Spectroscopic studies of peroxyacetic acid reaction intermediates of cytochrome P450cam and chloroperoxidase

Journal of Inorganic Biochemistry 91 (2002) 586–596 www.elsevier.com / locate / jinorgbio

Spectroscopic studies of peroxyacetic acid reaction intermediates of cytochrome P450cam and chloroperoxidase a, *, C. Jung b , J. Terner c , A.X. Trautwein a , R. Weiss d ¨ V. Schunemann a

¨ ¨ , Ratzeburger Allee 160, D-23538 Lubeck , Germany Institute of Physics, Medical University Lubeck b ¨ Max-Delbruck-Center for Molecular Medicine, D-13125 Berlin, Germany c Chemistry Department, Virginia Commonwealth University, Richmond, VA 23284 -2006, USA d Laboratoire de Cristallochimie, UMR 7513, Universite´ Louis Pasteur, 67070 Strasbourg, France Received 10 December 2001; received in revised form 1 May 2002; accepted 2 May 2002

Abstract It is generally assumed that the putative compound I (cpd I) in cytochrome P450 should contain the same electron and spin distribution as is observed for cpd I of peroxidases and catalases and many synthetic cpd I analogues. In these systems one oxidation equivalent resides on the Fe(IV)5O unit (d 4 , S51) and one is located on the porphyrin (S951 / 2), constituting a magnetically coupled ferryl iron-oxo porphyrin p-cation radical system. However, this laboratory has recently reported detection of a ferryl iron (S51) and a tyrosyl ¨ radical (S951 / 2), via Mossbauer and EPR studies of 8 ms-reaction intermediates of substrate-free P450cam from Pseudomonas putida, ¨ prepared by a freeze-quench method using peroxyacetic acid as the oxidizing agent [Schunemann et al., FEBS Lett. 479 (2000) 149]. In the present study we show that under the same reaction conditions, but in the presence of the substrate camphor, only trace amounts of the tyrosine radical are formed and no Fe(IV) is detectable. We conclude that camphor restricts the access of the heme pocket by peroxyacetic acid. This conclusion is supported by the additional finding that binding of camphor and metyrapone inhibit heme bleaching at room temperature and longer reaction times, forming only trace amounts of 5-hydroxy-camphor, the hydroxylation product of camphor, during peroxyacetic acid oxidation. As a control we performed freeze-quench experiments with chloroperoxidase from Caldariomyces fumago using peroxyacetic acid under the identical conditions used for the substrate-free P450cam oxidations. We were able to confirm earlier findings [Rutter et al., Biochemistry 23 (1984) 6809], that an antiferromagnetically coupled Fe(IV)=O porphyrin p-cation radical system is formed. We conclude that CPO and P450 behave differently when reacting with peracids during an 8-ms reaction time. In P450cam the formation of Fe(IV) is accompanied by the formation of a tyrosine radical, whereas in CPO Fe(IV) formation is accompanied by the formation of a porphyrin radical.  2002 Elsevier Science Inc. All rights reserved.

1. Introduction The enzymes of the cytochrome P450 superfamily play key roles in physiological processes of numerous organisms and are of considerable pharmacological interest. P450s catalyze a variety of reactions, such as aliphatic and aromatic hydroxylations, epoxidations, heteroatom oxidation, and N- and O-dealkylation, by activated oxygen atom transfer from the heme to substrates. Based upon studies of model compounds and other enzymes [1] an Fe(IV)=O porphyrin p-cation radical system has been generally assumed to occur within P450 *Corresponding author. Tel.: 149-451-500-4207; fax: 149-451-5004214. E-mail address: [email protected] (V. ¨ Schunemann).

reaction cycles, analogous to compound I (cpd I) of peroxidases and catalases. However, we recently detected a ferryl iron (S51) and a tyrosyl radical (S951 / 2) in a ¨ Mossbauer / EPR study on 8-ms reaction intermediates of substrate-free P450cam from Pseudomonas putida prepared by freeze-quenching using peroxyacetic acid (PA) as the oxidizing agent [2]. It is interesting that during experiments involving other reaction times we did not obtain indications for formation of a porphyrin p-cation radical. This observation was surprising since chloroperoxidase (CPO) from Caldariomyces fumago, which has a heme iron coordination sphere very similar to P450, reacts with peroxyacetic acid to form cpd I with one oxidation equivalent residing on the Fe(IV)=O unit (d 4 , S51), and a second one on the porphyrin (S951 / 2), constituting an antiferromagnetically coupled ferryl ironoxo porphyrin p-cation radical system [3,4]. Our observa-

0162-0134 / 02 / $ – see front matter  2002 Elsevier Science Inc. All rights reserved. PII: S0162-0134( 02 )00476-2

