Synergy not Antagonism in Antioxidant Defenses: The Unanticipated Effect of Electron Donors on Catalase-peroxidase Function

Abstract

Catalase-peroxidases, also called KatGs, have raised considerable interest due to their role in the activation of isoniazid, an anti-tubercular pro-drug for Mycobacterium tuberculosis. KatGs are heme-dependent enzymes and are primarily known as H2O2 scavengers. As the name suggests, they catalyze the decomposition of H2O2 by catalatic and peroxidatic mechanisms using a single active site. Despite the progress made during the past decade in understanding the structure-function relationships of KatGs, very little is known about the connectivity/interplay between the catalatic and peroxidatic functions. The initial step for both catalytic processes is the heterolytic reduction of H2O2 to H2O, giving rise to a two-electron oxidation of the ferric enzyme to a ferryl-porphyrin π cation radical intermediate, compound I (i.e., FeIV=O[porphyrin]•+). At this point, the reduction of compound I to the ferric state differs for both the catalase and peroxidase activities. To complete the catalase reaction, compound I returns to the ferric state by oxidation of another H2O2 to form O2 and H2O. However, the peroxidase reaction is completed by the subsequent reduction of compound I to compound II (i.e., FeIV=O), and subsequently, to the ferric state at the expense of two equivalents of an exogenous electron donor. This yields H2O and two equivalents of the corresponding radical of the donor. The ability of one activity to dominate over the other depends on several factors including: pH, the concentration of H2O2, and the availability of an appropriate exogenous electron donor. Catalase activity is optimal near neutral pH (i.e., pH ~ 7.5), whereas peroxidase activity is optimal under acidic conditions (i.e., pH ~ 4.5) and requires an exogenous electron donor. More so, peroxidase activity is favorable at low H2O2 concentrations. Conversely, catalase activity can sustain high concentrations of H2O2 without the enzyme inactivation. Clearly, conditions which favor the peroxidase activity do not coincide with those that favor the catalase activity. Thus, the simplest way to understand the mechanism of both functions in KatG has often been to superimpose both the catalatic and peroxidatic mechanisms. Based on this mechanism, both activities should be mutually antagonistic and peroxidatic electron donors should inhibit the catalase activity. In fact, it has been established that the common peroxidatic electron donor o-dianisidine does inhibit the catalase activity of KatG from Escherichia coli at pH 7.5. Strikingly, in this dissertation, we report the dramatic stimulation of the catalase activity of KatG from Mycobacterium tuberculosis (MtKatG) by up to 14-fold, in the presence of several common peroxidatic electron donors. The stimulatory effect was most prominent under conditions favorable to peroxidase activity (i.e., acidic pH and low H2O2 concentrations). In particular, we observed that aromatic amines were better stimulators than other donors like pyrogallol and ascorbate. In the absence of a peroxidatic electron donor, we observed a “low-KM” and “high-KM” component for catalase activity at pH 5.0. The inclusion of peroxidatic electron donors increased the apparent kcat for the “low-KM” component at pH 5.0. During stimulated catalase activity, very little of the donor (0.008 eq./H2O2 consumed) accumulated in its oxidized state. This is far less than the amount expected for normal peroxidatic turnover where two equivalents of oxidized donor is anticipated for every equivalent of H2O2 consumed. Evaluation of the dominant enzyme intermediates by stopped-flow spectroscopy revealed that a compound III-like (i.e., FeIII-O2•-[MYW]•) intermediate dominated during electron donor-stimulated catalase activity of MtKatG, and this intermediate converted directly to the ferric state upon depletion of H2O2. In the absence of the donor, a similar species persisted and returned slowly to the ferric state, but only long after H2O2 was fully consumed. Clearly, the catalase mechanism and the interrelationship between the catalase and peroxidase functions of KatG are much more complex than has been previously appreciated, and pH plays an important factor in both activities. Likewise, evidence is accumulating showing that pH also induce structural changes that markedly affect the catalase activity of KatG. Recent reports have shown that an invariant arginine residue (R418 in MtMatG numbering) is critical to the catalase activity, as it undergoes a conformational switch that is highly pH-dependent. KatG also possesses a unique covalent adduct (M255-Y229-W107 in MtKatG numbering) that is critical to its catalase activity. At pH 4.5, the arginine (R418) side chain is oriented away from the KatG-unique M255-Y229-W107 covalent adduct (i.e., “R” conformation). Whereas at pH 8.5, the side chain of this arginine is oriented toward the M255-Y229-W107 covalent adduct (i.e., “Y” conformation). The pH-dependence of the stimulatory effect and that of the R418 side chain and its role in catalatic turnover prompted us to evaluate the connection between the two. Substitution of R418 by alanine, R418A, produced an enzyme with almost no catalase activity at pH 7.0. However, catalase activity increased by nearly two orders of magnitude as pH was lowered to 5.0. Furthermore, at pH 5.0, peroxidatic electron donors such as 2,2’-azino-bis(3-ethyl-benzthiazoline-6-sulfonate) [ABTS] further stimulated the catalase activity of R418A by an order of magnitude. This was similar to the extent observed for wild-type and R418K KatG. Unlike the wild-type and R418K KatG, a greater amount of oxidized donor radical (i.e., ABTS•+) was produced per H2O2 consumed by R418A KatG. This was highly H2O2 concentration dependent such that far more donor radical accumulated at low H2O2 concentrations. Stopped-flow studies showed that a compound III-like intermediate dominated during electron donor-stimulated catalase activity of R418A and R418K KatG, and converted directly to the ferric state. Under these conditions, the time required for H2O2 consumption and return of the ferric state was 3 to 4 fold slower for R418A KatG than for wild-type and R418K KatG. However, in the absence of ABTS, the return of the ferric state lagged far behind the conclusion of H2O2 consumption by wild-type and both R418 variants. Even though the presence of ABTS seemed to resolve this problem for all three enzymes, it did so without a direct impact on the intermediates of the catalatic cycle. From our data we conclude that the stimulatory effect of electron donors is due to their ability to prevent the accumulation of intermediates which are inactive with respect to catalase activity. The R418A substitution, in addition to showing the conversion of the compound III-like intermediate to the ferric state, also appears to have a greater propensity for off-pathway electron transfers which produce catalase-inactive intermediates. Our results have potentially far reaching implications. The ability of peroxidatic electron donors to stimulate catalatic turnover instead of inhibition points toward a much more central role for peroxidase substrates in the unusual catalase mechanism of KatG. Clearly, the capacity of KatG to diffuse threats posed by H2O2 is far greater than first thought. The conditions which favor the stimulatory effect coincide with those observed during antimicrobial defenses such as the oxidative burst. This raises the importance of peroxidatic electron donors in accounting for the resistance of pathogens to H2O2 and stimulates the concern to investigate the identity and binding site of compounds which serve this capacity.