Catalytic Mechanism of Cofactor-Free Dioxygenases and How They Circumvent Spin-Forbidden Oxygenation of Their Substrates

Article

pubs.acs.org/JACS

Catalytic Mechanism of Cofactor-Free Dioxygenases and How They

Circumvent Spin-Forbidden Oxygenation of Their Substrates

Aitor Hernandez-Ortega,^†Matthew G. Quesne,^‡Soi Bui,^§Derren J. Heyes,^†Roberto A. Steiner,^§

́

Nigel S. Scrutton,*^,†and Sam P. de Visser*^,‡

^†Manchester Institute of Biotechnology and Faculty of Life Sciences, The University of Manchester, 131 Princess Street, Manchester

M1 7DN, United Kingdom

^§Randall Division of Cell and Molecular Biophysics, King’s College London, London SE1 1UL, United Kingdom

^‡Manchester Institute of Biotechnology and School of Chemical Engineering and Analytical Science, The University of Manchester,

131 Princess Street, Manchester M1 7DN, United Kingdom

S

* Supporting Information

ABSTRACT: Dioxygenases catalyze a diverse range of biological

reactions by incorporating molecular oxygen into organic substrates.

Typically, they use transition metals or organic cofactors for catalysis.

Bacterial 1-H-3-hydroxy-4-oxoquinaldine-2,4-dioxygenase (HOD)

catalyzes the spin-forbidden transfer of dioxygen to its N-

heteroaromatic substrate in the absence of any cofactor. We

combined kinetics, spectroscopic and computational approaches to

establish a novel reaction mechanism. The present work gives insight

into the rate limiting steps in the reaction mechanism, the eﬀect of

ﬁrst-coordination sphere amino acids as well as electron-donating/electron-withdrawing substituents on the substrate. We

highlight the role of active site residues Ser₁₀₁/Trp₁₆₀/His₂₅₁and their involvement in the reaction mechanism. The work shows,

for the ﬁrst time, that the reaction is initiated by triplet dioxygen and its binding to deprotonated substrate and only thereafter a

spin state crossing to the singlet spin state occurs. As revealed by steady- and transient-state kinetics the oxygen-dependent steps

are rate-limiting, whereas Trp₁₆₀and His₂₅₁are essential residues for catalysis and contribute to substrate positioning and

activation, respectively. Computational modeling further conﬁrms the experimental observations and rationalizes the electron

transfer pathways, and the eﬀect of substrate and substrate binding pocket residues. Finally, we make a direct comparison with

iron-based dioxygenases and explain the mechanistic and electronic diﬀerences with cofactor-free dioxygenases. Our

multidisciplinary study conﬁrms that the oxygenation reaction can take place in absence of any cofactor by a unique mechanism

in which the specially designed ﬁt-for-purpose active-site architecture modulates substrate reactivity toward oxygen.

Ring-cleaving dioxygenases play critical roles in the aerobic

biodegradation of aromatic compounds.⁵A large subgroup of

the ring-cleaving dioxygenases is the catechol dioxygenases that

contain nonheme iron as cofactor. On the basis of the cleavage

mode and redox state of the cofactor, these are classiﬁed in two

major groups: (i) the extradiol dioxygenases, which employ

Fe²⁺and cleave adjacent to the vicinal hydroxyl groups of

catechols (meta), and (ii) the intradiol dioxygenases, which use

an Fe³⁺cofactor and cleave in between the hydroxyl functions

(ortho) of the substrate. The catalytic mechanism of both

groups has been studied in detail by a combination of kinetic,⁶

spectroscopic,⁷crystallographic,⁸and computational studies.⁹

Although intra- and extradiol dioxygenases diﬀer in their O₂

activation mechanisms, both converge at an alkylperoxo

intermediate. However, since the substrate binding to the

iron diﬀers in both groups, the cleavage speciﬁcity is altered

through the coordination mode.^8b

INTRODUCTION

■

Dioxygenases are mechanistically intriguing enzymes since they

are able to perform a spin-forbidden reaction in which the

triplet spin ground-state of molecular oxygen reacts with either

a cofactor or an organic molecule. These enzymes catalyze the

incorporation of molecular oxygen into their organic substrates

as a means to initiate their metabolism. Dioxygenases perform

key functions for human health and participate in central

biochemical functions such as DNA base repair and hypoxia

responses.¹Furthermore, they have been used in the develop-

ment of new antitumor therapies.²In plants, dioxygenases are

essential in signaling mechanisms as well as in secondary

metabolism.³Bacterial dioxygenases, such as naphthalene

dioxygenase, are of great interest for their biotechnological

applications in contaminated soils, where they are involved in

the biodegradation of harmful compounds.⁴Interestingly, these

enzymes show diﬀerent cofactor requirements, ranging from

transition metals (typically iron with either a heme and

nonheme ligand environment) to ﬂavins; moreover they show

diverse structural frameworks to drive oxygenation chemistry.

Received: April 14, 2015

A

DOI: 10.1021/jacs.5b03836

J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society

Article

Scheme 1. Proposed Catalytic Cycle of the Cofactor-Free Dioxygenase HOD

Interestingly, one speciﬁc class of ring-cleaving dioxygenases

is able to catalyze the breakdown of catecholate-related

compounds via oxygen incorporation without the requirement

of any kind of cofactor.¹⁰One example of a cofactor-free

oxygenase, is 1-H-3-hydroxy-4-oxoquinaldine 2,4-dioxygenase

(HOD) from Arthrobacter nitroguajacolicus Ru61a, which

̈

catalyzes the oxygenolytic breakdown of its natural N-

heteroaromatic substrate (2-methyl-3-hydroxy-4(1H)-quino-

lone, MHQ) with concomitant release of carbon monoxide.

Although details of the catalytic cycle of HOD remain

uncertain, the hypothesized cycle based on experimental

studies is depicted in Scheme 1.¹¹The catalytic cycle starts

with MHQ binding and its deprotonation at oxygen atom O³.

Subsequently, dioxygen binds into the active site (structure R),

which is believed to trigger an electron transfer to form the

charge-transfer complex R_CT. The superoxo group attacks the

substrate to form the substrate-peroxo complex (I₁) followed

by a ring-closure step to form a bicyclic ring structure (I₂). The

latter releases CO and forms the N-acetyl-anthranilate product

(P).

HOD, together with Pseudomonas putida 33/1 1-H-3-

hydroxy-4-oxoquinoline 2,4-dioxygenase (QDO), constitute a

separate dioxygenase family. Structurally, they are not related to

any known cofactor-dependent dioxygenase.¹⁰Instead, they

belong to the α/β-hydrolase fold similar to hydrolases,^11band

for this reason, HOD and QDO represent fascinating examples

of divergent evolution of the α/β-hydrolase architecture from

ester hydrolysis to oxygenation reactions.^10,11b

Figure 1. Substrate (MHQ) and dioxygen bound HOD. Active-site

residues surrounding the substrates shown in stick representation. The

structure is based on the coordinates of the anaerobic HOD·MHQ

complex (2WJ4 PDB). The oxygen molecule was inserted at the

position of the bound chloride ion as seen in the 2WM2 PDB ﬁle.

but is essential for oxygen activation.^11b,12a,cVery little is known

about the catalytic mechanism after dioxygen binding and how

it diﬀers from iron-containing dioxygenases. However, recently,

a spin-trap experiment^11cfound evidence of a short-lived

superoxide radical that was tentatively assigned to structure R_CT

in Scheme 1.

