REVIEW ARTICLE Unraveling the catalytic mechanism of lactoperoxidase and myeloperoxidase A reflection on some controversial features Elena Ghibaudi and Enzo Laurenti Dipartimento di Chimica I.F.M., Universita ` di Torino, Italy Although belonging to the widely investigated peroxidase superfamily, lactoperoxidase (LPO) and myeloperoxidase (MPO) share structural and functional features that make them peculiar with respect to other enzymes of the same group. A survey of the available literature on their catalytic intermediates enabled us to ask some questions that remained unanswered. These questions concern controver- sial features of the LPO and MPO catalytic cycle, such as the existence of Compound I and Compound II isomers and the identification of their spectroscopic properties. After addressing each of these questions, we formulated a hypo- thesis that describes an integrated vision of the catalytic mechanism of both enzymes. The main points are: (a) a re-evaluation of the role of superoxide as a reductant in the catalytic cycle; (b) the existence of Cpd I isomers; (c) reci- procal interactions between catalytic intermediates and (d) a mechanistic explanation for catalase activity in both enzymes. Keywords: lactoperoxidase; myeloperoxidase; aminoacid radical; Compound I; Compound II; Compound III; catalytic intermediates. Introduction The catalytic cycle of peroxidases, including lactoperoxidase (LPO) and myeloperoxidase (MPO), is described usually as a sequence of three consecutive reactions, according to Scheme 1. Compound I (Cpd I), which arises from the reaction of the native enzyme with hydrogen peroxide (H 2 O 2 ), is two oxidizing equivalents above the resting state. It reacts with a substrate molecule and is converted into a secondary compound that has lost one equivalent, generally indicated as Compound II (Cpd II). A second substrate molecule recycles Cpd II into the resting enzyme. A large excess of H 2 O 2 converts Cpd I into the inactive intermediate, Com- pound III (Cpd III). The two oxidizing equivalents of Cpd I are on an iron ion, that assumes the formal oxidation state IV, and on the porphyrin ring, which becomes a cationic radical. Cpd II has been shown to contain Fe IV ¼O [1–3], whereas Cpd III is an enzyme adduct with superoxide [2–6]. Depending on the type of peroxidase, Cpd III formation may be reversible, whereby it can be reconverted into an active form of the enzyme, or irreversible, in which case it is associated with degradation of the enzyme. Moreover, a few peroxidases, e.g. haloperoxidases, can oxidize halides through the bielectronic reduction of Cpd I that is converted back to the resting state without forming Cpd II [3,7,8]. The generally accepted definition of the three intermedi- ates of this class of enzymes can be misleading. In fact, when comparing different peroxidases, the same name is applied to species with distinct electronic structures. Moreover, several peroxidase intermediates are known where the unpaired electron is localized onto an amino acid of the protein scaffold [9–14] and this aspect is not taken into account by the classical peroxidase cycle. Within this context, we propose to re-examine some of the literature data describing the catalytic cycle of two mammalianperoxidases,LPOandMPO,inorderto reconcile the apparent inconsistencies and to provide some new insights. MPO and LPO share functional and structural homology, reflecting their common phylogenetic origin [15] and participate in antimicrobial host defense, generating potent reactive species by the oxidation of halides or pseudohalides. Based on our survey of the experimental data concerning the reactivity of these enzymes, we formulated four questions that are focused on controversial features of the LPO and/or MPO catalytic cycle: (a) is formation of Cpd I reversible (or do mammalian peroxidases possess catalase activity); (b) does Cpd I exist in two isomeric forms, containing the porphyrin radical and the amino acid radical (aa +• ), respectively; (c) as the conversion of Cpd I fi Cpd II occurs spontaneously in the presence of peroxide, which is the reducing agent in this reaction step and (d) are the optical spectra of Cpd I–[Fe IV ¼O; aa +• ]andCpdII identical? Correspondence to E. Ghibaudi, Dipartimento di Chimica I.F.M., Universita ` di Torino, Via Giuria 7 – 10125 Torino, Italy. Fax: + 39 011 670 7855, Tel.: + 39 011 670 7951, E-mail: elena.ghibaudi@unito.it Abbreviations: LPO, lactoperoxidase; MPO, myeloperoxidase; Cpd I–III, Compound I–III. Enzymes: lactoperoxidase (EC 1.11.1.7); myeloperoxidase (EC 1.11.1.7). (Received 10 April 2003, revised 18 July 2003, accepted 23 September 2003) Eur. J. Biochem. 270, 4403–4412 (2003) Ó FEBS 2003 doi:10.1046/j.1432-1033.2003.03849.x After addressing each of these questions, we will propose a new hypothesis that describes an integrated vision of the catalytic mechanism of MPO and LPO. A brief survey of the absorption features of MPO, LPO and their intermediates The absorption spectrum of MPO is characterized by an intense Soret peak at k max ¼ 430 nm and weaker bands at 570, 620 and 690 nm, whereas at 370 nm and 496 nm, two shoulders are evident [16–20]. The Soret molar extinction coefficient (e)is89000 M )1 Æcm )1 per heme in human MPO [4,18,20–22] and 95 000 M )1 Æcm )1 perhemeincanineMPO [16]. These features change upon generation of Cpd I: it still absorbs at 430 nm, but e is about 50% lower than in native MPO [18,19] (Table 1); moreover, it shows bands at 572 and 625 nm [16]. Stopped-flow measurements are required to detect Cpd I spectrum, due to its short half-life (t 1/2  100 ms [16]). According to Harrison et al. [16], a 40-fold excess of peroxide is required to obtain a good yield. Other authors [18,19] claim that the amount of peroxide required in the reaction depends on the purity of the enzyme, as impurities in the preparation might be oxidized and thus contribute to additional consumption of H 2 O 2 .In the literature, there is good agreement on the fact that the conversion of native MPO into Cpd I is a bimolecular reaction; three k app values are reported and they are almost coincident (2.3 · 10 7 [23]; 1.4 · 10 7 [19]; 1.8 · 10 7 M )1 Æs )1 [18]). In the presence of peroxide, Cpd I converts sponta- neously to a secondary compound, called Cpd II. This is relatively stable (t 1/2  a few minutes) [18,23] and absorbs at 455 and 628 nm (Table 1). In the presence of large peroxide excess, Cpd I yields Cpd III, an intermediate that shares with Cpd II some spectroscopic features (e.g. bands at 450 nm and 625 nm) (Table 1), thus making it difficult to discriminate between these two species. The absorbance at 450 nm (Soret) is useless for distinguishing between Cpd II and