DNA Strongly Impairs the Inhibition of Cathepsin G by α1-Antichymotrypsin and α1-Proteinase Inhibitor*

This paper explores the possibility that neutrophil-derived DNA interferes with the inhibition of neutrophil cathepsin G (cat G) and proteinase 3 by the lung antiproteinases α1-proteinase inhibitor (α1PI), α1-antichymotrypsin (ACT), and mucus proteinase inhibitor (MPI). A 30-base pair DNA fragment (30bpDNA), used as a model of DNA, tightly binds cat G (K d , 8.5 nm) but does not react with proteinase 3, α1PI, ACT, and MPI at physiological ionic strength. The polynucleotide is a partial noncompetitive inhibitor of cat G whoseK i is close to K d . ACT and α1PI are slow binding inhibitors of the cat G-30bpDNA complex whose second-order rate constants of inhibition are 2300 m −1 s−1 and 21 m −1 s−1, respectively, which represents a 195-fold and a 3190-fold rate deceleration. DNA thus renders cat G virtually resistant to inhibition by these irreversible serpins. On the other hand, 30bpDNA has little or no effect on the reversible inhibition of cat G by MPI or chymostatin or on the irreversible inhibition of cat G by carbobenzoxy-Gly-Leu-Phe-chloromethylketone. The polynucleotide neither inhibits proteinase 3 nor affects its rate of inhibition by α1PI. These findings suggest that cat G may cause lung tissue destruction despite the presence of antiproteinases.

Neutrophils are phagocytic cells recruited at sites of infection and inflammation. Their azurophil granules, which participate in phagocytosis, store a number of hydrolytic enzymes including elastase, cat G, 1 and proteinase 3, three serine proteinases whose tertiary structure has been elucidated recently. These 25-30-kDa glycoproteins contain a large number of basic amino acid residues that are responsible for their cationic character. Their specificity is directed against small aliphatic amino acid residues (elastase and proteinase 3) or more bulky ones (cat G). In vitro, these enzymes are able to cleave lung extracellular matrix proteins such as elastin, collagen, fibronectin, and laminin. They also cause extensive lung tissue damage in the animal (1).
Ideally, digestion of phagocytosed material should take place within the neutrophil in a phagocytic vacuole filled with the above proteinases. Actually, however, part of the lysomal enzymes reaches the extracellular space as a result of incomplete closure of the phagolysosome or of frustrated phagocytosis of large particules. In addition, neutrophils are short lived cells that release the bulk of their proteinases when they die at sites of inflammation. Tissues are normally protected against these enzymes by a number of proteinase inhibitors. For example, lung secretions contain at least three antiproteinases, namely ␣ 1 -proteinase inhibitor (␣ 1 PI), ␣ 1 -antichymotrypsin (ACT), and mucus proteinase inhibitor (MPI) (2). The 53-kDa ␣ 1 PI and the 68-kDa ACT are glycoproteins synthesized in the liver and transported into the lung via the blood circulation. They belong to the serpins, a superfamily of proteins that have developed by divergent evolution over 500 million years. The serpins have a highly conserved secondary structure comprising nine ␣-helices and three ␤-sheets with a very flexible reactive site loop. They form denaturant-stable complexes with their target proteinases and behave kinetically like irreversible inhibitors. This irreversible binding is due to the formation of a nonhydrolyzable acyl-enzyme adduct between the serine residue of the enzyme's active site and the P 1 residue of the serpin. The P 1 residues of ␣ 1 PI and ACT are Met and Leu, respectively (3). ␣ 1 PI inhibits the three aforementioned proteinases, whereas ACT reacts only with cat G. The second-order rate constants for these serpin-proteinase associations range from 10 5 to 10 7 M Ϫ1 s Ϫ1 (1). MPI is a 11.7-kDa basic protein synthesized and secreted by bronchial epithelial cells. It is formed of a single chain of 107 amino acid residues organized in two domains of similar size and architecture. It has a Leu residue at the P 1 position of its reactive center. MPI belongs to the class of canonical inhibitors that have a rigid reactive site loop that forms a reversible "lock and key" complex with the target proteinase. This complex is stabilized by a large number of noncovalent bonds, which account for the high enzyme-inhibitor affinity (4). MPI is a tight binding reversible inhibitor of elastase (K i , 32 pM (5)). It is less efficient on cat G (K i , 35 nM (6)) and does not inhibit proteinase 3 (7).
