|Home | About | Journals | Submit | Contact Us | Français|
Catalase-peroxidases (KatGs) are unique bifunctional heme peroxidases with an additional posttranslationally formed redox-active Met-Tyr-Trp cofactor that is essential for catalase activity. On the basis of studies of bacterial KatGs, controversial mechanisms of hydrogen peroxide oxidation were proposed. The recent discovery of eukaryotic KatGs with differing pH optima of catalase activity now allows us to scrutinize those postulated reaction mechanisms. In our study, secreted KatG from the fungus Magnaporthe grisea (MagKatG2) was used to analyze the role of a remote KatG-typical mobile arginine that was shown to interact with the Met-Tyr-Trp adduct in a pH-dependent manner in bacterial KatGs. Here we present crystal structures of MagKatG2 at pH 3.0, 5.5, and 7.0 and investigate the mobility of Arg461 by molecular dynamics simulation. Data suggest that at pH ≥4.5 Arg461 mostly interacts with the deprotonated adduct Tyr. Elimination of Arg461 by mutation to Ala slightly increases the thermal stability but does not alter the active site architecture or the kinetics of cyanide binding. However, the variant Arg461Ala lost the wild-type-typical optimum of catalase activity at pH 5.25 (kcat = 6450 s–1) but exhibits a broad plateau between pH 4.5 and 7.5 (kcat = 270 s–1 at pH 5.5). Moreover, significant differences in the kinetics of interconversion of redox intermediates of wild-type and mutant protein mixed with either peroxyacetic acid or hydrogen peroxide are observed. These findings together with published data from bacterial KatGs allow us to propose a role of Arg461 in the H2O2 oxidation reaction of KatG.
Recently, the first structures of a eukaryotic bifunctional catalase-peroxidase (KatGs) in the ferric resting1 and oxoiron2 state were published. Similar to the prokaryotic counterparts,3−7 eukaryotic KatGs have two interdependent cooperating redox-active cofactors at their active site, namely a heme b and a unique posttranslationally and autocatalytically formed Met299-Tyr273-Trp140 (MYW) adduct [extracellular Magnaporthe grisea KatG (MagKatG2) numbering throughout]. KatG can be described as a catalase in peroxidase clothing8 because it has been engineered by Nature to catalyze efficient H2O2 dismutation similar to monofunctional catalases.9,10 Together with cytochrome c peroxidases, ascorbate peroxidases and hybrid-type peroxidases KatGs comprise Family I of the peroxidase-catalase superfamily.11,12 Besides the MYW adduct mentioned above, further KatG-specific structural peculiarities include (i) a two-domain structure with only the N-terminal domain containing the two redox cofactors and essential catalytic residues,1,3,4 (ii) loop insertions that contribute to the architecture of the long and restricted access channel, and (iii) a fully conserved aspartate at the entrance of the substrate channel to the heme cavity.13,14 Mutation of any of these KatG-typical structural features typically reduces the catalase activity without compromising the peroxidase activity.13−22
Both the heme and the Met299-Tyr273-Trp140 cofactor found in KatG are mutually dependent on one another to support efficient H2O2 dismutation. On the basis of (i) time-resolved spectroscopic studies, including freeze-quench electron paramagnetic resonance (EPR) spectroscopies, and (ii) hybrid quantum mechanics/molecular mechanics (QM/MM) computational studies, the following sequence of reactions has been proposed to account for the catalatic activity of KatG (see also Supplemental Figure 1).14,17−19,23−27 Initially, ferric KatG [MYW···Fe(III) Por] is oxidized by hydrogen peroxide to Compound I [MYW···Fe(IV)=O Por+•], which is typical of monofunctional peroxidases and catalases (reaction I).25,28,29 However, in KatG, the porphyryl radical is rapidly quenched by an internal electron-transfer process producing oxoiron(IV) and the adduct radical (MYW+•) (reaction II).14,23,24 This Compound I* [MYW+•···Fe(IV)–OH Por] reacts with the second H2O2 in a nonscrambling mechanism30 to directly form Compound III* (reaction III), a dioxyheme intermediate with distinct spectral features that may be considered analogous to oxymyoglobin, oxyhemoglobin, and peroxidase Compound III as a complex between ferrous heme and dioxygen [Fe(II)–O2], which is isoelectronic with superoxide bound to ferric heme [Fe(III)–O2–•].31 Monofunctional peroxidases have negligible catalase activity and are trapped in this Compound III state in the presence of H2O2 only.31 However, in KatG, the superoxyferric complex is formed in the presence of the adduct radical (i.e., Compound III*)14,19,24 (reaction III), thereby keeping an oxidation equivalent very close to superoxide in the Fe(III)–O2–• complex, which is the dominating electronic structure during KatG turnover.27 This guarantees rapid decomposition of Compound III* to the ferric resting state, thereby releasing solely triplet oxygen.23 The balance of one KatG turnover is dismutation of hydrogen peroxide (2H2O2 → 2H2O + O2), which is exactly the same as in monofunctional catalases. However, the mechanism of reaction is completely different and clearly reflects the fact that KatG is a posttranslationally modified peroxidase with a dominating catalatic besides a conventional peroxidase activity. With regard to reactions III and IV, an alternative mechanism has been postulated,26 including the transfer of (i) a hydrogen atom from the second H2O2 molecule to the oxyferryl species of Compound I* producing the hydroperoxide radical and restoring the ferric state as well as (ii) a proton from the adduct tryptophan to the [Fe(III)–OH] state thereby releasing water.26
Interestingly, the chemical properties of the KatG-typical adduct seem to be modulated by a fully conserved mobile arginine residue far from the heme (~20 Å) but closely related to the adduct. X-ray structures of bacterial KatGs showed that the Arg can switch between two distinct conformations (“in” vs “out”) depending on pH. Below pH 6.5 (i.e., the pH optimum of catalase activity as well as the proposed pKa of the adduct tyrosine in prokaryotic KatGs), the out conformation dominates, whereas under alkaline conditions, Arg points predominantly toward the adduct and forms an ionic interaction with the tyrosinate of the adduct. Elimination of this basic amino acid strongly reduces the catalase activity but not the peroxidase activity of KatG.14,33,34
Although several years of elaborate research have passed, there are still open questions about the catalatic reaction of KatGs. The recent finding that eukaryotic MagKatG2 has some novel structural features1,2,35 and a more acidic pH optimum (i.e., pH 5.25) for H2O2 dismutation36 now allows us to scrutinize whether the postulated catalatic mechanism (reactions I–IV) and the role of the mobile arginine in modulating the pH dependence are also valid for eukaryotic KatGs. In this work, we have determined the X-ray structures of MagKatG2 at pH 3, 5.5, and 7.0 and investigated the dynamics of Arg461 in ferric MagKatG2 by molecular dynamics simulation. We demonstrate the effect of exchange of Arg461 with alanine on active site geometry, thermal stability, and ligand binding as well as on the pH dependence of bifunctional activity. Furthermore, the kinetics of peroxyacetic acid- and H2O2-mediated interconversion of redox intermediates of wild-type MagKatG2 and the variant Arg461Ala are presented, discussed, and compared with those reported for prokaryotic catalase-peroxidases.