Cpd III in MPO. A system for evaluating the Ôrelative amountÕ of the two species has been proposed based on the A 625 /A 456 value [24]. Cpd II would produce a value of 0.20 at neutral pH and 0.25 at basic pH, whereas, Cpd III gives 0.52 at neutral pH. In no case is it possible to quanti- tate accurately Cpd III, as a residual amount of Cpd II is always present [24]. The optical spectrum for native LPO shows a Soret band at 412 nm (e ¼ 114 000 M )1 Æcm )1 ), and weaker absorptions at 501, 541, 589 and 631 nm [25]. Exposure to equimolar or twofold excess of H 2 O 2 produces Cpd I, whose spectros- copic features are controversial. According to reference [25], it absorbs at 410 nm (Soret) and 562, 600, 662 nm (Table 2), whereas, Doerge et al. [26] and Monzani et al. [27] reported that the porphyrin-radical Cpd I is characterized by absorp- tions at 420 and 416 nm, respectively. Such a change in k max with respect to the native form suggests that, in contrast with MPO, not only the transition probability (and thus e) but the energy of the porphyrin electronic levels is affected by Cpd I formation. The reaction is extremely favoured, as witnessed by its high second-order constant (1.8 · 10 7 [28]; 1.2Æ10 7 M )1 Æs )1 [29]), and is more influenced by lipophilicity and steric factors than by pH [28]. In the presence of peroxide only, LPO Cpd I converts spontaneously (within 200 ms) [29] into a relatively stable species (t 1/2  several minutes [30]) with spectroscopic features clearly distinct from those of thenativeenzyme(k Soret ¼ 430 nm). The electronic structure of such species is still uncertain. According to several authors [26,27,31–33], this unknown species would be a porphyrin- radical Cpd I isomer, having its electronic vacancy localized onto an amino acid residue instead of the heme ring. Such a compound would in turn generate conventional Cpd II that is characterized by absorptions at 430, 535 and 567 nm [3,26,27,30,32,34–37]. On the basis of kinetic, thus, indirect evidence, Monzani et al. [27]suggestthatCpdIIcan isomerize also, giving rise to a radical species with k Soret at 412 nm. Exposure to a slight peroxide excess rapidly inactivates LPO and generates Cpd III (with bands at 424, 550 and 588 nm) (Table 2). In contrast to MPO, LPO Cpd III spectral pattern is clearly distinct from that of Cpd II and can be easily distinguished from it. LPO is much more sensitive to peroxide inactivation than is MPO or other members of the animal peroxidase family [4,27]. Scheme 1. The catalytic cycle of peroxidases described as a sequence of three consecutive reactions. Peroxidases first react with H 2 O 2 , their first substrate, and generate a highly oxidizing intermediate, indicated as Cpd I. In the presence of peroxide excess, the intermediate Cpd III [a superoxide-Fe(III) adduct] can be generated. Cpd I is able to oxidize mono-electronically a second substrate, that can be either an organic or an inorganic molecule; it is simultaneously reduced to a second intermediate, named Cpd II – this can still oxidize a substrate mole- cule; it can also react with peroxide and give rise to Cpd III. Table 1. Summary of the spectroscopic properties of MPO, Compound I, II and III. Reference Compound I (nm) Compound II (nm) Compound III (nm) [18] 430 (e ¼ 88000) dimer 455 (e ¼ 160000) – [4] – – 450, 626 [1] – 628 (e ¼ 18000) at pH 7 – 635 (e ¼ 25000) at pH 11 [24] – 454 (452–458) 450 (449–452) 628 (622–630) 625 (622–625) [17] – 455 451 628 626 [16] 425 (e ¼ 52000) 455 – 572 (e ¼ 9700) 625 (e ¼ 9900) [39] – 452 – 622 4404 E. Ghibaudi and E. Laurenti (Eur. J. Biochem. 270) Ó FEBS 2003 Four unanswered questions Reversibility of Cpd I formation and catalase activity Catalase activity in peroxidases is strictly related with the reversibility of Cpd I formation. Although this reaction had been long considered strictly irreversible, several forms of evidence against this statement are now available. In the case of MPO, reversibility of Cpd I formation was first stated by Wever et al. [38] and subsequently confirmed by Marquez et al. [18] and Kettle et al. [22]. As for the presence of pseudocatalase or catalase activity, this was suggested by several authors [16,18,38–40] and demonstra- ted unambiguously by Kettle and Winterbourne [22]. They showed that MPO possess a true catalase activity, whose extent is so important, compared to other peroxidases (e.g. horseradish and chloroperoxidase) that one may look at MPO as a catalase/peroxidase. A first-order rate constant of 2.2 · 10 6 s )1 has been reported for the catalatic breakdown of Cpd I [22]. In order to explain their experimental findings, Kettle et al. [22] proposed a reaction scheme (Scheme 2) that unifies the previous mechanisms proposed by Iwamoto et al. [40] and Marquez et al. [18]. According to experimental findings, H 2 O 2 either reduces Cpd I to native MPO by a two-electron reaction or to Cpd II by a one-electron process. The latter reaction is about two orders of magnitude slower than the former. Superoxide contributes to maintain the catalase activity by preventing Cpd II accumulation and reducing this inter- mediate to the native form [41]. Kettle’s proposal [22] does not take into account the possibility (suggested by Marquez et al. [18]) that superoxide act as a reductant towards Cpd I as well, thus, generating Cpd II. This possibility has been included in Scheme 2. Further support for the presence of catalase activity comes from the E°¢ values of the redox couples H 2 O 2 /O 2 and Cpd I/native MPO, the former being 0.281 V [42] and the latter 1.16 V [19,43], from which it is evident that Cpd I can oxidize bi-electronically H 2 O 2 fi O 2 . MPO and LPO share so many structural and functional features that one would expect LPO to display catalase activity as well. Nevertheless, a final word on this item has not yet been mentioned. In fact, LPO has been proven by several authors to possess pseudo-catalase activity [3,30,36,44,45]. This derives from oxidation of H 2 O 2 by hypoiodite, which in turn is generated by the two-electron oxidation of iodide catalysed by LPO Cpd I. So, pseudo- catalase activity is related to the ability of certain substrates (i.e. halides and pseudo-halides) to undergo a two-electron oxidation, thus, preventing formation of Cpd II and converting Cpd I directly to native LPO [3,46]. As for the catalase reaction in LPO, the following evidence is available: Huwiler et al. [35] described oxygen release by LPO in the presence of a slight excess of peroxide, suggesting that LPO can exhibit catalase-like behaviour. Kohler et al. [30] report O 2 release and peroxide con- sumption in the stoichiometric ratio 1 : 2 (typical of catalase activity) both during the conversion of Cpd