The HOD reaction mechanism has been subject to a number

of biochemical studies during the past decade.^11b,c,12Site-

directed mutagenesis and kinetic studies have shown that HOD

employs a general base mechanism in which an active site

His₂₅₁/Asp₁₂₆dyad plays an essential role by deprotonation of

the hydroxyl group (O³) of the substrate (step 1 in Scheme 1).

By contrast, classical α/β-hydrolases react via a nucleophilic

reaction mechanism. Figure 1 displays the active site structure

of substrate-bound HOD (PDB code 2WJ4), where we inserted

the oxygen molecule at the chloride position in the 2WM2

PDB ﬁle.^11bThe substrate is located in a polar active site with

several histidine, tryptophan and aspartic acid residues, and

consequently is tightly bound through hydrogen bonding

To gain insight into the chemical details of the catalytic

mechanism of substrate oxygenation by cofactor-free dioxyge-

nases, we performed a combined experimental and computa-

tional study on the O₂dependent reaction steps of the catalytic

mechanism. We particularly focus on the roles played by the

conserved active site residues His₂₅₁and Ser₁₀₁, and the cap

domain residue Trp₁₆₀(Figure 1). The results highlight the

important contributions of His₂₅₁in the catalytic mechanism,

namely, as catalytic base for substrate activation, and

subsequently as the key residue during the O₂reaction. We

have identiﬁed important roles of additional active-site residues

(Ser₁₀₁, His₁₀₂and Trp₁₆₀) that stabilize substrate positioning in

the active site through hydrogen bonding interactions, but also

enable an eﬀective oxygenolytic substrate cleavage via

stabilization of the reaction intermediates through long-range

electrostatic interactions. The computational modeling studies

interactions including those with the aforementioned His₂₅₁

-

Asp₁₂₆dyad. The proton transfer step in the catalytic cycle

initiated by this dyad is not rate limiting for the overall reaction,

B

DOI: 10.1021/jacs.5b03836

J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society

Article

where v_obsis the observed turnover rate constant at certain oxygen

identify details of the reaction mechanism and the critical roles

of amino acid residues in the substrate binding pocket.

Moreover, they further explain how cofactor-free dioxygenases

are able to catalyze an energetically diﬃcult reaction, which is

otherwise not known to happen in nature without the

assistance of a cofactor.

concentration, k_catis the catalytic constant and K_mOis Michaelis

2

constant for O₂.

Transient-State Studies. Single-turnover experiments were

carried out under anoxic conditions using a TgK Scientiﬁc stopped-

ﬂow spectrometer maintained in an anaerobic glovebox at oxygen

concentrations <2 ppm (Belle Technology). Reactants were prepared

inside the glovebox and the enzyme was made anoxic by passing it

through a PD-5 column. To achieve a single-turnover reaction,

enzyme−substrate (ES) complexes were prepared using a small (20

μM substrate, 30 μM HOD) and large (15 μM substrate, 240 μM

HOD) enzyme excess. Oxygen concentrations were obtained by

mixing the desired amount of anaerobic buﬀer and oxygen saturated

buﬀer (prepared as above) in the syringe. The reaction was measured

by following the absorbance decrease at 340 nm upon mixing the ES

complex and oxygen in 0.1 M MTE, 5 mM NaCl and 2 mM EDTA at

20 °C. The oxygenation constant was obtained from eq 2, where, k_obs

is the apparent observed rate constant associated with substrate

oxygenation at any given concentration of oxygen, and k_oxis the

second-order rate constant for oxygenation.

MATERIALS AND METHODS

■

Materials. All chemicals were obtained from Sigma-Aldrich unless

otherwise stated.

DNA Techniques. The 1-H-3-hydroxy-4-oxoquinaldine 2,4

dioxygenase (HOD) cDNA was previously cloned into the pQE30

(Qiagen) vector,^12ccarrying the C69S mutation to prevent oxidative

protein dimerization^12awithout any eﬀect on HOD activity.^12bThe

corresponding pQE30-wtHOD construct was used to transform

Escherichia coli M15 [pREP4] strain (Qiagen).

Site-directed mutagenesis was performed following the QuikChange

Site-Directed Mutagenesis Kit (Stratagene) protocol, using pQE30-

wtHOD as template, and the following primers (direct sequences)

with mutations (underlined) at the corresponding triplets (brackets)

were used as primers: S101A, 5′-CTTCCGGTATCTCAT[GCG]-

CACGGCGGCTGGGTTC-3′; W160A, 5′-GCCTGTTCGACGT-

[CGC]GCTTGACGGGCATGACG-3′; W160F, 5′-CGGCC-

TGTTCGACGTC[GCG]CTTGACGGGCATGACG-3′; H251A, 5′-

GGGCGGGCCGACC[GCC]TTCCCCGCCATCGACG-3′. Muta-

tions were conﬁrmed by sequence analysis.

Expression and Puriﬁcation of HOD Samples. HOD

expression and puriﬁcation was carried out as described previously.^12c

Protein concentrations were determined spectrophotometrically using

an extinction coeﬃcient at 280 nm deduced from the amino-acid

sequence: 64 525 M⁻¹cm⁻¹for wt HOD, S101A and H251A; 59 025

M⁻¹cm⁻¹for W160F/A.

Substrates Analyses. The natural substrate, 3-hydroxy-2-methyl-

4(1H)-quinolone (MHQ) was synthesized by Sigma at 99.6% purity.

The 3-hydroxy-2-butyl-4(1H)-quinolone (BHQ) substrate was

synthesized by AF ChemPharm (Sheﬃeld, UK) at 97% purity. For

determining the molecular extinctions coeﬃcients at pH 8.5 (MHQ

ε₃₃₄= 9800 M⁻¹cm⁻¹, BHQ ε₃₃₆= 9700 M⁻¹cm⁻¹) the substrates

were ﬁrst dissolved in DMSO (200 mM, MHQ and 10 mM, BHQ)

and later diluted in 0.1 M MTE. MTE is a buﬀer solution containing 2-

(N-morpholino)-ethanesulfonic acid, Tris and ethanolamine. The

BHQ solubility in MTE is restricted to 100 μM, while MHQ solubility

limit is 1 mM. MHQ concentration at high absorbance was

determined by ε₃₅₇= 3300 M⁻¹cm⁻¹.

Steady-State Studies. Initial rates of MHQ/BHQ substrate

oxygenation by HOD were measured by following the absorbance

decrease at the corresponding wavelengths (using the above ε values).

To determine the kinetic parameters for the aromatic substrates

(MHQ/BHQ) the reactions were carried out under atmospheric

oxygen concentration at varying substrate concentrations (MHQ, 5−

400 μM and BHQ, 5−100 μM) in 0.1 M MTE buﬀer containing 5

mM NaCl and 2 mM EDTA at 20 °C. Reactions were initiated by

adding a small amount (5−10 μL) of concentrated HOD samples to

reach the desired concentrations: 15 nM (wt HOD), 18 nM (S101A),

40 nM (W160F), 360 nM (W160A) 8 μM (H251A) and 15 μM

(H102A/H251A). Steady-state kinetic constants were obtained by

ﬁtting the initial rates at diﬀerent substrate concentrations to the

Michaelis−Menten equation.

k_obs= k_ox[O₂]

(2)

Computational Studies. The calculations reported here were

done using density functional theory (DFT) methods as benchmarked

and calibrated previously.¹³All calculations use the Gaussian-09

program package,¹⁴with the unrestricted hybrid DFT method

UB3LYP in combination with 6-31G and 6-311++G** basis sets,

designated as BS1 and BS2, respectively.¹⁵Full geometry optimiza-

tions (without constraints for the oxygen and the aromatic substrate)

were performed followed by a frequency calculation at UB3LYP/BS1.

