ARCHIVES

Vol.

OF BIOCHEMISTRY

283, No. 2, December,

AND

BIOPHYSICS

pp. 351-355,

1990

Axial Ligand Coordination Masao

Ikeda-Saito*T’

*Department TLaboratory

Received

May

and Shioko

in Intestinal Peroxidase

Kimurat

of Physiology and Biophysics, Case Western Reserve University of Molecular Carcirwgenesis, National Cancer Institute, National

15, 1990, and in revised

form

August

10, 1990

EPR spectra

of intestinal peroxidase are reported for the first time. The resting state of intestinal peroxidase exhibits only a high spin EPR spectrum with pH-dependent rhombicity. Addition of chloride shifts the equilibrium between an acidic and a neutral form of the enzyme. In contrast, resting lactoperoxidase shows EPR spectra of both low spin and high spin species, indicating a different heme environment between these two peroxidases. The high spin signal of lactoperoxidase consists of multiple components; the major component exhibits pH-dependent rhombicity similar to intestinal peroxidase and the equilibrium between the acidic and the neutral forms is also shifted by chloride ion. EPR features of the low spin cyanide complex of intestinal peroxidase and lactoperoxidase are compared with those of other hemeproteins, whose proximal axial ligands are known to be histidine residues. The g-values of the cyanide adducts of the mammalian peroxidases are similar. The relationship between the g-value anisotropy and imidazolate charo 1990Academie acter of the proximal histidine is discussed. Prem,

Inc.

Intestinal peroxidase is a heme enzyme found in mucus of the small intestine (1). The enzyme showed optical spectral properties similar to those of lactoperoxidase, and these two enzymes show properties which are somewhat different from those of common iron protoporphyrin IX containing hemeproteins (l-3). Resonance Raman studies (4-6) indicated that intestinal peroxidase and lactoperoxidase possessthe same type of common heme group. The structure of this heme has been recently proposed to be a modified protoheme with a thiomethylene group at the 8 position (7). A “heme-linked” ionizable group controls the optical spectra and reactivities of both intestinal peroxidase and lactoperoxidase (2). The apparent pK, value of this ionizable group is lower in lactoperoxidase 1 Supported by NIH whom correspondence

Research Grants GM39492 should be addressed.

0003.9861/90 $3.00 Copyright 0 1990 by Academic Press, All rights of reproduction in any form

School of Medicine, Cleveland, Ohio 44106; and Institutes of Health, Bethesda, Maryland 20892

and GM39359.

To

(p& = 3.5) than in intestinal peroxidase (pK, = 4.8). In both peroxidases, addition of chloride raises the apparent pK, value of this acid-neutral transition. Since optical absorption and resonance Raman properties of intestinal peroxidase and lactoperoxidase are similar to each other (l-6), we expected that intestinal peroxidase also would show EPR spectral characteristics similar to those of lactoperoxidase. As EPR spectra of ferric heme proteins contain information related to the structure at the metal site (8), we have initiated an EPR study of intestinal peroxidase to probe its active site structure with reference to other heme proteins. We report herein the first EPR spectra of intestinal peroxidase and its complexes. EXPERIMENTAL