III into the native form and in the Cpd II fi native LPO reaction. In order to provide a mechanistic explanation for these experimental findings, they propose catalase activity to stem from a reaction loop involving Cpd III, Cpd II and a ferrous form of LPO. Although,theycannot exclude the intervention of H 2 O 2 as electron-donor. Such a reaction scheme is cited Table 2. Summary of the spectroscopic properties of LPO, Compound I, II and III. Reference Compound I p +• (nm) Compound I aa +• (nm) Compound II Ferryl (nm) Compound II Fe(III) aa +• (nm) Compound III (nm) [29] 410 (562, 600, 662) – 433 (537, 568) – 428 (551, 590) [32] – 430 430 – – [34] – – 430 – 423 [35] – – 430 – 423 [36] – – 430 (535, 565) – 423 (549, 588) [37] – – – – 423 [30] – – 430 (535, 567) – 424 (550, 588) [4] – – – – 420 (550, 589) [3] – – 430 (535, 567) – 424 (550, 588) [26] 420 430 430 – – [31] – 430 – – – [27] 416 430 430 412 424 Scheme 2. A reaction scheme that unifies the previous mechanisms proposed ([18,22] and [40]). Once MPO Cpd I is formed, H 2 O 2 can reduce Cpd I either to native MPO by a two-electron reaction or to Cpd II by a one-eletron process. The former reaction implies catalase activity. The latter reaction is about two orders of magnitude slower than the former and generates superoxide. This radical can either reduce Cpd II to the resting enzyme or Cpd I to Cpd II and generate dioxygen [18,22]. Ó FEBS 2003 Controversies on myelo- and lactoperoxidase mechanisms (Eur. J. Biochem. 270) 4405 again in reference [3] and a similar hypothesis is made by Jenzer et al. [36]. They hypothesize that peroxide acts as a reductant in the Cpd I fi Cpd II reaction, thus generating superoxide: this may in turn reduce Cpd II fi native LPO, while O 2 is released. According to this scheme, the stoichio- metric ratio between H 2 O 2 and O 2 would be 2 : 1, which is typical of catalases; although it seems to us that such an activity cannot be defined as truly catalatic, as it is superoxide-mediated. We would rather stress the fact that Jenzer et al. [36] observed a biphasic behaviour for peroxide consumption and O 2 release that is characterized by a quick initial step and a slow secondary one. This is an extremely interesting result in the light of the experimental findings by Kettle et al. on MPO [22] who observed a burst phase for peroxide consumption (as above). In fact, this set the stage for an extension to LPO of the reaction scheme adopted for MPO. Another argument relevant to catalase activity is the Cpd I redox potential. Unlike MPO, no E°¢ values for LPO Cpd I have been available until now; actually, its value is a point of considerable contention. Ohlsson et al. [28,47] reported the redox potential of LPO to be )191 mV at 25 °C and pH 7.0; however, this value refers to the Fe(III)/Fe(II) reduction and does not relate with the redox change undergone by the iron ion during the catalytic cycle. To our knowledge, no E°¢ value for the CpdI/native LPO redox couple has been reported until now. In spite of this fact, an estimate of the LPO Cpd I redox potential can be made, by considering that: (a) LPO cannot oxidize chloride whereas MPO does and (b) LPO Cpd I oxidizes phenolic substrates characterized by redox potentials that range from 760 to 1060 mV [27]. Consequently, an E°¢ value between 1.16 and 1.06 mV seems reasonable for LPO Cpd I-[Fe IV ¼ O; p +• ]; a value of  900 mV has been suggested for Cpd I-[Fe IV ¼ O; aa +• ] [27,46]. Such an estimate is consistent with the presence of catalase activity, based on similar arguments to those offered for MPO. In conclusion, experimental evidence for O 2 evolution concurrent to peroxide consumption are available for LPO; the estimated value of E°¢ for LPO Cpd I also supports the presence of catalase activity. As for the mechanistic details, we believe that the reaction mechanism adopted for MPO could be extended to LPO, although this should be unequivocally demonstrated by further studies. Isomeric forms of Cpd I It is well known that, in the absence of cosubstrates, MPO and LPO Cpd I decays spontaneously to a secondary product that is generically indicated as Cpd II [7,16]. Whether this intermediate derives from a mono-electronic reduction of Cpd I and, thus, is the actual Cpd II, or is a Cpd I isomer that contains the same number of oxidizing equivalents as Cpd I is matter of discussion. Several pieces of evidence suggest the existence of two Cpd I isomers both in MPO and LPO. In the case of MPO the evidence is: (a) the EPR spectrum of a trapped amino acid radical found by Lardinois et al. [48] in at least one catalytic intermediate of MPO and (b) the high redox potential of this intermediate, which is able to oxidize chloride. The E°¢ for MPO Cpd I has been found to be 1.16 V [43,49], a high enough value to make the oxidation of an amino acid residue on the polypeptide chain possible by intramolecular electron-transfer process [18]. Displacement of an oxidizing equivalent from the porphyrin ring to an amino acid would be expected to change the redox potential of the protein. The two Cpd I isomers are not expected to share the same redox properties, thus, agreeing with experimental evidence regarding the reactivity of MPO intermediates. The species obtained by spontaneous decay of Cpd I, which is usually called Cpd II regardless of its electronic structure, does not react with halides, whereas Cpd I can oxidize them [19]. On the other hand, this fact does not represent proof of the presence of one or two oxidizing equivalents on the decay product of Cpd I. The EPR spectrum of a spin-trapped radical in MPO Cpd I has been reported recently [48]. The signal, assign- able to an amino acid radical localized on the protein scaffold of MPO was generated only in the presence of H 2 O 2 . The nature of such a radical has not been ascertained, although the authors suggest that it might be a Tyr or a Trp. There is no unambiguous evidence implicating this radical intermediate in the catalytic mechanism of the enzyme and it might represent an alternative pathway, whose biological role is still unclear [48]. The same authors [48] suggest that the protein radical might stem from a vestigial process related to that responsible for the autocatalytic cross-linking of the heme groups to the protein [50]. Alternatively, they suggest that it could be involved in the oxidation of peroxidase substrates at the protein surface, akin to the activities of Cytochrome c oxidase or ligninase. Finally, one cannot exclude that this intermediate is part of a mechanism unrelated to the physiological role of MPO but that leads to eventual enzyme inactivation [48]. It is important to recognize that the EPR signal of the trapped radical in MPO clearly