All local minima had real frequencies only, and the transition states

had one imaginary frequency for the correct mode. Energies reported

in this work were obtained with basis set BS2 and contain zero-point

corrections. Gibbs free energies were calculated at 1 atm pressure and

298 K temperature, and contain entropic, zero-point and thermal

corrections to the energy at basis set BS2. In addition, solvent

corrections to the energies were included as obtained from a single

point calculation using the polarized continuum model with dielectric

constant of ε = 5.70 or ε = 78. Geometry optimizations of several

structures using a dielectric constant of ε = 78 gave little changes with

respect to the gas-phase optimized structures and energies to within 1

kcal mol⁻¹.

The oxygenation reaction was studied using models of various

molecular sizes (Scheme 2). The smallest chemical system (Model A;

shown with the purple circles in Scheme 2) only contained substrates

with diﬀerent substituents on the R-position with R = methyl, n-butyl,

H, F, Cl, NH₂and NO₂. This model was used in the studies of the

substituent eﬀect on the electron transfer step in the reaction

mechanism. Hence, we calculated the ionization energy (IE) for the

protonated (SubH) and deprotonated (Sub⁻) forms of these

substrates, by taking the energy diﬀerence between the ionized and

nonionized structures. In addition, we calculated the electron aﬃnity

2

•−

(EA) of molecular oxygen as the diﬀerence in energy between O₂

and ³O₂. We calculated the IE and EA values from the reaction

energies of eqs 3a, 3b, and 4:

SubH → SubH^+•+ e⁻; ΔG_eq3a= IE_SubH

(3a)

−

•

Sub → Sub + e⁻; ΔG_eq3b= IE_Sub

(3b)

To gain further information about HOD catalysis, the initial rates

were also determined in the presence of varying concentrations of

oxygen (35−1300 μM) at saturated MHQ/BHQ concentration in 0.1

M MTE buﬀer at 20 °C. Oxygen concentrations were obtained by

mixing the desired amount of anaerobic buﬀer and oxygen saturated

buﬀer (prepared by bubbling buﬀer with pure oxygen) in a screw-cap

cuvette. This procedure was performed inside an anaerobic glovebox.

The kinetic parameters were obtained by ﬁtting the initial rates to eq 1,

O₂+ e⁻→ O₂^•−; ΔG_eq4= EA_O

(4)

2

On the basis of the coordinates of the anaerobic HOD·MHQ and

the HOD·chloride complex (PDB entries 2WJ4 and 2WM2,

respectively) we created active site models of diﬀerent size: Models

B and C. The smallest model (B) includes the organic substrate, O₂,

the side chains of His₂₅₁, Asp₁₂₆, Ser₁₀₁, and Trp₁₆₀as well as the

peptide backbone of Trp₃₆(total number of atoms: 86). The largest

system (model C) studied here is constituted of model B

v

= k_cat/K [O₂]

obs

mO₂

(1)

supplemented with the imidazole groups of His₃₈, His₁₀₀, and His₁₀₂

,

C

DOI: 10.1021/jacs.5b03836

J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society

Article

Scheme 2. Structures of Model A (Purple Circle), Model B

(Green Circle) and Model C (All Residues Shown) Used in

Our DFT Calculations (See Main Text)

RESULTS

■

In this work we applied kinetics, spectroscopic, and computa-

tional methods to elucidate the mechanism of dioxygenation by

HOD, particularly on the oxygen-dependent steps. The

research is focused on the roles of the active-site residues

His₂₅₁, Ser₁₀₁and Trp₁₆₀by generating site-directed mutants

followed by steady- and transient-state studies. Details of the

reaction mechanism were established from density functional

theory modeling to determine the rate-determining step in the

reaction and the quantum mechanical features that aﬀect the

height of the barrier, and consequently the rate constant.

Steady-State Studies. First we determined the reaction

rate using two diﬀerent substrates, MHQ and BHQ, at ﬁxed

(atmospheric) oxygen concentration. The maximum reaction

rate was achieved at high substrate concentrations (Supporting

Information, Figure S1) and allowed us to determine the rate

constants. Apparent kinetic constants (k_catand k_cat/K_m) were

determined for active site single variants S101A, W160F,

W160A, and H251A and the double variant H102A/H251A for

MHQ and BHQ substrates, which enables us to establish the

eﬀect of the substituent on catalysis.

a

As shown in Table 1, MHQ binding is aﬀected by the S101A

substitution, which leads to a 48 fold lower k_cat/K_mvalue

although both k_catvalues are similar. The Trp₁₆₀replacement by

phenylalanine or alanine has a direct impact on overall catalysis,

as reﬂected on their reduced k_catvalues (4 and 65 fold lower

than wt, respectively) and k_cat/K_m(77 and 5200 fold lower than

wt, respectively). The most dramatic change is observed for the

H251A and H102A/H251A variants showing a k_cat/K_mreduced

by 7 × 10⁴and 4 × 10⁵fold, respectively. These kinetic

parameters and reaction rates are in good agreement with those

reported previously.^11b

In order to ﬁnd out how ring-substitution aﬀects the kinetics

of substrate activation, we attempted to synthesize additional

substrates where the MHQ methyl group was replaced by n-

butyl (BHQ). For wt HOD, the substitution of MHQ for BHQ

does not aﬀect the reaction rate (similar k_catvalues for MHQ

and BHQ). Each variant, however, shows a diﬀerent behavior.

The largest k_cat/K_mdecrease upon replacing the substrate with

BHQ is observed for S101A (16-fold) followed by the W160F

(7 fold) and W160A (2 fold).

a

Wiggly lines identify bonds that were cut from the crystal structure.

Several of the peptide backbone atoms were kept ﬁxed in the crystal

structure coordinates.

as well as three crystal water molecules (total number of atoms: 141).

Model C was also modiﬁed to get the corresponding variants: W160A

(whereby the Trp₁₆₀group was removed), W160F (Trp₁₆₀replaced by

a toluene group) and S101A (Ser₁₀₁replaced by methyl group).

Structure B was investigated with diﬀerent protonation states of His₂₅₁

,

whereby in B₁His₂₅₁is doubly protonated and Asp₁₂₆is deprotonated,

in model B₂His₂₅₁is singly protonated and lacks the Asp₁₂₆residue

and model B₃has a doubly protonated His₂₅₁and no Asp₁₂₆residue.

Transition state searches were started by initially running extensive

geometry scans on the singlet and triplet spin state surfaces between

various local minima, whereby one degree of freedom for the reaction

coordinate (i.e., the MHQ C²−O_tdistance) was ﬁxed and all other

degrees of freedom were fully optimized. The maxima of these scans

were then used as starting points for the full transition state searches.

Previous studies showed that within transition state theory these free

energies of activation can be converted into rate constants that were

found to reproduce experimentally determined rate constants of

oxygen atom transfer reactions to within 3 kcal mol⁻¹.¹⁶

In addition, we studied the eﬀect of oxygen concentration on

HOD catalysis, using the above substrates at saturating

concentration. The substrate concentration was diﬀerent for

a b

,

Table 1. Apparent Steady-State Kinetic Constants for wt HOD and Its Variants

k_cat(s⁻¹

)

K_m(μM)

k_cat/K_m(s⁻¹·mM⁻¹

)

MHQ

BHQ

WT

26

34

7

1

2.1 0.1

12500 600

265 19

S101A

1

127

41

8

2

W160F

W160A

H251A

H251A/H102A

WT

0.1

162 10

0.40 0.01

170 17

32

64 15

2.4 0.3

5.5 0.1 × 10⁻³

2.2 0.2 × 10⁻³

2

0.17 0.01

0.034 0.009

1680 190

33

1

20

101

57

2

5

2

S101A

1.7 0.1

17

23

1

2

W160F

W160A

H251A

1.3 0.1

0.8 0.1 × 10⁻²

1.1 0.1 × 10⁻³

75 10

57

1.0 0.1

7

0.20 0.01

a

Apparent kinetic constants were determined under atmospheric oxygen in 0.1 M MTE buﬀer at pH 8.5 supplemented with 2 mM EDTA and 5 mM

b

NaCl at 20 °C. Values reported are the calculated mean errors. The errors considered were taken larger than the standard deviation between four

replicates and the numerical errors after ﬁtting.

