A kinetic analysis of the interaction of human myeloperoxidase with hydrogen peroxide, chloride ions, and protons.

The effect of H2O2, Cl-, and pH on human myeloperoxidase activity has been examined. The Km for H2O2 is shown to be affected by the combined presence of Cl- and acid pH conditions. The Km for H2O2 is independent of pH in the absence of Cl- and dependent on pH in the presence of Cl-. Conversely, the dependence of the Km for H2O2 on Cl- concentration increases as the pH decreases. A model is proposed in which Cl- has a dual role, acting both as a substrate and as an inhibitor. According to this model, the inhibitor Cl- binding site must be protonated prior to the binding of Cl- and is distinct from the substrate Cl- binding site which is unaffected by pH. The rate equation derived from this model is used to further analyze the data presented. The values of Km for H2O2 predicted by the rate equation are in good agreement with the experimentally determined values.

' The abbreviations used are; TMB, tetramethylbenzidine; CTAB, cetyltrimethylammonium bromide; RZ, reinheit zahl. C1-concentrations over the pH range 4.4-7.4. Our results fit a model of myeloperoxidase catalysis in which the substrates H202 and C1-bind in any order and in which another C1interacts with myeloperoxidase at a protonated allosteric site resulting in competitive inhibition of the binding of HzOz.

Materials
TMB and dextran (Mr = 200,000-275,000) were purchased from Polysciences, Inc., Warrington, PA. Hydrogen peroxide (30%) was purchased from Fisher Scientific Co., Fair Lawn, NJ. Cetyltrimethylammonium bromide was purchased from Eastman Kodak Co., Rochester, NY. Sephadex G-75, superfine grade, was purchased from Pharmacia Fine Chemicals, Piscataway, NJ. Whole blood was provided by the New England Medical Center Hospital Clinical Pathology Laboratory.

Methods
Isolation of Human Neutrophils-Tubes of whole blood each containing about 3 ml of blood, were donated by the New England Medical Center Hospital Clinical Pathology Laboratory. To pooled whole blood, obtained within 12 h of being drawn, was added '/4 volume of 10% dextran in 0.3 M NaCI. The blood cells were allowed to sediment for 30 min at room temperature, after which the supernatant was removed and centrifuged for 15 min at 500 X g at 5 "C. The pellet was repeatedly exposed to 3 parts distilled H20 for 2 min followed by the addition of 1 part 0.6 M NaCl to lyse contaminating red blood cells. The cells were sedimented following each lysis step by centrifugation for 10 min at 500 X g at 5 "C. This method is a slight variation of the procedure of Levine et al. (13).
Isolation of Human Myeloperoxidase-Myeloperoxidase was isolated from human neutrophils as previously described (14) except that chromatography was performed on a Sephadex G-75, superfine grade, column. Myeloperoxidase isolated by this method had an RZ of 0.65. All detectable cetyltrimethylammonium bromide was shown to have been removed in the final dialysis step by the procedure of Jones (15).
Myeloperoxidase Assay-The oxidation of TMB by myeloperoxidase was followed at 655 mm (16) under various conditions of pH, NaCl concentration, and Hz02 concentration. Each assay was done in triplicate. The buffer used was either 0.12 M sodium acetate, pH 6.2, 5.6, 5.0, or 4.4 or 0.1 M sodium phosphate, pH 7.4. The NaCl concentrations used were 0, 0.001, 0.01, 0.05, 0.1, 0.15, 0.2, 0.3, and 0.4 M. Na2S04 was used to maintain a constant ionic strength as the NaCl concentration was varied. The Hz02 concentrations tested were 0.021, 0.046, 0.1, 0.21, 0.46, 1.0, 2.1, 4.6, 10, 21, and 46 mM. The reactions were performed at 37 "C and the resulting absorbance changes were converted to nanomoles of H202 utilized/min by using standard curves of TMB oxidation (16). The decay of oxidized species of TMB was corrected, as reported earlier (16). The Km for Hz02 was determined by the method of Lineweaver and Burk (17) using data obtained with H202 concentrations less than those causing substrate inhibition. The best linear fit for these data was determined using the Linear Regression Analysis program of a Texas Instruments model SR-56 calculator.