derives from the super- position of two different components that could correspond to two adducts of 3,5-dibromo-4-nitroso-benzenesulfonic acid with radicals having different mobility or to two forms of the protein present in solution simultaneously [48]. In the case of LPO, both spectral [51,52, E. Ghibaudi and E. Laurenti, unpublished observation] and kinetic data [26,27,31–33,53,54] suggest that a protein radical forms during the enzyme turnover. Lardinois et al. [51] used mass spectrometry of tryptic digests of LPO and EPR spectros- copy of spin-trapped species to demonstrate that there are two radical species, each of which might be a Tyr. One residue has been identified as the Tyr289, which is involved in LPO dimerization and has no catalytic role. Tyr289 could also be involved in H 2 O 2 -mediated cross-linking between LPO and myoglobin [55]. The precise identity of the second residue has not been determined but may play a functional role. In contrast, Goff et al. [52] have implicated two Phe radicals in the catalytic turnover of the enzyme. A multi- frequency (9 and 285 GHz band) EPR investigation at liquid helium temperature, performed by the present authors, has shown that the reaction of LPO in a slight excess of peroxide at 0 °C generates a protein radical, whose identity has yet to be defined (E. Ghibaudi and E. Laurenti, unpublished observation). The indirect evidence available on this problem comes essentially from two types of experiments: (a) titrations of the decay product of LPO Cpd I with ferrocyanide (a monoelectronic donor) and (b) kinetic studies on the role 4406 E. Ghibaudi and E. Laurenti (Eur. J. Biochem. 270) Ó FEBS 2003 of LPO in the frame of the thyroid hormone biosynthetic pathway. Courtin et al. [32,33] demonstrate, by titration with ferrocyanide, that the decay product of Cpd I-[Fe IV ¼O; p +• ], characterized by k Soret ¼ 430 nm, is still two oxidizing equivalents above the resting state of the enzyme and, thus, is isoelectronic with Cpd I-[Fe IV ¼O; p +• ]. By analogy to Cyt c peroxidase, they suggest that one equivalent is localized on the protein scaffold, designated Cpd I- [Fe IV ¼O; aa +• ]. At pH 6.1 and 7.3, reduction of Cpd I- [Fe IV ¼O; aa +• ] by ferrocyanide requires 2 mol of oxidant per mol of enzyme and is a biphasic process. This might indicate that such a reaction occurs in two mono-electronic steps, yielding first the genuine Cpd II, which is only one oxidizing equivalent above the resting state, and subsequently native LPO. The second source of indirect evidence comes from studies using LPO in place of thyroid peroxidase (TPO) in the biogenesis of thyroid hormone. This synthetic process occurs in two steps, tyrosine iodination followed by coupling of iodotyrosines to produce the hormone. Comparison of the kinetics of these two steps [31,32,54,56] catalysed by LPO or TPO shows that the reaction intermediate containing the porphyrin cationic radical Cpd I-[Fe IV ¼O; p +• ] catalyses both steps. On the contrary, its decay product is active only in the second step. This indicates that the two species possess different redox potentials, that of Cpd I-[Fe IV ¼O; p +• ]being higher than its decay product. Transfer of an oxidizing equivalent from the porphyrin ring to an amino acid residue is likely to reduce the redox potential of the protein [53] and, on this basis, the authors [31,32,54,56] claim that the species responsible for iodotyrosine coup- ling is the isomer Cpd I-[Fe IV ¼O; aa +• ]. A comparison between HRP, CcP and LPO [56] supports this statement as HRP, which can generate only Cpd I-[Fe IV ¼O; p +• ], can iodinate tyrosine, whereas CcP, which forms Cpd I-[Fe IV ¼O; aa +• ] only, cannot. The same argument is used to explain the reactivity of LPO towards sulphur- containing compounds [53], suggesting the presence of a reaction intermediate (Cpd I-[Fe IV ¼O; aa +• ]) that would be a weaker oxidant than is traditional Cpd I. In our opinion, these arguments are valid but still do not answer the question in a satisfactory fashion, as they do not exclude the possibility that the coupling of Tyr residues is mediated by Cpd II, whose oxidizing power is lower than that of Cpd I. Also unclear is the physiological role of the amino acid radical that forms only in the presence of an ÔactiveÕ form of LPO, as shown by Lardinois et al. [51]. It cannot be generated in the presence of inhibitors and has been demonstrated only when LPO was incubated with peroxide in the absence of other substrates, suggesting that it may not form under physiological conditions. Monzani et al. [27] suggest that its formation may represent a way for the enzyme to avoid undesired reactions and thereby reserve the oxidizing equivalents for the oxidation of thiocyanate. Lardinois et al. [51] posit that it is part of a Ôdead-endÕ metabolic pathway, aiming to protect LPO from the disrupting effect of H 2 O 2 or be part of a vestigial enzymatic pathway, now abandoned. O’Brien suggests its role is in the oxidation of physiologically relevant substrates, drugs and in the etiology of some pathologies [57]. In the light of the work by De Pillis et al. [50,58], we would not exclude that the radical plays a role in the autocatalytic binding of heme to the apoprotein, a post-translational modification that is typical of mammalian peroxidases. Cpd I fi Cpd II conversion The remarks made in the previous section set the stage for another question: why is Cpd I decay to Cpd II in the presence of H 2 O 2 spontaneous, as such a reaction would require a substrate able to reduce Cpd I? First, we should emphasize that in this paper we assign the term ÔCpd IIÕ to the reaction intermediate that actually lacks an oxidizing equivalent with respect to Cpd I and must be generated through a mono-electronic reduction. MPO behaviour has already been described above. Marquez et al. [18] and Kettle et al. [22] showed that, in the absence of exogenous substrates, peroxide acts as a mono-electronic reductant and converts MPO Cpd I fi Cpd II; as a result of peroxide oxidation, superoxide ions are produced. This species might be found either free in solution or bound to the enzyme, thus, giving rise to Cpd III. Superoxide (or Cpd III) would enter the cycle again, by acting in turn as a reductant (Scheme 2). This model is also supported by the fact that the kinetics of the Cpd I fi Cpd II reaction is either monophasic or biphasic, at low or high peroxide concentration, respectively. Consistent with this hypothesis, experimental evidence demonstrates that superoxide is generated during this chemical process in a pH-dependent fashion [18]. The kinetic constant for Cpd I fi Cpd II conversion at pH 7.0 has been reported to be 8.2 · 10 4 M )1 Æs )1 , which is