Because of their ability to cleave lung extracellular matrix proteins in vitro, neutrophil proteinases are thought to cause tissue destruction in inflammatory lung diseases. This raises, however, the following question: why does proteolysis occur despite the presence of the aforementioned inhibitors? Oxidative inactivation of proteinase inhibitors, transient excess of proteinase over inhibitor, and close neutrophil/matrix contact are conditions that favor proteolysis in the presence of inhibitors (1). Binding of proteinases to DNA released from neutrophils following cell death may also promote proteolysis in an inhibitory environment. We have recently shown that polynucleotides bind neutrophil elastase and impair its inhibition by MPI and ␣ 1 PI (8,9). Here we study the influence of neutrophil DNA on the inhibition of neutrophil cat G and proteinase 3 by ␣ 1 PI, ACT, and MPI. To this end, we use a 30-base pair DNA fragment ( 30bp DNA) as a handy model of DNA.

MATERIALS AND METHODS
Human neutrophil cat G was isolated and active site titrated as described previously (10). Human neutrophil proteinase 3 and human plasma ACT were purchased from Athens Research and Technology (Athens, GA). Human recombinant ␣ 1 PI and MPI were obtained through the courtesy of Dr. H.P. Schnebli, Novartis, Switzerland. ␣ 1 PI and MPI were active site titrated with elastase (11), whereas ACT was titrated with cat G (12). Proteinase 3 was titrated with ␣ 1 PI (7). Recombinant human pancreatic DNase I (Pulmozyme, Genentech, San Francisco, CA) was obtained from the Pharmacy of the University Hospital of Strasbourg. The p-nitroanilide and the thiobenzylester substrates came from Bachem (Switzerland) and E.S.P. (Livermore, CA), respectively. Stock solutions of the substrates and of 4,4Ј-dithiodipyridine (Sigma) were made in dimethylformamide (final concentration, 2% (v/v)). Z-Gly-Leu-Phe CH 2 Cl and chymostatin came from E.S.P. and Sigma, respectively. All experiments were done at 25°C in 50 mM Hepes, 150 mM NaCl, pH 7.4, a solution called the buffer.
Preparation of 30bp DNA-The bronchial secretions from cystic fibrosis patients were used as a source of neutrophil DNA. The viscous fluids were incubated for 3 h at 50°C with 0.5% SDS, 100 g/ml proteinase K (Sigma), 20 g/ml RNase (Sigma), and 5 mM EDTA. DNA was then extracted using phenol:chloroform:isoamylic alcohol (25:24:1) and precipitated with ethanol (13). The pellet was then dissolved in the buffer (2 mg/ml) and reacted with 290 nM DNase for 90 min at 25°C. After heating for 5 min at 65°C, the mixture was chromatographed on a 12 HR 10/30 Superose column (Amersham Pharmacia Biotech), calibrated with the pBr 322 plasmid previously digested with HaeIII (14). The fractions containing DNA fragments of about 30 base pairs (M r 20,000) were collected and concentrated. The purity was checked using polyacrylamide gel electrophoresis (Phastsystem, Phast Gel 20 and DNA silverstaining kit from Amersham Pharmacia Biotech).
Affinity Chromatography-Sepharose-bound 30bp DNA and MPI were prepared using epoxy-activated Sepharose 6B (Amersham Pharmacia Biotech) according to the manufacturer's instructions. The affinity supports were poured into HR 10/10 Amersham Pharmacia Biotech columns and equilibrated with 50 mM Hepes buffer, pH 7.4. cat G, proteinase 3, ␣ 1 PI, and ACT were dissolved in this buffer at a concentration of 0.2-0.3 mg/ml, loaded on the Sepharose-30bp DNA column and eluted with a linear NaCl gradient in the same buffer. Protein elution was followed at 280 nm. 30bp DNA (20 g/ml) was dissolved in the above buffer and chromatographed on the Sepharose-MPI column. NaCl gradient elution was followed at 260 nm.