¨ et al. / Journal of Inorganic Biochemistry 91 (2002) 586–596 V. Schunemann

tion on P450 makes the nature of cpd I still more puzzling in light of findings reported by other laboratories. Davydov et al. [5] reduced the dioxygen complex via g-irradiation at 77 and 6 K in both wild-type and mutant P450cam. These workers detected various intermediates within the reaction cycle via EPR and ENDOR by stepwise annealing of the reduced oxygen complex, increasing the temperature up to approximately 220 K. Although the formation of a hydroxylation product of camphor was observed, the proposed cpd I intermediate (high-valent iron-oxo porphyrin p-cation radical system) could not be identified. Similar preliminary observations had been made by Nyman and Debrunner (1989, personal communication: Observation of radical intermediates in the catalytic step of cytochrome P450cam). Transient state X-ray crystallographic studies by Schlichting et al. [6] have, however, indicated that a (Fe–O) species is formed after the reduction of the dioxygen complex by X-ray irradiation. On the other hand, resonance Raman studies under catalytic conditions revealed an O–O stretch signal very similar to the simple dioxygen complex. A Raman signal at ca. 774–790 cm 21 , characteristic of Fe–O stretching, was not observed [7]. Pederson et al. [8], Wagner et al. [9] and Egawa et al. [10] used m-chloroperoxybenzoic acid (mCPBA) as an oxidizing agent in stopped-flow studies of P450cam in efforts to spectroscopically characterize reaction intermediates. An electronic absorption spectrum obtained as one component in the single value component analysis of a timedependent spectrum of the P450cam-mCPBA mixture was assigned to a high-valent iron-oxo porphyrin p-cation radical system [10] due to its similarity to the electronic absorption of CPO [11]. Wagner et al. [9] and Sligar et al. [12] reported optical spectra of an intermediate in the reaction with peroxyacetic acid obtained in stopped-flow studies at subzero (Celsius) temperatures which were, however, difficult to interpret and did not match each other completely, and thus did not unequivocally indicate the formation of an intermediate Fe(IV)=O porphyrin p-cation radical system. Sligar et al. [12] had reported that 5-exohydroxycamphor is formed to a significant extent. However, Kobayashi et al. [13] observed the formation of exoand endo-5-hydoxycamphor by adding KO 2 to ferrous cytochrome P450cam under anaerobic conditions. Thus it is not clear as to whether a cpd I intermediate is formed during the reaction cycle of cytochrome P450. To clarify the mechanism of cpd I formation in the reaction of P450cam with peroxyacetic acid we have performed freeze-quench experiments with camphor-bound P450cam, analyzed the reaction intermediate by EPR and ¨ Mossbauer spectroscopy, and checked by gas chromatography–mass spectrometry as to whether 5-hydroxycamphor was formed. Only trace amounts of the tyrosine radical could be observed, no Fe(IV) was detectable and almost 99% of the protein remained in the resting state. Furthermore, only trace amounts of 5-hydroxycamphor were found, in contrast to the report by Sligar et al. [12]. We

587

conclude that camphor restricts the access of the heme pocket by peroxyacetic acid, which is required for binding of peroxyacetic acid to the heme iron and formation of cpd I. This conclusion is supported by the finding that binding of camphor and metyrapone inhibits the porphyrin ring cleavage reflected in Soret band bleaching which appears at room temperature and at extended reaction times when peroxyacetic acid is used as the oxidant. As control experiments we performed freeze-quenching on chloroperoxidase from Caldariomyces fumago with peroxyacetic acid oxidant under identical conditions as was used for substrate-free P450cam. We were able to verify earlier findings by Rutter et al. [3,4], that an antiferromagnetically coupled Fe(IV)=O porphyrin p-cation radical system is formed.

2. Material and methods

2.1. Proteins Cytochrome P450cam (CYP101) from Pseudomonas putida, expressed in Escherichia coli strain TB1 was isolated as previously described [14]. The procedure for the preparation of the 57 Fe-derivative, by 56 Fe-heme replacement, followed the protocol published by Wagner et al. [15], as described in Ref. [2]. Putidaredoxin and putidaredoxin reductase expressed in the Escherichia coli strain DH5a were isolated and purified according to procedures described in Refs. [16,17]. CPO was purified as previously described [18].

2.2. Freeze-quench experiments Rapid freeze-quench experiments [19] were performed with a System 1000 apparatus from Update Instruments by mixing |1 mM cytochrome P450cam and 5 mM peroxyacetic acid in 100 mM potassium phosphate buffer, pH 7. A Delrin cup with 4 mm inner diameter and a volume of 50 ml was attached to a quartz tube, which was in turn connected to a funnel that was completely immersed in an isopentane bath at T52100 8C. The reaction mixtures were quenched by spraying into cold isopentane. The thus obtained frozen material was packed into the bottom of a Delrin cup with a Teflon packing rod. This procedure ¨ allowed the recording of Mossbauer and EPR spectra from the same sample.

2.3. EPR measurements EPR spectra were recorded with a conventional X-Band spectrometer (Bruker 200D SRC) equipped with a He-flow cryostat (ESR 910, Oxford Instruments) at a temperature of 20 K. EPR spectra of P450cam samples as well as those of ferric CPO were simulated using effective g values and angular-dependent Gaussian line shapes, according to the

¨ et al. / Journal of Inorganic Biochemistry 91 (2002) 586–596 V. Schunemann

588

procedure of Beinert and Albracht [20]. Spin quantitations were performed by double-integration of simulated absorption-derivative subspectra. Differences in g values were taken into account using the Aasa g-factor [21] for corrections. ¨ 2.4. Mossbauer measurements ¨ Mossbauer spectra were recorded in constant-acceleration mode using a home-built spectrometer. Isomer shifts are given relative to a-Fe at room temperature. The samples were contained in a He-bath cryostat (Oxford Instruments), equipped with permanent magnets. The typical measurement time was 3 weeks per spectrum. Magnetically split spectra were simulated with the spin Hamiltonian formalism [22] utilized below for the EPR data.

2.5. Spin Hamiltonian simulations EPR spectra of the oxoferryl porphyrin -radical complex of CPO were simulated using the spin Hamiltonian He 5 D[S z2 2 2 / 3 1 E /D(S x2 2 S y2 )] 1 mB (Sg 1 g9S9) B 2 SJS9 ]] ]]

(1)

which includes zero-field splitting (ZFS), electronic Zeeman interaction and exchange in the intrinsic spin representation of the coupled S51 (Fe(IV)=O) and S95 1 / 2 (p-cation radical) system. The electronic coupling tena sor ]J51? a trace]] J0 1J]] consistsaof an isotropic part J0 and ] less anisotropic tensor ]J . The components of ]J a as well as ] ferryl iron and its ZFS ] parameter the local g tensor of the D and the rhombicity parameter E /D were taken from Rutter et al. [3]. The radical spin g9 was fixed at 2.0. The EPR spectra were simulated with Eq. (1) using a ‘g strain’ model for the line shape [23], which is based on a Gaussian distribution of spin-Hamiltonian parameters leading to an angular variation of Gaussian linewidths. ¨ The magnetic Mossbauer spectra of 57 Fe-enriched P450cam were simulated using Eq. (1) with the electronic coupling terms (J) being zero, and Eq. (2) for the nuclear Hamiltonian HN [22] HN 5 eQVzz (4I(2I 2 1))21 [3I 2z 2 I(I 1 1) 1 h(I 2x 2 I 2y )] 2 gN mN I B 1 kSlAI ] ]

