Stimulation of the Activity of Horseradish Peroxidase by Nitrogenous Compounds*

A variety of nitrogenous compounds broaden the activity versus pH profile for the peroxidation of dianisidine catalyzed by horseradish peroxidase (HRP), but not by myeloperoxidase, chloroperoxidase, Escherichia coli hydroperoxidase I, methemoglobin, or micro- peroxidases. The peroxidation of dianisidine catalyzed by cytochrome c peroxidase was affected by the nitro- genous compounds, but to a lesser extent than was the action of HRP. The peroxidations of a variety of phen- ols by HRP exhibited broad activity vers1(8 pH profiles and were unaffected by the nitrogenous compounds. The energy of activation for the peroxidation of dianisidine by HRP was unaffected by changes of pH in the range 6.5-8.5 and was unchanged by the presence of the nitrogenous compounds. The nitrogenous com- pounds markedly increased V,,, for the peroxidation of dianisidine by HRP, but did not change the slope of Lineweaver-Burk plots of kinetic data. These results are accommodated by a mechanism in which nitrogenous compounds hydrogen-bond to the distal histidine of HRP and in so doing raise its pK,. Since the acid form of the distal histidine is thought to facilitate peroxidations catalyzed by HRP by hydrogen bonding to the ferryl oxygen of compound 11, raising its pK, broadens the activity versus pH profile for the peroxidation of anilino substrates, such as dianisidine. We propose that phenolic substrates hydrogen-bond directly to the ferryl oxygen, thus

The mechanism of action of horseradish peroxidase (HRP)' appears to involve divalent oxidation of the ferriheme prosthetic group by HzOz, followed by successive univalent reductions, by the electron donor substrate, as follows (1,2). HRP The rate constant k7 is much greater than k4, and the overall process may be made dependent upon k1 when [ H202 J is ratelimiting or upon k, when the concentration of the electron donor substrate (DHz) is rate-limiting.
We have previously reported (3) that the peroxidation of anilino substrates, such as dianisidine' or p-phenylenediamine, by HRP could be markedly stimulated (boosted) by a variety of nitrogenous compounds, such as ammonia, imidazole, pyridine, and alkylamines; whereas the peroxidation of phenolic substrates, such as guaiacol, was unaffected. This boosting effect was seen only when k4 was rate-limiting and only in the alkaline pH range (pH 7.0-10.0). Indeed, the boosting compounds had the effect of broadening the activity uersus pH curve into the alkaline region, as though the pK. of some activity-limiting ionizable group were being raised. Boosting by nitrogenous compounds required an unshared pair of electrons on nitrogen since ammonia and the mono-&-, and trimethylamines were all active, but tetramethylammonium was not. More recently, the boosting effect of variously substituted imidazoles has similarly been explored (4, 5), but the mechanism of this effect remains a mystery. We now describe further studies of this phenomenon and propose a mechanism.
MATERIALS AND METHODS HRP (RZ = 1.88, 241 units/mg) was purchased from Cooper Biomedical, Inc. and was either used as received or purified to RZ = 3-3.6 by chromatography over Bio-Gel A-0.5m. Purification did not affect responsiveness to nitrogenous boosters. Acidic isozymes of HRP (types VI1 and VIII) and a basic isozyme (type IX) were from Sigma. HRP isozyme C (type I-C, RZ = 3.13) was a gift from Toyobo Biochemicals. Microperoxidases 8, 9, and 11 and chloroperoxidase were from Sigma. The catalase/peroxidase (hydroperoxidase I) of Escherichia coli was isolated as previously described (6,7). Myeloperoxidase was purified from HL-60 cells as described by Anderson et al. (8). The HL-60 cells were generously provided by Dr. W. Lynn. Human hemoglobin (Pentex Biochemicals) was oxidized to the met state with ferricyanide and then dialyzed. Cytochrome c peroxidase was prepared from bakers' yeast (9,10). It exhibited A 4 M n m / A m n m = 0.85. A stock solution of A40nm = 3.3 was diluted a total of 106-fold when ferrocytochrome c was used as the electron donor substrate, but only 250-fold when dianisidine was used. o-Dianisidine from Eastman was twice recrystallized from 95% ethanol and was stored in the dark. Imidazole was from Aldrich, and H,O, was from Mallinckrodt Chemical Works. The concentration of solutions of H,O, was based upon E 2 4 0 n m = 43.6 M" cm" (11). Other reagents were commercially available and of analytical grade.