PROCEDURES

Hog intestinal peroxidase was prepared as described previously (1, 2). The enzyme preparation had A11Znm/A280nmof 0.9, which was comparable to 0.906 reported previously (1). Lactoperoxidase (from cow milk) was purified from a commercial preparation (Sigma, L-2005). Lactoperoxidase in 10 mM phosphate buffer, pH 6, was charged to a column of CM Sepharose CL-6B equilibrated with the same buffer, and eluted with a linear gradient from 10 mM phosphate buffer, pH 6, to 100 mM phosphate buffer, pH 7. The greenish brown-colored fractions by with &Z m/A~80 nm values larger than 0.8 were pooled, concentrated ultrafiltration (Amicon Centricon 30), and gel-filtered on a column of Sephacryl S-200 equilibrated with 0.1 M phosphate buffer, pH 7. Fractions with A 412nm/Am nm values larger than 0.9 were pooled and concentrated as above. The enzyme preparation used in this study had an A 1x2 m/Am nm of 0.92, which was similar to or better than the A,12nm/ Amnm values of preparations used for EPR studies by others (6,10-12). Both enzyme preparations exhibited a single band on the sodium dodecyl sulfate-polyacrylamide gel electrophoresis analysis. EPR spectra were recorded by a Varian E-109 spectrometer operating at 9.32 GHz with a field modulation of 0.5 mT at 100 kHz. Incident power was 2 mW. An Air-Products liquid helium flow cryostat was used to attain low temperatures. High spin derivatives were measured at 7 K, whereas low spin species were recorded at 10 K. Microwave frequency was monitored by an EPI Model 391 frequency counter and the magnetic field was calibrated by standards with known g-values. Buffers used were 0.1 M NJ-bis(2-hydroxyethyl)glycine (Bicine)’ at pH 8,0.1 M N-

’ Abbreviations used: Bicine, 2-(IV-morpholino)ethanesulfonic

N,N-bis(2-hydroxyethyl)glycine; acid.

Mes,

351 Inc. reserved.

352

IKEDA-SAITO g=6.49 Intestinal

I

Peroxidose.

9.32GHz

1 g=5.45

g=t.sa

7 150

g=2.94

I

gb1.57

g=2.24

mTI

MIS-1

62

FIG. 1. EPR spectra

of intestinal peroxidase (spectrum A) and its cyanide complex (spectrum B). Intestinal peroxidase was dissolved in 0.1 M Bicine buffer, pH 8.

2-hydroxyethylpiperazine-N’-2-ethane sulfonic acid (Hepes) at pH 7, 0.1 M 2-(N-morpholino)ethanesulfonic acid (Mes) at pH 6, and 0.1 M citrate at pH 5. Enzyme solutions (60 pM for intestinal peroxidase and 250 pM for lactoperoxidase) were rapidly frozen by immersing EPR tubas into a 5 to 1 (v/v) mixture of 2-methylbutane and methylcyclohexane cooled to about 170 K.

RESULTS

AND

DISCUSSION

The EPR spectrum of resting (as prepared) intestinal peroxidase in 0.1 M Bicine buffer, pH 8, is shown in Fig. 1 (spectrum A). The resting enzyme exhibited a rhombic high spin EPR signal with g-values of 6.49,5.45, and 1.98 without any trace of low spin component.3 Measurements on the same sample at 20 K with incident power of 0.5 mW did not detect any low spin signals, indicating that the absence of low spin signal in the 7 K spectrum is not due to saturation behavior often encountered in EPR experiments on low spin ferric heme proteins at low temperatures. As the resonance Raman results showed that intestinal peroxidase is in a high spin state at room temperature (4), we conclude that the heme iron in the resting intestinal peroxidase is in a high spin state both at ambient and at cryogenic temperatures without thermal mixing of low spin and high spin states. The EPR spectrum of the resting intestinal peroxidase is comparable to that of eosinophil peroxidase, which exhibited a high spin EPR signal with g-values of 6.50, 5.40, and 1.98 at pH 7.2 without traces of low spin signals (10). In contrast, lactoperoxidase shows an EPR spectrum consisting of both high spin (Fig. 3) and low spin species (g = 3.12, 2.13, and 1.4 at pH 8; data not shown) (6, 11, 12). The 3 Spin quantitation using met myoglobin fluoride estimated that the high spin component corresponds total enzyme.

