A radical on the Met-Tyr-Trp modification required for catalase activity in catalase-peroxidase is established by isotopic labeling and site-directed mutagenesis

Abstract

A transient tyrosyl-like radical with a narrow doublet X-band EPR signal is present during catalase turnover by M. tuberculosis catalase-peroxidase (KatG). Labeling of KatG with β-methylene deuterated tyrosine causes a collapse of the doublet to a singlet, while for 3,5- ring-deuterated tyrosine labeled enzyme, no changes occur in the EPR signal. Except for the replacement Tyr229Phe, all other single tyrosine mutants of KatG exhibit the same narrow doublet EPR signal and catalase activity similar to wild-type enzyme. These findings confirm that this catalytically competent radical is associated with Tyr229, whose 3′, 5′ protons are replaced due to crosslinks with neighboring Met255 and Trp107 side chains in the post-translationally modified enzyme containing a distal side Met255Tyr229Trp107 adduct.

M. tuberculosis (Mtb) catalase-peroxidase (KatG) has attracted great interest because of its role in the activation of the anti-tubercular pro-drug isoniazid (INH) and because of widespread resistance to INH due to KatG mutations in the tuberculosis pathogen.^1–2 Aside from a broad peroxidase activity, KatG enzymes also exhibit robust catalase activity (k_cat/K_M ~ 10⁶ M⁻¹s⁻¹),³ dismutating hydrogen peroxide into dioxygen and water in a non-scrambling mechanism despite a lack of sequence homology with mono-functional heme catalases.⁴ One key structural feature that differentiates KatG from other heme peroxidases is the conserved distal side tri-amino-acid adduct (Met²⁵⁵-Tyr²²⁹-Trp¹⁰⁷, MYW) (Figure 1), a posttranslational modification essential for catalase activity.^5–9 What makes KatG such an efficient catalase is intriguing because peroxidases are generally very poor catalases.

Figure 1.

The distal side Met-Tyr-Trp adduct in Mtb KatG (PDB entry code: 2CCA).

We recently reported that a transient tyrosyl-like radical forms during the reaction of KatG with H₂O₂ and persists during the time interval for turnover of excess peroxide.¹⁰ Mutations that disrupt MYW adduct formation (e.g., M225A, Y229F, or W107F) abolish this radical species and cause complete loss of catalase activity without interfering with peroxidase activity.¹⁰ According to the correlation of EPR spectral parameters for this radical with structural features of tyrosine residues in Mtb KatG, and with DFT calculations of the electronic structure of a neutral MYW radical modeled after the crystal structure of Mtb KatG, the tyrosyl-like radical was tentatively assigned to the MYW adduct. Confirmation of the radical site is essential for understanding KatG function. For the present study, twenty-one Mtb KatG tyrosine mutants were prepared by site-directed mutagenesis and two types of deuterium-labeled wild-type KatG were also prepared such that the β-methylene positions or the 3,5,-ring positions carried ²H, in an approach to identify the site of the catalytically competent radical.

When WT KatG reacts with a large excess of H₂O₂, a narrow doublet EPR signal is found (Figure 2) (peak to trough line width of 17 G, g = 2.0034), which we previously identified as a tyrosyl-like radical based on the g-value anisotropy evaluated using high-field EPR.¹⁰ Here, isotopically labeled Mtb KatG was overexpressed in E. coli BL21 grown in M9 minimal medium supplemented with either of two deuterium labeled tyrosine molecules. Purified enzymes were reacted with excess H₂O₂ and immediately frozen. For KatG containing β-methylene-deuterated tyrosine, a singlet signal with a line width of 11 G is seen (Figure 2). Replacement of hydrogen with deuterium reduces electron-nuclear hyperfine interactions because of the ~ 6-fold smaller g-factor for ²H vs. ¹H. The change in the X-band EPR spectrum demonstrates that the narrow doublet signal in wild-type KatG is associated with a tyrosine residue as we previously assigned. In contrast to this, the EPR signal from KatG labeled with ring-deuterated tyrosine (3,5-d₂-Tyr) is unchanged compared to unlabeled KatG. Replacement with deuterium of these two ring hydrogens in a tyrosyl radical would also alter hyperfine splittings and/or reduce the line width of the EPR signal, neither of which was observed. The absence of a change suggests that the radical species lacks 3,5-ring hydrogens. Alternatively, the unchanged narrow doublet could indicate unsuccessful labeling of KatG in the case of the ring-deuterated tyrosine. That the KatG used here was in fact labeled with 3,5-d₂-Tyr was demonstrated by taking advantage of our previous observations that in reactions of KatG with alkyl peroxides,^11–12 tyrosyl radical(s) remote from the MYW adduct are formed. As shown in Figure 3, the line width of the singlet signal typical of wild-type KatG treated with PAA for 5 sec (using a rapid freeze-quench technique) is reduced from 23.0 G to 18.8 G for the 3,5-d₂-Tyr labeled enzyme. These observations prove that 3,5-d₂-Tyr was successfully incorporated into KatG. For enzyme labeled with Tyr deuterated at the β-methylene sites, the line width was reduced to 14.1 G, again demonstrating that the majority species giving these EPR signals upon turnover with PAA is a tyrosyl radical(s) (and that the radical(s) formed using PAA is different from that formed during catalase turnover ¹³).