The peroxidation of dianisidine was followed at 25 "C at 460 nm in reaction mixtures containing 50 mM buffer, 0.52 mM H,O,, 0.36 mM dianisidine rf: boosters. The concentration of peroxidase was varied to compensate for the effect of boosting compounds so that a convenient rate was measured in all cases. Buffering was achieved with sodium acetate (pH 2.3-5.5), sodium phosphate (pH 5.8-8.0), sodium pyrophosphate or borate (pH 8.5-10.0), and sodium carbonate (pH 10 and above).
Dianisidine should be handled with care since it is a carcinogen.  Fig, 1 which presents the results on reciprocal coordinates, demonstrates that imidazole, while sharply increasing the activity of the enzyme, did not increase its apparent affinity for dianisidine. The parallel pattern of lines in Fig. 1 suggests that imidazole could bind to HRP only in the presence of the dianisidine and that it caused parallel increases in the K, value for dianisidine and V, .

Do
Is Boosting a General Property of Dianisidine Peroxidases?-The effect of imidazole on the peroxidation of dianisidine was examined when HRP was replaced by a variety of catalysts of this reaction. Since, in the case of HRP, the boosters have the effect of shifting the alkaline limb of the activity uersus pH profile (3) and since the optimum pH was likely to differ from one catalyst to the next, we examined each catalyst over a range of pH. Myeloperoxidase catalyzed the peroxidation of dianisidine over a wide range of pH values, with optima at pH 5.5 and 9.0; and imidazole exerted only marginally significant effects on its activity at any pH, as shown in Fig. 2. HRP was compared with E. coli hydroperoxidase I; and as shown in Fig. 3, HRP was much more susceptible to boosting by imidazole than was hydroperoxidase I. Thus, lines 1 and 4 in Fig. 3 are the activity uersus pH profiles for HRP without and with imidazole, respectively; whereas lines 2 and 3 are the corresponding profiles for hydroperoxidase I. Methemoglobin can catalyze the peroxidation of dianisidine, albeit not as efficiently as does HRP, and the data in Fig. 4 demonstrate that imidazole slightly suppressed the activity of methemoglobin over most of the pH ranges examined. The effects of imidazole on the activities of chloroperoxidase and microperoxidases 8, 9, and 11 were examined at pH 8.0 and 10.0; and as shown by Fig. 5, in no case was boosting seen. The peroxidation of ferrocytochrome c by the yeast cytochrome c peroxidase was also explored. A sharp optimum in activity was noted at pH 5.0, but imidazole at 10.0 mM exerted only small effects on the alkaline limb of the activity uersus pH profile. These results are shown in Fig. 6. Cytochrome c peroxidase is also able to catalyze the peroxidation of dianisidine, albeit not as well as it does that of cytochrome c. This was examined at pH 8.0 and 9.0 without and with 10 mM imidazole. As shown in Fig. 7, imidazole boosted approximately 3-fold at pH 8.0, whereas there was very little activity at pH 9.0 with and without imidazole. We may note by way of comparison that this level of imidazole stimulates HRP approximately 50-fold at pH 9.0. HRP exists as a family of isozymes (12). It should also be noted that acidic (Sigma type VIII) and basic (Sigma type IX) isozymes of HRP exhibited  virtually identical responses to imidazole over the pH range 6.0-9.5 (data not shown).
It is apparent, from the above, that HRP is almost unique in its response to nitrogenous compounds. Of the several catalysts of the peroxidation of dianisidine which were examined, HRP showed, by far, the greatest boosting by imidazole with cytochrome c peroxidase a poor second. It follows that boosting by nitrogenous compounds reflects some feature of the active site of HRP, rather than an interaction with an intermediate common to the peroxidation mechanism, however catalyzed.
Does Boosting Involve a Change of Mechanism?-One way to approach this question would be to measure energy of activation in the absence and presence of imidazole. The data in Fig. 8, plotted according to the Arrhenius equation, in which case the slope is 2.3 times the energy of activation, demonstrate that the boosting effect of imidazole is not accompanied by a change in energy of activation and therefore in enthalpy of activation. It follows that boosting compounds do not make available a new reaction pathway, with a lower energy of activation, but rather allow the same mechanism, Activation of Horse which occurs optimally at pH 5.5, to occur over a wider range of pH. Arrhenius plots of data collected at pH 6.5, 7.1, and 8.5 demonstrated that the energy of activation for the peroxidation of dianisidine by HRP was invariant with pH in this range (data not shown).