(2)

In Eq. (2) I is the nuclear spin, Q is the nuclear quadrupole moment of the excited nuclear state of 57 Fe, Vzz is the main component of the electric-field gradient tensor, and h 5 (Vxx 2Vyy ) /Vzz is the asymmetry parameter of the electric field gradient. ] A denotes the hyperfine coupling ] g-factor, m is the nuclear tensor, gN is the nuclear N magneton and B is the applied field.

2.6. Stability study The effect of peroxyacetic acid on the chemical stability of P450, i.e., bleaching due to porphyrin ring cleavage, was followed by monitoring the Soret band using a Shimadzu UV2101PC spectrometer. The concentrations for substrate-free P450cam and peroxyacetic acid were 3 and 15 mM, respectively, giving the same concentration ratio peroxyacetic acid–P450 of 5:1 as used for the freezequench experiments in 50 mM potassium phosphate buffer, pH 7. The effects of camphor (500 mM) and metyrapone (104 mM) were studied under the same buffer and concentration conditions.

2.7. Product analysis Product (5-exo-hydroxy-camphor) formation in the native reconstituted system with putidaredoxin and putidaredoxin reductase, and in the shunt pathway with peroxyacetic acid, was analyzed by gas chromatography– mass spectrometry (GC–MS). The analyses were made with a software (Class 5000) controlled Shimadzu GC– MS-QP5050A equipped with a 30-m / 0.25-mm Optima-1 fused-silica capillary column (0.25-mm coating thickness, Macherey-Nagel) using helium as carrier gas. For EI (electron impact) spectra the oven temperature was held for 1 min at 110 8C and then raised to 220 8C at a rate of 5 8C / min. Under these conditions the retention times of 1R-camphor and 5-exo-hydroxycamphor were 4.7 and 8 min, respectively. The mass spectra for these retention times show molecular ions (M1) at m /z5153 and m /z5 168 for camphor and hydroxycamphor, respectively, corresponding to the signals for the pure compounds. Pure 5-exo-hydroxycamphor, used as a reference, was kindly supplied by John H. Dawson (University of South Carolina, Columbia, SC, USA). Camphor was obtained from Sigma. The assay for the native system consisted of 1 mM putidaredoxin, 0.13 mM putidaredoxin reductase, 0.037 mM P450cam, 1 mM 1R-camphor and 360 mM NADH in 50 mM potassium phosphate buffer, pH 7. The reaction time was 20 min at 25 8C. For the assay with peroxyacetic acid, 3 mM P450cam and 1 mM 1R-camphor in 50 mM potassium phosphate buffer, pH 7, were reacted with 110 mM or 1.1 mM peroxyacetic acid for different time periods of up to 130 min at 25 as well as at 4 8C. The product was extracted from the reaction solution with chloroform and concentrated by a stream of nitrogen gas, and then applied to the GC–MS system.

3. Results

3.1. Freeze-quench and EPR measurements The EPR traces of both the native (ferric) enzymes cytochrome P450cam from Pseudomonas putida and CPO

¨ et al. / Journal of Inorganic Biochemistry 91 (2002) 586–596 V. Schunemann

from Caldariomyces fumago exhibit an anisotropic EPR signal within the g|2 region corroborating their S51 / 2 state (2.62, 2.27, 1.843 for CPO [3,4], and 2.45, 2.26, 1.91 for substrate-free P450cam [24]). The EPR spectrum obtained from substrate-free P450cam, which reacted with peroxyacetic acid within 8 ms, is shown in Fig. 1a (see Table 1). The spectrum does

589

Table 1 Spin-Hamiltonian parameters used for the simulation of the EPR trace of chloroperoxidase compound I shown in Fig. 1b DFe( IV ) 21 (cm )

E /DFe( IV ) a

g ]]Fe( IV )

J0 21 (cm )

Ja ]] 21 (cm )

gb ]]

36

0.035

2.265 2.295 1.96

35

20.5 20.5 1.0

1.65 1.75 1.97

The spin on the Fe(IV)=O unit is S51 and that of the porphyrin radical is S951 / 2. a s (E /D)50.030. The values s (E /D) represent the widths of the Gaussian distributions (2ps 2 )21 / 2 exp(2x 2 / 2s 2 ) of the distributed parameter E /D derived from the simulations of the EPR spectra. b Effective g values of the spin coupled system with total spin 1 / 2 obtained from the spin Hamiltonian simulation.

Fig. 1. EPR spectra of (a) substrate-free cytochrome 57 Fe-P450cam freeze-quenched with peroxyacetic acid (PA) (8-ms reaction time). 100 mM potassium phosphate buffer, pH 7. PA (5 mM):P450 (1 mM) ratio of 5:1 was mixed to give half of each concentration in the mixture. The dashed-dotted line is a simulation assuming two components: the signal of the starting material ( g values 1.91, 2.26 and 2.45 and relative contribution 85%) and the signal of a radical ( g52; relative contribution 15%). Experimental parameters: T520 K; P580 mW; n 59.64 GHz; modulation amplitude 5 G; modulation frequency 100 kHz. (b) EPR spectra of 56 Fe-CPO freeze-quenched with PA (8-ms reaction time). Five mM sodium acetate buffer, pH 4.8. PA (5 mM) to CPO (1 mM) ratio of 5:1 were mixed to give half of each concentration in the mixture. The dashed-dotted line is a simulation assuming two components: the signal of the starting material ( g values 1.84, 2.27 and 2.62 and relative contribution 30%) and the signal of cpd I (spin-Hamiltonian simulation with parameters given in Table 1; relative contribution 70%). For experimental parameters see Fig. 1a.

