Presence of endogenous calcium ion and its functional and structural regulation in horseradish peroxidase.

The endogenous calcium ion (Ca2+) in horseradish peroxidase (HRP) was removed to cause substantial changes in the proton NMR spectra of the enzyme in various oxidation/spin states. The spectral changes were interpreted as arising from the substantial alterations in the heme environments, most likely the heme proximal and distal sides. The comparative kinetic and redox studies revealed that these conformational changes affect the reduction process of compound II, resulting in the decrease of the enzymatic activity of HRP. It is also revealed from the ESR spectrum and the temperature dependences of the NMR and optical absorption spectra of the Ca2+-free enzyme that the heme iron atom of the Ca2+-free enzyme is in a thermal spin mixing between ferric high and low spin states, in contrast to that of the native enzyme. These results show that Ca2+ functions in maintaining the protein structure in the heme environments as well as the spin state of the heme iron, in favor of the enzymatic activity of HRP.

The endogenous calcium ion (Ca"') in horseradish peroxidase (HRP) was removed to cause substantial changes in the proton NMR spectra of the enzyme in various OxidationJspin states. The spectral changes were interpreted as arising from the substantial alterations in the heme environments, most likely the heme proximal and distal sides. The comparative kinetic and redox studies revealed that these conformational changes affect the reduction process of compound 11, resulting in the decrease of the enzymatic activity of HRP. It is also revealed from the ESR spectrum and the temperature dependences of the NMR and optical absorption spectra of the Ca2+-free enzyme that the heme iron atom of the Ca2+-free enzyme is in a thermal spin mixing between ferric high and low spin states, in contrast to that of the native enzyme. These results show that Ca2+ functions in maintaining the protein structure in the heme environments as well as the spin state of the heme iron, in favor of the enzymatic activity of HRP.
Horseradish peroxidase (HRPl) is a heme enzyme catalyzing the oxidation of a wide variety of aromatic molecules by hydrogen or alkyl peroxides. The enzymatic reaction cycle has been established and normally proceeds by the following mechanism (1,2), HRP + HZ02compound I + Hz0 (1) compound I + AH2compound I1 + AH ' k7 (2)

kd
(3) 2AH ' -A2Hz or A + AH2 kd (4) where compound I and compound I1 represent the enzymatic reaction intermediates, and AH, the reducing substrate. This enzyme has been established to contain iron(II1)-protoporphyrin IX as a noncovalently bound prosthetic group. The fifth ligand of the heme iron is now confirmed to be a histidyl imidazole (3,4), and the sixth coordinated position has been suggested to be occupied by a water molecule (5). The distal histidyl imidazole situates within the hydrogen-bonding dis-kl compound I1 + AH2 -HRP + AH' *This research was supported by Grants-in-Aid 60540285 and 60790122 for Scientific Research from the Ministry of Education, Japan. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
However, a recent x-ray crystallographic study on cytochrome-c peroxidase (5,6), an analogous protein to HRP, revealed that the geometry of the distal histidine in peroxidase is not proper for the hydrogen bonding to the water coordinated to the heme iron, while the distal imidazole in metMb is normally hydrogen bonded to the bound water molecule. Indeed, there is reason to believe that the water ligand at the sixth position in HRP is less strongly bound and in a somewhat different environment than it is in metMb. For example, results from absorption (7), ESR (8,9), resonance Raman (lo), and nuclear magnetic relaxation (11) measurements have indicated that in HRP the axial water ligand is either absent or loosely bound. However, as was displayed on the Evans and Sutherland Picture System, the distal histidine of peroxidase is oriented to readily hydrogen bond with the ironcoordinated peroxides, resulting in a heterolytic cleavage of the peroxide 0-0 bond by its acid-alkaline catalysis. Furthermore, the x-ray analysis also showed that in peroxidase the proximal histidine contacts and probably hydrogen bonds with the side chain of glutamine, which in turn interacts with the buried carboxylate group of glutamic acid, while in metMb the residue is hydrogen bonded to a backbone carbonyl oxygen atom. This former set of interactions in peroxidase may impart sufficient aniopic character to the proximal histidine to stabilize higher oxidation states of the heme iron during the peroxidase catalytic cycle (6, 12). These arguments seem to indicate that the specific structures in the heme environments may regulate the activities of the heme prosthetic group in peroxidases.