as a standard (9) to 95 + 10% of the

AND KIMURA

appearance of a low spin species in the frozen lactoperoxidase solution has been characterized as a freezing artifact by Manthey et al. (6). Sievers et al. (11) concluded that the distal imidazole group became the sixth ligand of the heme iron at cryogenic temperature, and this ligand change is responsible for the low spin state. The lack of a freezing-induced low spin form in intestinal peroxidase indicates that there are differences in the distal heme pocket between lactoperoxidase and intestinal peroxidase. The cyanide adduct of intestinal peroxidase exhibited an EPR spectrum typical of a low spin ferric hemeprotein, as shown in spectrum B in Fig. 1. g-Values of the cyanide adduct of intestinal peroxidase are compared (Table I) with those of other cyanide complexes of hemeproteins which have been known to have a histidine residue as a fifth ligand of the heme iron. It is apparent that the gvalues of cyanide adducts of mammalian peroxidases are similar, and that they are distinctly different from those of metmyoglobin. It should also be pointed out that the EPR signals of cyanide adducts of mammalian peroxidases are readily detected under the present conditions, namely measurements near 10 K with 2 mW of incident power, whereas EPR signals of cyanide derivatives of metmyoglobin and methemoglobin are known to saturate under these conditions. The electronic structure of low spin iron differs between these peroxidases and globins despite the same heme ligand structure, imidazole-Fecyanide. Similar variations of low spin EPR properties have been seen in bis-imidazole complexes such as between cytochrome b5 and imidazole derivatives of metmyoglobin and methemoglobin (16). Peisach et al. (17) demonstrated that the electronic structure of the axial imidazole ligand dictates the EPR properties of the bisimidazole heme derivatives; the g-value anisotropy of the bis-imidazole complex is reduced upon deprotonation of the axial imidazole ligand. Since the g-value anisotropy of cyanide complexes of mammalian peroxidase is smaller than that of globins, we propose that the distinctive features in EPR spectra of cyanide adducts in animal peroxidases are indicative of the imidazolate character of the proximal axial ligand. This proposal is consistent with

TABLE EPR g-Values

I

of Low Spin Cyanide Adducts g-Values

Enzyme Metmyoglobin Horseradish peroxidase Intestinal peroxidase Lactoperoxidase

3.45 3.05

2.94 2.92 2.96 2.93 2.87

Eosinophil peroxidase Myeloperoxidase

2.91 2.83

of Hemeproteins Refs.

1.89

0.93

2.1 2.24 2.25 2.26 2.24 2.21 2.25 2.25

1.2 1.57 1.56 1.54 1.57 1.53 1.58 1.66

Hori (13) Blumberg et al. (14) This paper This paper Sievers et al. (11) Bolscher et al. (10) Lukat et al. (12) Bolscher et al. (10) Ikeda-Saito (15)

AXIAL g:

=

LIGAND

6.49

9.32

I A

9;

=

5.45

COORDINATION

GHz 7K

I 1 20

mT

1

FIG. 2. Low field portion Spectrum A, pH 6; spectrum of 0.1 M chloride; spectrum

911, 2

5.23

MIS-1

63

of the EPR spectra of intestinal peroxidase. B, pH 8; spectrum C, pH 8 in the presence D, pH 6 in the presence of 0.1 M chloride.

the conclusions reached by 15N and proton NMR studies of the cyanide complex of lactoperoxidase (18, 19). An NMR study by de Ropp et al. (20) has established the presence of axial imidazolate ligation in cyano-horseradish peroxidase and imidazole ligation in cynanometmyoglobin. The imidazolate character in the cyanide complex of horseradish peroxidase is considered to be attained by formation of a strong hydrogen bond between the axial imidazole ligand and a nearby amino acid residue (19,21). We think that such a hydrogen bond is also formed in mammalian peroxidases bearing imidazolate character of the proximal histidine. g-Value anisotropy for the cyanide complex of horseradish peroxidase is smaller than that of globins but larger than that of lactoperoxidase and eosinophil and intestinal peroxidases. The same trend is seen in the magnitude of the upfield bias of 15N chemical shift for C15N adducts of globins, horseradish peroxidase, and lactoperoxidase (18): globins > horseradish peroxidase > lactoperoxidase. There may be several factors that contribute to the differences in the EPR properties of horseradish peroxidase and mammalian peroxidases. The proximal histidine in mammalian peroxidases may have more imidazolate character than that in horseradish peroxidase, as proposed for lactoperoxidase (18). Other factors which could contribute to EPR g-values (also possibly to upfield bias of 15N resonance of the bound C15N) may include (i) geometry of proximal histidine, (ii) geometry of the bound cyanide, and (iii) the hydrogen bond formed between the bound cyanide and a nearby amino acid, such as distal histidine and arginine residues. A proton NMR study (20) showed that the orientation of the imidazole group of the proximal histidine residue in the cyanide complex is the same for lactoperoxidase and horseradish peroxidase, ruling out possibility (i) above. A recent resonance Raman study suggested a bent iron cyanide structure in myeloperoxidase, which is different from those seen in horseradish peroxidase and model iron porphyrin and chlorin complexes (22). Proton NMR and EPR stud-