not show a signal attributable to a magnetically coupled cpd I species. Instead, the pattern of the starting material is superimposed on a strong radical signal. The latter exhibits hyperfine structure that is indicative of a tyrosine radical (see Fig. 4a). Due to its relaxation behavior this tyrosine ˚ from the has been assigned to Tyr96 [2], which is 9.4 A heme iron [25,26]. This assignment has been recently verified by high-field EPR at 285 GHz which reveals g tensor components of: gx , 2.0078–2.0064; gy , 2.0043; and gz , 2.0022 (2002, manu¨ script in preparation: V. Schunemann, F. Lendzian, C. Jung, A.-L. Barra, A.X. Trautwein). These g values are fingerprints for a tyrosine radical in a polar environment [27–29]. The presence of a protein based tryptophanyl radical can be presently excluded on the basis of recently reported g values of gx 52.0033, gy 52.0024, and gz 5 2.0021 for this radical, which were found for W111* of the Y122F mutant of the subunit R2 protein of E. coli ribonucleotide reductase (RNR) [28]. Besides, a tryptophanyl radical exchange-coupled with the heme iron can also be ruled out, because of the same argument used for Tyr96 [2]: the nearest tryptophan, Trp42, has a distance of ˚ to the nearest porphyrin meso-carbon and 16.4 ca. 13.3 A ˚A to the heme iron (the corresponding distances for Tyr96 ˚ respectively). Such distances would be are 7.4 and 9.4 A, too large to enable sizable Heisenberg exchange interaction. In contrast, EPR spectra of CPO, when reacted with peroxyacetic acid within 8 ms, do clearly show signals typical for cpd I (Fig. 1b). This is similar to the earlier results of Rutter and Hager [4]. Seventy percent of the spin density exhibited an axial S51 / 2 signal with gi |2.00 and g' |1.75. Such a signal is typical for a magnetically coupled ferryl iron-oxo porphyrin p-cation radical system with one oxidation equivalent residing on the Fe(IV)=O unit (d 4 , S51), and one on the porphyrin (S951 / 2). The g values can be rationalized by considering a Heisenberg exchange interaction, which antiferromagnetically couples the spin 1 of the ferryl iron and the spin 1 / 2 of the porphyrin p-cation radical resulting in a net system of spin 1 / 2. The strength of the Heisenberg exchange interaction

590

¨ et al. / Journal of Inorganic Biochemistry 91 (2002) 586–596 V. Schunemann

is described by the exchange coupling parameter J0 , which in the present case, is of similar magnitude as the ZFS parameter D of the ferryl iron S51 system. The spinHamiltonian simulation of the cpd I signal shown in Fig. 1b yields J0 /D50.97. Within experimental error this is the same as the value obtained from J0 and D in Ref. [3]. It is remarkable that the EPR pattern of our 8 ms freezequenched CPO samples are similiar to those reported in Ref. [4], where mCPBA was used as an oxidizing agent. On the other hand Rutter et al. [3] using peroxyacetic acid, observed a strong additional radical signal which is absent in the present study (Fig. 1b). We consider the difference in reaction times of the present study (8 ms) and that used by Rutter and Hager [4] (12.5 ms) to be insignificant. Therefore we prepared additional freeze-quenched samples after reacting CPO with peroxyacetic acid for 40 and 200 ms, respectively. The EPR-pattern of these samples resembles that shown in Fig. 1b, yet a small radical signal with relative spin density of ca. 1% appears, which remained up to reaction times as long as 3 min (data not shown). The EPR-spectrum of the substrate-bound 57 FeP450cam, which has reacted with peroxyacetic acid for 8 ms is shown in Fig. 2. The spectrum consists of a mixture of ferric high-spin (46%) and ferric low-spin (52%) iron centers. The radical signal at g52 corresponds to a contribution of only 0.6% to the total spin density of the sample. This signal includes a small radical contribution (|0.2%) from the sample holder material (Delrin). Thus, we conclude that camphor prevents the protein from reacting with peroxyacetic acid. To obtain higher yields of the radical intermediate we carried out freeze-quench experiments on camphor-bound 56 Fe-P450cam with differing ratios of peroxyacetic acid to camphor-bound P450. The amount of radical formed increases with increasing ratio, but the maximum radical concentration achieved was only 4% of the total spin density (Fig. 3). Fig. 4a shows EPR spectra of the tyrosyl radical formed by the reaction of peroxyacetic acid with substrate-free P450 [2] in comparison to the radical of the substratebound enzyme (Fig. 4b). Despite the lower signal to noise ratio of the latter which is caused by the low concentration of the radical (|20 mM), the traces show comparable hyperfine patterns. This is thus strong evidence that both radicals originate from the same amino acid residue, presumably Tyr96. Our attempt to prepare even higher radical concentrations were not successful. At a peroxyacetic acid concentration of 200 mM (ratio 200:1) the protein precipitated, and was bleached, indicating oxidative cleavage of the porphyrin ring. ¨ 3.2. Mossbauer measurements ¨ A Mossbauer spectrum obtained after reaction of the substrate-free P450cam with peroxyacetic acid (reaction

Fig. 2. EPR spectrum of camphor-bound cytochrome 57 Fe-P450cam freeze-quenched with PA (8-ms reaction time). The solid lines are simulations with the parameters: (1) S55 / 2; g5(3.95,7.80,1.78); 46%; (2) S51 / 2; g5(2.45,2.25,1.907); 23%; (3) S51 / 2; g5 (2.41,2.23,1.970); 29%; (4) S51 / 2; g5(2.0); 0.6%; (5) S51 / 2; g5 (2.06); 1.4%. Component (5) represents a Cu(II) impurity present in the cavity. For experimental parameters and reaction conditions see Fig. 1a.