RESULTS
The oxidation of TMB by myeloperoxidase can occur either directly by enzymatic oxidation or indirectly through the production of HOCl when myeloperoxidase peroxidizes chloride. HOCl has the capacity to chemically react with TMB and form a product which is spectrally identical with that formed by the direct enzymatic oxidation of TMB by horseradish peroxidase (16). Thus, by using TMB to monitor myeloperoxidase activity, the assay can detect and quantitate the total utilization of H202 by myeloperoxidase. We have examined the rate of H202 utilization by myeloperoxidase as a factor of HzOz concentration varying both pH and C1-concentration. In these experiments, HZ02 was varied between 21 PM and 46 mM, NaCl between 0 and 0.4 M, and pH between 4.4 and 7.4.
Under all conditions of pH and C1-concentration tested, myeloperoxidase was found to be inhibited by excess HZ02 . As shown in Fig. la, for an experiment carried out at pH 6.2 in the absence of C1-, the maximum effectiveness of H202 occurs at approximately 0.46 m. The marked inhibition at higher concentrations of H202 is shown more clearly in Fig. lb. This type of inhibition has been reported previously (9,18) and has been ascribed to the destruction of the heme prosthetic groups of myeloperoxidase by the high concentrations of H202 (9). In order to visualize more readily the effect of varying pH and [ClF] on the activity of the myeloperoxidase, the data in Fig. 1 have been replotted on a semilog curve in Fig. 2. Again, the activity of the enzyme with respect to utilization of H202 can be seen to be maximal at 0.46 mM and to display substrate inhibition.
The effect of varying pH and [Cl-] on this system is to alter the activity of the enzyme such that the maximum activity occurs at different concentrations of H2O2. Examples of this alteration in activity are shown in Fig. 3. In the absence of C1-, the maximal activity of myeloperoxidase is shifted from 0.21 mM H202 at pH 7.4 to 0.46 mM a t pH 5.0. Of additional interest is the observation that the addition of C1-has no effect on the shape or maximum of the activity curve at pH 7.4, but induces a dramatic shift in activity at pH 5.0, with the maximum now observed a t 4.6 mM H202. In essence, the interrelationship of pH and C1-can be observed only under acidic  conditions and not under neutral conditions. This linked effect of pH and C1-can be seen in more detail in an analysis of the K,,, values of H202, as a function of pH and [Cl-1. As shown in Fig. 4, pH has very little effect on the K , of H202 in the absence C1-, but pH becomes increasingly important as the [Cl-] increases. This same effect is shown in Fig. 5, where the effect of varying [Cl-] on the K, is readily discerned. As the pH is decreased, the affinity of myeloperoxidase for Hz02 becomes more sensitive to the presence Cl-.
During the course of analyzing the K , values for H202 under varying conditions of pH and C1-, an observation was made relating the slope of the Lineweaver-Burk curves (Km/ VmaX) with [Cl-1. As shown in Fig. 6, the slope of the Lineweaver-Burk analysis clearly displays a linear relationship to the C1-concentration at each pH value tested. In addition, this linear dependency varies with the pH values studied.
In an attempt to explain the interacting effect of pH and C1-on the K , for H~OZ, the model shown in Fig. 7 is proposed. This model is based on the assumption that there are two distinct C1-binding sites on myeloperoxidase. One site can accept C1as a substrate for the production of HOC1. In the other site, C1-acts as a competitive inhibitor for the binding of H202. In addition, this model assumes that the inhibitor C1-can only bind to the protonated enzyme.
The rate equation, derived under "Appendix," in Lineweaver-Burk format, which mathematically expresses this model is: In this equation, KI is the dissociation constant of C1-binding to the inhibitor site and KH+ is the dissociation constant of a proton binding to the inhibitor site.
As can be seen from the rate equation, the slope of the Lineweaver-Burk plot must have a linear dependence upon the C1-concentration. This is precisely what was demonstrated in Fig. 6. Solving the equations generated by this analysis, KI was calculated to be 1.