Recently, Haschke and Friedhoff (13) and present authors (14) reported that HRP contains 2 mol of Ca2+/mo1 of enzyme and that removal of the bound Ca2+ from the enzyme causes a 2-fold decrease in its enzymatic activity. They also studied the effect of Ca2+ binding on the thermal &ability of the protein. These results allowed us to expect that Ca2+ binding by HRP is essential in maintaining the protein structure which is suitable for the peroxidase activity. However, more detailed studies including kinetic and spectroscopic measurements have not yet been carried out to see the calcium binding to HRP in relation to its catalytic mechanism. Furthermore, the effect of the calcium binding to the enzyme on the spin state of the heme iron and heme microenvironmental structure have remained open to further studies.
We have attempted here to study kinetics and redox potential measurements of Ca2+-free HRP and to delineate the Ca2+-binding effect on the catalytic process of the enzyme in more detail by comparing these results with those previously reported for the native enzyme. The relevance of the present results to the structural effect of calcium binding revealed by 'H NMR, ESR, and absorption spectral measurements is also discussed to shed light on the relationship of HRP between the specific structure in the heme environment and its characteristic function.

MATERIALS AND METHODS
Enzyme Preparation-Horseradish peroxidase type VI (RZ = 3.2) was purchased as a lyophilized salt-free powder from Sigma. Calcium removal from the enzyme was achieved by the method of Haschke and Friedhoff (13). The Caz+-reconstituted enzyme was prepared by adding CaClz to the Ca*+-free enzyme solution or incubation of the Ca2+-free enzyme against CaC12 solution, and purified by chromatography on CM-52. The concentration of the enzyme was determined spectrophotometrically by using molar absorptivity of 102 cm" mM" at 403 nm and pH 7 (15). The cyanide or benzohydroxaminic acid complex of the enzyme was prepared by adding a 5-fold excess amount of the ligand or the substrate to the enzyme solution, respectively. The ferrous enzyme was generated by addition of sodium dithionite under N,.
The specific activities of the enzymes were measured spectrophotometrically by following the 480-nm absorption accompanied by the oxidation of o-aminophenol in the presence of hydrogen peroxide at 20°C.
Kinetics and Redox Potential Measurements-The reaction rate constants of the enzymes, kl, k,, and k,, were measured with a Union Giken rapid reaction analyzer RA-1300 at 20 "C. The measurements of the k7 and k4 values were carried out within 3 min after compounds I and I1 were generated. The spectrophotometric wavelength used to monitor the reaction was an isosbestic point between the other two.
PABA, which was obtined from Nakarai Chemical Co. as its potassium salt, was used as a substrate for compound I1 formation.
Its concentration was monitored spectrophotometrically in phosphate buffer with the molar absorptivity of 14.5 mM" cm" at 265 nm and pH 7.6 (16). Hydrogen peroxide was stored and analyzed as described previously (17). Deionized water obtained through Organo G-10 with extremely small electric conductivity was distilled twice for kinetic measurements.
The oxidation-reduction potentials for the ferric-ferrous couple of HRP were measured by the methods described previousIy (18). Sodium dithionite and ferricyanide were employed as a reductant and an oxidant, respectively. Spectral Measurements-Proton NMR spectra were recorded at 300 MHz on a Nicolet NTC-300 Fourier transform NMR spectrometer equipped with a NMC-1280 computer system. For recording the typical proton spectra, 5,000-15,000 transients were accumulated to obtain the Fourier transformed spectra with 8,000 data points and a 5.791s 90" pulse. Large proton peaks of the solvent water were suppressed by a 500-ps low power 180" pulse prior to the high power observe pulse. Proton chemical shift is referenced with respect to the proton signal of H20 in the protein solution, assigning a positive value for the low-field resonance.
All the sample solutions for NMR spectral measurements were prepared in appropriate buffer solutions between pH 5 and 10 such as PIPES except phosphate buffer, which contaminates calcium ion.
