essential for the inhibitory activity of TIMP-1. To further study whether modifying the N-terminal cysteine might account for the loss of TIMP-1 inhibitory function, we examined the relationship between the latter and Cys 1 -88 formation. The yield of Cys 1 -88 correlated strongly with the loss of TIMP-1 inhibitory activity (R 2 ϭ 0.95) (Fig.  4D). In contrast, the production of P2 containing methionine sulfoxide showed little association with the inhibitory activity of TIMP-1.
These observations indicate that P2 is more sensitive to oxidation than P1 but that its modification does not significantly diminish the inhibitory activity of TIMP-1. In contrast, oxidative modification of P1 strongly associated with loss of TIMP-1 activity and therefore likely plays a critical role in TIMP-1 inactivation.
Proposed Structure of Cys 1 -88-To investigate the molecular structure of the Cys 1 -88 that appears in P1-88 when TIMP-1 is oxidized by the MPO-H 2 O 2 -chloride system, we used a synthetic peptide, (AcC)CT(AcC)CVPPH, that duplicates the sequence and disulfide bond pattern of the N-terminal sequence of P1. To generate this peptide, N-acetyl-L-cysteine was cross-linked with disulfide bonds to the thiol groups in both Cys 1 and Cys 3 . The synthetic peptide was used to mimic the oxidation of TIMP-1 undergoing tryptic digestion before LC-MS/MS detection. When the synthetic peptide was exposed to the complete MPO system, LC-MS detected [(AcC)CT(AcC)CVPPH-88] as a major oxidative product. MS/MS analysis of this product revealed the same mass spectrum (Fig. 5) as the Cys 1 -88 from MPO-oxidized TIMP-1, suggesting that identical products have been formed in the oxidized synthetic peptide and TIMP-1 protein. Furthermore, detection of unmodified y 6 (m/z 710.1) confirmed that Cys 1 had been modified. We used this model peptide to investigate the structure of the Cys 1 -88 product. MS analysis of control peptide detected both the doubly charged ion [M ϩ 2H] ϩ2 , m/z 435.8, and the singly charged ion [M ϩ H] ϩ1 , m/z 870.3 (Fig. 5, A1), but the doubly charged ion was the predominant ion. In contrast, MS of Cys 1 -88 revealed only the singly charged ion, m/z 782.0 (Fig. 5, B1), suggesting modification of the N-terminal amino group. Consistent with this proposal, acetylation of the N-terminal ␣-amino group of the peptide blocked the oxidative loss of 88 atomic mass units when the peptide was exposed to the complete MPO system (data not shown).
MS/MS analysis of the Cys 1 -88 product (m/z 782.0) demonstrated that the base peak exhibited an m/z of 738.1, which is consistent with the loss of 44 atomic mass units from the precursor ion, [M Ϫ 44 ϩ H] ϩ , suggesting that Cys 1 -88 might release CO 2 under the acidic conditions and elevated temperature required for MS/MS analysis. These observations indicate that the product likely lost the ␣-amino group and gained a heat-unstable carboxylic group on Cys 1 . According to the chemical properties and molecular mass of Cys 1 -88, we proposed that the possible structure of Cys 1 -88 product was as 2-oxo acid (Scheme 1) (36 -38). It is important to note that the disulfide-bonded Cys 1 in the native TIMP-1 was reduced and alkylated by iodoacetamide with a gain of 58 atomic mass units on Cys 1 before trypsin digestion. The net mass loss from native disulfide-bonded Cys 1 to product 2-oxo acid is 30 mass units.