three to fourfold the k of the bielectronic reduction of Cpd I fi MPO by chloride [19]. Nonetheless, we would like to emphasize that the above- mentioned hypothesis does not exclude the possibility that Cpd I-[Fe IV ¼O; p +• ] decays first to Cpd I-[Fe IV ¼O; aa +• ], whichinturnisreducedtoCpdII. LPO behaviour is less straightforward. Cpd I spontane- ously converts to a relatively stable species [4,30,35,36] that absorbs at 430 nm but the reducing agents responsible for this process and the subsequent reaction step have yet to be defined. In fact, Cpd II slowly converts back to native LPO [30] through a mono-electronic reaction. A comparison of the experimental behaviours of MPO and LPO in the presence of peroxide might help to solve some of these issues. The mono-electronic oxidation of H 2 O 2 generates super- oxide that can co-ordinate to the Fe(III) ion and yield Cpd III. LPO is much more sensitive than MPO to inactivation by peroxide [4,27] and produces Cpd III even in the presence of slight H 2 O 2 excess, although its inacti- vation is partially reversible [3,30,34,35,37]. By analogy to MPO, we suggest that the superoxide ion or, more probably, its adduct with the enzyme, that is Cpd III, reduces both Cpd I, regardless of its electronic structure, and Cpd II according to Scheme 3. Dioxygen release from the conversion of Cpd II fi LPO [30,36], presumably by the oxidation of superoxide anion, provides experimental data to support this model. In the first step, Cpd II would be produced, whereas native LPO would be generated in the second (Schemes 2 and 3). Ó FEBS 2003 Controversies on myelo- and lactoperoxidase mechanisms (Eur. J. Biochem. 270) 4407 Two groups have independently suggested that LPO Cpd II exists in two isoelectronic forms, Cpd II-[Fe IV ¼O] and Cpd II-[Fe III ;aa +• ], whose stability would be strongly pH-dependent [27,32,33]. The individual isomers could be distinguished by their optical spectrum as they contained iron in two different oxidation states. At acidic pH, Cpd I reduction with an equimolar amount of ferrocyanide yielded Cpd II-[Fe III ;aa +• ] whose optical spectrum is identical to that of the native enzyme [31,33]. At neutral or basic pH, the amino acid site would be preferentially reduced, giving rise to Cpd II-[Fe IV ¼O], which lacks radicals and absorbs at 430 nm. This species would reconvert into Cpd II-[Fe III ;aa +• ], by changing pH, whereas the reverse reaction would not be possible. Addition of a second ferrocyanide molecule reconverts the enzyme to the native form. Such a mechanism, if confirmed, is analogous to that proposed for Cyt c peroxidase that is thought to generate two different intermediates, each being one oxidizing equivalent above the resting enzyme [32,59]. Nevertheless, we believe that the evidence for two Cpd II isomers is not compelling and merits additional investiga- tion as the correlation between iron redox state and optical spectrum of the protein is not straightforward. For example, the conversion of MPO from the resting state, containing Fe(III), to Cpd I, containing Fe(IV), affects e without shifting the Soret band. Most kinetic data offered to support this model comes from studies using different experimental conditions [27]. All experimental evidence of radical LPO intermediates [32,33] was obtained exclusively in the pres- ence of H 2 O 2 alone, a non-physiological condition. We believe that the absence of direct evidence of the radical in the presence of cosubstrates limits the validity of extrapo- lating the model proposed by Courtin et al. [32,33] to other experimental conditions. Optical spectra of Cpd I-[Fe IV =O; aa +• ] and Cpd II Once the existence of Cpd I-[Fe I ¼O; aa +• ] is accepted, its optical features remain to be defined. In the light of the experimental data available on MPO, one can think of two possibilities (Scheme 4). First, the unstable intermediate Cpd I-[Fe IV ¼O; p +• ] decays to the relatively stable Cpd II, which is characterized by the bands at k max 455 and 628 nm. Were this hypothesis correct, the protein radical observed by Lardinois et al. [48] would not participate in the traditional catalytic cycle but rather belong to an alternative or to Ôdead endÕ pathway. Alternatively Cpd I-[Fe IV ¼O; p +• ] (t 1/2 ¼ 100 ms) decays first to Cpd I-[Fe IV ¼O; aa +• ], whose t 1/2 is still unknown, and subsequently to Cpd II (t 1/2  10 min) [4]. In the former case, the optical features of the intermediates are clearly assigned: Cpd I-[Fe IV ¼O; p +• ] absorbs at 430 nm and its decay to Cpd II is followed by a shift of the Soret band towards 455 nm. Such a band is stable for several minutes, a feature consistent with Cpd II’s t 1/2 .In the latter case, the peak absorption at 455 nm has yet to be assigned: it might be associated with either Cpd I-[Fe IV ¼O; aa +• ]orCpdII[Fe IV ¼O] or with both of them. It seems reasonable to speculate that the optical spectra of both intermediates Cpd I-[Fe IV ¼O; aa +• ] and Cpd II- [Fe IV ¼O] are extremely similar, if not identical. In fact, the two species differ only in the presence of a protein radical aa +• , a species not likely to dramatically modify the spectrum relative to that of the holoprotein. Neither the Fe redox state nor the electronic configuration of the porphyrin ring are expected to change upon conversion of one intermediate into the other. It is reasonable to assume that the transient species Cpd I-[Fe IV ¼O; aa +• ] shows an absorption at 455 nm and that its decay to Cpd II leaves the spectrum unaltered. Scheme 3. LPO reacts with peroxide to generate a porphyrin-radical intermediate that can isomerize spontaneously. By analogy to MPO, we suggest that the superoxide ion or, more likely, its adduct with the enzyme (Cpd III) reduces Cpd I and Cpd II, by losing an oxidising equivalent and gives rise to dioxygen, according to evidence available in the literature [30,36]. Some authors suggest that Cpd II isomerize as well [27]. Scheme 4. Two hypothesis for explaining the optical changes observed in MPO solutions upon addition of peroxide. Does the spectrum observed upon decay of the porphyrin radical Cpd I (430, 572 and 625 nm bands) corres- pond to the real Cpd II? In such a case, the optical features of Cpd I isomer have to be defined. The second hypothesis assigns the 455 and 625 nm bands to this isomer. In such a case, the spectral properties of Cpd II remains undefined. 4408 E. Ghibaudi and E. Laurenti (Eur. J. Biochem. 