Protein concentration was about 1.0-2.0 mM in 90% Hz0 and 10% 'H20. The pH titration was performed by direct addition of 0.1 N NaOH or HCl to the enzyme solution in the NMR sample tube. The pH values were measured with a radiometer model PHM-100 pH meter, equipped with an Ingold microcombination glass electrode.
The visible absorption spectra were recorded with a Union-Giken SM-401 spectrometer by using a cell with a I-cm path length.

RESULTS
Kinetic and Redox Measurements-The bound Ca2+ in HRP can be readily removed from the enzyme by incubation of the enzyme in the presence of guanidine hydrochloride and EDTA as confirmed by atomic absorption spectroscopic measurements. Before and after the removal of the bound Ca2+, we measured the enzymatic activity of HRP at pH 7. The activity of the enzyme was reduced to approximately 40% after calcium removal and returned to 80% of the initial activity upon reconstitution with Ca". When the enzyme was purified by the use of CM-cellulose after addition of Ca2+ (Ca2+-reconsti- tuted enzyme), the activity was almost restored.' This result is the same as that reported by Haschke and Friedhoff (13).
In order to clarify the effect of Ca2+ removal on the enzymatic reaction of HRP, the kinetic measurements were made at pH 7 for the native (Ca2+-bound), Ca2'-free, and Ca'+reconstituted enzymes of HRP and analyzed under pseudofirst order approximation (19)(20)(21). The rate constants, kl, k7, and k4, determined by dividing the pseudo-first order constants by substrate concentrations are listed in Table I. Inspection of the table shows that significant differences are formed for all the constants between the native and the Ca"free enzymes and were almost recovered upon Ca2+ reconstitution. With these rate constants, we calculated the specific activities of these enzymes on the basis of the expression relating the specific activity and the rate constants under the steady-state condition (22, 23).

2klk7k,[HRP][PABA][Hz02]
The calculated value for the Ca2'-free enzyme is 47% of the native one, which is quite consistent with those obtained from the static activity measurement. This result indicates that the differences in the rate constants arise from Ca'+ removal.
Upon Ca" removal, substantial change of kg, which was reduced to 44% of the initial value, is more remarkable, compared with kl and k,. Keeping in mind that the k4 process is a rate-determining step in the enzymatic cycle of HRP and thus governs the overall activity, it seems reasonable to conclude that the decrease in the k4 may be predominantly responsible for a loss of the enzymatic activity upon Ca'+ removal. Then, in order to see the effect of Ca2+ removal on k h value in more detail, the oxidation of PABA by compound I1 was examined for the native, Ca"-free, and Ca2+-reconstituted enzymes over a wide range of substrate concentration   Michaelis constant), are compiled in Table 11. In this analysis, significant changes are exerted for k4.1 and k4.2 at pH 7.0 before and after the Ca2+ removal, while those at pH 4.4 are little altered. However, it is worthy to note that I/&, which represents the binding constant of PABA to the enzyme, was unchanged at both pH values upon the Ca2+ removal.
We also measured the redox potentials of HRP values. Fig.   1 shows the potentiometric titration curves of the ferrous/ ferric couple for Ca2+-free and native HRPs. The value of E' for the Fe(III)/Fe(II) couple was increased from -273 to -244 mV upon the Ca2+ removal. The E' value for the native enzyme was reproduced upon incorporation of CaZ+ into the Ca2+-free enzyme. To our regret, we failed to measure E' for the compound II/ferric couple which directly relates to the value of kq, due to the instability of compound I1 and the resting enzyme of Ca2+-free HRP in an alkaline pH region.
