Formation of a tyrosyl radical intermediate in Proteus mirabilis catalase by directed mutagenesis and consequences for nucleotide reactivity

Monofunctional Heme-Catalases

The review focuses on four issues that are critical for the understanding of monofunctional catalases. How hydrogen peroxide (H2O2) reaches the active site and outcompetes water molecules to be able to function at a very high rate is one of the issues examined. Part of the answer is a gate valve system that is instrumental to drive out solvent molecules from the final section of the main channel. A second issue relates to how the enzyme deals with an unproductive reactive compound I (Cpd I) intermediate. Peroxidatic two and one electron donors and the transfer of electrons to the active site from NADPH and other compounds are reviewed. The new ascribed catalase reactions are revised, indicating possible measurement pitfalls. A third issue concerns the heme b to heme d oxidation, why this reaction occurs only in some large-size subunit catalases (LSCs), and the possible role of singlet oxygen in this and other modifications. The formation of a covalent bond between the proximal tyrosine with the vicinal residue is analyzed. The last issue refers to the origin and function of the additional C-terminal domain (TD) of LSCs. The TD has a molecular chaperone activity that is traced to a gene fusion between a Hsp31-type chaperone and a small-size subunit catalase (SSC).

Flavonoid-induced conversion of catalase to its inactive form--Compound II

Flavonoids (FlaOHs), plant polyphenols, are ubiquitous components of human diet and are known as antioxidants. However, their prooxidant activity has also been reported. We have recently found that FlaOHs inhibit catalase, the heme enzyme which catalyzes the decomposition of hydrogen peroxide (H2O2) into water and molecular oxygen. The catalytic cycle proceeds with the formation of the intermediate, Compound I (Cpd I), an oxoferryl porphyrin π-cation radical, the two-electron oxidation product of a heme group. Under conditions of low H2O2 fluxes and in the presence of an appropriate substrate, Cpd I can undergo one-electron reduction to inactive Compound II (Cpd II), oxoferryl derivative without radical site. Here we show that in vitro, under low fluxes of H2O2, FlaOHs reduce Cpd I to inactive Cpd II. Measurable amounts of Cpd II can be formed even in the presence of reduced nicotinamide adenine dinucleotide phosphate (NADPH) at concentration comparable with the investigated FlaOHs. Possible mechanisms of electron transfer from FlaOH molecule to the heme are discussed.

Ascorbic acid reduction of compound I of mammalian catalases proceeds via specific binding to the NADPH binding pocket

Mammalian (Clade 3) catalases utilize NADPH as a protective cofactor to prevent one-electron reduction of the central reactive intermediate Compound I (Cpd I) to the catalytically inactive Compound II (Cpd II) species by re-reduction of Cpd I to the enzyme's resting state (ferricatalase). It has long been known that ascorbate/ascorbic acid is capable of reducing Cpd I of NADPH-binding catalases to Cpd II, but the mode of this one-electron reduction had hitherto not been explored. We here demonstrate that ascorbate-mediated reduction of Cpd I, generated by addition of peroxoacetic acid to NADPH-free bovine liver catalase (BLC), requires specific binding of the ascorbate anion to the NADPH binding pocket. Ascorbate-mediated Cpd II formation was found to be suppressed by added NADPH in a concentration-dependent manner, for the achievement of complete suppression at a stoichiometric 1:1 NADPH:heme concentration ratio. Cpd I → Cpd II reduction by ascorbate was similarly inhibited by addition of NADH, NADP(+), thio-NADP(+), or NAD(+), though with 0.5-, 0.1-, 0.1-, and 0.01-fold reduced efficiencies, respectively, in agreement with the relative binding affinities of these dinucleotides. Unexpected was the observation that although Cpd II formation is not observed in the presence of NADP(+), the decay of Cpd I is slightly accelerated by ascorbate rather than retarded, leading to direct regeneration of ferricatalase. The experimental findings are supported by molecular mechanics docking computations, which show a similar binding of NADPH, NADP(+), and NADH, but not NAD(+), as found in the X-ray structure of NADPH-loaded human erythrocyte catalase. The computations suggest that two ascorbate molecules may occupy the empty NADPH pocket, preferably binding to the adenine binding site. The biological relevance of these findings is discussed.