time 8 ms) is shown in Fig. 5a and was discussed in more detail in Ref. [2]. Approximately 87% of the spectral area corresponds to ferric low-spin iron with the same g factors and hyperfine parameters as the starting material, upon which is superimposed a quadrupole doublet with d 50.13 mm s 21 and DEQ 51.94 mm s 21 . The isomer shift of this doublet is indicative of ferryl iron with S51 [2]. The fact that this doublet does not exhibit magnetic hyperfine splitting within a moderate field of 1 T [30] is consistent with an S51 system with positive ZFS parameter D and indicates that the ferryl iron is not taking part in a spincoupled (ferromagnetic or antiferromagnetic) system. ¨ A Mossbauer spectrum of substrate-bound 57 FeP450cam, which has reacted for 8 ms with peroxyacetic acid, is shown in Fig. 5b (see Table 2). The solid lines are simulations resembling the resting state of substrate-bound P450cam. It is known that the iron center of substratebound P450cam exhibits a high-spin (S55 / 2) / low-spin (S51 / 2) equilibrium [31] as reflected by the two com¨ ponents of the Mossbauer spectrum at 4.2 K shown in Fig.

¨ et al. / Journal of Inorganic Biochemistry 91 (2002) 586–596 V. Schunemann

591

Fig. 3. EPR spectra of camphor-bound cytochrome 56 Fe-P450cam (a) freeze-quenched with different excess of PA as indicated (8-ms reaction time). One hundred mM potassium phosphate buffer, pH 7, 0.5 mM camphor, camphor-bound P450cam 1 mM before mixing. For further experimental parameters see Fig. 1a; (b) g|2 region of the EPR spectra shown in (a). An increase in radical signal with respect to ferric low-spin iron is clearly visible. (c) Percentage of radical observed in freeze-quench experiments of camphor-bound cytochrome 56 Fe-P450cam (8-ms reaction time) at different ratios of PA to P450.

¨ 5b. Within the detection limits of Mossbauer spectroscopy (|5%), iron(IV) is not formed within a reaction time of 8 ms. This is in accordance with the small amount of tyrosine radical (0.4%) detected by EPR.

3.3. Stability studies In order to check whether denaturation of the protein might play a role within a time scale of up to 200 ms, and at a peroxyacetic acid to P450 ratio of 5:1 (for substratefree P450cam as well as camphor-bound P450cam) we followed the time-dependent disappearance (bleaching) of the Soret band. Fig. 6 shows the decrease of the P450 concentration calculated from the Soret-band absorbance. We have estimated from the initial rate (|0.027 s 21 ) for a time range up to 20 s that within the first 8 ms only 0.005%, and within 200 ms only |0.1% of P450 isbleached, indicating that P450 bleaching due to peroxyacetic acid plays no role in the freeze-quench experiments.

Metyrapone, which binds to the heme iron and forms a hydrogen bond to Tyr96 [25], actually inhibits bleaching to a remarkable extent. The most dramatic effect is seen in the presence of camphor which also inhibits bleaching (Fig. 6). When using a peroxyacetic acid to P450 ratio of 38:1, bleaching is observed with an initial rate of |0.001 s 21 (data not shown). Also in this case the contribution of bleached protein within 8 and 200 ms is negligibly small (|0.0006 and |0.014%, respectively) indicating that in the freeze-quench experiments with camphor-bound P450cam, a small amount of porphyrin ring cleavage reflected by bleaching of the Soret band is not important.

3.4. Product analysis In order to check whether the product 5-exo-hydroxycamphor was formed in the reaction mixture we subjected thawed freeze-quenched samples (reaction time 5 min) to gas chromatography–mass spectrometry analysis. No such

¨ et al. / Journal of Inorganic Biochemistry 91 (2002) 586–596 V. Schunemann

592

product was found. In separate activity studies with different reaction times at 4 and 25 8C we searched for conditions for any detectable product formation. We found that product formation by peroxyacetic acid is detectable for a reaction time of 130 min at 5 8C (peroxyacetic acid:P450 ratio of 38:1). The yield was quite low (Fig. 7a) and could not be quantified. Peroxyacetic acid in the absence of P450cam, however, does not yield 5-exohydroxycamphor (data not shown). For comparison, 5-exohydroxycamphor formation in the native reconstituted system is possible, clearly indicating product formation (Fig. 7b), demonstrating that the experimental conditions for the GC–MS were appropriate for product analysis.

Fig. 4. EPR spectra of the radical formed (a) after reacting substrate-free 57 Fe-P450cam with PA for 8 ms (PA:P450 ratio 5:1) [2]. Experimental parameters: T5140 K; P5200 mW; n 59.64 GHz; modulation amplitude 0.05 mT; modulation frequency 100 kHz. (b) EPR spectrum of the radical formed after reacting camphor-bound 56 Fe-P450cam with PA for 8 ms (PA:P450 ratio 50:1). Experimental parameters: T577 K; P5800 mW; n 59.64 GHz; modulation amplitude 0.10 mT; modulation frequency 100 kHz.

4. Discussion Heme iron-oxo species are intermediates in many heme proteins. In cytochrome P450 this species is regarded as the catalytically relevant oxygen species that attacks the substrate. Thus far, the electron and spin distribution of such a species has not been experimentally identified

¨ Fig. 5. Mossbauer spectra of (a) substrate-free 57 Fe-P450cam the EPR spectrum of which is shown in Fig. 1a. The solid line is a simulation with two components: 87% of the starting material (parameters given in Table 2) and 13% of a doublet which exhibits d 50.13 mm / s and DEQ 51.94 mm / s and which is attributed to Fe(IV) [2]; (b) of camphor-bound cytochrome 57 Fe-P450cam which was freeze-quenched with peroxyacetic acid (8-ms reaction time); PA (5 mM) to P450cam (1 mM) of 5:1 were mixed to give half of each concentration in the mixture. The solid lines are simulations calculated with the parameters given in Table 2.