Equilibrium Dissociation Constant K d of the cat G-30bp DNA Complex-Fluorescent labeling of cat G was done by reacting 1 mol of cat G with 10 mol of 5-(4, 6-dichlorotriazinyl)aminofluorescein (Molecular Probes, Eugene, OR) in 0.2 M carbonate buffer, pH 8.4. After 15 min of continuous stirring at room temperature, labeled cat G was isolated using a PD-10 gel filtration column (Amersham Pharmacia Biotech) equilibrated with a buffer formed of 50 mM Hepes and 150 mM NaCl, pH 7.4. Constant concentrations of labeled cat G were mixed with increasing concentrations of 30bp DNA, and the fluorescence intensity of the mixtures was determined with a Shimadzu RF 5000 spectrofluorimeter ( ex , 495 nm; em , 514 nm) The data were fitted to Equation 1 by nonlinear regression analysis, which gave the best estimate of K d and its confidence interval.
where D stands for 30bp DNA, ⌬F and ⌬F min are the absolute variations of fluorescence intensity at a given 30bp DNA concentration and at saturating 30bp DNA concentrations, respectively.
Inhibition of cat G and Proteinase 3 by 30bp DNA-The inhibition of cat G by 30bp DNA was monitored using a synthetic substrate as described in the legend to Fig. 3. The inhibition of 30 nM proteinase 3 by concentrations of 30bp DNA up to 1 M was assessed using 2 mM methoxysuccinyl-Lys(2pico)-Ala-Pro-Val-pNA (14). The effect of 30bp DNA on the elastolytic activity of these two proteinases was tested using Remazol Brilliant Blue elastin as described previously (8). The final concentration of the proteinases was 0.5 M. One single concentration of 30bp DNA (15 M) was used in replicate assays.
Kinetics of Inhibition of cat G-Rate constants of inhibition were measured by adding cat G Ϯ 30bp DNA to a mixture of inhibitor ϩ substrate Ϯ 30bp DNA and recording the release of product as a function of time. When the progress curves lasted for more than 10 min, an ordinary spectrophotometer (Uvikon 941, Kontron) was used to record them. When the reaction was terminated in 2 min or less the Uvikon spectrophotometer was equipped with a SFA-11 fast mixing accessory (High-Tech Scientific, Salisbury, UK). When faster reactions were followed, mixing of the reagents and data acquisition were performed with a PQ/SF-53 stopped flow apparatus (High Tech Scientific). All experiments were done under pseudo-first-order conditions, that is The concentration of product at the end of the inhibition experiments was always lower than 5% of the initial substrate concentration so that the latter was virtually not depleted during the inhibition process. Under these two conditions the progress curves describing the enzyme-catalyzed hydrolysis of substrate in the presence of irreversible inhibitors such as ACT, ␣ 1 PI, or Z-Gly-Leu-Phe-CH 2 Cl may be described by Equation 2 (15,16), where [P] is the product concentration at any time t, v z is the velocity at t ϭ 0, and k is the pseudo-first-order rate constant of inhibition. The progress curves were fitted to Equation 2 by nonlinear least square analysis to obtain the best estimates of k. The variation of k as a function of inhibitor or substrate concentration was analyzed assuming either that the inhibition conforms to a simple bimolecular reaction E ϩ k ass EI or that it takes place in two-steps.
The rate constant of inhibition is given by Equation 4 (one-step inhibition) or Equation 5 (two-step inhibition).