Mutagenesis of the recombinant extracellular catalase-peroxidase from M. grisea and expression of the wild-type and mutant protein in Escherichia coli strain BL21(DE3) Star (Invitrogen) were performed as described recently.1,2 Briefly, the point mutation was introduced with a polymerase chain reaction-based site-directed mutagenesis kit (QuikChange Lightning Kit from Agilent Technologies) using two complementary oligonucleotides with the planned mutation in the middle of their sequence (5′-AACTGCTGCATGCTGATCTGGGCCCGACCACC-3′ and 3′-GGTGGTCGGGCCCAGATCAGCATGCAGCAGTTT-5′). For protein expression, transformed E. coli cells were grown in M9ZB medium [supplemented with ampicillin (100 μg mL–1)] and 40 mg L–1 hemin prior to induction. After induction by 0.5% (w/v) lactose or 0.5 mM isopropyl β-d-1-thiogalactopyranoside, the cultivation temperature was lowered to 16 °C during overnight protein expression. Cells were harvested by centrifugation and homogenized by multiple ultrasonication cycles. The crude homogenate was clarified by centrifugation (>20000g), and the clear supernatant was loaded onto a 20 mL Chelating Sepharose Fast Flow Column (GE Healthcare). His-tagged proteins were eluted with a linearly increasing gradient of imidazole and immediately checked by UV–vis spectroscopy. The KatG-containing fractions were pooled, concentrated by centrifugation, and further purified using a hydroxyapatite column (25 mL, Bio-Rad) following a stepwise elution [step 1 being 50 mM phosphate buffer (pH 7.7) and then step 2 being 200 mM phosphate buffer (pH 8.0)].
For evaluation of purification success, protein electrophoresis was performed with NuPAGE Novex Bis-Tris gradient gels (4 to 12%) with 15 wells (Invitrogen) (running buffer being MOPS with SDS). Gels were stained with Coomassie Brilliant Blue (Sigma-Aldrich) or immunoblotted onto a nitrocellulose membrane (Amersham Biosciences) for detection of MagKatG2 by using a penta-His antibody.
Crystals of poly-His-tagged MagKatG2 were obtained using the sitting-drop vapor-diffusion method at 4 °C with a protein concentration of 5 mg/mL in 5 mM phosphate buffer (pH 7.5). The mother liquor contained 15% PEG 4000 and 0.1 M sodium acetate (pH 4.6). Crystals were soaked in buffers with varying pH values (phosphate-citrate or phosphate) containing elevated levels of PEG 4000 (35%) as a cryoprotectant. At beamline ID23eh1 (ESRF, Grenoble, France), diffraction data up to 1.4 Å were obtained using flash-cooled crystals. Crystals belonged to space group P212121, and a biological dimer was present in the asymmetric unit. A molecular replacement solution was found for the newly obtained data sets using the wild-type structure of MagKatG2 [Protein Data Bank (PDB) entry 3UT2] as a search model. Data collection and structural refinement statistics for all presented structures are summarized in Table 1. Structure factors and coordinates have been deposited in the Protein Data Bank as entries 5JHX, 5JHY, and 5JHZ.
UV–visible spectra of wild-type MagKatG2 and its variants were routinely recorded with a Hitachi U-3900 spectrophotometer at 25 °C. The molar absorption coefficient of the ferric, high-spin MagKatG2 at Soret maximum (ε404 = 102600 M–1 cm–1) was determined recently2 and used for calculation of KatG concentrations throughout this work. The absorption coefficient of MagKatG2 at 280 nm (ε280) was calculated to be 147800 M–1 cm–1 using the protein sequence (Expasy server). The spectra were recorded in either 50 mM citrate-phosphate buffer (pH 4.0–5.5) or 50 mM phosphate buffer (pH 7.0–8.5).
Electronic circular dichroism (ECD) spectra were recorded on a Chirascan spectrophotometer (Applied Photophysics, Leatherhead, U.K.). The instrument was flushed with nitrogen with a flow rate of 5 L min–1. The instrument allowed simultaneous UV–vis and ECD monitoring and was equipped with a Peltier element for temperature control. For recording of far-UV spectra (260–190 nm), the quartz cuvette had a path length of 1 mm, and for recording visible spectra, a path length of 10 mm was used. Instrumental parameters were set as follows: spectral bandwidth, 5 nm; step size, 1 nm; scan time, 15 s per point.