Detection of the Oxidized N-terminal Cys 1 Residue of TIMP-1 in Vivo-Acute lung injury, a major cause of respiratory failure in critically ill patients, is characterized by an intense neutrophil-dominated inflammatory response (39,40). Such a response could upset the balance between MMP activity and TIMP inhibition if neutrophil products inactivated TIMP. To investigate whether HOCl generated by MPO might oxidize the N-terminal Cys 1 of TIMP-1 in vivo, we analyzed tryptic digests of BAL fluid from ARDS patients and normal control subjects, using nano-LC-MS/MS. With selected reaction monitoring, we were able to detect P1 (m/z 792.0, triply charged ion) in both ARDS and normal BAL fluids (Fig. 6A). Importantly, we detected Cys 1 -88 (the oxidized P1 product, m/z 1142.6, doubly charged ion) in pooled ARDS BAL fluid, but not in normal BAL fluid (Fig. 6B). MS/MS analysis of the ions of m/z 792.0 and 1142.6 (Fig. 6, A1 and B1) confirmed the sequences of the unmodified and oxidized peptides. By comparing the relative peak areas of unmodified and oxidized P1, we estimated that ϳ7% of TIMP-1 was oxidized from the pooled ARDS BAL fluid. The observations suggest that HOCl produced by the MPO system of activated phagocytes oxidizes Cys 1 of TIMP-1 in vivo.

DISCUSSION
We demonstrated that HOCl, generated by the MPO-H 2 O 2chloride system, inactivates TIMP-1. Tandem MS analysis of proteolytic digests of TIMP-1 exposed to the MPO-H 2 O 2 -chloride system revealed that Cys 1 and multiple methionine residues in the N-terminal domain of the protein were significantly modified. Loss of TIMP-1 inhibitory activity associated strongly with modification of Cys 1 but not with oxidation of the methionine residues. The Cys 1 region of the TIMP-1 molecule, which is shaped like a wedge (5), inserts itself into the active site cleft of MMPs in a manner similar to that of substrates. The N-terminal cysteine, located at the tip of the wedge, is highly accessible to solvent. Our results indicate that the N-terminal cysteine (Cys 1 ) is a specific oxidation target in TIMP-1 and suggest that MPO inactivates TIMP-1 by modify-ing Cys 1 to a novel product that loses 88 mass units. We named this product Cys 1 -88.
Two lines of evidence demonstrate that Cys 1 oxidation associates strongly with loss of inhibitory activity. First, impairment of activity correlated quantitatively only with loss of peptide P1, which contains Cys 1 . Moreover, there was strong linear correlation between Cys 1 -88 formation and loss of inhibitory activity. We cannot exclude the possibility that other modifications of TIMP-1 contributed to this loss, but our observations suggest that modification of N-terminal Cys 1 is particularly important. Second, MS/MS analysis indicated that Cys 1 lost its N-terminal amino group when TIMP-1 was exposed to oxidant. The crystal structure of the TIMP-1-MMP-3 (⌬C) complex reveals that the N-terminal ␣amino and carbonyl groups of Cys 1 coordinate the catalytic Zn 2ϩ of MMP-3. In addition, a hydrogen bond joins the N-terminal amino group of TIMP-1 with the carboxylate oxygen of Glu 202 in the active site of MMP-3 (15,41). Modifying Cys 1 would disrupt the hydrogen bond and impair the ability of the residue to coordinate the catalytic Zn 2ϩ of MMPs (Fig. 7). Moreover, cleavage between the ␣and ␤carbons of Cys 1 may dramatically change the overall structure of the N-terminal domain. Collectively, these observations strongly suggest that modifying the N-terminal Cys 1 perturbs the interaction of TIMP-1 with MMPs, impairing the inhibitory activity of TIMP-1.