IN

INTESTINAL

353

PEROXIDASE

ies on model heme complexes showed that hydrogen bond formation to the heme bound cyanide affects magnetic anisotropy (23). Thus, possibilities (ii) and/or (iii) above could be responsible for the smaller g-value anisotropy in the low spin EPR of cyanide complexes of mammalian peroxidases with respect to horseradish peroxidase. Since Kimura and Yamazaki (2) reported a pH-dependent change in the optical spectral properties of intestinal peroxidase and lactoperoxidase, we compared EPR spectra of these two enzymes between pH 5 and 8. Figure 2 shows the low field portion (g 6 region) of the EPR spectrum of the resting intestinal peroxidase at pH 6 and 8 in the absence and presence of 0.1 M chloride. The spectrum of intestinal peroxidase at pH 6, in the absence of chloride, (spectrum A) is a mixture of two sets of rhombic high spin species; species I has a smaller rhombicity with higher signal amplitude (g: = 6.49 and ga = 5.45) and species II has a larger rhombicity with reduced signal amplitude (gi’ = 6.66 and gil = 5.23). At pH 8 (spectrum B), the signal amplitude of species I increases and that of species II decreases. Measurements below pH 6 were not feasible due to instability of the enzyme in the absence of chloride. The addition of chloride at pH 8 (spectrum C) slightly alters the spectrum due to the appearance of a small amount of species II. A drastic change is observed for the spectrum at pH 6 in the presence of chloride (spectrum D), which shows conversion to species II. The straightforward interpretation of the observed spectral change would be that intestinal peroxidase assumes neutral and acidic forms, and the equilibrium between these two forms is modulated by chloride ion, supporting the original proposal of Kimura and Yamazaki (2). Both neutral and acidic forms are high spin derivatives; the former is characterized by the g-values of 6.49, 5.45, and 1.98 (species I), and the latter by those of 6.66, 5.23, and 1.97 (species II).

&?: = 6.43

9.32GHz.

I

g = 7.16

7K

g = 5.99 $=

5.51

1

> 0

1

! !

g!I = 6.58 2

110

mT/

I g’I 1, 5.35 2

FIG. 3. Low field portion of the EPR spectra of lactoperoxidase. Spectrum A, pH 8; spectrum B, pH 5; spectrum C, pH 8 in the presence of 0.1 M chloride; spectrum D, pH 5 in the presence of 0.1 M chloride. The signal marked by * in the spectrum D is theg = 6.3 and 5.72 species.