Equations 4 and 5 show that a linear or a hyperbolic plot of k versus [I] o will diagnose one-step or two-step inhibition, respectively. Two-step inhibition does not, however, necessarily yield a hyperbolic dependence of k versus The progress curves for the inhibition of 30bp DNA-bound cat G by the reversible inhibitors MPI and chymostatin may be described by Equation 6 (15,16), where v s is the steady-state velocity, and the other symbols have the same meaning as in Equation 2. The variation of k as a function of inhibition concentration was analyzed assuming one-step inhibition: for which k is given by: Kinetics of Elastin Solubilization by cat G Ϯ ␣ 1 PI Ϯ 30bp DNA-One volume of a buffered solution of cat G Ϯ 30bp DNA was mixed with an equal volume of a continuously stirred suspension of Remazol Brilliant Blue elastin Ϯ ␣ 1 PI Ϯ 30bp DNA. While stirring was continued at 37°C, 0.5-ml aliquots were withdrawn after given periods of time and diluted with 0.5 ml of 0.75 M acetate buffer, pH 5.0, to stop the reaction. After centrifugation at 10,000 ϫ g for 15 min, the absorbances were read at 595 nm. The absorbances of appropriate controls were substrated. The buffer was 50 mM Hepes, 150 mM NaCl, pH 7.4. The final reactant concentrations were: cat G ϭ ␣ 1 PI ϭ 1 M, 30bp DNA ϭ 2 M, elastin ϭ 3 mg/ml.

RESULTS
Binding of cat G to 30bp DNA- Fig. 1A shows that at low ionic strength (50 mM Hepes, pH 7.4) Sepharose-30bp DNA binds ACT and cat G but does not significantly bind ␣ 1 PI. MPI was weakly bound and was eluted with a low NaCl concentration (not shown). Whereas cat G was tightly bound to the affinity support, ACT was eluted at a NaCl concentration lower than that contained in the buffer used for the enzyme kinetic assays, namely 50 mM Hepes, 150 mM NaCl, pH 7.4. The weak binding of ACT to the affinity column is not surprising, because ACT is known as a DNA-binding protein whose binding domain has been identified recently (17). To further confirm the weak binding of MPI to the polynucleotide, we have chromatographed 30bp DNA on a Sepharose-MPI column. Fig. 1B shows that the polynucleotide elutes from this column with a NaCl concentration of about 100 mM. Thus, cat G is the only protein that binds 30bp DNA in the aforementioned buffer.
In an attempt to quantitate the cat G-30bp DNA affinity we have measured the effect of polynucleotide concentration on the fluorescence intensity of fluorescently labeled cat G. Fig. 2A shows that this intensity decreases up to a limiting value F min . Fig. 2B is a replot of these data in accordance with Equation 1 from which K d , the equilibrium dissociation constant of the cat G-30bp DNA complex, was calculated by nonlinear regression analysis. K d was found to be 8.5 Ϯ 3.2 nM.
Partial Inhibition of cat G Activity by 30bp DNA-Reaction of constant amounts of cat G with increasing amounts of 30bp DNA resulted in partial inhibition of the enzymatic activity on Suc-Ala 2 -Pro-Phe-pNA, whether the enzyme concentration was low (24 nM) or high (300 nM) (Fig. 3). When the inhibition of 24 nM cat G was measured using different substrate concentrations, similar results were obtained (data not shown). Partial inhibition may be analyzed using the scheme adapted from Ref. 18, where E, S, and I stand for enzyme, substrate, and 30bp DNA, respectively, K i is the inhibition constant and ␣ and ␤ are dimensionless numbers. The inhibition is said to be partially competitive if ϱ Ͼ ␣ Ͼ 1 and ␤ ϭ 1 and partially noncompetitive if ␣ ϭ 1 and 0 Ͻ ␤ Ͻ 1 (18). To decide between these two mechanisms we have measured the kinetic parameters for the hydrolysis of Suc-Ala 2 -Pro-Phe-pNA by free cat G to get k cat and K m and by 30bp DNA-saturated cat G to obtain ␤k cat and ␣K m . The following constants were found: k cat , 3.1 Ϯ 0.3 s Ϫ1 ; ␤k cat , 1.4 Ϯ 0.1 s Ϫ1 ; K m , 2.5 Ϯ 0.2 mM; ␣K m , 2.8 Ϯ 0.2 mM. Thus, ␣ ϭ 1.1 Ϯ 0.2 and ␤ ϭ 0.45 Ϯ 0.09, which strongly suggests partially noncompetitive inhibition, that is 30bp DNA does not alter the binding of Suc-Ala 2 -Pro-Phe-pNA to the active site of cat G but hinders its breakdown into products.