The reaction mechanisms of heme catalases: an atomistic view by ab initio molecular dynamics

Catalases are ubiquitous enzymes that prevent cell oxidative damage by degrading hydrogen peroxide to water and oxygen (2H(2)O(2) → 2H(2)O+O(2)) with high efficiency. The enzyme is first oxidized to a high-valent iron intermediate, known as Compound I (Cpd I, Por(·+)-Fe(IV)=O) which, at difference from other hydroperoxidases, is reduced back to the resting state by further reacting with H(2)O(2). The normal catalase activity is reduced if Cpd I is consumed in a competing side reaction, forming a species named Cpd I*. In recent years, Density Functional Theory (DFT) methods have unraveled the electronic configuration of these high-valent iron species, helping to assign the intermediates trapped in the crystal structures of oxidized catalases. It has been demonstrated that the a priori assumption that the H(+)/H(-) type of mechanism for Cpd I reduction leads to the generation of singlet oxygen is not justified. Moreover, it has been shown by ab initio metadynamics simulations that two pathways are operative for Cpd I reduction: a His-mediated mechanism (described as H·/H(+) + e(-)) in which the distal His acts as an acid-base catalyst and a direct mechanism (described as H·/H·) in which the distal His does not play a direct role. Independently of the mechanism, the reaction proceeds by two one-electron transfers rather than one two-electron transfer, as previously assumed. Electron transfer to Cpd I, regardless of whether the electron is exogenous or endogenous, facilitates protonation of the oxoferryl group, to the point that formation of Cpd I* may be controlled by the easiness of protonation of reduced Cpd I.

Classical catalase: ancient and modern

This review describes the historical difficulties in devising a kinetically satisfactory mechanism for the classical catalase after its identification as a unique catalytic entity in 1902 and prior to the breakthrough 1947 analysis by Chance and co-workers which led to the identification of peroxide compounds I and II. The role of protons in the formation of these two ferryl complexes is discussed and current problems of inhibitory ligand and hydrogen donor binding at the active site are outlined, especially the multiple roles involving formate or formic acid. A previous mechanism of NADPH-dependent catalase protection against substrate inhibition is defended. A revised model linking the catalytic ('catalatic') action and the one-electron side reactions involving compound II is suggested. And it is concluded that, contrary to an idea proposed in 1963 that eukaryotic catalase might be a 'fossil enzyme', current thinking gives it a central role in the redox protective processes of long term importance for human and other eukaryotic and prokaryotic life.

A density functional investigation of hydrogen peroxide activation by high-valent heme centers: implications for the catalase catalytic cycle

Catalases employ a tyrosinate-ligated ferric heme in order to catalyze the dismutation of hydrogen peroxide to O2 and water. In the first half of the catalytic cycle, H2O2 oxidizes Fe(III) to the formally Fe(V) state commonly referred to as Compound I. The second half of the cycle entails oxidation of a second hydrogen peroxide molecule by Compound I to dioxygen. The present study employs density functional (DFT) calculations to examine the nature of this second step of the catalatic reaction. In order to account for the unusual choice of tyrosinate as an axial ligand in catalases, oxidation of hydrogen peroxide by an imidazole-ligated Compound I is also examined, bearing in mind that imidazole-ligated hemoproteins such as myoglobin or horseradish peroxidase tend to display little, if any, catalatic activity. Furthermore, in order to gauge the importance of the cation radical of Compound I in peroxide activation, the performance of Compound II (the one-electron reduced version of Compound I, formally Fe(IV) ), is also examined. It is found that hydrogen peroxide oxidation occurs in a quasi-concerted manner, with two hydrogen-atom transfer reactions, and that the tyrosinate ligand is in no way required at this stage. We propose that the role of the tyrosinate is purely thermodynamic, in avoiding accumulation of the much less peroxide-reactive ferrous form in vivo – all in line with the predominantly thermodynamic role of the cysteinate ligands in enzymes such as cytochromes P450.

Spectroscopic description of an unusual protonated ferryl species in the catalase from Proteus mirabilis and density functional theory calculations on related models. Consequences for the ferryl protonation state in catalase, peroxidase and chloroperoxidase