Table 2 ¨ Mossbauer and Spin-Hamiltonian parameters of cytochrome

57

Fe-P450cam used for the simulations shown in Fig. 5

S

d (mm / s)

DEQ (mm / s)

h

g ]]

A /g m ] N N ] (T)

D 21 (cm )

E /D

Ref.

1/2 5/2

0.38 0.44

2.85 0.79

21.8 0.6

(1.91,2.26,2.45) (2.0,2.0,2.0)

(245,10,19) (218,218,218)

– 3.5

– 0.087

[24] [24]

¨ et al. / Journal of Inorganic Biochemistry 91 (2002) 586–596 V. Schunemann

593

Fig. 6. Effect of peroxyacetic acid on the time-dependent change of cytochrome P450cam concentration due to bleaching at 22 8C. The concentrations were calculated from the Soret-band absorption of P450cam in the absence (at 417.5 nm, ´mM 5125 mM 21 cm 21 ; lower trace) and in the presence (at 390 nm, ´mM 5102 mM 21 cm 21 ; upper trace) of camphor and of substrate-free P450 plus metyrapone (at 421 nm, ´mM 5125 mM 21 cm 21 ). P450 (3 mM), peroxyacetic acid (15 mM), camphor (500 mM), metyrapone (104 mM), 50 mM potassium phosphate buffer, pH 7. The inset shows the spectrum for substrate-free P450cam at different times (1 cm optical pathlength).

within the native reaction cycle. In CPO complementary ¨ EPR and Mossbauer studies on freeze-quenched samples of CPO reacted with peroxyacetic acid [3,4] led to the conclusion that cpd I has one oxidation equivalent residing on the Fe(IV)=O unit (d 4 , S51), and one on the porphyrin (S951 / 2), constituting an antiferromagnetically coupled ferryl iron-oxo porphyrin p-cation radical system. This electronic structure has been generally assumed to occur in cytochrome P450 due to the same cysteine proximal hemeiron ligand as in CPO. Our studies, however, indicate that this is in fact not the case when intermediates are detected after approximately 8 ms. In contrast to an expected porphyrin p-cation radical, a tyrosine radical, presumably Tyr96, was observed when cytochrome P450cam reacts with peroxyacetic acid via a shunt pathway [2]. The iron is in the Fe(IV) state, as expected. Such an Fe(IV)=O tyrosine-radical system is formed after 8 ms to only 10– 15% in the absence of substrate. The remaining 85–90% of P450 remains in the native Fe(III) state. This observation is remarkable since it calls into question the nature of the P450 intermediate as a Fe(IV)=O porphyrin p-cation radical system, as well as the interpretation of stoppedflow experiments [9,10] which span the same reaction time of 6–9 ms, as our own freeze-quench studies (8 ms). The important finding of the present study is that in the presence of camphor only minor amounts of tyrosine radical (ca. 0.4%) are formed (reaction time 8 ms and concentration ratio peroxyacetic acid to P450cam of 5:1). The linear increase in the amount of radical formation with increasing ratio of peroxyacetic acid to camphor-bound P450cam (Fig. 5) indicates that the access of the active center for peroxyacetic acid is restricted by camphor. It

also explains that only trace amounts of hydroxylation product are found (Fig. 7a). This is supported by the degradation behavior of P450cam in the presence of peroxyacetic acid as shown in Fig. 6. In the absence of camphor and with a peroxyacetic acid to P450 ratio of 5:1, the Soret band of P450cam is bleached with a rate of |0.027 s 21 , and this is slowed down significantly in the presence of camphor. At a high peroxyacetic acid to P450 (camphor-bound) ratio of 38:1, bleaching was observed with a rate of only |0.001 s 21 . The inhibitory effect of metyrapone on bleaching indicates that peroxyacetic acid reacts with the heme iron before degradation occurs. Cleavage of the porphyrin ring follows the formation of the iron-oxo species. The kinetic data, however, show that for reaction times of 8–200 ms P450 degradation by peroxyacetic acid is negligible and plays no role in formation of the tyrosine radical and the Fe(IV) state. Recently, Davydov et al. [5] found various intermediates within the reaction cycle by stepwise annealing of the reduced oxygen complex by increasing the temperature up to approximately 220 K. Although product formation was observed, the proposed cpd I (high-valent iron-oxo porphyrin p-cation radical system) could not be identified. Unfortunately, the region around g|2 had been omitted in their EPR spectra, therefore a tyrosine radical, similar to our observations, was not confirmed by their experiments. On the other hand we have found no indication of the presence of ‘end-on ferric peroxo’ species with g5(2.24, 2.14, 1.96) and hydroperoxyferri-protein complexes with g5(2.30, 2.18, 1.94) as observed by Davydov et al. [5]. Our finding of an intermediate tyrosine radical in

594

¨ et al. / Journal of Inorganic Biochemistry 91 (2002) 586–596 V. Schunemann

Fig. 7. GC–MS spectra of chloroform extracts of the reaction mixtures of camphor-bound cytochrome P450cam with peroxyacetic acid and of the native reconstituted P450cam monooxygenase system; (a) top: gas chromatogram for reaction of peroxyacetic acid with camphor-bound P450cam in a ratio of 38:1, reaction time of 130 min at 5 8C, 1 mM camphor; middle: trace of the gas chromatogram for the signal m /z5152.15 (1R-camphor, expanded by 2) and for the signal m /z5168 (5-exo-hydroxycamphor, expanded by 100); bottom: mass spectrum at the integrated retention time from 7.95 to 8.38 min in the gas chromatogram. (b) Top: gas chromatogram for the native reconstituted monooxygenase system P450cam, putidaredoxin and putidaredoxin reductase, 20-min reaction time at 25 8C; bottom: mass spectrum at the integrated retention time from 8.06 to 8.21 min in the gas chromatogram.