Szedlacsek et al. (19) have derived an equation that allows the estimation of K i from partial tight binding inhibition such as that shown in Fig. 3. This complex equation (not shown) contains five unknown parameters: K i , k cat , K m , ␣, and ␤. We have taken the experimentally determined k cat , K m , ␣, and ␤ terms as known constants of Szedlacsek's equation to calculate K i by nonlinear regression analysis. With the data collected with 24 nM cat G in Fig. 3, K i was 6.7 Ϯ 0.5 nM, whereas with cat G, 300 nM, K i ϭ 11 Ϯ 5 nM. These figures are in excellent agreement with the fluorometrically measured equilibrium dissociation constant. 30bp DNA also partially inhibited the activity of cat G on insoluble elastin; the percentage of cat G inhibition was 16.2 Ϯ 4.1.
Effect of 30bp DNA on the Inhibition of cat G by ACT and ␣ 1 PI- Fig. 4 shows that 30bp DNA sharply depresses the rate constant for the inhibition of cat G by ACT and ␣ 1 PI. With ACT the maximal decrease was 210-fold, whereas it was 3100-fold in the case of ␣ 1 PI. The inhibition of cat G by ACT was also tested with variable concentrations of full-length DNA (0.5-10 g/ml). The rate constant decreased 200-fold as with 30bp DNA (data not shown), indicating that 30bp DNA is a valuable model of fulllength DNA.
To study the influence of substrate concentration on the inhibition of 30bp DNA-bound cat G by ACT and ␣ 1 PI, we used constant concentrations of enzyme and inhibitor (30 nM cat G ϩ 0.5 M 30bp DNA ϩ 0.5 M ACT or 10 nM cat G ϩ 0.5 M 30bp DNA ϩ 50 M ␣ 1 PI) and variable concentrations of Suc-Ala 2 -Pro-Phe-pNA (0.8 -6 mM). We found that 1/k was linearly related to [S] o , as predicted by Equations 4 and 5.
As shown in Fig. 5, the rate constant for the inhibition of 30bp DNA-bound cat G varies hyperbolically with the ACT concentration, suggesting two-step inhibition. Indeed, the data could be fit to Equation 5 by nonlinear regression analysis. In contrast, the linear variation of k with the ␣ 1 PI concentrations indicates one-step inhibition. The kinetic constants are reported in Table I together with those collected previously with free cat G (12). It can be seen that 30bp DNA lowers the secondorder rate constant for the inhibition of cat G by ACT and ␣ 1 PI by a factor of 195 and 3200, respectively. Two experiments were run to further illustrate the dramatic effect of 30bp DNA on the rate inhibition of cat G by ␣ 1 PI. First, we have incubated constant concentrations of free or polynucleotide-bound cat G with increasing concentrations of ␣ 1 PI for 1 h, a time largely sufficient to ensure full inhibition of the free enzyme. The inhibitor yielded a linear inhibition curve with free cat G, whereas the 30bp DNA-cat G complex was fully resistant to inhibition (data not shown). Second, we have studied the influence of 30bp DNA on the inhibition of the elastolytic activity of cat G by ␣ 1 PI. We have found that ␣ 1 PI fully inhibited the elastolytic activity of cat G in the absence of 30bp DNA, whereas in the presence of the polynucleotide, there was only 28% inhibition.
Effect of 30bp DNA on the Inhibition of cat G by MPI- Fig. 6 shows that the progress curves run in the presence of cat G ϩ MPI Ϯ 30bp DNA are biphasic, as predicted for reversible inhibition (Equation 6). The rate constant of inhibition is linearly related to the MPI concentration, indicating one-step inhibition (Equation 8). The kinetic constants k ass , k diss , and K i were found to be 1.5 ϫ 10 5 M Ϫ1 s Ϫ1 , 1.7 ϫ 10 Ϫ2 s Ϫ1 , and 1.2 ϫ 10 Ϫ7 M, respectively. MPI also inhibits free cat G in a one-step reaction characterized by k ass ϭ 1.0 ϫ 10 5 M Ϫ1 s Ϫ1 , k diss ϭ 3.5 ϫ 10 Ϫ3 s Ϫ1 and K i ϭ 3.5 ϫ 10 Ϫ8 M. Thus, the polynucleotide has virtually no effect on k ass and moderately increases k diss and K i .