DISCUSSION
The study reported here utilized TMB to examine the effect of pH, HzOz, and C1-on the activity of human myeloperoxidase. TMB can be oxidized by myeloperoxidase both directly and through the intermediate production of HOCl (16). Thus, the role of C1-as a substrate could not be assessed, but the role of C1-as an inhibitor was revealed.
Under all conditions of pH and C1-concentrations tested, myeloperoxidase exhibited substrate inhibition with respect to H202 (Fig. 2). The HzOz concentration required for maximal activity was only sensitive to C1-at low pH values (Fig. 3). This linked effect of C1-and pH was also observed in the analysis of the K,,, of myeloperoxidase for Hz02 (Figs. 4 and 5). A simiiar interaction had been observed previously for the myeloperoxidase-catalyzed chlorination of diethanolamine (11).
We have introduced a model for the interaction of myeloperoxidase with HzOz, C1-, and H' (Fig. 7) which is consistent with the results reported here. An important feature of this model is that it proposes that there are two C1-binding sites, one a substrate binding site leading to the production of HOCl and the other an inhibitor binding site which alters the binding of HzOz . Binding of C1-to the inhibitor binding site requires the prior protonation of that site, as the effect is only observed at acid pH values. The rate equation derived from this model predicts that the K,,, for H2O2 will be a function of both pH and C1-concentration with the following stipulations: 1) in the absence of C1-, pH will have no effect upon the K , for H202, and 2) the effect of C1-concentration on the K,,, for H20p will diminish as the pH is raised. These general aspects of the rate equation are consistent with the data presented in Figs. 4 and 5.
A more specific prediction of the rate equation is that there should be a linear dependence of the slopes of the Lineweaver-Burk plots, used to determine the K,,, values for H2O2, on C1concentration. This prediction is also shown to be fully consistent with the data (Fig. 6). Using this analysis, the parameters KI and KH+ were calculated to be 1.2 mM and 30 p~ respectively, the latter value corresponding to a pK, of 4.5.
The pK, value of 4.5 is in agreement with the results of Stelmaszynska and Zygliczynski (12) who followed the spectral changes induced by the titration of an ionizable group on myeloperoxidase. They determined a pK, for the group of 4.4-4.7. Previous models of myeloperoxidase catalysis (10-12) have suggested only one C1-binding site, which requires protonation and binds the substrate chloride. According to these models, the ability of myeloperoxidase to catalyze the oxidation of C1-would decrease as the pH is raised. This is not consistent with the fact that myeloperoxidase can effectively catalyze the chlorination of taurine at pH 7.4 (ll), 3 pH units removed from the pK, of the titratable group of myeloperoxidase. The model presented here, suggesting 2 distinct C1-binding sites, would allow for the full ability of myeloperoxidase to oxidize C1-at pH 7.4, but proposes that it is the inhibition by C1F which decreases with increased pH.
Investigations of several heme-containing proteins have shown a pH-dependent C1-effect on activity and many of these have been related to a pH-dependent C1-effect on the absorption spectrum . Human hemoglobin (19) and Limulus polyphemus hemocyanin (20) have been shown to have a pHdependent C1-inhibition of 0 2 binding, with maximum effectiveness at approximately pH 6.0 and pH 8.5, respectively. Catalase (21), lactoperoxidase (22), and intestinal peroxidase (22), all of which use H202 as a substrate, each demonstrate a pH-dependent C1-inhibition of activity and spectral changes with pK, values of 4.4, 3.5, and 4.75, respectively. The chlorination activity of chloroperoxidase appears to depend upon the protonation of a substrate Cl-binding site (23).
This sampling of heme-containing proteins shows that a pH-dependent C1-effect on activity is not peculiar to myeloperoxidase. Of the peroxidases cited, only chloroperoxidase can also utilize C1-as a substrate (24). The suggestion that the substrate C1-binding site of chloroperoxidase must be protonated is supported by evidence that the chlorinating activity of chloroperoxidase drops off with increasing pH (23). Lactoperoxidase and intestinal peroxidase cannot use C1-as a substrate and yet exhibit a pH-dependent C1-inhibition of activity. Once these enzymes have been fully titrated with H' , no further spectral changes are induced by the addition of C1-(22). This is not the case with myeloperoxidase. Further spectral changes occur with fully protonated myeloperoxidase upon the addition of C1-(12). This is consistent with the hypothesis that myeloperoxidase has two C1-binding sites, a pH-dependent inhibitor C1-binding site analogous to those of intestinal peroxidase and lactoperoxidase and a pH-independent substrate C1-binding site.
The data presented here suggest how the polymorphonuclear leukocyte intraphagosomal environment might affect myeloperoxidase activity. The neutrophil employs multiple mechanisms in the process of killing ingested organisms (25). Of importance in this discussion are the 02-dependent mechanisms which include the production of the oxidants, HZ02 (26), OH. (27), and 0 2 -(28), as well as the production of HOC1 by myeloperoxidase (6). The physiological conditions of the phagosome include a pH of 4.5-5.0 (29) and a C1concentration of 0.1 M (30). While myeloperoxidase does not require an acid pH to kill bacteria (31), an acid pH in the presence of C1-allows myeloperoxidase to function at a higher Hz02 concentration than is possible at a neutral pH. The conditions of pH and C1-concentration in the phagosome may serve to allow greater concentrations of oxidants to accumulate without inactivating myeloperoxidase and thereby maximizing the contribution of each of these O5-dependent mechanisms to the ability of the neutrophil to destroy ingested organisms.
The interaction of myeloperoxidase with H202, C1-, and H' is a complicated system. We have not examined the substrate inhibition by H202 nor does our system allow investigation of the parameters of substrate C1-binding. For some time now, it has been apparent that pH and C1-have a linked effect on myeloperoxidase activity. We have suggested that the simplest explanation of all the data involves the hypothesis that two distinct C1-binding sites exist. The binding of C1-to its substrate binding site is proton independent and leads to the production of HOCI, whereas the binding of inhibitor C1-to its site requires prior protonation of that site and leads to competitive inhibition of H202 binding to myeloperoxidase. Thus, the combined actions of C1-and H' permit myeloperoxidase to function at HzOz concentrations which would be inhibitory under neutral conditions.