To monitor the thermal unfolding, 5 μM wild-type or mutant KatGs were incubated in 5 mM phosphate buffer (pH 7.0) with a stepwise increasing temperature (1 °C min–1) ranging from 20 to 90 °C. Single-wavelength scans were performed with instrumental parameters set as follows: spectral bandwidth, 1 mm (Soret) or 0.5 mm (far-UV); scan time per point, 12 s. Melting temperatures (Tm) were determined by sigmoidal fitting of obtained graphs using Pro-Data Viewer from Applied Photophysics (version 4.1.9).
Thermal denaturation of MagKatG2 and its variant was monitored using a VP-DSC MicroCal LLC calorimeter from GE Healthcare. The machine uses a cell volume of 137 μL and is equipped with a cooled autosampler for 96-well plates and automated cell cleaning and sample injection. The protein concentration of all samples was 5 μM in 5 mM phosphate buffer (pH 7.0–8.5) or 5 mM citrate-phosphate buffer (pH 4.5–5.5). The differential heat capacity (ΔCp) was recorded between 20 and 90 °C with a heating rate of 60 °C h–1. Baseline correction of the recorded thermograms was done by subtraction of buffer runs or rescans of each respective sample. After normalization for the exact protein concentrations, the acquired peaks were fitted by a non-two-state unfolding model using the LLC-Cap-DSC Add-on of Origin 7.0 (OriginLab).
Apparent kinetic parameters of catalase activity were determined polarographically using a Clark-type electrode from Hansatech (Oxygraph plus). The temperature was held constant at 30 °C by a thermostat-controlled water bath. The electrode was equilibrated by flushing the stirred reaction chamber with either dioxygen (O2) (100% saturation) or pure nitrogen (N2) (0% saturation). Reactions were started in oxygen-free 50 mM citrate-phosphate (pH 4.0–7.0) or 50 mM phosphate buffer (pH 7.0–8.5) by addition of enzyme. One unit of catalase activity was defined as the amount of enzymes that decomposes 1 μmol of hydrogen peroxide per minute in a 5 mM hydrogen peroxide solution at 30 °C. Final protein concentrations were 1 and 10 nM for the wild-type enzyme and Arg461Ala variant, respectively.
Peroxidatic activity was monitored spectrophotometrically at 470 nm by using 1 mM peroxyacetic acid or hydrogen peroxide as oxidants and guaiacol (ε470 = 26.6 M–1 cm–1) as the one-electron donor (1 mM) (all from Sigma-Aldrich). One unit of peroxidase activity was defined as the amount of enzyme that oxidizes 1 μmol of selected electron donor per minute at room temperature (25 °C).
Kinetic measurements of the direct reaction between the ferric proteins with cyanide, hydrogen peroxide, and peroxyacetic acid (PAA) at 25 °C were performed using SX-18MV or Pi*-180 stopped-flow machines from Applied Photophysics. Diluted PAA solutions were routinely incubated with low nanomolar concentrations of bovine liver catalase for degradation of H2O2. The path length of the optical cells was 10 mm and the volume 20 μL. Peroxide solutions were prepared freshly and diluted in 50 mM citrate-phosphate (pH 4.0–7.0) or 50 mM phosphate buffer (pH 7.0–8.5).
Calculation of pseudo-first-order rate constants (kobs) was performed with the Pro-Data Viewer software (Applied Photophysics, version 4.1.9). Second-order rate constants were determined from plots of kobs versus substrate concentration.
Molecular dynamics simulations of a MagKatG2 monomer (PDB entry 3UT2(1)) were performed using the GROMOS11 software package37 with the 54A7 force field.38 The simulations were conducted for 50 ns at a constant temperature (300 K) and a constant pressure (1 atm) and essentially performed as described previously.2 In the work presented here, the aim was to study the effect of the protonation state of the Tyr273 side chain on the conformational preference of mobile Arg461. Consequently, we conducted the following four independent MD simulations with Tyr273 either protonated (HO-Tyr273) or deprotonated (−O-Tyr273) and Arg461 in the “in” or “out” conformation: HO-Tyr/Arg“out”, HO-Tyr/Arg“in”, –O-Tyr/Arg“out”, and –O-Tyr/Arg“in”.
The relative free energies of the mobile Arg461 were calculated according to the linear interaction energy (LIE) method using the following equation:39
where angular brackets indicate ensemble averages, calculated over one of the four simulations. VEL and VVdW represent the electrostatic and van der Waals interactions, respectively. The subscripts start and end refer to hypothetical processes involving protonation of Tyr273 and a conformational change of Arg461 from the “out” to the “in” conformation, respectively. β and α are parameters of the LIE equation and were set to 0.5 and 0.18, respectively.40,41Figure Figure33 graphically depicts the processes that were studied. Only the first 10 ns of the four simulations were used for the free energy estimates because Arg461 changed its conformation from “out” to “in” in simulation –O-Tyr/Arg“out” afterward.
Before analysis of the functional role of Arg461 in ferric MagKatG2, we aimed to investigate its conformation in protein crystals at various pH values. In prokaryotic KatGs crystallized so far, this mobile amino acid showed a clear sensitivity to pH with mixed conformations around pH 6.0, 100% “out” conformation at pH <5, and 100% “in” conformation at pH >7.0.3,4,7,32 Crystals of MagKatG2 were obtained using the sitting-drop vapor-diffusion method at pH 4.6 as described recently1 and soaked in buffers with varying pH values. X-ray structures of MagKatG2 at pH 3, 5.5, and 7.0 were obtained at 1.4, 1.4, and 1.7 Å resolution (Figure Figure11 and Table 1).
Comparative analyses show that the overall structure remains unchanged at these pH values with respect to the first reported MagKatG2 structure (1.55 Å, PDB entry 3UT2)1 with two exceptions, namely the modification on the indole nitrogen of the adduct tryptophan (Trp140) and the conformation(s) of the flipping arginine (Arg461). Also, the electronic density maps show some extra density 15 Å from the first traced residue (Thr50) of one of the subunits, which could account for some of the missing residues of this flexible N-terminus.