Effect of 30bp DNA on the Inhibition of cat G by Z-Gly-Leu-Phe-CH 2 Cl and Chymostatin-Z-Gly-Leu-Phe-CH 2 Cl is an irreversible cat G inhibitor, whereas chymostatin inhibits the enzyme reversibly (20). The rates of inhibition were measured using 5-10 nM cat G, 0.5 M 30bp DNA, 0.6 mM Suc-Ala 2 -Pro-Phe-thiobenzylester, and variable concentrations of inhibitor. For both compounds the rate constant of inhibition k was found to be linearly related to the inhibitor concentration, indicating one-step inhibition. Z-Gly-Leu-Phe-CH 2 Cl inhibited cat G with a k ass of 25 M Ϫ1 s Ϫ1 and 23.9 M Ϫ1 s Ϫ1 in the absence and presence of polynucleotide, respectively. The kinetic parameters for the cat G-chymostatin interaction were found to be: k ass ϭ 2.4 ϫ 10 4 M Ϫ1 s Ϫ1 , k diss ϭ 1.2 ϫ 10 Ϫ2 s Ϫ1 , in the absence of 30bp DNA and k ass ϭ 1.8 ϫ 10 4 M Ϫ1 s Ϫ1 , k diss ϭ 1.6 ϫ 10 Ϫ2 s Ϫ1 in its presence. The polynucleotide does not, therefore, significantly affect the inhibition of cat G by these synthetic inhibitors.
Effect of 30bp DNA on the Activity and Inhibition of Proteinase 3-Proteinase 3 did not bind to the Sepharose-30bp DNA column even at low ionic strength (50 mM Hepes, pH 7.4). Also, its activity on methoxysuccinyl-lysyl (2-picolinoyl)-Ala-Pro-Val-pNA and on elastin was unaffected by 30bp DNA. On the other hand the polynucleotide did not significantly change its rate of inhibition by ␣ 1 PI (data not shown). DISCUSSION We have explored the possibility that DNA released from neutrophils at sites of inflammation interferes with the inhibition of neutrophil cat G and proteinase 3 by their endogenous inhibitors. A 30bp DNA fragment used as a model of DNA was able to tightly bind cat G to form an enzymatically active complex. This complex was almost as active on synthetic substrates and on elastin as free car G and could be easily inhibited by MPI and synthetic cat G inhibitors. In contrast, the 30bp DNA-cat G complex was virtually resistant to inhibition by the two serpins ACT and ␣ 1 PI. ACT inhibits both free (12) and 30bp DNA-bound cat G via a two step reaction (see Eq. 3).
The polynucleotide decreases both the affinity of the Michaelis-type complex EI * (K i * increases 42-fold) and the rate constant for its conversion into the final complex EI (k 2 decreases ϳ15-fold) ( Table I). Whereas ␣ 1 PI inhibits free cat G in two steps (12), it reacts in one step with the 30bp DNA-cat G complex. The value of k ass (21 M Ϫ1 s Ϫ1 ) is, however, several orders of magnitude lower than the maximum rate constant for a bimolecular diffusion-controlled reaction (21). It may, therefore, be assumed that the reaction involves an intermediate even if the latter is not seen kinetically. This assumption predicts that the largest inhibitor concentration used (50 M) must be lower than K i * (1 ϩ [S] o /K m ) (see Equation 5). Hence, K i * Ͼ 17 M and k 2 Ͼ 3.6 ϫ 10 Ϫ4 s Ϫ1 . Comparison of these limits with K i * and k 2 previously determined with free cat G (12) indicates that 30bp DNA increases K i * at least 21-fold and decreases k 2 at least 153-fold. Thus, the dramatic decrease in the rate of inhibition of cat G by ACT and ␣ 1 PI is due in both cases to an unfavorable effect of 30bp DNA on K i * and k 2 .