Figure 2.

The X-band EPR spectrum (77 K) of the radical formed during catalase turnover in Mtb KatG becomes an apparent singlet for enzyme labeled with β-methylene-deuterated tyrosine but is unchanged in KatG labeled with 3,5-deuterotyrosine. Enzymes were manually mixed with 8000-fold molar excess of H₂O₂ at pH 8.5 and rapidly frozen in liquid N₂.

Figure 3.

The EPR spectrum (77 K) of the radical formed in Mtb KatG reacted with peracetic acid is narrowed for enzyme containing β-methylene- or 3,5-ring deuterotyrosine. Enzymes were mixed with a 3-fold molar excess of peracetic acid and frozen after 5 sec using a rapid freeze-quench apparatus.

In an attempt to explore possible roles of specific tyrosine residues in radical formation and the catalase mechanism, we mutated each of the 21 tyrosine residues in Mtb KatG. As given in Table S1 (Supporting Information), except for Tyr229Phe, each mutant enzyme exhibited the narrow doublet EPR signal when reacted with H₂O₂ and had close to normal catalase activity. Do these observations directly confirm residue 229 as the radical site? It is conceivable that Tyr229 or an intact MYW adduct is required for formation of a tyrosyl radical(s) on some other site or sites during catalase turnover. However, the observation of the same narrow doublet in each mutant would then mean that upon mutation of one of the residues other than Tyr229, yet another Tyr radical that exhibits the same EPR signal is formed. Furthermore, a scenario in which different tyrosyl residues beyond Y229F could be radical sites is unreasonable given the conserved catalase activities, since a catalytic function of the radical in the active site containing oxyferrous heme ^9,14 must be maintained in each functional mutant. These observations taken together- the unchanged narrow doublet EPR signal found during catalase turnover in the 3,5-d₂-Tyr labeled enzyme and the presence of the narrow doublet in each mutant except Y229F is then best explained by assigning the catalytic radical to the MYW adduct in which the tyrosine ring lacks 3,5 hydrogens. These hydrogens are replaced by sulfur or carbon in the crosslinks to neighboring Met225 and Trp107 side chains (Figure 1).

The evidence for a radical associated with the MYW adduct and catalase activity in catalase-peroxidase is supported by the observation that the EPR signal intensity can be as high as 0.5 spins per heme in rapid freeze-quench samples frozen after a few milliseconds of reaction with H₂O₂ (not shown). According to DFT calculations, the majority of unpaired spin density is contained within the tyrosyl ring in this structure.¹⁰

Unlike other peroxidases, KatG forms Compound III (oxyferrous heme) upon reaction with excess H₂O₂ yet rapidly turns over massive excesses of peroxide. In mutants such as M255A, Y229F, or W107F, the formation of a stable Compound III requires only a very small excess of H₂O₂ and similar to monofunctional peroxidases, no catalase turnover occurs.^7,9–10 The rapid catalase activity of KatG therefore depends on efficient recycling of Compound III. Consistent with this idea is our earlier evidence that the radical must participate in a reaction with oxyferrous heme, enabling dioxygen release and yielding ferric KatG ^9–10 as proposed by Jackopitsch et al.¹⁴ Therefore, it is reasonable that superoxide and not dioxygen is dissociated from oxyferrous heme and the unpaired spin is quenched by the MYW radical during catalase turnover by KatG (Scheme 1).

Scheme 1.

Formation of adduct radical and its possible role in the catalase activity of KatG. The cartoon depicts heme, the distal Met-Tyr-Trp (MYW) adduct/radical and protein sites (Y, W residues) where other amino acid radicals may form.