The indole nitrogen from Trp140 is not modified in any of the structures analyzed in this work, whereas it was modified with a peroxide group in the first MagKatG2 structure and also in many of the reported BpKatG and MtKatG structures. This suggests that formation of this modification seems to be affected by subtle changes in the environmental conditions. The conformation of Arg461 changes from fully “out” to fully “in” between pH 3 and 7 (Figure Figure11 and Table 1). Consistently, at pH 5.5, the side chain of Arg461 shows a poor electron density suggesting mixed “in” and “out” conformations, though in the first MagKatG structure (at pH 4.6), Arg461 showed ~100% “in” conformation. The mobility of Arg461 at pH 5.5 is additionally underlined by the fact that the electron density for the “out” conformation is split (Figure Figure11).
The “in” conformation of Arg461 exhibits a hydrogen bond between one of the nitrogens of its guanidinium group and the tyrosinate oxygen of the MYW adduct (Figure Figure11). In the “out” conformation, Arg461 is making a hydrogen bond with the amide oxygen of Gln625 close to a conserved region that includes three arginines (Arg528, Arg533, and Arg540) (Figure Figure22 and Supplemental Figure 2). These observations confirm the conformational versatility of Arg461 in MagKatG2 with a pH dependency that appears to be shifted with respect to prokaryotic KatGs.
Molecular dynamics simulation underlines the existence of H-bonds between the guanidinium nitrogens of Arg461 in the “out” conformation and the amide oxygen of Gln625 (32 and 36% occurrence in the 50 ns simulation) (Figure Figure22). In the known X-ray structures of prokaryotic KatGs, the corresponding residue is a serine (Ser590 in BpKatG or Ser583 in MtKatG and Ser564 in SeKatG1) and Arg in its “out” conformation comes closest to the oxygen of its side chain at distances of ~3.5 Å (Supplemental Figure 2).32 Additionally, in its “out” conformation, the backbone oxygen of Arg461 is H-bonded to the guanidinium group of Arg533 (30, 18, and 39% occurrence) (Figure Figure22). This residue is also found in prokaryotic KatGs (Supplemental Figure 2). Split electron density for the side chain of Arg533 in crystal structures as well as MD simulations suggests increased flexibility of its side chain (see below). In any case, the N-terminal domain of MagKatG2 ends with Pro475, clearly demonstrating that mobile Arg461 in the “out” conformation interacts solely with residues from the C-terminal domain (Figure Figure22) that forms an accessible cavity for this flexible basic amino acid.
In summary, the crystal structures of MagKatG2 underline the mobility of Arg461 and demonstrate that, in contrast to prokaryotic KatGs, the “in” conformation is also found in crystal states under more acidic conditions. This correlates with the more acidic pH optimum of the catalatic activity (i.e., pH 5.25) of MagKatG2 compared to that of prokaryotic KatGs (see below).
Next, we explored the role of the protonation state of adduct Tyr273. The starting MagKatG2 coordinates used in the MD simulations were derived from the high-resolution (1.55 Å) structure (PDB entry 3UT2).1 Although MagKatG2 is a homodimer, the protomers (each being composed of an N-terminal heme-containing domain and a C-terminal domain) work independently. Thus, for the sake of computational convenience, models were constructed with only one protomer, half of the MagKatG2 molecule.
Four 50 ns MD simulations were performed with four different starting situations with Tyr273 either protonated (HO-Tyr273) or deprotonated (−O-Tyr273) and Arg461 in the “in” or “out” conformation: HO-Tyr/Arg“out”, HO-Tyr/Arg“in”, –O-Tyr/Arg“out”, and –O-Tyr/Arg“in”. Within 50 ns, changes in the orientation of the Arg461 side chain were seen only in the –O-Tyr/Arg“out” model (Figure Figure22). During this simulation, the side chain of Arg461 rotates from the “out” conformation to the “in” conformation. In the final (50 ns) structure, the side chain is slightly displaced with respect to the experimental crystallography data, so that the distance between the guanidinium group and the tyrosinate oxygen is significantly longer (5.4 ± 0.3 Å), suggesting the existence of a bridging water molecule (Figure Figure22). Furthermore, it was noticed that the side chain of Arg533 might move away from Arg461 when the latter is in the “in” conformation.
Next, we calculated the differences in the free energy (ΔG) for Arg461 among four 10 ns MD simulations of the starting models HO-Tyr/Arg“out”, HO-Tyr/Arg“in”, –O-Tyr/Arg“out”, and –O-Tyr/Arg“in”. Figure Figure33 compares the four hypothetical processes, clearly demonstrating the favorable interaction between Arg461 and the adduct Tyr273 in its deprotonated state. Protonation of Tyr273 forces Arg461 to adopt the “out” conformation. Because in the crystal state the “in” conformation was even present at pH 4.6 (see above), this might suggest a very low pKa value for the side chain of Tyr273 in MagKatG2.
The UV–vis spectrum of recombinant dimeric MagKatG2 (monomer size of 85 kDa) at pH 7.0 has been published recently.2 At room temperature, the high-spin ferric protein shows a Soret maximum at 404 nm, Q-bands at 505 and 539 nm, and a broad CT1 band at 640 nm. Figure Figure44B demonstrates that these spectral signatures are very similar in the pH range of 4.5–8.5, suggesting unchanged heme cavity architecture. This is reflected by almost identical coordinates of residues in the vicinity of the prosthetic group as seen in the X-ray structures. Upon exchange of Arg461 with Ala, the absorbance maximum of the mutant was only slightly red-shifted but still reflects the presence of high-spin ferric protein (Figure Figure44A), suggesting very similar active site geometries of wild-type and mutant protein. Identical UV–vis spectra of Arg461Ala were obtained between pH 4.5 and 8.5 with a Soret maximum at 406 nm, Q-bands at 506 and 541 nm, and a broad CT1 band at 642 nm.