Spectral Measurements- Fig. 2 illustrates the hyperfineshifted proton NMR spectra of the Ca2'-free and the native enzymes in a ferric resting state in HzO. For the native enzyme, the four heme peripheral methyl proton signals are observed at 77.6, 70.1, 67.0, and 50.1 ppm downfield from the H20 resonance, while the signals for the Ca2+-free enzyme are located at 70.9, 63.3, 61.1, and 46.1 ppm. We also observed a broad single-proton peak at 95.2 ppm in the spectrum of native HRP in HzO, which was assigned to the proximal histidyl imidazole NIH by La Mar and de Ropp (3). In the case for the Ca2+-free enzyme, this exchangeable proton signal was observed at 88.3 ppm. The NMR spectrum of the native enzyme was reproduced upon an addition of 1 mol of Ca2+/ mol of the Ca2+-free enzyme. (This will be discussed in detail in a subsequent paper.) The NMR spectrum of Ca2+-free HRP exhibited a different temperature-dependent behavior from that for the native enzyme, showing that the spin state of the heme iron of the Caz+-free enzyme is much different from that of the heme iron of the native one (14). In order to clarify the iron spin state of Ca2+-free HRP, its ESR and optical absorption spectra at low temperature were examined. Fig. 3 shows the ESR spectra of the Ca2+-free and the native enzymes at liquid helium temperature. In the spectrum of the Ca2+-free enzyme, the ferric high spin signal at g = 6 and the low spin signals at g = 2.96, 2.07, and 1.59 are concomitantly observed, in sharp contrast with that of the native enzyme characteristic of ferric hemoproteins in predominantly high spin state. When Ca2+ was added to the Ca2+-free enzyme, the low spin signals disappeared and the high spin signals grew up again, yielding the spectrum of the native enzyme. In Fig. 4 is illustrated temperature dependences of the optical absorption spectrum of Ca*+-free HRP. As the temperature rises from 83 to 193 K, the intensities of the high spin bands at 498 and 638 nm increase and that of the low spin band at 535 nm decreases. The spectrum reversibly changed with three isosbestic points with respect to raising or lowering temperature. On the contrary, the spectrum of the native enzyme was insensitive to changes in temperature between 83 and 193 K.
The ESR and absorption spectral features of Ca2+-free HRP are closely similar to those of the derivatives of metMb, catalase, and cytochrome-c peroxidase where the heme irons are in a thermal spin equilibrium between ferric high and low spin states. To confirm this view, we attempted to obtain the thermodynamic parameters for the equilibrium of the Ca2+free enzyme by analyzing the absorption temperature curve according to the method described previously (24). The analysis afforded the equilibrium constant for this spin conversion, which is defined as (low spin)/(high spin), to be 63 at 83 K and 0.95 at 193 K. The thermodynamic quantities A H o and ASo, were also computed to be -1200 cal/mol and -6.4 e.u., respectively. The values of A H o and ASo for Ca*+-free HRP appear to be close to those for hemoprotein derivatives where the heme iron is in the thermal spin mixing state, e.g. metMb. OH-(-1200 cal/mol, -5.8 e.u.), metMb-Ng (-2740 cal/mol, -6.8 e.u.), and metMb-imidazole (-3040 cal/mol, -5.8 ea.) (25).
In order to gain further insight into Ca2+ binding by HRP, we obtained the proton NMR spectra in various spin and oxidation states. proton peaks of the cyanide complex of the Ca2+-free enzyme are observed at 21.0 and 26.8 ppm downfield from the Hz0 resonance, whereas those of the cyanide complex of the native enzyme are located at 21.5 and 26.6 ppm. Some minor spectral differences are also noticeable between these spectra in the 5-20 ppm range. Upon addition of Ca2+ to the cyanide complex of the Ca2+-free enzyme, the signals of the Ca2+-free and the native enzymes were concomitantly observed when a halfsaturated amount of Ca2+ was added. In Fig. 6A are compared the proton NMR spectrum of ferrous Ca2+-free HRP with that of the native enzyme. The heme peripheral proton signals of the enzyme are observed at 22.6, 17.9, 14.2, and 8.5 ppm for the native enzyme, while at 20.0, 15.8, 9.2, and 6.3 ppm for the Ca2+-free one. The broad and single-proton peak for the ferrous native enzyme, assigned to proximal histidyl NIH signal, is observed at 74.0 ppm (26,27), whereas the one for ferrous Ca2+-free enzyme appears at 79.7 ppm. Fig. 6B shows the corresponding absorption spectral change in which the Soret peak shifted from 437 to 428 nm upon calcium removal. The NMR and visible spectral changes were reversible with respect to removal or recombination of Ca2+.