D

DOI: 10.1021/jacs.5b03836

J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society

Article

each sample, depending on the K_mvalue reported above, and

the oxygen concentration was varied in the sealed cuvette. In

this case, we were not able to observe any reaction rate

saturation at high oxygen concentrations (Supporting Informa-

tion, Figure S2) even when we employed the maximum oxygen

dissolved concentration that could be experimentally achieved

at 20 °C (1.3 mM), implying that bimolecular association of

oxygen to HOD contributes to reaction rate limitation. For that

In a second set of experiments, the reaction mixtures were

prepared using a large excess of enzyme (240 μM) compared to

the substrate (15 μM). A single-exponential decay was observed

for all samples tested, and, hence, no distinctive reaction

intermediates were observed. However, in these experiments

the reaction rates were substantially higher as compared to

those observed above (Supporting Information Table S1).

The observed oxygenation rates (k_obs) are linearly dependent

on the oxygen concentration (Figure 2) for both MHQ and

reason, we could only determine the k_cat/K_mOconstants shown

2

in Table 2. These values not only conﬁrm the eﬀect of the

Table 2. Steady- and Transient-State Oxygen Constants for

ab

,

wt HOD and Its Variants

c

d

k_cat/K_mO(mM⁻¹·s⁻¹

)

k_ox(mM⁻¹·s⁻¹

)

2

MHQ

BHQ

WT

51

52

13

2

3

1

102

57

5

3

1

S101A

W160F

W160A

H251A

H251A/H102A

WT

18

2.0 0.1

1.0 0.1

1.7 0.1 × 10⁻²

3.1 0.2 × 10⁻³

2.1 0.1 × 10⁻²

7.3 0.3 × 10⁻³

50

4

116

6

S101A

2.0 0.2

3.8 0.2

W160F

W160A

H251A

2.0 0.1

4.2 0.3

0.13 0.02

1.8 0.1 × 10⁻³

0.38 0.03

2.2 0.1 × 10⁻³

a

Constants were obtained in 0.1 M MTE buﬀer at pH 8.5

b

supplemented with 2 mM EDTA and 5 mM NaCl at 20 °C. Values

reported are the calculated mean error. The errors for transient-state

constants were taken larger than the standard deviation between six

c

replicates and the numerical errors after ﬁtting. Steady-state constants

were determined at saturated MHQ or BHQ concentration and

d

varying O₂concentration in the cuvette. Transient-state constants

were determined by reacting the ES complexes at diﬀerent O₂

concentration using an stopped-ﬂow equipment.

mutation observed, but also reﬂect the above eﬀect of the

aromatic substrate (MHQ vs BHQ) on the reaction rate. In all

cases, the apparent k_cat/K_mvalues (Table 1) are larger than the

Figure 2. Oxygenation rate dependence on oxygen concentration for

wt HOD and its variants. The observed oxygenation rates (k_obs) were

obtained by reacting the anaerobic enzyme-MHQ (A) and enzyme-

BHQ (B) complexes with buﬀer at diﬀerent oxygen concentrations

using a stopped-ﬂow instrument: wt HOD (black), S101A (green),

W160F (blue), W160A (red) and H251A (white). All reactions were

measured in 0.1 M MTE buﬀer pH 8.5, 5 mM NaCl and 2 mM EDTA

at 20 °C. The errors considered in the measured rates were taken as

the standard deviation of six replicates. Values of k_ox(Table 2) were

calculated from the gradients of these plots.

k_cat/K_mOvalues (Table 2, left), which implies that the oxygen-

dependent reaction steps are rate-limiting. Furthermore, the

2

diﬀerence in the k_cat/K_mOvalues observed for the variants

2

(compared to wt HOD) are quite similar to the changes

observed for the apparent k_catunder atmospheric oxygen

concentration, although the absent K_mOvalues compromise

2

detailed comparison between samples.

Transient-State Studies. To gain further insight into the

oxygen dependent steps during HOD catalysis, we measured

the transient absorbance decrease at λ_max= 340 nm upon

mixing anaerobic samples of enzyme−substrate complex with

aerobic buﬀer in a stopped-ﬂow instrument. Two diﬀerent

experimental designs were employed to trap and characterize

short-lived intermediates formed during the reaction and to

ensure the maximum amount of enzyme−substrate (ES)

complex is formed. In the ﬁrst approach, the ES mixtures

contained only a slightly higher amount of enzyme (30 μM)

than substrate (20 μM) to explore the contribution of ES

complex formation during catalysis. Using this strategy, no

distinctive new species could be detected during the oxygen-

ation reaction, which implies that either these species have a

spectroscopic signature parallel to the substrate or their lifetime

is too short to be detected.

BHQ substrates, for this reason only a second-order rate

constant (k_ox) could be determined (Table 2). The oxygen

constants k_cat/K_mO(steady-state) and k_ox(transient-state),

2

shown in Table 2, are of the same order of magnitude.

Consequently, the oxygen-dependent steps are rate-limiting for

overall catalysis.

Theoretical Calculations. Density functional theory

(DFT) methods were employed to gain additional information

about the reaction mechanism and identify the rate-limiting

step in the overall catalysis. Furthermore, the calculations

establish the role of active site residues in the reaction

mechanism and their function to position and stabilize the

substrate. The oxygen transfer reaction was studied for wt

E

DOI: 10.1021/jacs.5b03836

J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society

Article

Figure 3. Reaction energy proﬁle and DFT optimized structures of key intermediates and transition states for wt HOD. The reaction energy proﬁle

gives Gibbs free energies (gas-phase) and solvent corrected values (in parentheses). DFT optimized geometries of ³I₁and ¹TS₃are given in the inset

and include distances in angstroms and the imaginary frequency in the transition state in wavenumbers. Values shown were determined using

UB3LYP/6-311++G** on model C.

HOD as well as the H251A, S101A, W160F and W160A

mutants. Additionally, we studied the eﬀect of the nature of the

substrate on chemical catalysis through the use of ring-

substituted HOD models (HOD^R), whereby the methyl

group at position C²was replaced with another substituent R

= n-Bu, F, and NO₂.

described as an [³O₂−Sub⁻] complex with a spin density of

approximately 1.6 on the O₂moiety and the rest on the

substrate: ρ_O3= 0.13, ρ_C2= 0.10, ρ_O4= 0.05 and ρ_C4= 0.04 (see

Supporting Information Table S10). Therefore, close approach

of O₂onto the MHQ group leads to a minor redistribution of

charges and electronic spins, but retains most of the original

electronic description of the reactants. To further conﬁrm that

the charge-transfer state (right-hand-side of eq 5) is an excited

state, we swapped molecular orbitals and attempted to

converge the wave function of the reactant complex to the

charge-transfer complex, R_CT. However, during the SCF

convergence the wave function converged back to the reactant

conﬁguration with a triplet spin O₂complexed to a closed-shell

singlet substrate anion. Therefore, even with the protein

environment included, the substrate-dioxygen complex is

unable to form spontaneously by electron transfer, and,

consequently any reactivity of O₂will come from dioxygen

and not from the stabilization of a radical pair.