270) Ó FEBS 2003 The observed effect of pH on the band at 628 nm of Cpd II [1,24] might reflect protonation of an amino acid found in the heme crevice and involve Fe IV ¼O. As this group should be present in Cpd I as well, one would expect similar pH dependence for the spectrum of such inter- mediate. Unfortunately, no data are currently available, partly due to the short t 1/2 of Cpd I. As for LPO, the decay product of Cpd I absorbs at 430 nm [probably due to the presence of Fe(IV)] and is stable for several minutes [26,30,34–36] and is not associated with an EPR signal assignable to a radical species. On the basis of these facts, the spectrum is unlikely to correspond to Cpd I- [Fe IV ¼O; aa +• ], which is highly unstable at room tempera- ture and can only be detected at low temperature after rapid freezing of the sample (E. Ghibaudi and E. Laurenti, unpublished observation). The same arguments offered earlier for MPO pertain to LPO. Cpd I-[Fe IV ¼O; aa +• ] and Cpd II-[Fe IV ¼O] should form sequentially and their spectra are likely identical as these two intermediates differ only for the presence of the radical aa +• in the former. Based on the evidence presented above, it should be clear that the catalytic cycles of both LPO and MPO have unusual features with respect to the classical peroxidase mechanism and interactions with H 2 O 2 . In the light of the experimental data presented above, we offer the following conclusions: (a) both enzymes are claimed to possess catalase activity [16,18,19,22,26]. In the case of MPO, this wasproventobeassociatedwithCpdIreconversiontothe resting state [22]. As for the catalase function in LPO, it is supported by stoichiometric oxygen release measured during catalytic cycling in the absence of cosubstrates and by structural resemblance to MPO. As catalase activity recycles Cpd I into native LPO, one should expect accu- mulation of the resting enzyme to take place in the exclusive presence of peroxide, unless the extent of catalase vs. peroxidase activity is low. This seems to be the case for LPO, where the competition of catalase- vs. peroxidase- activity is weaker than in MPO and shifts all catalytic equilibria towards accumulation of Cpd II rather then native LPO. This should also explain the higher sensitivity of LPO towards peroxide inactivation with respect to MPO. Unanswered is the question: which factors are discrimi- nant in determining a peroxidase to possess catalase activity? Although the individual enzyme-peroxide adducts are characterized by different redox potentials, this fact alone does not explain the wide catalase activity-spectrum among members of the animal peroxidase protein family. Structural factors (such as the presence of a sulphonium linkage in MPO, hydrophobicity of the substrate access channel, distorsion of the heme plane and size of the catalytic pocket) must be crucial in playing on kinetic equilibria of these enzymes and determining whether the peroxidase or catalase function has to prevail. The role of water ÔsittingÕ in the active site has also to be taken into account, as suggested by a recent work by Peter Jones [60]. (b) Both LPO and MPO have been shown to generate an aminoacid radical as decay product of Cpd I-[Fe IV ¼O; p +• ], whose identity has not yet been ascertained. Its contribution to the catalytic cycle of both enzymes is not clear. Ferrocyanide titration on LPO intermediates showed that it is isoelectronic with Cpd I-[Fe IV ¼O; p +• ] and can be reduced to a secondary species that has lost an oxidizing equivalent (Cpd II). The fact that such a product does not react with halides indicates that its redox potential is lower than that of Cpd I, although it cannot be correlated with the number of oxidizing equivalents localized on it. Of note, all detections of the amino acid radical were made in the absence of cosubstrates, far from the conditions in which the enzyme displays its physiological catalytic activity. These arguments suggest that the radical may not have a functional role. (c) The model proposed for MPO by Marquez et al. [18] and Kettle et al. [22] concerning the spontaneous reduction of Cpd I-[Fe IV ¼O; p +• ] in the presence of peroxide stresses the role of superoxide in the catalytic cycle of peroxidases and may be extended to LPO. Accordingly, the one electron oxidation of peroxide by Cpd I yields superoxide. This superoxide anion could either be free in solution or, more likely, bound to the enzyme and give rise to Cpd III, which is responsible for a reversible inactivation mechanism. Cpd III might in turn participate to the reduction of Cpd I fi Cpd II and to the subsequent step Cpd II fi native enzyme, generating molecular oxygen, consistent with the experi- mental data reported by Jenzer et al. [36] (Scheme 5). The involvement of superoxide in the reduction of Cpd I, through a ÔfeedbackÕ cycle, provides a mechanism for the mono- phasic/biphasic kinetics observed by Marquez et al. [18]. Scheme 5. The catalytic cycle of LPO and MPO. In this scheme, the role of H 2 O 2 as an oxidant is shown in steps 1 and 2; catalase activity is displayed in step 1;CpdIisomeri- zation is taken into account (step 2); the reducing effect of superoxide ions, in the form of Cpd III, is shown in steps 5, 6 and 7 and goes along with dioxygen release. The optical absorption peaks of each intermediate are shown on top of the scheme for both enzymes. Ó FEBS 2003 Controversies on myelo- and lactoperoxidase mechanisms (Eur. J. Biochem. 270) 4409 (d) The optical absorption data relative to the species obtained by spontaneous decay of Cpd I can be explained by positing that Cpd I-[Fe IV ¼O; aa +• ]andCpd IIformsin turn and share the same UV-Vis spectroscopic features. The radical intermediate is short-lived and can be detected by EPR only after stabilization with a spin-trap at room temperature or after rapid freezing at liquid helium temperature. The radical, as a transient species, quickly evolves towards a secondary species that corresponds to Cpd II and remains for several minutes at RT. Conclusions A summary of LPO and MPO catalytic features is found in Scheme 5. This model assumes that the enzymes share the same behaviour: a reasonable assumption, in the light of the common phylogenetic origin of LPO and MPO and of their similar physiological role. It also provides a partial answer to the four questions addressed in this paper. According to Scheme 5, the native enzyme, once incuba- tedwithH 2 O 2 , gives rise to Cpd I-[Fe IV ¼O; p +• ], which is able to oxidize peroxide to molecular oxygen, thereby exhibiting catalase activity (step 1). Cpd I-[Fe IV ¼O; p +• ] may also isomerize to Cpd I-[Fe IV ¼O; aa +• ](step2).Inthe absence of other substrates, the conversion of Cpd I fi Cpd II is still mediated by peroxide, which acts as a mono-electronic reductant and generates superoxide (step 3). This ion is able to co-ordinate to the Fe(III) of native enzyme and to form Cpd III (step 4), which can still act as a reductant. Cpd III may either