Is the Peroxidation of Phenols Unresponsive to Boosters?-We had previously noted that peroxidation of a phenol, such as guaiacol, by HRP was unresponsive toward boosters (3). This was re-examined using phenols chosen to vary considerably in acidity and in standard redox potential. The peroxidation of these phenols was examined over a wide range of pH with and without imidazole. The data in Fig. 9 demonstrate that HRP-catalyzed peroxidation of phenols was entirely unresponsive toward the boosting effect of imidazole. The data also show that the activity versus pH profiles for the oxidation of phenols extended much further into the alkaline range than was the case for the peroxidation of dianisidine. Indeed, the profiles for the peroxidation of phenols resembled that seen with dianisidine in the presence of imidazole (see Fig. 9).
Nucleophilicity and Steric Factors in Boosting-A variety of compounds were examined for their ability to boost the peroxidation of dianisidine by HRP. Thiocyanate and iodide, tested at 10 mM, were totally ineffective; and acetate was only marginally active (data not shown). Several nitrogenous compounds were tested at 12 mM and were found to differ markedly in their activities. Thus, as shown in Table I, imidazole was most effective, followed by pyridine and dipyridyl;    whereas bipiperidine and quinuclidine were inactive.
Because an unshared pair of electrons on nitrogen appears essential for boosting (3), only the free base form of nitrogenous boosters should be active. Compounds with different basicities would therefore appear to differ in effectiveness merely on this basis. There is also a possible distinction between maximum boosting, seen at an effectively infinite concentration of the boosting agent, and its apparent affinity for HRP, as reflected in the concentration needed to achieve half of the maximum boosting. When the increment in activity was plotted as a function of the concentration of the boosting agent on reciprocal coordinates, straight lines were obtained whose extrapolated ordinate intercept gave the maximum boost and whose slope divided by the ordinate intercept gave the concentration of booster needed for half-maximal boosting ( K J . K , so obtained reflects both the free base and the conjugate acid forms of the booster, and this could be corrected to the value for the free base form by applying the Henderson-Hasselbalch equation.
Data collected in this way are presented in Table 11. When maximum boosting is considered, the compounds which were very active were all comparable. These active compounds were imidazole, pyridine, ammonia, cyclohexylamine, and trimethylamine; whereas piperidine was substantially less active. When apparent affinity of the free base form of the boosting agent for HRP was considered, large differences were apparent. It thus appears that apparent affinity varies from booster to booster, much more than do their maximal effects; and there was no obvious relationship between pK. and K,.
Role of Calcium-HRP has been reported to contain tightly bound Ca2+ which, in the case of isozyme C, could be removed by treatment with guanidinium chloride plus EDTA, with loss of approximately half of the enzymic activity, and which could subsequently be replaced, with restoration of activity (13-16).
Since Ca2+ removal was reported to decrease k4 when measured at pH 7.0, but not at pH 4.4 (15), it appeared possible that ligation of boosting compounds to this Ca2+ on the enzyme might be involved in their stimulatory activity. Treatment of HRP isozyme C with guanidinium chloride f EDTA, followed by dialysis, as described (13-X), did cause loss of activity; but a proportional loss of heme was also evident, in terms of a decrease in absorbance at 403 nm. Moreover, the lost activity could not be restored by treatment with Ca2+, and the residual activity of the Ca2+-depleted enzyme was as susceptible to boosting by imidazole as was the native enzyme. Atomic absorption spectrophotometry was used to verify that the treatment with guanidinium chloride did result in calcium removal from the HRP. These data are summarized in Table  111. It is not clear why we failed to achieve selective removal of Ca2+ by the published procedure, but the parallel loss of heme is not surprising in view of the harsh treatment (6.0 M guanidinium chloride) used in this procedure. The activity which survived this treatment was as responsive to boosting by imidazole as was the native enzyme.

Do Boosting Compounds Shift thepK. of an Activity-limiting
Ionization?-We previously suggested (3) that the action of nitrogenous boosters was to raise the pK, of an activitylimiting ionization. At that time, the only booster which was examined over a wide range of pH was imidazole. It appeared possible that nitrogenous bases with very different pK. values might shift the activity-limiting ionization to different degrees. Pyridine and ammonia differ markedly in pK, and in the concentration of the free base form required for halfmaximal boosting. Accordingly, the boosting actions of ammonia and pyridine were examined from pH 6 to 11. At each pH, the concentration of booster was varied so that the boosting effect could be extrapolated to provide values of the maximum rate possible, as described for Table 11. The data presented in Fig. 10 indicate that ammonia and pyridine shifted the apparent activity-limiting ionization in the same way and to the same limit.