The catalase from Proteus mirabilis peroxide-resistant bacteria is one of the most efficient heme-containing catalases. It forms a relatively stable compound II. We were able to prepare samples of compound II from P. mirabilis catalase enriched in (57)Fe and to study them by spectroscopic methods. Two different forms of compound II, namely, low-pH compound II (LpH II) and high-pH compound II (HpH II), have been characterized by Mössbauer, extended X-ray absorption fine structure (EXAFS) and UV-vis absorption spectroscopies. The proportions of the two forms are pH-dependent and the pH conversion between HpH II and LpH II is irreversible. Considering (1) the Mössbauer parameters evaluated for four related models by density functional theory methods, (2) the existence of two different Fe-O(ferryl) bond lengths (1.80 and 1.66 A) compatible with our EXAFS data and (3) the pH dependence of the alpha band to beta band intensity ratio in the absorption spectra, we attribute the LpH II compound to a protonated ferryl Fe(IV)-OH complex (Fe-O approximately 1.80 A), whereas the HpH II compound corresponds to the classic ferryl Fe(IV)=O complex (Fe=O approximately 1.66 A). The large quadrupole splitting value of LpH II (measured 2.29 mm s(-1) vs. computed 2.15 mm s(-1)) compared with that of HpH II (measured 1.47 mm s(-1) vs. computed 1.46 mm s(-1)) reflects the protonation of the ferryl group. The relevancy and involvement of such (Fe(IV)=O/Fe(IV)-OH) species in the reactivity of catalase, peroxidase and chloroperoxidase are discussed.

Mammalian catalase: a venerable enzyme with new mysteries

Mammalian catalase has been the subject of many classic biochemical studies. Despite our detailed knowledge of its functional mechanisms and its three-dimensional structure, however, several unexpected features of mammalian catalase have been recently discovered. For example, some mammalian catalases seem to have oxidase activity and produce reactive oxygen species when exposed to UVB light. In addition, bovine catalase uses unbound NAD(P)H to prevent substrate inactivation without displacing catalase-bound NADP(+). Coupled with the earlier discovery of catalase-bound NADPH, these developments indicate that serendipity and new investigative approaches can reveal unexpected features, even for an enzyme that has been studied for over 100 years.

Non-oxygen-forming pathways of hydrogen peroxide degradation by bovine liver catalase at low hydrogen peroxide fluxes

Heme catalases are considered to degrade two molecules of H(2)O(2) to two molecules of H(2)O and one molecule of O(2) employing the catalatic cycle. We here studied the catalytic behaviour of bovine liver catalase at low fluxes of H(2)O(2) (relative to catalase concentration), adjusted by H(2)O(2)-generating systems. At a ratio of a H(2)O(2) flux (given in microM/min(- 1)) to catalase concentration (given in microM) of 10 min(- 1) and above, H(2)O(2) degradation occurred via the catalatic cycle. At lower ratios, however, H(2)O(2) degradation proceeded with increasingly diminished production of O(2). At a ratio of 1 min(- 1), O(2) formation could no longer be observed, although the enzyme still degraded H(2)O(2). These results strongly suggest that at low physiological H(2)O(2) fluxes H(2)O(2) is preferentially metabolised reductively to H(2)O, without release of O(2). The pathways involved in the reductive metabolism of H(2)O(2) are presumably those previously reported as inactivation and reactivation pathways. They start from compound I and are operative at low and high H(2)O(2) fluxes but kinetically outcompete the reaction of compound I with H(2)O(2) at low H(2)O(2) production rates. In the absence of NADPH, the reducing equivalents for the reductive metabolism of H(2)O(2) are most likely provided by the protein moiety of the enzyme. In the presence of NADPH, they are at least in part provided by the coenzyme.

Cold adapted features of Vibrio salmonicida catalase: characterisation and comparison to the mesophilic counterpart from Proteus mirabilis

The gene encoding catalase from the psychrophilic marine bacterium Vibrio salmonicida LFI1238 was identified, cloned and expressed in the catalase-deficient Escherichia coli UM2. Recombinant catalase from V. salmonicida (VSC) was purified to apparent homogeneity as a tetramer with a molecular mass of 235 kDa. VSC contained 67% heme b and 25% protoporphyrin IX. VSC was able to bind NADPH, react with cyanide and form compounds I and II as other monofunctional small subunit heme catalases. Amino acid sequence alignment of VSC and catalase from the mesophilic Proteus mirabilis (PMC) revealed 71% identity. As for cold adapted enzymes in general, VSC possessed a lower temperature optimum and higher catalytic efficiency (k (cat)/K (m)) compared to PMC. VSC have higher affinity for hydrogen peroxide (apparent K (m)) at all temperatures. For VSC the turnover rate (k (cat)) is slightly lower while the catalytic efficiency is slightly higher compared to PMC over the temperature range measured, except at 4 degrees C. Moreover, the catalytic efficiency of VSC and PMC is almost temperature independent, except at 4 degrees C where PMC has a twofold lower efficiency compared to VSC. This may indicate that VSC has evolved to maintain a high efficiency at low temperatures.