General Catalytic Reaction Mechanism. Our work was

initially focused on the catalytic reaction mechanism for MHQ

breakdown into N-acetyl-anthranilate and carbon monoxide.

Upon substrate binding the His₂₅₁-Asp₁₂₆dyad abstracts the

hydroxyl proton from O³, for which we previously calculated a

reaction energy ΔG = −3.9 kcal mol⁻¹.^12cIt has been suggested

that binding of O₂into the substrate binding pocket leads to an

electron transfer from substrate to O₂and the formation of a

superoxo radical anion, as shown in eq 5. We calculated an

electron aﬃnity (EA) of ³O₂of 0.60 eV in the gas phase, which

is within 4 kcal mol⁻¹from the value reported in the

literature.¹⁷With solvent corrections included an electron

transfer free energy ΔG_eq5= 11.2 kcal mol⁻¹(ε = 5.7), whereas

a ΔG_eq5= 8.0 kcal mol⁻¹is found when the dielectric constant

of water is used. This implies that the electron transfer is an

endothermic process and unlikely to happen spontaneously in

the protein or in solvent.

Subsequently, we calculated the mechanism of dioxygen

transfer to MHQ as described in Scheme 1 using models B and

C with DFT methods. The obtained results starting from the

3

triplet spin reactant complex R are schematically depicted in

Figure 3. The reaction starts with attack of dioxygen on carbon

C²of MHQ, via a barrier ³TS₁, to form the superoxide complex

³I₁. The latter is a very shallow minimum below ³TS₁. However,

detailed geometry scans indicated ³TS₁to be close in energy to

•−

−

•

2

³O₂+ Sub → Sub + O₂

(5)

To investigate how the protein and, in particular, the local

environment aﬀects the electron transfer in reaction 5, we set

up a set of models with increasing size and protein description.

Our simplest models take substrate and oxygen in isolation

(model A), but in subsequent models the local protein

environment is taken into consideration using models B and

C (Scheme 2). Thus, in both reactant models B and C the

calculations converge to an electronic ground state that is best

3

³I₁(see Supporting Information Figure S4). Electronically, I₁

can be described as a biradical with radical character on the

terminal oxygen atom of the peroxo group (ρ = 0.69), while

there is spin of 0.32 on the distal oxygen atom and the rest on

the substrate: ρ_O3= 0.30, ρ_C3= 0.14, ρ_O4= 0.21 and ρ_C4= 0.24.

3

Upon formation of I₁from R, therefore, a minor increase of

radical character on O³is observed that reduces its charge and

F

DOI: 10.1021/jacs.5b03836

J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society

Article

weakens the hydrogen bond between O³and the proton of

The Function of His₂₅₁. In previous work, we identiﬁed an

important role of the His₂₅₁-Asp₁₂₆dyad in regulation the

protonation state of the substrate and thereby initiating the

catalytic cycle.¹²To explore the role of His₂₅₁during the

oxygen-dependent reaction steps, we did a further set of

calculations on model B, whereby His₂₅₁was chosen in diﬀerent

protonation states: (i) Model B₁contains the His₂₅₁-Asp₁₂₆

groups and has the Nδ1 atom of His₂₅₁protonated (proton

shared with Asp₁₂₆); (ii) Model B₂has the Nδ1 atom of His₂₅₁

deprotonated and excludes Asp₁₂₆; (iii) Model B₃has the Nδ1

atom of His₂₅₁protonated and does not contain the Asp₁₂₆

residue. Figure 4 displays the calculated reaction mechanism of

oxygen insertion into MHQ substrate by models B₁, B₂and B₃

as well as the full model C.

3

His₂₅₁(Nε2), which elongates from 1.64 Å in R to 1.72 Å in

3

³TS₁. The other geometrical change from R to I₁relates to

elongation of the bonds associated with carbon C², as it is

rehybridized from sp²to sp³. It may very well be that the

experimentally trapped radical^11cis the ³I₁state described here.

The formation of ³I₁encounters the highest barrier along the

reaction mechanism (ΔG^‡= 17.4 kcal mol⁻¹in a solvent

model), and hence is the rate limiting step for wt HOD

catalysis. As shown in Figure 3 the substrate is located in a tight

substrate binding pocket and is stabilized via several hydrogen

bonding interactions with active-site residues, such as His₂₅₁

,

Ser₁₀₁and His₁₀₂as well as several water molecules. The ring-

closure step on the triplet spin state surface to form ³I₂via ³TS₂

is energetically high in energy (>50 kcal mol⁻¹, see Supporting

Information Figure S4), and therefore, the mechanism will not

proceed further on the triplet spin state surface. However, a

lower energy singlet spin pathway exists and a spin crossing

3

1

from I₁to I₁continues the mechanism on the singlet spin

state surface. Although we made attempts to calculate the

3

1

minimum energy crossing point from I₁to I₁, we were not

able to locate it precisely, but predict it to be around +10 kcal

mol⁻¹above ³R (Figure S4). We, therefore, expect a fast

equilibration and conversion from triplet to singlet in or around

³I₁. Structure ¹I₁has a strongly elongated O−O distance of 1.56

Å, but only minor geometric changes in the substrate moiety

3

are found with respect to I₁.

In the next reaction step, the terminal oxygen atom (O_t) of

the peroxy group attacks the keto carbon atom C⁴with a small

barrier (¹TS₂) of 1.9 kcal mol⁻¹in solvent to form the bicyclic

ring-structure, ¹I₂. The transition state (¹TS₂) has a small

imaginary frequency of i67.5 cm⁻¹for the C⁴−O_tbond

formation, which implicates that the energetic barrier is broad.

At the same time as the C−O bond formation there is also a

motion for the C³−O³group that bends it out of the plane of

the substrate group due to changes in hybridization of C³from

sp²to sp³and will eventually leave as a CO molecule. In ¹I₂the

dioxygen bond has weakened to 1.61 Å, while the C⁴−O⁴

distance has grown to 1.36 Å. In this endoperoxide

intermediate the dioxygen bond is also weakened similar to

what is seen in nonheme iron dioxygenases¹⁸and breaks to

1

form products P via transition state TS₃.

Figure 4. Free energy proﬁles of oxygen inserting into MHQ as

aﬀected by a varying protonation state on His₂₅₁. Values shown were

determined using UB3LYP/6-311++G** with solvent corrections

included. At the bottom we highlight the structural diﬀerences of the

catalytic triad for models B₁, B₂and B₃.

The last reaction step, namely release of CO and substrate

breakdown to the corresponding product, takes place through a

small energetic barrier (5 kcal mol⁻¹) from I₂. The imaginary

1

mode for this transition state (¹TS₃) represents the

simultaneous dioxygen bond cleavage (O−O stretch) and

C³−C⁴bond breaking, although the C²−C³distance stays

1

virtually intact at a similar value as in I₂. As such, the CO

In general, model C, B₁and B₂give the same overall reaction

mechanism with similar free energy values of local minima and

transition states. The only noteworthy changes are a slight

1

release from complex I₂will be accomplished by a stepwise

bond breaking process of the initial C³−C⁴bond followed by

the C²−C³bond. However, no local minimum was found for

this process and indeed a geometry scan showed a gradual

decrease in energy for the CO release. Structurally (inset in

Figure 3), the O−O bond has broken (distance of 2.08 Å) but

3

increase of I₁(by just over 4 kcal mol⁻¹) and a somewhat

1

larger increase of TS₃, which becomes competitive with the

initial barrier. Therefore, second sphere amino acids, such as

His₃₈, His₁₀₀and His₁₀₂as well as the additional solvent water

molecules, have a stabilizing eﬀect on the overall structure.