In order to reveal the heme microenvironmental structural changes induced by calcium removal, we have also examined pH dependence of the spectra for Ca2+-free HRP with recourse to proton NMR measurements. In Fig. 7A is shown the downfield portions of proton NMR spectra of reduced HRP in H 2 0 as a function of pH from 5.2 to 9.0. The pH-dependent spectral changes for Ca2+-free HRP are similar to those for the native HRP reported by La Mar and de Ropp (27). There are two pH-induced interconvertible protein conformers as is manifested by pH-induced spectral alterations. However, integration of individual methyl resonances or the proximal histidyl NIH signal as a function of pH yields a pK of 5.5, which is associated most probably with the ionization of the distal histidine. With further raising of the pH, all hyperfine shifted NMR signals decreased in intensities and finally disappeared at pH 9.0. These spectral changes were reversible with respect to raising or lowering pH. Disappearance of the hyperfine shifted resonances above pH 9 may suggest that the heme iron in ferrous Ca2+-free HRP becomes diamagnetic in the alkaline region, which is in sharp contrast to ferrous native HRP of which proton NMR spectrum was insensitive to the pH variation in this pH region. Corresponding pHdependent alteration in absorption spectrum is shown in Fig.  7B. With raising pH, there is a steep increase in intensity of low spin band at 425,530, and 560 nm. The resultant spectrum at pH 9.7 bears a strong resemblance to that of diamagnetic pyridine hemochromogen, which is in good agreement with the proton NMR result stated above (28). Fig. 8 illustrates the effects of Ca2+ removal from the enzyme on the proton NMR spectrum of the HRP-substrate complex. In the spectrum of the benzohydroxaminic acid complex of the native enzyme, the heme methyl proton signals are observed at 81.5, 70.0, and 51.5 ppm downfield from the Hz0 resonance, while they are located at 78.9, 68.7, 66.7, and 50.6 ppm for the benzohydroxaminic acid complex of the Ca2+-free enzyme. We also found that the proximal histidyl NIH signal at 97.9 ppm is shifted upfield by 0.9 ppm upon removal of Ca2+ from the enzyme.

DISCUSSION
A current fundamental question to be addressed by structural investigations on hemoproteins is how the protein structures regulate the properties of the heme prosthetic group, e.g. peroxidase activity in one case and reversible oxygen binding capability in the other. The x-ray crystallographic studies of cytochrome-c peroxidase and metMb revealed that the structure in the heme environment is subtly different between peroxidase and oxygen-carrying proteins (5, 6). For example, the proximal histidine in peroxidase has uniquely an anionic character resulting from the stronger hydrogen bonding of its NIH with the nearby proton acceptor residue, probably glutamate. Furthermore, it is also noteworthy that the water ligand at the heme sixth site in peroxidase is either absent or loosely bound, while it occupies the heme sixth site of metMb. Such a subtle conformational difference is suggested to contribute to the differences in the physicochemical properties of these hemoproteins. Here, a second question may arise as to how the specific structures in the heme environment of peroxidase can be maintained to favor the peroxidase activity. In spite of the extensive studies for peroxidases such as cytochrome-c peroxidase, HRP, and so on, a knowledge on this point is not sufficiently accumulated for understanding the structure-function relationship of peroxidases.
The present spectroscopic data for Ca2+-free HRP such as proton NMR, ESR, and optical spectra revealed that the heme environmental structure of Ca2+-free HRP is substantially different from that of the native enzyme in various iron spin/oxidation states , and kinetic results showed that Caz+ removal from HRP affects all the rate constants in the catalytic cycle of the enzyme (Tables I and 11). These results allow us to expect that Ca2+ binding is essential for the structural and functional characteristics of HRP. Bearing in mind that the Ca2+ removal from HRP reduces its enzymatic activity, thermal stability (13), and compound I stability (14), all the pn%ent results and discussion may provide us with a clue to unveil the relationship between the heme microenvironmental structures and the catalytic function of HRP. The spectral and kinetic observations will be discussed in structural terms relevant to the peroxidase activity.