DISCUSSION
Compound I of HRP is thought to contain an Fe(IV)=O in a heme r-cation radical (17-21), and donation of an electron to compound I reduces the r-cation radical, leaving an Fe(IV)=O heme, as compound I1 (20, 22-27). The rate of reduction of compound I1 by electron donor substrates (i.e. k4) decreases with increasing pH (28-32), and this effect of pH seems due to titration of a catalytically important residue, rather than to changes in the net charge on the enzyme (33). One histidine residue is thought to be ligated to the heme iron and is referred to as the proximal histidine. Another histidine residue appears to hydrogen-bond to the ferryl oxygen atom of compound I1 and is called the distal histidine (34-36). Hydrogen bonding from the imidazolium ring of the distal histidine facilitates rapid exchange of the ferryl oxygen with water, and the rate of this exchange correlates with the rate of the catalytic process (34). Absence of hydrogen bonding to the ferryl oxygen, as shown by lack of an effect of pH on the Fe(IV)=O stretching frequency, is seen with metmyoglobin and correlates with a lack of significant peroxidase activity

(35).
It thus appears that hydrogen bonding from the distal imidazolium to the ferryl oxygen of compound 11 is important for the reduction of compound I1 by the electron donor substrate, and the hydrogen from the distal imidazolium probably leaves with the ferryl oxygen as OHwhen electron transfer from the donor substrate has occurred. The decrease in k, with increasing pH may thus be attributed to ionization of the distal imidazolium. Since the effect of nitrogenous boosters was seen only on the alkaline limb of the activity versus pH curve, these compounds must elevate the pK. of the distal imidazolium. They might exert this effect by hydrogen bonding to it, as shown in Fig. 1lA. In this figure, pyridine is shown hydrogen-bonded to the distal imidazolium of compound 11, and the electron donor substrate (aniline) is shown donating an electron via the heme edge. The electron from the anilino donor substrate would be conducted to the ferryl oxygen through the heme, and the hydrogen-bonded hydrogen from the distal imidazolium would then leave with that oxygen as OH-.
This scheme can account for the boosting of the peroxidation of dianisidine and other anilino substrates by nitrogenous compounds, but not for the lack of boosting seen with phenolic substrates. Paul and Ohlsson (37) have suggested that substrates for HRP fall into two groups, i.e. phenolic and anilino, with the former class of substrates exhibiting greater affinity for the enzyme. They interpreted this as being due to hydrogen bonding between the phenol and the enzyme. Roles for hydrophobic interactions and for hydrogen bonding in the binding of phenols to HRP were also proposed by Schejter et al. (38). In studies of the reaction of phenols with HRP compound II, Dunford and Adeniran (39) noted that p-aminophenol did not fit the Hammett plot for phenols and suggested that its preferred binding orientation favored electron donation from the amino, rather than from the phenolic, group.
We apply these findings to our data by suggesting that phenolic substrates displace the distal histidine from the ferry1 oxygen and, in turn, hydrogen-bond directly to that oxygen, as depicted in Fig. 11B. This displacement of the distal histidine has the consequence that ionization of its imidazolium is no longer rate-limiting. Hence, the enzyme should be active over a wider range of pH with phenolic substrates than with anilino substrates; and the reduction of HRP compound I1 by phenols should not be subject to boosting by nitrogenous compounds which can hydrogen-bond to the distal imidazolium since its pK, is no longer relevant. Fig. 11 appears to accommodate the available data, yet one would expect that nitrogenous boosters of different basicity should shift the pK. of the distal histidine to different degrees. The data in Fig. 10, however, show that ammonia and pyridine appeared to shift the activity-limiting ionization to the same degree. This apparent discrepancy between the expectations from the proposed mechanism and the data can be accommodated. Thus, suppose that both of these boosters shift the pK, of the distal histidine to such a high value that ionization of one or more other groups then becomes rate-limiting before titration of the distal histidine occurs. Fig. 11 explains the kinetic data shown in Fig. 1. Thus, the compound I1 form of HRP would be produced only in the presence of HzOZ plus dianisidine. Since it is only this form whose catalytic activity should be increased by the booster, it would appear that the booster (imidazole) could bind only in the presence of the donor substrate. Moreover, since a ratelimiting step of the catalytic cycle other than substrate binding was increased by the boosting compound, one would expect a parallel increase in K,,, for dianisidine and V,. Fig. 11 also explains the constancy of energy of activation f booster, which was shown in Fig. 8 In effect, at any given pH in the alkaline range, the booster increases the fraction of HRP compound I1 whose distal histidine is protonated and so increases rate, but does not change mechanism. The inactivity of some nitrogenous bases, such as quinuclidine and bipiperidine (Table I), and the very wide range of apparent affinities of boosters for HRP compound I1 (Table 11) could be explained in terms of ease of access of these compounds to the active site.