The major diﬀerence, however, is obtained when model B₃is

selected, and the most plausible reason is because this model is

charge neutral whereas the other models have an overall charge

of −1. These additional Coulombic interactions are expected to

give major diﬀerences to the energetics. This raises all local

minima and transition states dramatically, and, hence, the

protonation state of the His₂₅₁residue is crucial for a correct

also the C³−C⁴bond is elongated strongly (1.73 Å). Both I₁

3

1

and TS₃are placed in a similar orientation relative to most

active site residues, except for His₂₅₁which, in the ¹TS₃bridges

O³and O⁴at almost equal distances (2.1 Å).

To further establish the roles of the substrate and active site

residues on the reaction mechanism and catalysis, we created

computational mutants and substrate-analogues and repeated

the calculations for the catalytic cycle.

G

DOI: 10.1021/jacs.5b03836

J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society

Article

description of the enzyme reaction mechanism. During the

reaction, the chemical structures of local minima and transition

states are similar, which suggests that the energy diﬀerences

between model B₁, B₂and B₃are related with the His₂₅₁

protonation for each system during the reaction. All together

these results indicate that His₂₅₁not only acts as a catalytic base,

but, furthermore, it acts as catalytic conjugated acid for eﬀective

oxygen insertion into MHQ.

The Function of Trp₁₆₀and Ser₁₀₁. In order to ﬁnd out

what the function of active site residues Trp₁₆₀and Ser₁₀₁is, we

created computational mutants of models C for the following

variants: W160A (the Trp₁₆₀description was removed), W160F

(indole group representing Trp₁₆₀was replaced by toluene) and

S101A (hydroxyl was replaced by methyl). For each of these

variants the full mechanism of MHQ oxygenation and

breakdown was studied computationally.

eﬀective in substrate breakdown in comparison with wt HOD.

Thus, the rate-determining step in the reaction mechanism for

wt and W160F is the ﬁrst barrier (³TS₁) and attack of ³O₂onto

the substrate. By contrast, the active site mutations W160A and

1

S101A considerably destabilize the CO release from I₂and

¹TS₃is raised in energy. In particular, the hydrogen bond of

Ser₁₀₁to O⁴in wt leads to a charge donation toward the

substrate (Q_O4= −0.68), which drops to Q_O4= −0.61 in

S101A. At the same time the charge on O³is more negative in

S101A as compared to wt (Q_O3= −0.32/−0.38 in wt/S101A)

and this makes the release of CO more diﬃcult and hence

1

raises TS₃in energy.

The comparison of structures ³I₁and ¹TS₃(Figure 5B) gives

further information on the changes in relative energies along

the mechanism. The structure of ³I₁is similar for wt HOD and

W160F systems, whereas geometric changes are encountered

for S101A and W160A mutants. In the calculations for the

W160A system, due to removal of a bulky tryptophan group,

the substrate reorients and the Ser₁₀₁−O⁴hydrogen bond is

lost. In summary, the active site mutations reveal the

importance of Ser₁₀₁and Trp₁₆₀for positioning and

constraining the substrate into a speciﬁc orientation.

Furthermore, the Ser₁₀₁−O⁴hydrogen bond donates negative

charge to O⁴and, thereby, pulls CO away from the complex.

Figure 5 shows the free energy proﬁles for model C as

calculated for wt, W160A, W160F and S101A. Active site

1

In structure TS₃, the substrate positioning is shifted for

S101A and W160A systems, where the distances between O⁴

and Ser₁₀₁/His₁₀₂/His₂₅₁have changed. As above, the Trp₁₆₀

and Ser₁₀₁residues play an important role in the reaction

mechanism by stabilizing the reaction intermediate. Addition-

ally, the His₂₅₁establishes a new interaction with O⁴, and

together with His₁₀₂enables the intermediate to breakdown

into the corresponding products.

The calculated potential energy proﬁles of wt and mutants

compare well with the experimental rate constants shown in

Table 1. Thus, in wt the rate-determining step (via ³TS₁) has a

free energy of activation of 17.4 kcal mol⁻¹, whereas it is 22.2

kcal mol⁻¹for W160F. By contrast, for S101A and W160A, the

1

rate-determining step becomes TS₃with barriers of 21.7 and

29.5 kcal mol⁻¹, respectively. As such, the calculations predict

that the rate-determining steps will increase by 4.3, 4.8, and

12.1 kcal mol⁻¹for S101A, W160F and W160A, respectively.

Using transition state theory, the experimental rates from Table

1 implicate increase of the free energy of activation for these

mutants by 2.3, 2.6, and 5.1 kcal mol⁻¹, respectively. The

calculations, therefore, reproduce the experimental trends well

and predict similar rate decreases.

Eﬀect of Substrate on the Reaction Steps. In a ﬁnal set

of calculations, we decided to explore the eﬀects of electron

donating/electron withdrawing groups on the N-heteroaro-

matic substrate, whereby the methyl group at C²was replaced

by either n-butyl, ﬂuoro or nitro groups. The reaction

mechanism of substrate dioxygenation was recalculated for

model C with these three alternative substrates HOD^X, X = n-

butyl, F or NO₂. All reactions follow the same mechanism as in

the previous sections, but the relative energies of the various

structures ﬂuctuate dramatically, Table 3. Part of this is due to

electrostatic interactions and partially due to changes in

electron donating properties of the substrate.

Figure 5. Reaction energy proﬁles and DFT structures for wt HOD

and its variants. (A) Reaction energy proﬁles for: wt HOD (black),

3

S101A (green), W160F (blue) and W160A (red). (B) I₁(left) and

¹TS₃(right) optimized geometries (W160F system) including

distances between the intermediates (green), active site residues and

water (CPK colors). Distances compiled for all systems using same

colors as in (A). Optimized geometries and energy values were

determined using BS1 and BS2, respectively.

mutants give diﬀerences in chemical reaction and particularly

regarding the rate-determining step. The energy proﬁles for the

3

W160F/W160A systems give elevated free energies for TS₁

and ¹TS₃of 4.8/5.4 kcal mol⁻¹and 5.0/14.8 kcal mol⁻¹,

respectively, as compared to wt HOD. In the case of the S101A

mutant, the peroxo bound complexes (³I₁and ¹I₁) are stabilized

in energy, while ¹TS₃is raised by 7.0 kcal mol⁻¹with respect to

wt HOD. Consequently, the S101A system will be more

eﬃcient during the C−O bond formation, but much less

Table 3 gives relative free energies of key intermediates along

the reaction mechanism for the four calculated substrates. As

can be seen, the rate-determining step is via ³TS₁for wt as well

as HOD^NO, whereas the release of CO becomes rate

2

determining for HOD^nBuand HOD^F. Note that in all cases

H

DOI: 10.1021/jacs.5b03836

J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society

Article

applications, since they do not require cofactors/coenzymes,

even if their restricted substrate speciﬁcity limits their

applicability.²⁰In particular, their catalytic mechanism is

chemically intriguing and in this work, a combined

experimental and computational study, is presented on the

reaction mechanism of a cofactor independent dioxygenase. In

addition, a comparison will be made between cofactor-

independent/dependent dioxygenases.