354

IKEDA-SAITO

Figure 3 illustrates the g 6 region of the EPR spectra of lactoperoxidase at pH 5 and 8 in the absence and presence of 0.1 M chloride. At pH 8 (spectrum A), two sets of high spin EPR signals were observed, consisting of a major component with g-values of 6.43 and 5.51 (species I) and a minor component with g-values of 7.16 and 4.75. (The g 4.75 feature is outside the spectral window of Fig. 3.) Upon lowering the pH to 5 (spectrum B), the apparent linewidth of the major component was broadened due to the appearance of another high spin signal (species II, gi’ = 6.58 andg;i = 5.35). At the same time, the rhombic minor component (gl = 7.16 and g, = 4.75) disappeared and a new minor axial signal (g x 6) appeared. Species II was discernible as the major component at pH 5 in the presence of chloride (spectrum D) together with two minor components; one an axial g = 6 signal and the other a rhombic high spin form with apparent g-values of 6.3 and 5.72. The pH dependency of the major high spin EPR component of lactoperoxidase (species I and II) is similar to that observed in intestinal peroxidase; the equilibrium between a neutral and an acidic form is shifted by chloride. The origin of the minor components is unknown, although the enzyme preparation appeared to be homogeneous as judged by the sodium dodecyl sulfate-polyacrylamide gel electrophoresis analysis. Kimura and Yamazaki (2) originally attributed the pH dependence of characteristic features of intestinal peroxidase to an alteration of the coordination state of the heme iron. They suggested that the heme iron is hexacoordinated with a water molecule as the sixth ligand in neutral pH, and that the water molecule dissociates from the iron resulting in the pentacoordinated heme structure when a distal base with a pK, value of 4.8 is protonated. In ferric heme proteins, the change from pentacoordinated high spin to hexacoordinated high spin states is expected to have a considerable effect on their Soret region optical absorption and EPR spectra, as documented for horseradish peroxidase (24) and for cytochrome c peroxidase (25). The pH effects seen in the EPR and optical spectra (2) of lactoperoxidase and intestinal peroxidase are small in comparison with the spectral changes associated with transitions from pentacoordinated to hexacoordinated ferric high spin species in horseradish peroxidase (24) and cytochrome c peroxidase (25). This indicates that the heme irons in lactoperoxidase and intestinal peroxidase likely have the same coordination number both in neutral and in acidic pH. Although spectroscopic studies have established pentacoordinated heme-iron in ferric horseradish peroxidase and cytochrome c peroxidase (24, 25), resonance Raman results on intestinal peroxidase (4) and lactoperoxidase (5, 6) showed the hexacoordinated high spin heme iron, possibly with a water molecule as the sixth ligand in these peroxidases. On the basis of these Raman results (4-6), we therefore conclude that the heme irons in these peroxidases are hexacoordinate high spin both in the acidic and in the neutral forms. Chloride com-

AND

KIMURA

petitively inhibits both peroxidase activity and cyanide binding to intestinal peroxidase (2), as seen in myeloperoxidase (26, 27). However, the effects of chloride on the spectral properties of intestinal peroxidase and lactoperoxidase differ from those observed for myeloperoxidase. Myeloperoxidase forms a spectroscopically distinguishable chloride complex, in which chloride is considered to bind to the iron center (15, 26). In the case of intestinal peroxidase or lactoperoxidase, chloride shifts the equilibrium between neutral and acidic forms of these enzymes without changing their EPR spectral properties. This behavior is consistent with the proposal that chloride interacts with the distal base responsible for the acid-neutral transition rather than binding to the heme iron (2). As Thanabal and La Mar (19) showed the presence of the distal histidine residue in lactoperoxidase, we think that a distal histidine residue is the most likely candidate for the functional group responsible for the pH and chloride effects in both intestinal peroxidase and lactoperoxidase. In these enzymes, upon lowering pH or upon addition of chloride, the hydrogen bond formed between the distal histidine residue and a water molecule at the sixth coordination position of the heme iron would break due to a protonation of the distal histidine residue. The breakage of the hydrogen bond would change the relative disposition of the bound water, the heme iron, and the porphyrin macrocyle, and such a change could further lower the symmetry of the electronic environment of the d electron system, resulting in an increase in the rhombicity of the high spin EPR spectrum of the acid form of the enzymes. The difference between intestinal peroxidase and lactoperoxidase seen in pK, values of heme-linked ionization and EPR g-values (Figs. 2 and 3) probably come from the difference in the distal heme pocket structure such as amino acid residues around the distal arginine and histidine, and the geometry of these residues to the heme. ACKNOWLEDGMENT We thank

T. Ohnishi

for the use of her EPR

facilities.