conjunction with an iron(IV) is in line with various recent observations for cpd I in other heme proteins. While cpd I in cytochrome c peroxidase with its Fe(IV)=O tryptophan radical system had been regarded as an exception for many years, there are now several recent examples of where a radical on the protein, in particular a tyrosine radical, has been observed. Miller et al. [32] found that for the Phe172Tyr mutant of horseradish peroxidase in the reaction with H 2 O 2 , approximately 10% of the mutant carries a protein radical, presumably Tyr172. Ivancich and co-workers [27,33,34] observed a tyrosine radical by EPR when bovine liver catalase reacts with peroxyacetic acid. This radical is directly formed by electron transfer to the porphyrin p-cation radical. These workers also reported the formation of a tyrosine radical after treatment of turnip peroxidase isozymes with hydrogen peroxide at pH 7.7, accompanied by the formation of a ferryl iron-oxo porphyrin p-cation radical system [29]. Tyrosine radicals had been found earlier for prostaglandin H synthase [35] and cytochrome c oxidase [36]. Even for cytochrome c per-

oxidase the oxidation of tyrosine is favored over tryptophan oxidation when tryptophan is removed or its interaction with the environment (Asp235) is perturbed [37]. Our studies on CPO with reaction times up to 3 min indicate the formation of yet a minor organic radical (|1% relative spin density). Due to its low concentration we were not able to assign it to a specific group in the protein. Summarizing our observations in the light of these results on other heme proteins, we note that in P450cam intramolecular electron transfer occurs from a nearby tyrosine to the porphyrin p-cation radical. This might be formed transiently in a time period significantly shorter than 8 ms. In the natural reaction cycle such electron transfer would compete with the decay of cpd I due to oxygen transfer to the substrate bound close to the heme iron-oxo species. However, in the absence of a substrate, electron transfer might be the initial step in the decay cascade of cpd I. The distance of the tyrosine to the porphyrin, and the reorganization energy for removal of an electron (in concert with hydrogen atom abstraction) from

¨ et al. / Journal of Inorganic Biochemistry 91 (2002) 586–596 V. Schunemann

GC mCPBA MS NADH PA P450 P450cam

ZFS

595

gas chromatography m-chloroperoxybenzoic acid mass spectrometry nicotine amide adenine dinucleotide, reduced form peroxyacetic acid cytochrome P450 cytochrome P450cam from Pseudomonas putida (CYP101) which catalyzes the hydroxylation of 1R-camphor zero field splitting

Acknowledgements

Fig. 8. Structure of the active site of cytochrome P450cam iron-oxo complex (PDB entry code 1DZ9 [6]). The dotted lines indicate the electron pathway from Tyr-96-OH via a through-space jump to Thr-101CA, along the bonds to Thr101-OH and from there in a through-space jump to the CA-carbon of the propionic acid side group of the heme. The ˚ The pathway has been calculated distance along this pathway is |8 A. using the Pathway program HARLEM [41]. The substrate camphor has been removed from the structure 1DZ9 because the radical was observed for the substrate-free protein. In the presence of camphor the pathway goes through the camphor directly to the iron-oxo group.

tyrosine then become the relevant parameters [38–40] which determine the detectable electronic structure of cpd I in cytochrome P450cam. The direct distance of the oxygen atom of Tyr96, presumably the source of the radical, is ˚ away from the heme iron and only |7.4 A ˚ away |9.4 A from the closest porphyrin meso-carbon atom. The application of the pathway program of Beratan and co-workers [41] yields an intramolecular pathway from Tyr96-OH via a through-space jump to Thr101-CA, along the bonds to Thr101-OH and from there in a through-space jump to the CA-carbon of the propionic acid side group of the heme ˚ and is (Fig. 8). The distance along this pathway is |8 A thus similar to the direct distance. In conclusion a yet to be identified cpd I of P450cam decays faster than 8 ms via an electron transfer from a nearby tyrosine residue, presumably Tyr96. This result motivates freeze-quench studies of P450 mutants where such an electron transfer is disturbed or even hindered. Such studies are presently underway in our laboratories and may help to elucidate the nature of cpd I in cytochrome P450s.

5. Abbreviations cpd I CPO ENDOR EPR

compound I chloroperoxidase from Caldariomyces fumago electron nuclear double resonance electron paramagnetic resonance

¨ We are grateful to Horst Honeck (Max-Delbruck-Center for Molecular Medicine, Berlin) for running the gas chromatography–mass spectrometry analysis. We thank John H. Dawson (University of South Carolina, Columbia, SC, USA) for the generous supply of 5-exo-hydroxy¨ camphor, Andrei Kariakin (Max-Delbruck-Center for Molecular Medicine) for purifying putidaredoxin and putidaredoxin reductase which was used for the activity studies, and Julian A. Peterson (University of Texas Southwestern Medical Center, Dallas, TX, USA) for providing the plasmids for putidaredoxin and putidaredoxin reductase to CJ. David N. Beratan is kindly acknowledged for providing the HARLEM program to CJ for electron transfer pathway calculations. Financial support is acknowledged from the Deutsche Forschungsgemeinschaft (Ju229 / 4-1,2 and TR97 / 26-1,2) and from the NIH (GM 57042).