The structural changes of HRP in the heme vicinity upon Ca2+ removal are sensitively manifested in the proton NMR spectra as the spectral shift of the hyperfine-shifted proximal histidyl NH resonance for a ferrous high spin form (27,29). As shown in Fig. 6A, the downfield shift by 6 ppm of this signal induced by Caz+ removal from the enzyme significantly reflects the subtle perturbation at the heme proximity. Since the heme iron of ferrous form enzyme has no sixth coordination ligand and eventually has no direct interaction with the distal histidine, the spectral change in the NH signal may result from the changes in the binding profiles of the proximal imidazole such as bond compression or tilting of the ironimidazole bond or modulation of the NH hydrogen bonding. The structural change in the proximal side was also supported by the NMR spectral change of the proximal histidyl CZH signal by 2 ppm for the ferric low spin cyanide complexes (Fig. 5), which has been suggested to be a probe for the extent of the proximal histidine NIH hydrogen bonding (30, 31). Furthermore, the NMR spectral shift for the proximal NIH of the HRP-benzohydroxaminic acid complex in a pure ferric high spin state appears to be also consistent with this suggestion. This implies that the conformational changes in the heme proximal side associated with the Ca" removal from the enzyme could occur irrespective of spin/oxidation states of the heme iron.
These results remind us of the previous studies by Kastner and co-workers (32-34) which have emphasized the influence of protein structure on the spin equilibrium through a control of axial ligation. More recently, Neya and Morishima (35) suggested on the basis of paramagnetic NMR shifts for heme.N;. (1-or 2-methylimidazole) model systems that the  binding nature of proximal ligand to the heme iron could some change of the binding nature of proximal histidine is modulate the spin equilibrium of the iron. Therefore, the induced by removal of Ca2+ from the enzyme. Other structural structural change in the heme proximal side, most probably factors such as van der Waals interaction between the heme the proximal imidazole, may be responsible for the spin equiperipheral and peptide chains and effects from the distal side librium of the Caz+-free enzyme in a ferric spin state, in which may account for the iron spin state change of the ferric enzyme, and these possibilities could not be necessarily ruled out at the present stage. In any event, it is evident that Ca2+ removal elicits the structural perturbation in the heme proximal side. Further insight into the conformational change in the heme vicinity was gained on the basis of the pH dependence of the NMR spectra of the reduced enzyme. It has been well known that the spectral change of native HRP in a ferrous high spin is attributable to the ionization of the distal histidine (27). The pK value of the distal histidine is lowered by 2 pH units although the NMR spectral feature of ferrous Ca2+-free HRP associated with its ionization is similar to that of the native enzyme. Furthermore, it is to be noted that the drastic spectral change is encountered in the alkaline region as is manifested by proton NMR and absorption spectra (Fig. 7). This result indicates that some internal nitrogenous ligand from the distal amino acid residue, most likely the imidazole group of the distal histidine, occupies the sixth coordination position of the heme iron, thereby producing the diamagnetic species like pyridine hemochrome. The above spectral change was not observed for the native enzyme in the same pH region, and the Ca2+ reconstitution converted the above spectral changes to those for native HRP. It is thus tempting to suggest that Ca2+ removal causes the conformational changes in the heme distal side, which appears to be more drastic than that in the proximal side. These conformational changes in the heme vicinity induced by Ca2+ removal may substantially affect the enzymatic activity of HRP.