We have previously shown that polynucleotides also depress the rate of neutrophil elastase inhibition by ␣ 1 PI. For instance, tRNA and polydeoxycytosine decrease the second-order rate constant of inhibition by factors of 3 and 31, respectively (9). We have confirmed these mild effects using 30bp DNA, which depressed the second-order rate constant by a factor of 9 (data not shown). Thus, polynucleotides affect much less the inhibition of elastase by ␣ 1 PI than that of cat G by ␣ 1 PI. This is reminiscent of previous findings showing that heparin, another anionic ligand, decreases the rate of inhibition of cat G by ␣ 1 PI by a factor of 400 (22), whereas it lowers the rate of inhibition of elastase by ␣ 1 PI by a factor of only 5 (23). DNA and heparin have thus similar effects on the two proteinase/␣ 1 PI systems; they enormously decrease the rate of inhibition of cat G but only marginally affect the rate of inhibition of elastase. This difference may be related to differences in the localization of the arginine residues in cat G and elastase. Arginine residues have the potential to form salt bridges with the anionic groups of heparin and DNA. Unlike elastase (24), cat G has three arginine residues in the immediate vicinity of its active site (25). As a consequence, heparin or 30bp DNA might bind closer to the active site of cat G than to that of elastase thus causing a more steric hindrance to the access of ␣ 1 PI in cat G than in elastase. ACT a Ϫ two-step 6.2 ϫ 10 Ϫ8 2.8 ϫ 10 Ϫ2 4.5 ϫ 10 3 ACT ϩ two-step 2.6 ϫ 10 Ϫ6 6.0 ϫ 10 Ϫ3 2.3 ϫ 10 3 ␣ 1 PI a Ϫ two-step 8.1 ϫ 10 Ϫ7 5.5 ϫ 10 Ϫ2 6.7 ϫ 10 4 ␣ 1 PI ϩ one-step Ͼ1.7 ϫ 10 Ϫ5b Ͼ3.6 ϫ 10 Ϫ4b 21 a From Duranton et al. (12). b Assuming that in fact the inhibition occurs in two steps with K i *(1 ϩ [S] o /K m ) greater than the maximum inhibitor concentration used in the experiment (see also text).
FIG. 6. Effect of 30bp DNA on the inhibition of cat G by MPI at pH 7.4 and 25°C. A, progress curves for the hydrolysis of Suc-Ala 2 -Pro-Phe-thiobenzylester by cat G alone (curve 1), cat G ϩ 30bp DNA (curve 2), or cat G ϩ 30bp DNA ϩ 400 or 800 nM MPI (curves 3 or 4). The cat G and substrate concentrations were 5 nM and 0.6 mM, respectively. The progress curves recorded in the presence of MPI were used to calculate k, the pseudo-first-order rate constant of inhibition. B, plot of k versus MPI concentration. Proteinase 3 is not inhibited by ACT or MPI but is rapidly inactivated by ␣ 1 PI (k ass ϭ 8 ϫ 10 6 M Ϫ1 s Ϫ1 (7)). This reaction rate is not altered by 30bp DNA. Thus, the inhibition of the three neutrophil serine proteinases by ␣ 1 PI is diversely affected by 30bp DNA; cat G almost fully resists inhibition, elastase is inhibited with a moderately reduced rate, and the inhibition of proteinase 3 is not affected at all by the polynucleotide.
We believe that our kinetic data have pathological bearing. In chronic bronchitis there is a continuous recruitment of neutrophils in airways. During activation or phagocytosis, neutrophils release part of their cat G, elastase, and proteinase 3 content in airway secretions. In addition, when these short lived cells die in situ they release both their proteinase and DNA content (26). We have shown that DNA forms a tight complex with cat G, and so it renders this proteinase virtually resistant to inhibition by the fast acting serpins ␣ 1 PI and ACT. Thus, DNA promotes cat G-mediated proteolysis of lung matrix proteins in an inhibitory environment.