References [1] D.F.V. Lewis, in: Cytochromes P450-Structure, Function and Mechanism, Taylor & Francis, London, 1996. ¨ [2] V. Schunemann, C. Jung, A.X. Trautwein, D. Mandon, R. Weiss, FEBS Lett. 479 (2000) 149–154. [3] R. Rutter, L.P. Hager, H. Dhonau, M. Hendrich, M. Valentine, P. Debrunner, Biochemistry 23 (1984) 6809–6816. [4] R. Rutter, L.P. Hager, J. Biol. Chem. 257 (1982) 7958–7961. [5] R. Davidov, T.M. Makris, V. Kofman, D.E. Werst, S.G. Sligar, B.M. Hoffman, J. Am. Chem. Soc. 123 (2001) 1403–1415. [6] I. Schlichting, J. Berendzen, K. Chu, A.M. Stock, S.A. Maves, D.E. Benson, R.M. Sweet, D. Ringe, G.A. Petsko, S.G. Sligar, Science 287 (2000) 1615–1622. [7] T. Egawa, T. Ogura, R. Makino, Y. Ishimura, T. Kitagawa, J. Biol. Chem. 266 (1991) 10246–10248. [8] T.C. Pederson, R.H. Austin, I.C. Gunsalus, in: V. Ullrich, I. Roots, A. Hildebrandt, R.W. Estabrook (Eds.), Microsomes and Drug Oxidations, Pergamon, Oxford, 1977, pp. 275–283. [9] G.C. Wagner, M.M. Palcic, H.B. Dunford, FEBS Lett. 156 (1983) 244–248. [10] T. Egawa, H. Shimada, Y. Ishimura, Biochem. Biophys. Res. Commun. 201 (1994) 1464–1469. [11] M.M. Palcic, R. Rutter, T. Araiso, L.P. Hager, H.B. Dunford, Biochem. Biophys. Res. Commun. 94 (1980) 1123–1127.

596

¨ et al. / Journal of Inorganic Biochemistry 91 (2002) 586–596 V. Schunemann

[12] S.G. Sligar, B.S. Shastry, I.C. Gunsalus, in: V. Ullrich, I. Roots, A. Hildebrandt, R.W. Estabrook (Eds.), Microsomes and Drug Oxidations, Pergamon, Oxford, 1977, pp. 202–209. [13] K. Kobayashi, T.I. Iwamoto, K. Honda, Biochem. Biophys. Res. Commun. 201 (1994) 1348–1355. ¨ [14] C. Jung, G. Hui Bon Hoa, K.-L. Schroder, M. Simon, J.P. Doucet, Biochemistry 35 (1992) 12855–12863. [15] G.C. Wagner, M. Perez, W.A. Toscano Jr., I.C. Gunsalus, J. Biol. Chem. 256 (1981) 6262–6265. [16] I.C. Gunsalus, G.C. Wagner, Methods Enzymol. 52 (1978) 166–188. [17] J.A. Peterson, M.C. Lorence, B. Amarneh, J. Biol. Chem. 265 (1990) 6066–6073. [18] C.M. Hosten, A.M. Sullivan, V. Palaniappan, M.M. Fitzgerald, J. Terner, J. Biol. Chem. 269 (1994) 13966–13978. [19] D.P. Ballou, Methods Enzymol. 54 (1978) 85–93. [20] H. Beinert, S. Albracht, Biochim. Biophys. Acta 683 (1982) 245– 277. ˚ [21] R. Aasa, T. Vanngard, J. Magn. Reson. 19 (1975) 308–315. [22] A.X. Trautwein, E. Bill, E.L. Bominaar, H. Winkler, Struct. Bond. 78 (1991) 1–95. [23] C.E. Schulz, R. Rutter, J.T. Sage, P.G. Debrunner, L.P. Hager, Biochemistry 23 (1984) 4743–4754. [24] M. Sharrock, P.G. Debrunner, C. Schulz, J.D. Lipscomb, V. Marshall, I.C. Gunsalus, Biochim. Biophys. Acta 420 (1976) 8–26. [25] T.L. Poulos, A.J. Howard, Biochemistry 26 (1987) 8165–8174. [26] T.L. Poulos, B.C. Finzel, A.J. Howard, Biochemistry 25 (1986) 5314–5322. [27] A. Ivancich, T.A. Mattioli, S. Un, J. Am. Chem. Soc. 121 (1999) 5743–5753.

[28] G. Bleifuss, M. Kolberg, S. Poetsch, W. Hofbauer, R. Bittl, W. Lubitz, A. Graeslund, G. Lassmann, F. Lendzian, Biochemistry 40 (2001) 15362–15368. [29] A. Ivancich, G. Mazza, A. Desbois, Biochemistry 40 (2001) 6860– 6866. ¨ [30] V. Schunemann, C. Jung, A.X. Trautwein, R. Weiss, Hyperfine Interactions (2002) accepted. [31] G.H. Loew, D.L. Harris, Chem. Rev. 100 (2000) 407–419. [32] V.P. Miller, D.B. Goodin, A.E. Friedman, C. Hartman, P.R. Ortiz de Montellano, J. Biol. Chem. 270 (1995) 18413–18419. [33] A. Ivancich, H.M. Joue, J. Gaillard, J. Am. Chem. Soc. 118 (1996) 12852–12853. [34] A. Ivancich, H.M. Joue, B. Sartor, J. Gaillard, Biochemistry 36 (1997) 9356–9364. [35] R. Karthein, R. Dietz, W. Nastainzyk, H.H. Ruf, Eur. J. Biochem. 171 (1988) 313–320. [36] G. Lang, K. Spartalian, T. Yonetani, Biochim. Biophys. Acta 451 (1976) 250–258. [37] L.A. Fishel, M.F. Farnum, J.M. Mauro, M.A. Miller, J. Kraut, Biochemistry 30 (1991) 1986–1996. [38] C.C. Moser, J.M. Keske, K. Warnke, R.S. Farid, P.L. Dutton, Nature 355 (1992) 796–802. [39] C.C. Page, C.C. Moser, X. Chen, P.L. Dutton, Nature 402 (1999) 47–52. [40] C.T. Tommos, J.J. Skalicky, D.L. Pilloud, A.J. Wand, P.L. Dutton, Biochemistry 38 (1999) 9495–9507. [41] A.J.A. Aquino, P. Beroza, D.N. Beratan, J.N. Onuchic, Chem. Phys. 197 (1995) 277–288.