REFERENCES 1. Stelmaszynska,

T., and Zglinczynski,

J. M. (1971)

Eur.

J. Biochem.

19,56-63. 2. Kimura, 14-19.

S., and Yamazaki,

3. Kimura, S., and Yamazaki, 580-588. 4. Kimura, S., Yamazaki, I., 20,4632-4638. 5. Kitagawa, T., Hashimoto, H., and Hosoya, T. (1983) 6. Manthey, Biol.

J. A., Bold,

Chem.

7. Nichol, Biochem.

I. (1978)

Arch.

Biochem.

Biophys.

189,

I. (1979)

Arch.

Biochem.

Biophys.

198,

and Kitagawa, S., Teraoka, Biochemistry

N. J., Bocian,

T. (1981) J., Nakamura, 22,2788-2792.

D. F., and Chan,

Biochemistry S., Yajima, S. I. (1986)

J.

25,6734-6741.

A. W.,

Angle,

L. A., Moon,

T., and

Clezy,

P. S. (1987)

J. 247,147-150.

8. Palmer, G. (1979) in The IV, pp. 313-353, Academic

Porphyrins (Dolphin, Press, New York.

D. H., Ed.),

Vol.

AXIAL 9. Aasa, R., and Vangaard,

T. (1975)

10. Bolscher, B. G. J. M., Plat, phys. Acta 784, 177-186.

LIGAND J. Magn.

12. Lukat,

G. S., Rodgers,

Z&son.

H., and Wever,

11. Sievers, G., Gadsby, P. M. A., Peterson, (1980) Biochim. Biophys. Acta 785, 7-13. K. R., and Goff,

COORDINATION

19,308-315.

R. (1984)

Biochim.

J., and Thomson,

H. M.

IN

(1987)

19. Thanabal, Bio-

A. J.

Biochemistry

26,6927-6932. 13. Hori,

H. (1971)

Biochim.

Biophys.

Acta

251, 227-235.

14. Blumberg, W., Peisach, J., Wittenberg, B. A., and J. B. (1968) J. Riol. C&m. 243, 1854-1862. 15. Ikeda-Saito,

M. (1985)

J. Biol.

Chem.

260,

Wittenberg,

11,688-11,696.

16. Ikeda, M., Iizuka, T., Takao, H., and Hagihara, B. (1974) Biochim. Biophys. Acta 336, 15-24. 17. Peisach, J., Blumberg, W. E., and Adler, A. (1973) Ann. New York Acad. Sci. 206, 310-327. 18. Behere, D. V., Gonzales-Vergara, Biophys. Acta 832,319-325.

E., and Goff,

H. M. (1985)

INTESTINAL

Biochim.

355

PEROXIDASE V., and La Mar,

G. N. (1989)

Biochemistry

28,

7038-

7044. 20. de Ropp, J. S., Thnanabal, V., and La Mar, Chem.Soc. 107,8268-8280.

G. N. (1985)

J. Amer.

21. Thanabal, Chem.Soc.

G. N. (1988)

J. Amer.

V., de Ropp,

J. S., and La Mar,

110,3027-3035.

22. Lopez-Garriga, J. J., Oertling, W. A., Kean, R. T., Hoogland, H., Wever, R., and Babcock, G. T. (1988) Proceedings of the 11th International Conference on Raman Spectroscopy (Clark, R. J. H., and Long, D. A., Eds.), pp. 665-666, Wiley, New York. 23. La Mar, G. N., Del Gaudio, phys. Acta 496,422-435. 24. Dunford, 25. Yonetani,

H. B. (1982) T., and Anni,

J., and Frey,

Adu. Znorg. E. (1987)

26. Bolscher, B. G. J. M., and Wever, 788, l-10. 27. Ikeda-Saito,

M. (1988)

Biochemistry

J. S. (1977)

Chem.

4,41-68.

J. Biol.

Chem.

R. (1984)

Bio-

262,9547-9554.

Biochim.

26,4344-4349.

Biochim.

Biophys.

Acta