Importantly, our observations indicate that HOCl generated by MPO converts N-terminal Cys 1 to Cys 1 -88, a novel oxidative product. LC-MS and MS/MS studies of the model peptide (AcC)CT(AcC)CVPPH, which mimics the N-terminal region of TIMP-1, confirmed that HOCl specifically oxidizes the N-terminal cysteine residue. The molecular mass of the product was 88 mass units less than that of the precursor; Cys 1 -88 also lacked the N-terminal amino group and readily lost CO 2 under MS/MS conditions. Based on our observations and the known reactivity of HOCl (34,36,42,43), we propose that Cys 1 -88 is a 2-oxo acid (Scheme 1). Although the precise reaction pathway is unclear, a plausible mechanism would involve the initial formation of an N-terminal glyoxylyl peptide intermediate, followed by loss of water. The resulting ␣-oxo alde- hyde would be further oxidized into the stable 2-oxo acid. It should be noted that oxidation of a 1, 2-amino alcohol to glyoxylyl by periodate is a well recognized reaction. Indeed, this reaction converts N-terminal Ser, Thr, or Cys residues to N-terminal glyoxylyl peptides (36,37). HOCl might oxidize Cys 1 to the 2-oxo acid product by a similar pathway.
Dysregulation of MMPs and TIMP-1 has been implicated in the pathogenesis of inflammatory diseases. Both oxidative activation of MMPs (32) and inactivation of TIMP-1 could alter the balance between MMP and TIMP-1, resulting in excess proteolytic activity and eventual tissue damage. Quantitative studies suggest that activated neutrophils (5 ϫ 10 6 per ml) can generate Ͼ100 M HOCl (44). Considering the large number of neutrophils that accumulate in inflamed tissues (e.g. lungs of patients with ARDS), the concentrations of oxidants used in our studies are physiologically relevant. Importantly, we were able to detect the proposed 2-oxo acid in the complex mixture of proteins in BAL fluid from ARDS patients. Thus, HOCl generated by phagocytes at inflamed sites might inactivate TIMP-1 in vivo by modifying Cys 1 to the novel product. MMPs and TIMPs are detectable in BAL fluids of patients with ARDS, and the concentration of the TIMPs sometimes exceeds that of the MMPs (45). Our findings suggest that oxidative inactivation of TIMPs could shift the physiological balance toward increased MMP activity and increased proteolysis in the lungs.
Methionine, cysteine, and tyrosine residues, tryptophan, and disulfide bonds are known to be sensitive to oxidation (24, 43, 46 -48). Our results showed that all 3 methionine residues in TIMP-1 are readily oxidized. Alkylated thiol groups of Met 42 , Met 45 , and Met 66 residues were almost completely converted to the sulfoxide when they were exposed to the MPO-H 2 O 2 -chloride system. However, oxidation of methionine residues showed little  Myeloperoxidase Inactivates TIMP-1 NOVEMBER 2, 2007 • VOLUME 282 • NUMBER 44 association with inactivation of TIMP-1. It is important to note that these methionine residues are located on the N-terminal domain of TIMP-1 (49), a critical region for inhibiting MMP activity, suggesting that they might function as endogenous antioxidants for TIMP-1 (50,51). At a high concentration of oxidant (50 ϳ 100 M), we also detected a low level of oxygenated tryptophan residues (Trp 105 , Trp 147 , and Trp 176 ) and chlorinated tyrosine residue (Tyr 120 ) (data not shown) on the Cterminal domain. It is unlikely that these modifications impair the inhibitory activity of TIMP-1 because of the low yield of the products. Because these residues are located in the C-terminal domain, their modification might impair other functions (10).
In summary, our observations indicate that HOCl generated by the MPO-H 2 O 2 -chloride system inactivates TIMP-1 and that loss of activity associates strongly with modification of Cys 1 . The reaction pathway may be broadly relevant because all TIMPs have a conserved N-terminal sequence, CT/SCXPXHPQ, even though they may have different functions (11). Moreover, we detected Cys 1 -88 in BAL fluid of humans suffering from ARDS, suggesting that activated phagocytes, the cellular hallmark of inflammation, can impair TIMP function in vivo. Because an imbalance between the proteolytic activity of MMPs and the inhibitory activity of TIMPs is implicated in many pathological conditions, our findings could have broad implications for understanding the regulation of MMP function in tissue injury.