Figure Figure55 shows that elimination of the mobile arginine significantly modifies the catalase activity. In contrast to the wild-type enzyme that shows a maximal activity at pH 5.25 (kcat = 6450 s–1), the variant Arg461Ala exhibits a significantly lower specific activity and shows a plateau between pH 4.5 and 7.5 (Figure Figure55A). At pH 5.5, the corresponding kcat value was determined to be 270 s–1, which is 24 times smaller than that of the wild-type protein at its pH optimum. In prokaryotic KatGs, the impact of exchange of mobile Arg with Ala was very similar (20–26-fold decrease in kcat at the pH optimum),14,33 but the catalase activity of Arg461Ala is still more than 2 orders of magnitude higher than that of conventional (monofunctional) peroxidases31 or KatGs with a disrupted Met-Tyr-Trp adduct.14,17−19,23−27 Interestingly, the specific catalase activity of the mutant increases with a decrease in pH from 4.5 to 3.0 (see the inset of Figure Figure55A), but it is important to note that in this very acidic pH range H2O2 degradation was incomplete, suggesting inactivation of the enzyme.
These polarographically determined data are fully confirmed by following the degradation of hydrogen peroxide spectrophotometrically at 240 nm in the stopped-flow apparatus. Figure Figure55B shows two representative time traces for the degradation of 10 mM H2O2 by 2 μM Arg461Ala at pH 5.5 and 7.0. It is clearly shown that (i) there is no pH dependence and (ii) the mutant is able to fully degrade H2O2 (10 mM H2O2 corresponds to ΔAbs240 ~ 0.4) without being inhibited ([substrate]/[enzyme] = 5000). For comparison, Figure Figure55C shows the degradation of 10 mM H2O2 by the wild-type enzyme at pH 5.25 and 7.0. Both panels demonstrate the higher catalatic activity as well as the pH dependence of the wild-type enzyme.
In contrast to the catalatic activity, the peroxidase activity is slightly enhanced upon exchange of Arg461 with Ala in a manner independent of whether peroxoacetic acid or H2O2 is used for initiation of reaction (Supplemental Figure 3). With guaiacol as the one-electron donor, wild-type MagKatG2 shows a pH optimum around pH 7.0 with both peroxides, whereas oxidation of guaiacol by Arg461Ala with H2O2 shows a broad optimum around pH 6.5 and with PAA a plateau between pH 5.5 and 6.5.
Next, we were interested in the impact of mobile Arg461 on the thermal stability of MagKatG2. At room temperature, both wild-type MagKatG2 and the variant Arg461Ala are dimeric in solution and show a single band at 85 kDa in sodium dodecyl sulfate–polyacrylamide gel electrophoresis under reducing conditions (not shown). Under nonreducing conditions, one band at 170 kDa is observed for both proteins, suggesting that the MagKatG2-typical interchain disulfide bridges1 between Cys55 and Cys74 of the N-terminal domains are intact in both enzymes. The far-UV circular dichroism (CD) spectra at pH 7.0 of both proteins are identical (insets of Supplemental Figure 4), suggesting that elimination of Arg461 has no impact on the overall secondary structure composition of the protein.
Figure Figure66 compares the temperature-mediated unfolding pathways of wild-type MagKatG2 and the mutant Arg461Ala between pH 4.5 and 8.5 followed by differential scanning calorimetry (DSC). It has to be mentioned that at pH ≤4.0 both proteins started to precipitate with an increase in temperature. At the pH optimum of catalatic activity of the wild-type enzyme, unfolding of both proteins follows a two-state unfolding pathway with two main and well-separated transitions. The mutant protein has a slightly higher thermal stability (Tm values at 53.2 and 63.2 °C) compared to that of the wild-type protein (52.6 and 61.2 °C). Thermal unfolding followed by CD spectroscopy at 208 nm (typical minimum in ellipticity of α-helices) and at the Soret minimum of 395 nm (Supplemental Figure 4) clearly demonstrates that the N-terminal heme-containing and cross-linked domain in MagKatG2 is much more stable than the C-terminal domain. Loss of heme ellipticity occurs between 60 and 66 °C. In prokaryotic KatGs, the N-terminal domain has a much lower thermal stability and both endotherms merge.42
The thermal stability of the N-terminal domain of both wild-type MagKatG2 and the variant Arg461Ala is highest between pH 5.5 and 7.0 and slightly decreases under more acidic and basic conditions. The symmetric shapes of the endotherms clearly suggest cooperative unfolding of the heme-containing domain at all investigated pH values. By contrast, the C-terminal domain seems to be much more susceptible to denaturation. Apart from pH 5.5, unfolding of the C-terminal domain becomes uncooperatively reflected by a broad peak tailing, which is more pronounced in wild-type MagKatG2 (Figure Figure66). The DSC data suggest that at pH 4.5 the C-terminal domain of wild-type MagKatG2 starts to unfold already at 35 °C.
Manipulation of the mobile Arg461 does not influence the binding kinetics of the low-spin ligand cyanide, again reflecting identical heme cavity architecture and accessibility. At pH 7.0, the binding rates and calculated KD values are (5.4 ± 0.2) × 105 M–1 s–1 and 14.4 μM (wild-type) and (3.5 ± 0.3) × 105 M–1 s–1 and 10.2 μM (Arg461Ala), respectively (Supplemental Figure 5).
Figure Figure77 shows the changes in the electronic absorption spectra of wild-type MagKatG2 and the variants Tyr273Phe and Arg461Ala upon reaction of the respective ferric proteins with peroxyacetic acid (PAA) at pH 7.0. The mutant Tyr273Phe was included in this study, because it represents the monofunctional variant without the Met-Tyr-Trp adduct and thus without catalatic activity. The reaction of Tyr273Phe with equimolar PAA leads to formation of a Compound II-like spectrum with bands at 412, 530, and 560 nm within 0.5 s, which most probably represents a Compound I* resulting from internal electron transfer from remote aromatic amino acids to transiently formed Compound I (Figure Figure77B). Both the spectral signatures and the kinetics of interconversion were almost identical at pH 5.5.