Structures of the high-valent metal-ion haem–oxygen intermediates in peroxidases, oxygenases and catalases

Peroxidases, oxygenases and catalases have similar high-valent metal-ion intermediates in their respective reaction cycles. In this review, haem-based examples will be discussed. The intermediates of the haem-containing enzymes have been extensively studied for many years by different spectroscopic methods like UV-Vis, EPR (electron paramagnetic resonance), resonance Raman, Mössbauer and MCD (magnetic circular dichroism). The first crystal structure of one of these high-valent intermediates was on cytochrome c peroxidase in 1987. Since then, structures have appeared for catalases in 1996, 2002, 2003, putatively for cytochrome P450 in 2000, for myoglobin in 2002, for horseradish peroxidase in 2002 and for cytochrome c peroxidase again in 1994 and 2003. This review will focus on the most recent structural investigations for the different intermediates of these proteins. The structures of these intermediates will also be viewed in light of quantum mechanical (QM) calculations on haem models. In particular quantum refinement, which is a combination of QM calculations and crystallography, will be discussed. Only small structural changes accompany the generation of these intermediates. The crystal structures show that the compound I state, with a so called pi-cation radical on the haem group, has a relatively short iron-oxygen bond (1.67-1.76A) in agreement with a double-bond character, while the compound II state or the compound I state with a radical on an amino acid residue have a relatively long iron-oxygen bond (1.86-1.92A) in agreement with a single-bond character where the oxygen-atom is protonated.

Role of radical formation at tyrosine 193 in the allene oxide synthase domain of a lipoxygenase-AOS fusion protein from coral

Coral allene oxide synthase (cAOS), a fusion protein with 8R-lipoxygenase in Plexaura homomalla, is a hemoprotein with sequence similarity to catalases. cAOS reacts rapidly with the oxidant peracetic acid to form heme compound I and intermediate II. Concomitantly, an electron paramagnetic resonance (EPR) signal with tyrosyl radical-like features, centered at a g-value of 2.004-2.005, is formed. The radical is identified as tyrosyl by changes in EPR spectra when deuterated tyrosine is incorporated in cAOS. The radical location in cAOS is determined by mutagenesis of Y193 and Y209. Upon oxidation, native cAOS and mutant Y209F exhibit the same radical spectrum, but no significant tyrosine radical forms in mutant Y193H, implicating Y193 as the radical site in native cAOS. Estimates of the side chain torsion angles for the radical at Y193, based on the beta-proton isotropic EPR hyperfine splitting, A(iso), are theta(1) = 21 to 30 degrees and theta(2) = -99 to -90 degrees. The results show that cAOS can cleave nonsubstrate hydroperoxides by a heterolytic path, although a homolytic course is likely taken in converting the normal substrate, 8R-hydroperoxyeicosatetraenoic acid (8R-HpETE), to product. Coral AOS achieves specificity for the allene oxide formed by selection of the homolytic pathway normally, while it inactivates by the heterolytic path with nonoptimal substrates. Accordingly, with the nonoptimal substrate, 13R-hydroperoxyoctadecadienoic acid (13R-HpODE), mutant Y193H is inactivated after turning over significantly fewer substrate molecules than required to inactivate native cAOS or the Y209F mutant because it cannot absorb oxidizing equivalents by forming a radical at Y193.

High-resolution structure and biochemical properties of a recombinant Proteus mirabilis catalase depleted in iron

Heme catalases are homotetrameric enzymes with a highly conserved complex quaternary structure, and their functional role is still not well understood. Proteus mirabilis catalase (PMC), a heme enzyme belonging to the family of NADPH-binding catalases, was efficiently overexpressed in E. coli. The recombinant catalase (rec PMC) was deficient in heme with one-third heme and two-thirds protoporphyrin IX as determined by mass spectrometry and chemical methods. This ratio was influenced by the expression conditions, but the enzyme-specific activity calculated relative to the heme content remained unchanged. The crystal structure of rec PMC was solved to a resolution of 2.0 A, the highest resolution obtained to date with PMC. The overall structure was quite similar to that of wild-type PMC, and it is surprising that the absence of iron had no effect on the structure of the active site. Met 53 close to the essential His 54 was found less oxidized in rec PMC than in the wild-type enzyme. An acetate anion was modeled in an anionic pocket, away from the heme group but important for the enzymatic reaction. An alternate conformation observed for Arg 99 could play a role in the formation of the H-bond network connecting two symmetrical subunits of the tetramer.