Kinetic Constants. From the comparison of the steady-

state kinetic constants k_cat/K_mvalues for the aromatic substrates

and O₂, it is observed that the oxygen dependent steps are rate

limiting the overall catalysis in agreement with former studies^12a

and other dioxygenases.²¹In previous studies using diﬀerent

approaches, whereby both substrates concentrations (MHQ

and O₂) were simultaneously varied, a ternary complex

Table 3. Solvent Corrected Gibbs Energy Values for Model

a

C HOD^R

R

³TS₁

¹I₁

¹TS₃

¹P

Me

n-Bu

F

17.4

26.8

11.1

26.9

9.3

14.7

33.6

18.0

25.2

−100.7

−103.1

−92.9

16.6

1.1

NO₂

16.0

−87.0

a

Energy ΔG(solv) values for model C, HOD^Rwhere R is Me, n-Bu, F,

NO₂, were calculated using UB3LYP and BS2.

3

1

the energy diﬀerence between TS₁and I₁stays virtually the

same, so that substrate substitution does not aﬀect the spin

state change from triplet to singlet and the spin state ordering,

but only aﬀects substrate binding to the active site pocket.

Addition of electron withdrawing groups to the HOD

mechanism and a K_mOvalue ∼900 100 μM were determined

1

substrate strongly aﬀects the charge of carbon C²in TS₃,

2

for wt HOD.^12aThis K_mOvalue is similar to that reported for

whereby Q_C2= 0.29 for HOD^Me, but values of 0.57, 0.45, and

2

human²²and bacterial²³dioxygenases. However, in all cases

0.30 are found for HOD^F, HOD^NOand HOD^nBu, respectively.

2

In particular, the latter substrate (BHQ) encounters strongly

electrostatic interactions with protein residues in the substrate

binding pocket and gives elevated energies and transition state

energies along the full reaction mechanism as compared to the

natural substrate. This is in agreement with the drop in rate

constant reported in Tables 1 and 2.

mentioned, the K_mOvalue was close to the maximum

2

concentration of O₂that could be experimentally achieved,

hence these K_mOvalues are likely subject to large experimental

2

errors. In any case, the K_mvalues for the aromatic substrate

were 2 orders of magnitude lower than the values for O₂, clearly

indicating that the ternary complex formation is limited by the

O₂availability. Furthermore, the O₂kinetic constants obtained

Finally, Figure 6 displays a Hammett plot for the free energy

19

1

for TS₃with respect to the Hammett parameter σ_p. A linear

under multiple-turnover (k_cat/K_mO) and single-turnover (k_ox)

2

conditions are similar, conﬁrming again that the O₂dependent

steps are rate-limiting in the overall catalysis and in addition

suggest that other catalytic steps may have minor eﬀect during

the overall turnover, in agreement with previous studies.¹²As

discussed above, the observed rate values (k_obs) under transient-

state conditions are linearly dependent on O₂concentration.

Similar tendencies were observed in ﬂavin-dependent mon-

oxygenases²³and nonheme iron-dependent dioxygenases.^6a,8b

Surprisingly, the second-order rate constants calculated for the

above enzymes are similar (∼10⁵M⁻¹s⁻¹), which highlights the

eﬃcient oxygenation chemistry performed by HOD even in the

absence of any cofactor.

Using the stopped-ﬂow technique we could successfully

monitor a single-turnover reaction by mixing an anaerobically

preformed ES complex with O₂. The absorbance decrease

observed was assigned to the substrate oxygenolytic cleavage.

This decay perfectly ﬁtted to a single-exponential decay,

indicating that none of the reaction intermediates were

detected, the reason might be their transient lifetime or that

their absorbance spectra is not substantially diﬀerent (as

suggested from the measurements using the photodiode array

detector). By contrast, the reaction intermediates have been

successfully characterized and detected in nonheme iron

dioxygenases, where these intermediates have distinguishable

kinetic and spectroscopic features,^6b,24or were trapped in

crystal structures.⁸Hence, we performed several computational

calculations to further investigate the reaction intermediates

and rate limiting steps.

1

Figure 6. Hammett plot of the free energy for TS₃as a function of

Hammett parameter for diﬀerent substrates HOD^R, R = CH₃, F and

NO₂.

correlation is observed for substitution of the methyl group of

MHQ by either F or NO₂. The correlation implies that the

barriers TS₃are aﬀected by the nature of the substituent on

1

the C²position. Moreover, the linear correlation implicates that

substitution of methyl by ﬂuoro or nitro incurs no stereo-

chemical interactions with the protein and only aﬀects the

reaction through electron-donation. The substrate binding

pocket, therefore, can accommodate substrates with the methyl

group of MHQ replaced by ﬂuoro or nitro, but a substitution

with a much larger group, such as, n-butyl distorts the protein

substrate interactions considerably.

Catalytic Mechanism. We suggest a mechanism that starts

with an initial deprotonation of the substrate O³position and

DISCUSSION

■

3

binding of O₂at the active site pocket. The mechanism after

The majority of ring-breakdown oxygenases require an organic

or metallic cofactor for their catalytic activity. However, several

dioxygenases have been found to perform similar reactions in a

cofactor-independent fashion. These class of dioxygenases and

monooxygenases are interesting enzymes for biotechnological

O₂binding is summarized in Scheme 3. This work gives no

evidence of an initial radical pair formation (long-range electron

3

transfer between O₂and MHQ) but instead the O₂attacks

MHQ C²atom to form a C−O bond as a peroxo in a triplet

I

DOI: 10.1021/jacs.5b03836

J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society

Article

diagrams were used previously to rationalize the rate constant

of hydrogen atom abstraction and double bond epoxidation

reactions by metal-oxo complexes, and are particularly useful to

understand the electron transfer migrations during reaction

mechanisms.²⁵The VB curve crossing diagram starts from our

reactant complex (³R), which contains deprotonated MHQ and

triplet oxygen. The substrate (Sub) is in a singlet spin state and

the π-bond that is being activated in the reaction mechanism is

highlighted with the two dots along the bond. The product VB

state for this reaction step is the singlet spin peroxo complex ¹I₁

Scheme 3. Proposed Catalytic Mechanism of HOD on the

Basis of Computational, Kinetics and Previous Structural

Studies

1

(Figure 7, right) and has a wave function Ψ_Ps. In comparison

to the reactants structure, one electron from C²forms a bond

with one of the unpaired electrons of O₂, and at the same time,

the π electron from C³forms a π-bond with O³. Because of this

reorganization one electron has transferred from O³to the

terminal oxygen atom of the peroxo group. These VB

descriptions can now be used to estimate the barrier height

1

for the direct conversion of reactants to I₁through curve

crossings. Thus, the product wave function connects to an

1

excited state in the reactant geometry (wave function Ψ_CT)

and the reactant wave function connects to an excited state in

the product geometry (wave function Ψ_P*). The point where

these two curves cross leads to an avoided crossing and a

transition state for the reaction. Hence, the crossing of the blue

and green lines in Figure 7 leads to a transition state for

3

1

spin state (³I₁, peroxy-biradical). Subsequently, an internal

electron transfer leads to a closed-shell singlet (¹I₁) peroxide

intermediate. The terminal oxygen atom attacks carbon C⁴to

form an endoperoxide intermediate (¹I₂) that releases CO and

leads to products. The detailed reasoning for this mechanism is

discussed below.

conversion of R into I₁. However, in our diagram there is

another, lower-lying VB curve (purple line) that bisects the

3

1

curves between R and I₁and generates a local minimum for

³I₁and leads to a stepwise process.