display a feedback action, supporting the Cpd I fi Cpd II conversion (steps 5 and 6), or induce Cpd II reduction to the resting form. Both reaction steps are mono-electronic and are associated with release of molecular oxygen (step 7). This hypothesis would explain both the reversible nature of Cpd III- mediated inactivation and the evolution of small quantities of dioxygen, reported by several authors. It’s worth noting that accumulation of a specific intermediate is never observed in LPO or MPO: this suggests that k values for each reaction step are comparable, thus, preventing a single species to convert quantitatively into the following one. As a consequence, several chemical intermediates coexist and sometimes interact mutually, thus, justifying the complexity of the scheme. Despite their common mechanism, significantly different catalytic features are found in LPO and MPO. They dealwith the the major importance of catalase activity in MPO with respect to LPO, the different sensitivity of the enzymes to Cpd III inactivation and the redox potential of the chemical intermediates. This means that a distinct balance is estab- lished between catalytic steps of MPO and LPO cycles. This is due to differences in the kinetic constant values and likely reflects distinct structural features of the two heme pockets. A survey of the experimental data available in the literature allowed us to highlight some unresolved questions concerning the catalytic behaviour of LPO and MPO. A critical examination of these data resulted in a model that offer a solution to some of these opened questions. It re-evaluates the role of superoxide ion as a reductant in the cycle; it includes Cpd I isomers in it; it highlights the presence of mutual chemical interactions between enzyme intermedi- ates and provides and explanation for catalase activity. Other questions, such as the localization of the amino acid radical in Cpd I, its role and spectroscopical properties need further investigations to be solved. Acknowledgements The authors thank Prof. Ivano Bertini, Prof. Bill Nauseef and Prof. R.P. Ferrari for the very helpful comments and remarks and for critically reading the paper. This work was supported by Italian ÔMinistero per lÕUniversita ` e la Ricerca scientifica e Tecnologica’ (MURST) [PRIN 2000 – prot. MM03185591–005]. References 1. Oertling, W.A., Hoogland, H., Babcock, G.T. & Wever, R. (1988) Identification and properties of an oxoferryl structure in Myelo- peroxidase Compound II. Biochemistry 27, 5395–5400. 2. Chance, B., Powers, L., Ching, Y., Poulos, T., Schonbaum, G.R. & Yamazaki, I. (1984) X-ray absorption studies of intermediates in peroxidase activity. Arch. Biochem. Biophys. 235, 596–611. 3. Kohler, H. & Jenzer, H. (1989) Interaction of lactoperoxidase with hydrogen peroxide. Free Rad. Biol. Med. 6, 323–339. 4. Metodiewa, D. & Dunford, B. (1989) The reactions of horseradish peroxidase, lactoperoxidase and myeloperoxidase with enzi- matically generated superoxide. Arch. Biochem. Biophys. 272, 245– 253. 5. Hu, S. & Kincaid, J.R. (1991) Resonance Raman structural characterization and the mechanism of formation of lactoper- oxidase Compound III. J. Am. Chem. Soc. 113, 7189–7194. 6. Kettle, A.J., Sangster, D.F., Gebicki, J.M. & Winterbourne, C.C. (1988) A pulse radiolysis investigation of the reactions of myelo- peroxidase with superoxide and hydrogen peroxide. Biochim. Biophys. Acta 956, 58–62. 7. Dunford, H.B. (1999) Heme Peroxidases, Wiley-VCH Publishers, New York. 8. Fenna, R.E. (2001) Handbook of Metalloproteins (Messersch- midt,A.,Huber,R.,Poulos,T.,&Wieghardt,K.,eds),Vol.1p. 219. Wiley-VCH Publishers, New York. 9. Stubbe, J. & Van der Donk, W.A. (1998) Protein radicals in enzyme catalysis. Chem. Rev. 98, 705–762. 10. Ivancich, A., Mazza, G. & Desbois, A. (2001) Comparative EPR study of radical intermediates in turnip peroxidase isoenzymes. Biochemistry 40, 6860–6866. 11. Miller, V.P., Goodin, D.B., Friedman, A.E., Hartmann, C. & Ortiz de Montellano, P.R. (1995) Horseradish peroxidase Phe172/ Tyr mutant. J. Biol. Chem. 270, 18413–18419. 12. Rutter, R. & Haber, L.P. (1982) The detection of two electron paramagnetic resonance radical signals associated with chloro- peroxidase compound I. J. Biol. Chem. 257, 7958–7961. 13. Fishel, L.A., Farnum, M.F., Mauro, J.M., Miller, M.A. & Kraut, J. (1991) Compound I radical in site directed mutants of CCP as probed by EPR and ENDOR. Biochemistry 30, 1986–96. 14.Erman,J.E.,Vitello,L.B.,Mauro,J.M.&Kraut,J.(1989) Detection of an oxoferryl porphyrin cation radical intermediate in the reaction between H 2 O 2 and a mutant yeast cytochrome c peroxidase. Biochemistry 28, 7992–7995. 15. De Gioia, L., Ghibaudi, E.M., Laurenti, E., Salmona, M. & Ferrari, R.P. (1996) A theoretical three-dimensional model for LPO and eosinophil peroxidase built on the scaffold of the MPO X-ray structure. J. Biol. Inorg. Chem. 1, 476–485. 16. Harrison, J.E., Araiso, T., Palcic, M. & Dunford, B. (1980) Compound I of MPO. Biochem. Biophys. Res. Comm. 94, 34–40. 17. Svensson, B.E., Domeij, K., Lindvall, S. & Rydell, G. (1987) Peroxidase and peroxidase-oxidase activities of isolated human myeloperoxidase. Biochem. J. 242, 673–680. 4410 E. Ghibaudi and E. Laurenti (Eur. J. Biochem. 270) Ó FEBS 2003 18. Marquez, L.A., Huang, J.T. & Dunford, H.B. (1994) Spectral and kinetic studies on the formation of MPO Compound I and II: roles of hydrogen peroxide and superoxide. Biochemistry 33, 1447– 1454. 19. Furtmuller, P.G., Burner, U. & Obinger, C. (1998) Reaction of myeloperoxidase Compound I with chloride, bromide, iodide and thiocyanate. Biochemistry 27, 17923–17930. 20. Odajima, T. & Yamazaki, I. (1972) Myeloperoxidase of the leu- kocyte of normal blood. V. The spectral conversion of myelo- peroxidase to a cytochrome oxidase like derivative. Biochim. Biophys. Acta 284, 368–374. 21. Bakkenist, A.R.J., Wever, R., Vulsma, T., Plat, H. & Van Gelder, B.F. (1978) Isolation procedure and some properties of myelo- peroxidase from human leucocytes. Biochim. Biophys. Acta 524, 45–54. 22. Kettle, A.J. & Winterbourn, C.C. (2001) A kinetic analysis of the catalase activity of myeloperoxidase. Biochemistry 40, 10204– 10212. 23. Floris, R. & Wever, R. (1992) Reaction of myeloperoxidase with its product HOCl. Eur. J. Biochem. 207, 697–702. 24. Hoogland, H., van Kuilenburg, A., van Riel, C., Muijsers, A.O. & Wever, R. (1987) Spectral properties of myeloperoxidase Com- pounds II and III. Biochim. Biophys. Acta 916, 76–82. 25. Ohtaki, S., Nakagawa, H., Nakamura, M. & Yamazaki, I. (1982) Reaction of purified hog thyroid peroxidase with H 2 O 2 ,tyrosine and methylmercaptoimidazole in comparison with bovine lacto- peroxidase. J. Biol. Chem. 257, 761–766. 