We then discuss these structural changes upon Ca2+ removal in the heme environment with relevance to the unique changes in the reaction rates of HRP; k4 experiences a substantial change as compared with kl and k7. The k4 process involves two reaction steps, the binding of a substrate to the protein moiety and the subsequent reduction of the heme iron from ferry1 to ferric states. It is reasonable to assume that the

Structural Regulation by Ca2+ in HRP 9389
former and/or the latter steps are disturbed by the structural change induced by calcium removal. In other words, the Ca2+ removal may alter the structures relevant to these processes. Indeed, the redox property of HRP was significantly altered upon Ca2+ removal, as manifested as a cathodic shift by 30 mV of the redox potential for the Fe(III)/Fe(II) couple. However, a little change in the binding constant of PABA by compound I1 at pH 7 (I/&) allows us to expect that the protein structural change around the substrate-binding site may be insignificant (Table 11). This is also supported by the measurement of the binding constant of indolepropionic acid to HRP on the basis of the proton NMR spectral shifts of the heme methyl of the enzyme (36), which was not affected by calcium removal. These findings thus suggest that the change of the reaction rate k4 induced by calcium removal, which is directly related to the loss of the specific activity, is essentially caused at the reduction step of the heme iron rather than at the step of substrate binding. The redox potential of HRP for the Fe(III)/Fe(II) couple is characteristically low (-273 mV), as compared with metMb (+50 mV) (37). This feature is invariant for peroxidases (38). Mechanistically this requirement for a lowered redox potential in peroxidases may be understandable, since the Fe(1V) and/or porphyrin *-cation radical in compound I must be stabilized (39). These unique redox properties of peroxidases have been primarily explained as due to the anionic character of the proximal imidazole. The validity of this view could be assessed by Doeff and his co-workers (40) in the study of the model complex hemin-(imidazole)2 and 1,lO-phenanthroline system, where the redox potential of the heme iron is modulated by variable strength of a hydrogen bond of the ironcoordinated imidazolyl NH. Thereby, the increase of redox potential of HRP upon Ca2+ removal from -273 to -244 mV ( Fig. 1) may be interpreted in terms of the structural change in the proximal side such as a decrease of an anionic character of proximal imidazole. This conformational change, if any, appears to be reasonably responsible for instability of compound I of Ca2+-free HRP relative to native compound I (14).
However, one may ask why there is no significant difference in k4 at pH 4.4 between native and Ca2+-free HRPs (Table 11) in spite of the different nature of the heme fifth ligand as inferred from the different proton NMR spectra. Thus, we must take into account another possible cause which is more predominantly affected upon Ca2+ removal and governs the redox property of HRP rather than the proximal histidine. Yamazaki and his co-workers (39,41,43) have postulated that the protonation of the distal His is significantly related to the redox property of HRP on the basis of the extensive studies including proton balance, redox potential and kinetic measurements for the individual reaction intermediates of HRP. It has been also revealed that distal histidine is associated with the pH-dependent features of E' (compound 11/ ferric) and the rate of the reaction of compound I1 with substrate. As mentioned above, the Ca2+ removal drastically altered the structure of HRP at the heme distal side and resulted in a decrease in the basicity of the distal histidine. This structural change may be responsible for the change of the redox potential of the enzyme and eventually for a decrease of the rate constant k4.
A variety of studies including the x-ray analysis of cytochrome-c peroxidase (5, 6) suggested that the distal histidine of peroxidase plays a crucial role in its enzymatic reaction. The proposed mechanism for peroxidase-catalyzed heterolytic cleavage of the peroxide 0-0 bond to generate compound I suggested that the distal histidine serves as a acid-alkaline catalyst to facilitate the transfer of a Droton from one oxwen 6. 7.

19.
atom of peroxide to the other one (6, 12). One might expect the histidine residue to form a hydrogen bond with the oxygen atom of peroxide. In fact, disruption of hydrogen-bonded lattice caused by selective and reversible modification of its imidazole group with p-chlorobenzoylchloride decreases or abolishes the enzymatic activity of HRP (9,44), although the modification induces no significant structural changes in the heme environment (45). Thus, the structural changes in the heme distal side as well as the change of the imidazole basicity of the distal histidine possibly affect kl and k4. Moreover, it is also likely that these structural alterations are accompanied by the change in the distance between the heme and the substrate as is manifested as a different ratio of k4.1 to k4.* between Ca2+-free and native HRP (Table 11).
As mentioned so far, the endogenous Ca2+ binding maintains the heme environmental structures of HRP, the heme proximal and distal side, which is relevant to its activity. The Ca2+ titration to the Ca2+-free enzyme followed by proton NMR spectra indicates that one Ca2+ is essential for the protein structure of HRP, although the enzyme contains 2 mol of Ca2+/mo1 of the enzyme. It thus follows that one Ca2+ is a structural factor to favor the peroxidase activity of HRP. However, there is no evidence for identification of the Ca2+binding sites. To gain insights into the different roles of two Ca2+ ions and the structures at the Ca2+ site, the metal substitutions for Ca2+ in HRP were carried out and will be shown in the subsequent paper.