APPENDIX
The kinetic model depicted in Fig. 7 is proposed as the simplest model to explain the experimental data determined for native myeloperoxidase. There are several salient features to this model, described below.
1. ES, CES, HES, and HCES are all assumed to be capable of producing product. ES and HES oxidize TMB directly while CES and HCES produce HOC1 which can itself oxidize TMB.

2.
The rates of product formation from ES, CES, HES, and HCES are all taken to be equal.
3. The equilibrium constant for the binding of any ligand to the enzyme is assumed to be unaffected by the presence of any other bound ligand.
4. The inhibitor C1-and the substrate C1-bind to the enzyme a t different sites. The inhibitor C1-can bind only to the protonated enzyme while the substrate C1-can bind to either the protonated or the unprotonated enzyme.
For reasons of clarity, two facets of the enzyme have not been included in this model as they will not affect the analysis of the kinetic data: 1) substrate inhibition by HzOz and 2) binding of TMB to the enzyme for direct oxidation by the ES and HES complexes.
In determining the rate equation which mathematically describes this model, the rapid equilibrium assumptions were made. These assumptions state that the free species and the enzyme-ligand complexes are in rapid equilibrium and that the rate determining step is the formation of product.
The rate equation was derived following the Michaelis-Menten derivation for the simpler model: E + S + ES "+ E + P; as described by Segel (32).
In the present derivation, substrate C1-will be symbolized by C, inhibitor C1-by I, proton by H, and Hz02 by S.
The total enzyme concentration is equal to the concentration of the free enzyme plus the sum of the concentration of each enzyme-ligand complex: The rate of the reaction depends upon the concentration of four enzyme-ligand complexes and the rate constant, kp v = kplEsl + %icEsl + kpIHES1 + %[rrasl