When wild-type MagKatG2 or the mutant Arg461Ala was mixed, completely different spectral transitions were monitored compared to those of Tyr273Phe, clearly suggesting participation of the KatG-typical adduct in the redox reactions. Several distinct phases could be observed. In wild-type MagKatG2, an increase in intensity and a narrowing of the Soret band were observed, followed by a “lag phase” and, finally, by a decrease in the Soret absorbance maximum at 405 nm, leading to an intermediate with absorbance bands at 407, 543, 598, and 640 nm. The latter conversion follows an apparent bimolecular rate constant of 7.5 × 103 M–1 s–1 at pH 7.0 (and 6.7 × 103 M–1 s–1 at pH 5.0) but might represent a steady-state shift to classical Compound I (blue spectrum in Figure Figure77A). The length of the lag phase decreased with an increase in PAA concentration (see insets of Figure Figure77A). Both the spectral signatures and the kinetics of interconversion were similar at pH 5.5 (not shown).
The kinetics of absorbance changes of the mutant Arg461Ala mixed with PAA also showed an initial increase at Soret absorbance followed by a decrease in Soret intensity (inset of Figure Figure77C). In contrast to the case for wild-type MagKatG2, the lag phase and the hypochromicity of the Soret absorbance of the resulting intermediate were less pronounced (Figure Figure77C). Similar kinetic and spectral transitions were observed at pH 5.5.
Next, we elucidated the spectral signatures of the redox intermediate of MagKatG2 that dominates during H2O2 degradation (Figure Figure88). Upon reaction of 3.6 μM wild-type ferric protein with 10 mM H2O2 at pH 5.25 within 1 ms, a low-spin species was formed showing a slighly red-shifted Soret band, two peaks at 545 and 580 nm in the Q-band region, and disappearance of the high-spin CT band at 640 nm. This spectrum resembles that of Compound III, which is observed when monofunctional peroxidases (e.g., APx) are mixed with millimolar H2O2.31 With MagKatG2, the reaction is very fast, and within 500 ms, 10 mM H2O2 is fully degraded and the ferric enzyme state is restored (inset of Figure Figure88A). These observed kinetics and spectral transitions during reaction of the eukaryotic wild-type catalase-peroxidase with hydrogen peroxide are reminiscent of the corresponding reactions found in prokaryotic KatGs below pH 7.0 as first described by Jakopitsch et al.23 With both prokaryotic and eukaryotic KatGs, in the alkaline pH region (pH >7.0), the spectral features of the dominating redox intermediate during H2O2 degradation are different. Together with a red-shifted Soret peak, a broad shoulder around 520 nm and loss of the high-spin CT band are typically observed immediately after mixing with hydrogen peroxide.2,23
Both the kinetics of degradation of hydrogen peroxide and the accompanying spectral transitions of the variant Arg461Ala with hydrogen peroxide revealed significant differences compared to those of the wild-type enzyme. Panels B (pH 5.5) and C (pH 7.0) of Figure Figure88 show the corresponding reactions between 2 μM Arg461Ala and 10 mM H2O2. Similar to that of the wild-type protein, the first monitored spectrum immediately after mixing in the stopped-flow apparatus shows a red-shifted Soret band and spectral features in the visible region that are reminiscent of the species seen in the wild-type protein at pH 5.25, i.e., peaks at 545 and 585 nm (green spectra in panels B and C of Figure Figure88). In contrast to the wild-type protein, 1 ms after mixing the mutant protein still shows some spectral features of the ferric state, reflecting its slower reactivity toward H2O2.
An important difference between wild-type and mutant protein concerns the dominating redox intermediate during H2O2 degradation. In wild-type MagKatG2, the intermediate immediately seen after mixing with H2O2 is fully present and visible during H2O2 degradation until (after complete peroxide depletion) the ferric spectrum is recovered. In case of the mutant protein, a steady-state shift to a mixture of species including Compound I (as seen in the reaction with peroxoacetic acid) is observed at both pH values (gray spectra in panels B and C of Figure Figure88), until (after complete peroxide depletion) the resting state is re-formed. Restoration of the ferric state after H2O2 depletion is slightly faster at pH 5.5 than at pH 7.0 (compare insets of panels B and C of Figure Figure88). Very similar spectral transitions are also seen at pH 3 and 8.5 (not shown). Thus, the dominating redox intermediate of Arg461Ala at alkaline pH still exhibits Compound III-like spectral features, which is in contrast to wild-type MagKatG2 and the prokaryotic counterparts (see above).