26. Doerge, D.R., Taurog, A. & Dorris, M. (1994) Evidence for a radical mechanism in peroxidase-catalysed coupling: single turn- over experiments with HRP. Arch. Biochem. Biophys. 315, 90–99. 27. Monzani, E., Gatti, A.L., Profumo, A., Casella, L. & Gullotti, M. (1997) Oxidation of phenolic compounds by lactoperoxidase. Evidence for the presence of a low-potential compound II during catalytic turnover. Biochemistry 36, 1918–1926. 28. Ohlsson, P.I., Paul, K.G. & Wold, S. (1984) The formation of compound I of lactoperoxidase and horseradish peroxidase. A comparison. Acta Chem. Scand. B38, 853–859. 29. Kimura, S. & Yamazaki, I. (1979) Comparisons between hog intestinal peroxidase and bovine LPO-compound I formation and inhibition by benzohydroxamic acid. Arch. Biochem. Biophys. 198, 580–588. 30. Kohler, H., Taurog, A. & Dunford, H.B. (1988) Spectral studies with lactoperoxidase and thyroid peroxidase: Interconversions between native enzyme, compound II, and compound III. Arch. Biochim. Biophys. 264, 438–449. 31. Taurog, A., Dorris, M.L. & Doerge, D.R. (1996) Mechanism of simultaneous iodination and coupling catalysed by thyroid per- oxidase. Arch. Biochem. Biophys. 330, 24–32. 32. Courtin, F., Deme, D., Virion, A., Michot, J.L., Pommier, J. & Nunez, J. (1982) The role of LPO-H 2 O 2 compounds in the cata- lysis of thyroglobulin iodination and thyroid hormone synthesis. Eur. J. Biochem. 124, 603–609. 33. Courtin,F.,Michot,J.L.,Virion,A.,Pommier,J.&Deme,D. (1984) Reduction of LPO-H 2 O 2 compounds by ferrocyanide: indirect evidence of an apoprotein site for one of the two oxidising equivalents. Biochem. Biophys. Res. Comm. 121, 463–470. 34. Jenzer, H. & Kohler, H. (1986) The role of superoxide radicals in LPO-catalysed H 2 O 2 -metabolism and in irreversible enzyme inactivation. Biochem. Biophys. Res. Comm. 139, 327–332. 35. Huwiler, M., Jenzer, H. & Kohler, H. (1986) The role of com- pound III in reversible and irreversible inactivation of LPO. Eur. J. Biochem. 158, 609–614. 36. Jenzer, H., Jones, W. & Kohler, H. (1986) On the molecular mechanism of LPO-catalysed H 2 O 2 metabolism and irreversible enzyme inactivation. J. Biol. Chem. 261, 15550–15556. 37. Jenzer,H.,Kohler,H.&Broger,C.(1987)Theroleofhydroxyl radicals in irreversible inactivation of lactoperoxidase by excess H 2 O 2 . Arch. Biochem. Biophys. 258, 381–390. 38. Bolscher, B.G.J.M. & Wever, R. (1984) A kinetic study of the reaction between human myeloperoxidase, hydroperoxidase and cyanide. Biochim. Biophys. Acta 788, 1–10. 39. Agner, K. (1963) Studies on myeloperoxidase activity: spectro- photometry of the MPO- H 2 O 2 compound. Acta Chem. Scand. 17, S332–S338. 40. Iwamoto, H., Kobayashi, T., Hasegawa, E. & Morita, Y. (1987) Reaction of human myeloperoxidase with hydrogen peroxide and its true catalase activity. J. Biochem. 101, 1407–1412. 41. Kettle, A.J. & Winterbourne, C.C. (1988) Superoxide modulates the activity of myeloperoxidase and optimizes the production of hypochlorous acid. Biochem. J. 252, 529–536. 42. Sawyer, D.T. (1988) Oxygen Complexes and Oxygen Activation by Transition Metals. (Martell, A.E. & Sawyer, D.T., eds), p. 131. Plenum Publishing Co, New York. 43. Arnhold, J., Furtmuller, P.G., Regelsberger, G. & Obinger, C. (2001) Redox properties of the couple compound I/native enzyme of myeloperoxidase and eosinophil peroxidase. Eur. J. Biochem. 268, 5142–5148. 44. Huwiler, M. & Kohler, H. (1984) Pseudo-catalytic degradation of hydrogen peroxide in the lactoperoxidase/ H 2 O 2 /iodide system. Eur. J. Biochem. 141, 69–74. 45. Magnusson, P.R., Taurog, A. & Dorris, M.L. (1984) Mechanism of thyroid peroxidase- and lactoperoxidase-catalysed reactions involving iodide. J. Biol. Chem. 259, 13783–13790. 46. Furtmuller, P.G., Jantschko, W., Regelsberger, G., Jakopitsch, C., Arnhold, J. & Obinger, C. (2002) Reaction of lactoperoxidase compound I with halides and thiocyanate. Biochemistry 41, 11895–11900. 47. Ohlsson, P.I. & Paul, K.G. (1983) The reduction potential of lactoperoxidase. Acta Chem. Scand. B37, 917–921. 48. Lardinois, O.M. & Ortiz de Montellano, P. (2000) EPR spin- trapping of a myeloperoxidase protein radical. Biochem. Biophys. Res. Comm. 270, 199–202. 49.Furtmuller,P.G.,Arnhold,J.,Jantschko,W.,Pichler,H.& Obinger, C. (2003) Redox properties of the couples compound I/ compound II and compound II/native enzyme of human myeloperoxidase. Biochem. Biophys. Res. Commun. 301, 551–557. 50. De Pillis, G.D., Ozaki, S., Kuo, J.M., Maltby, D.A. & Ortiz de Montellano, P.R. (1997) Autocatalytic processing of heme by lactoperoxidase produces the native protein bound prosthetic group. J. Biol. Chem. 272, 8857–8860. 51. Lardinois, O.M., Medzihradszky, K.F. & Ortiz de Montellano, P.R. (1999) Spin-trapping and protein cross-linking of the lacto- peroxidase protein radical. J. Biol. Chem. 274, 35441–35448. 52. Goff, H.M. & Chen, K. (1999) Protein radicals and formation and reactivity of the lactoperoxidase heme prosthetic group. J. Inorg Biochem. 74,143. 53. Doerge, R.D. (1986) Oxigenation of organosulphur compounds by peroxidases: evidence of an electron transfer mechanism for lactoperoxidase. Arch. Biochem. Biophys. 244, 678–685. 54. Taurog,A.,Dorris,M.&Doerge,D.R.(1994)Evidencefora radical mechanism in peroxidase-catalysed coupling steady-state experiments with various peroxidases. Arch. Biochem. Biophys. 315, 82–89. 55. Lardinois, O. & Ortiz de Montellano, P.R. (2001) H 2 O 2 -mediated cross-linking between lactoperoxidase and myoglobin. J. Biol. Chem. 276, 23186–23191. 56. Deme, D., Virion, A., Michot, J.L. & Pommier, J. (1985) Thyroid Hormone Synthesis and Thyroglobulin Iodination related to the peroxidase localization of oxidizing equivalents. Studies with Cyt c eHRP.Arch. Biochem. Biophys. 236, 559–566. Ó FEBS 2003 Controversies on myelo- and lactoperoxidase mechanisms (Eur. J. Biochem. 270) 4411 57. O’Brien, P.J. (2000) Peroxidases. Chem. Biol. Interact. 129, 113–139. 58. Colas, C., Kuo, J.M. & Ortiz de Montellano, P.R. (2002) Asp-225 and Glu-375 in autocatalytic attachment of the prosthetic heme group of lactoperoxidase. J. Biol. Chem. 277, 7191–7200. 59. Coulson, A.F.W., Erman, J.E. & Yonetani, T. (1971) Studies on cytochrome c peroxidase. J. Biol. Chem. 246, 917–924. 60. Jones, P. (2001) Roles of water in heme peroxidase and catalase mechanisms. J. Biol. Chem. 276, 13791–13796. 4412 E. Ghibaudi and E. Laurenti (Eur. J. Biochem. 270) Ó FEBS 2003 . ARTICLE Unraveling the catalytic mechanism of lactoperoxidase and myeloperoxidase A reflection on some controversial features Elena Ghibaudi and Enzo Laurenti Dipartimento. model assumes that the enzymes share the same behaviour: a reasonable assumption, in the light of the common phylogenetic origin of LPO and MPO and of their similar