Heme peroxidases like cytochrome c peroxidases (CcP) or ascorbate peroxidases (APx) are known to be irreversibly inhibited by excess hydrogen peroxide in the absence of exogenous one-electron donors.31 Despite the fact that the core of the catalytic domain of bifunctional KatG bears a strong resemblance to CcP and APx,11 catalase-peroxidases are able to efficiently dismutate hydrogen peroxide without being inhibited. Moreover, in KatG, the catalase activity clearly dominates over the peroxidase activity and even rivals the activity of monofunctional catalases.9,10 In the past decade, the structural peculiarities like KatG-typical insertions (loops) and the presence of an autocatalytically and posttranslationally formed redox-active Met-Tyr-Trp adduct close to heme b have been analyzed in detail and brought into context with the mechanism of H2O2 dismutation (see recent reviews43,44). However, there is still an ongoing debate about the exact mechanism of the H2O2 oxidation reaction.43,44
So far, almost all experimental and computational data are derived from work on prokaryotic KatGs. The recent discovery of eukaryotic KatGs1,2,12,35,36 now opens the possibility of querying postulated mechanisms and/or specifying so far unsolved mechanistic details. The selected model enzyme MagKatG2 has conformational and thermal stability much higher than those of the prokaryotic counterparts and, interestingly, has a pH optimum of catalase activity at pH 5.25 compared to the range of pH 6.0–6.5 reported for prokaryotic KatGs.43,44
In this work, we have focused on the perhaps most intriguing modulator of catalase activity of KatG, i.e., the mobile arginine that is fully conserved in both prokaryotic and eukaryotic KatGs.35,36 Data from prokaryotic KatGs and this work clearly show that it is integral to the catalase activity because substitution with anything but Lys greatly diminishes the catalase but not the peroxidase activity.14,32−34 Similar to the case in prokaryotic KatGs, Arg461 in MagKatG2 is mobile as demonstrated by X-ray crystallography and MD simulations; however, in contrast to prokaryotic KatGs, the “in” conformation (i.e., ionic interaction with adduct Tyr273) is present also at acidic pH values as seen in the original structure at pH 4.62 and the novel structure obtained from soaking at pH 5.5. Only at a very acidic pH, the “out” conformation dominates (Figure Figure11). In the “out” conformation, Arg461 interacts exclusively with residues from the C-terminal (catalytically inactive) domain. In MagKatG2, Arg461 in its “out” conformation is hydrogen bonded with the amide oxygen of Gln625, whereas in BpKatG, MtKatG, and KatG from Synechococcus elongatus, it is H-bonded with the side chain of a serine. In this context, it is important to note that the C-terminal domain in MagKatG2 has a low conformational and thermal stability at pH ≤4.5 in contrast to the N-terminal domain (Figure Figure66). It is well-known that an intact heme cavity architecture of KatG depends to some extent on the gene-duplicated C-terminal domain. KatG lacking this domain or parts of it displays exclusively low-spin heme and loses both the catalase and peroxidase activity.20,45 Considering the very low stability of the C-terminal domain in the acidic pH region, it is reasonable to assume that the binding site of Arg461 in the “out” conformation collapses in solution at very low pH values (despite the fact that it was still intact in the crystal soaked with buffer at pH 3.0). This suggests that enzyme activity data obtained from measurements at pH <4.0 might be problematic.
In any case, we could demonstrate that the “in” conformation of Arg461 dominates in the crystal structures of MagKatG2 at pH ≥4.5. Additionally, the performed 50 ns MD simulations and calculations of differences in the free energy among four MD simulations of various starting models clearly show that Arg461 is mobile and that deprotonation of Tyr273 thermodynamically favors the “in” conformation, whereas its protonation promotes Arg461 to adopt the “out” position. These data might suggest that the pKa of Tyr273 is lower compared to those of prokaryotic KatGs, but it is also important to keep in mind that the position of the mobile arginine itself modulates the pKa of the adduct tyrosine. As shown in this paper, the mobile arginine in the “out” position has different interaction partners in prokaryotic and eukaryotic KatGs, and it is tempting to speculate whether these differences favor a stronger tendency in MagKatG2 to adopt the “in” conformation and give rise to a lower pKa. Additionally, it is important to note that the adduct is oxidized during enzyme turnover (reflected in the intermediate states Compound I* and Compound III*), which in addition promotes formation of tyrosinate and ionic interaction with Arg461 (see below).
Similar to the case for prokaryotic KatGs, elimination of Arg461 by mutation to Ala partially impairs the catalase activity. Additionally, the pH dependence is modified, and similar catalatic activity is observed between pH 4.5 and 7.5. The turnover number of Arg461Ala at pH 5.5 is only 4% of the kcat of the wild-type enzyme at its pH optimum. By contrast, the peroxidase activity is almost unaffected by elimination of Arg461. This clearly underlines the fact that for conventional peroxidase activity neither an intact adduct2,16−19 nor the presence of the mobile arginine is needed, but both fully conserved structural KatG-typical features are essential for efficient oxidation of hydrogen peroxide.
Similar to the case for cytochrome c peroxidase, the characteristic green porphyrin π cation radical Compound I (typical for horseradish peroxidase or ascorbate peroxidase)46 does not accumulate in KatG when it is mixed with H2O2. In CcP, Compound I* carries the proximal Trp191●+ radical.46 In KatG, transiently formed Compound I (reaction I) is also rapidly converted to Compound I* (reaction II) as demonstrated by UV–vis stopped-flow spectroscopic studies23,24 and freeze-quench EPR.14,24 Studies of KatG from Mycobacterium tuberculosis (MtKatG) demonstrated that the Met-Tyr-Trp adduct radical14,24 is formed immediately after mixing with H2O2 and persists during the time interval corresponding to the H2O2 consumption kinetics. Importantly, the spin concentration could not be increased at higher H2O2 concentrations,14 and deletion of the mobile arginine had no impact on reaction II; i.e., the corresponding mutant produces a similar yield of adduct radical, and the persistence of the radical signals coincided with the period required for H2O2 consumption.14
Formation of Compound I* should be independent of the oxidant used for the two-electron oxidation of the ferric resting state. The effects of the produced acid notwithstanding, the chemistry of the peroxide (PAA vs H2O2) determines the kinetics of reaction I but has no direct impact on the internal electron transfer represented by reaction II. It was interesting to see that, with both the wild-type protein and the mutant Arg461Ala, the reaction with peroxacetic acid was biphasic, reflected by an increase in absorbance followed by hypochromicity at the Soret maximum (Figure Figure77A,C). A similar kinetics is also seen in prokaryotic KatGs, and the increase was interpreted as formation of a complex between the respective ferric proteins and PAA.47,48 However, in the absence of the covalent adduct (e.g., in Tyr273Phe) but with the distal tryptophan in place (as in CcP and APx)2,49 and identical substrate channel architecture, the PAA-mediated reaction was always accompanied by a monophasic decrease in Soret absorbance only. Thus, one might speculate that the first fast phase of the reaction of ferric wild-type MagKatG2 or Arg461Ala with PAA reflects the rate-limiting step in Compound I* formation that is followed by a steady-state shift to conventional Compound I. Because H2O2 is absent, Compound I* is relatively stable and can be reduced to the resting state only by electron transfer from the protein matrix that becomes oxidatively damaged.48 As a consequence, reaction II becomes more and more impaired and (classical) Compound I accumulates with time. In contrast to the H2O2-mediated conversion, the adduct radical has a significantly longer life span in the KatG/PAA system.
Elimination of the mobile arginine has no impact on the H2O2-mediated oxidation of the enzyme to Compound I* (reactions I and II) or formation of Compound III* (reaction III), but it impairs the effective adduct radical utilization in the second half-reaction, i.e., the release of O2 (reaction IV). Stopped-flow kinetic investigations of both prokaryotic and eukaryotic KatGs mixed with an excess of H2O2 around the pH optimum and in the acidic pH region typically show the immediate (<1 ms) formation of an intermediate with a red-shifted Soret band and two typical bands at 542–545 and 580–585 nm2,14,19,23,24 that strongly resemble the spectral signatures of Compound III (oxyheme species) of peroxidases, including CcP50 and APx,51 or KatG variants with a disrupted adduct.2,14,19 In contrast to monofunctional peroxidases31 in KatG, (i) oxidation of H2O2 according to reaction III is several orders of magnitude faster52 and (ii) oxyheme formation occurs in the presence of the adduct radical thus denoted Compound III*.14,24 The fact that carbon monoxide (which binds to only ferrous heme as found in Compound III*) interferes with the catalase reaction of KatG (but not with monofunctional catalases) under both acidic and basic conditions14 suggests that Compound III* formation occurs throughout the pH range of the catalase reaction. As a conclusion, we propose that elimination of the KatG-typical arginine has no impact on reactions I–III and thus Compound III* formation as shown for both MtKatG14 and MagKatG2 (Figure Figure88).
The question remains about the role of Arg461 in reaction IV, i.e., the release of O2? Importantly, significant differences between wild-type MagKatG2 and the mutant Arg461Ala in the H2O2-mediated spectral conversions were observed. There was always a coincidence between H2O2 consumption and the persistence of the dominating low-spin intermediate in wild-type MagKatG2. This was also observed with prokaryotic KatGs.14,23 Moreover, for MtKatG, it could be demonstrated that this coincidence included also the persistence of the adduct radical. In the absence of the mobile arginine upon addition of H2O2, the same redox intermediate is formed immediately after mixing but was then replaced in the time course of reaction by an intermediate with spectral signatures similar to that of (classical) Compound I (Figure Figure88B,C). It has been demonstrated by DFT calculations14 that the ionic interaction between the side chain of the mobile arginine and the adduct Tyr favors the redox reaction between the adduct radical and the adjacent superoxide/dioxyheme of the Compound III* intermediate through electronic effects. In wild-type KatG, this effect will be favored at all pH values where the mobile arginine is in the “in” conformation. In the absence of the basic amino acid, the electron transfer between superoxide and the adduct radical is hampered, and consequently, the turnover of Compound III* is significantly retarded. Similar to the PAA-mediated reaction under this condition, the protein might act as a poor electron donor, thus providing a Met-Tyr-Trp adduct radical escape pathway. This is supported by the observation of pronounced formation of dityrosine cross-linked oligomers of KatGs that lack the mobile arginine.14 This also suggests that the large amount of tryptophans and tyrosines found in the N-terminal domains of KatGs have an antioxidant role. In any case, as a consequence, (classical) Compound I accumulates during H2O2 degradation by Arg461Ala. This might also explain why in conventional (hand-mixing) EPR measurements an oxoiron(IV) porphyryl radical intermediate could be detected.33
In summary, in the case of MagKatG2, the pH optimum of H2O2 dismutation was shown to be 5.25. Around the pH optimum and above, Arg461 is predominantly in the “in” conformation. Below pH 5.25, the catalase activity decreases most probably because Tyr273 is protonated and its interaction with Arg461 is disfavored and quenching of the adduct radical by the dioxyheme intermediate becomes slower. Additionally, the fact that the C-terminal domain starts to collapse below pH 4.5 must be taken into account, which additionally diminishes the catalatic activity. Above pH 5.25, the capability of MagKatG2 to dismutate H2O2 decreases, too, despite the favorable interaction between Arg461 and Tyr273. At pH >7, the dominating redox intermediate during H2O2 degradation does not exhibit the Compound III-like spectral features2,14,23 while the adduct radical is still present at normal abundance.14 This might suggest that in wild-type MagKatG2 at alkaline pH formation of Compound III* (reaction III) is slower than its turnover (reaction IV). As a consequence, Compound III* does not accumulate during turnover. It is known from monofunctional peroxidases that Compound III formation is pH-dependent and disfavored under alkaline conditions.52 In contrast to the case for wild-type MagKatG2, in the variant Arg461Ala Compound III* also accumulates at alkaline pH values, because in the mutant reaction IV is even slower than reaction III.
In conclusion, the role of the mobile Arg in MagKatG2 and prokaryotic KatGs is to guarantee optimal electron transfer between the superoxide/dioxygen intermediate and the adduct radical. Its suppresses unspecific escape pathways of the adduct radical and oxidative damage of the protein. As a consequence, KatGs efficiently dismutate H2O2 and release O2. Upon supporting the turnover of Compound III*, the mobile arginine helps to avoid both the release of superoxide and irreversible enzyme inhibition, which is typically observed when monofunctional peroxidases are incubated with H2O2 in the absence of one-electron donors.31
Atomic coordinates and structure factors have been deposited in the Protein Data Bank as entries 5JHX, 5JHY, and 5JHZ.
This project was supported by the Austrian Science Fund, FWF [Doctoral program BioToP-Biomolecular Technology of Proteins (W1224) and Projects P23855 and P25270] and the Vienna Science Technology Fund (LS08-QM03).
The authors declare no competing financial interest.