The height of the barrier (ΔE^‡) can be estimated from the

crossing point of the two curves and was shown to be

proportional to a fraction (f) of the excitation energy (G) in the

reactant geometry from the ground state to the product state

minus a correction for the resonance energy (B) for the change

in geometry from reactants to transition state.²⁶A careful

Valence Bond Rationalization of Mechanism. In order

to fully understand the details of the initial step in the reaction

3

1

mechanism, i.e., the steps from R → I₁→ I₁, we set up a

valence bond (VB) curve crossing diagram (Figure 7). Similar

Figure 7. Valence bond curve crossing diagram for the formation of a singlet spin peroxo complex from reactants. Dots represent relevant valence

electrons and lines between dots are bonds.

J

DOI: 10.1021/jacs.5b03836

J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society

Article

analysis of the VB structures for the reactants (³Ψ_R) and the

excited state products that connect to ³I₁, i.e., ³Ψ_R*, reveals the

intrinsic chemical properties that determine the barrier height.

Thus, the transition from Ψ_Rto Ψ_R* represents the breaking

of the π-bond of the C²−C³bond (E_π) as well as the pairing of

the two radicals from C²and one of the oxygens of O₂(E_OC),

eq 6.

the later reaction with oxygen? To address this question, we

compared the ionization potentials of the protonated and

deprotonated substrates and found that those values are

considerably higher in energy for the protonated substrates

than for deprotonated ones. Thus, the ionization potential of

neutral MHQ in solvent is calculated to be 5.94 eV, whereas

that for the deprotonated structure is only 3.67 eV under the

same conditions. Therefore, the deprotonation step in the

catalytic cycle is essential for weakening the π-system of the

substrate, which then enables subsequent oxygen addition.

In addition, when we compared the reaction energies in the

presence of diﬀerent protonation states of His₂₅₁(Model B) we

observed higher energies and activation barriers for the neutral

system (Model B₃) than those for a negatively charged system

(Models B₁and B₂). The eﬀect is the strongest for the ﬁrst

reaction barrier (³TS₁) for C−O bond formation, while little

changes are observed in the later parts of the mechanism (e.g.,

3

G_triplet= E_π+ E_OC

(6)

3

1

The alternative direct pathway from R to I₁will incur a

much larger barrier due to higher energy excitation energy

3

1

G_singletfrom Ψ_Rto Ψ_CT. A comparison of the VB structures

3

1

for the wave functions Ψ_Rand Ψ_CTgives excitation energy

(G_singlet) for the ³O₂electron aﬃnity and the substrate

ionization potential. In MHQ, G_singletis higher in energy than

G

_triplet, and, consequently the reaction will be stepwise with an

initial C−O bond formation to form ³I₁followed by an internal

for ¹TS₃). Our results implicate multiple roles for His₂₅₁

,

1

electron transfer to form I₁. Everything together suggests that

namely as catalytic base to drive substrate deprotonation, but

also to polarize the C³−O³bond which activates the substrate

C²−C³π-bond and triggers the electron transfer to O₂.

the spin forbidden reaction between ³O₂and ¹Sub⁻is bypassed

through the approach of both substrates in the active site,

which only leads to an electron transfer from ¹Sub⁻to ³O₂after

the peroxy-biradical intermediate is formed. The latter seems to

be the actual species formed in this reaction step instead of a

caged radical pair (superoxide and substrate radical) as

proposed previously.^11c

When the active site is mainly neutral, as in several

ﬂavooxidases, it has been shown that either a double-

protonated His, positively charged Lys residue or substrates

are essential during the initial electron transfer from reduced

ﬂavins to O₂.²⁸In these enzymes, a neutral environment around

the ﬂavin cofactor helps to control the O₂electrostatics by

creating speciﬁc polarization sites around the ﬂavin C^4aand N⁵

atoms. On the other hand, in HOD, several hydrogen bonding

interactions between the aromatic substrate and polar residues

(N¹−Trp₃₆, O³−His₂₅₁and O⁴−Ser₁₀₁) are indeed modulating

the substrate electrostatics and hence O₂reactivity, in which the

O³−His₂₅₁bond plays a major contribution.

Substituted Substrates. On the basis of Figure 7, it can be

anticipated that changing the substituent at C²from methyl to

an electron-withdrawing or electron-donating group should

aﬀect the ionization potential of the substrate and consequently

the π-energy E_π. Indeed, the calculated ionization energies of

C²-substituted substrates follow a linear correlation with

Hammett σ_p-values (Supporting Information, Figure S3). In

contrast, the fact that DFT calculations on HOD^X, X = CH₃, n-

In this study, we investigated the role of the conserved

residue Ser₁₀₁and the active site residue Trp₁₆₀. The rate

constants for the S101A variant were similar to those observed

for wt HOD, which suggests that the residue plays a minor role

during catalysis. However, the increase in the K_mvalue for

MHQ (60 fold) and the reactivity decrease against BHQ (30

fold) suggest that this residue might be involved in the

positioning/stabilization of the reaction intermediates. When

we compare diﬀerent model C structures, it is found that when

Ser₁₀₁undergoes hydrogen bonding interactions to O⁴(as in

WT and W160F models) the whole reaction proﬁle is low in

energy. This hydrogen bond seems to facilitate the endoper-

3

C₄H₉, F and NO₂, found no linear correlation for barrier TS₁

with the Hammett σ_p-value, implicates that both E_πand E_COare

inﬂuenced by the substituent on C². As reported above,

addition of an n-butyl substituent (BHQ) had neither eﬀect on

the charge nor the electron aﬃnity of the substrate. Instead, we

observed strong electrostatic interactions between BHQ and

the active site residues which gave elevated energies along the

whole reaction proﬁle compared to the natural substrate. The

latter computational observations can explain the shift in the

rate constants observed here when the BHQ substrate was

assayed, in agreement with the literature.^11c

1

Role of His₂₅₁, Ser₁₀₁and Trp₁₆₀. To gain further insight

into the reaction mechanism, we studied the role of several

oxide breakdown (larger value of TS₃in the S101A system)

instead of contributing during the initial reaction steps.

Noticeably, in the S101A system the negative charge in the

intermediates, i.e., in ³I₁and ¹I₂, is stabilized via interactions of

the O⁴atom with an H₂O molecule, which suggests that this

new interaction can partially mimic the role of Ser₁₀₁and

thereby explains the similarity of the rate constants of wt HOD

and S101A. In the α/β fold hydrolases, the role of the catalytic

triad serine residue diﬀers, i.e., in α/β-fold lipases and serine-

proteases this residue is performing a nucleophilic attack on the

substrate, and has a central role during the catalysis as shown in

mutagenesis studies.²⁹In other cases, as in MhpC and BphD

hydrolases, the serine residue either stabilizes the oxyanion

intermediate via hydrogen bond or contributes to its

protonation.³⁰Importantly, the structure comparison between

HOD (PDB 2WJ4) and MhpC (PDB 1U2E) reveals that the

distance between His₂₅₁(Nε) and Ser₁₀₁(Oγ) is similar (3.39

and 3.26 Å, respectively) while this distance in typical serine

active site residues, including the conserved residues His₂₅₁and

12

Ser₁₀₁

,

which correspond to the histidine/nucleophile

residues²⁷of the catalytic triad in the α/β fold superfamily,

and the active site residue Trp₁₆₀. The kinetic constants

obtained for the H251A variant were 4 orders of magnitude

lower than wt HOD, in line with previous experimental

observations, which concluded that His₂₅₁is crucial for catalysis

acting as catalytic base.^11b,15cAs the reaction in H251A is

aﬀected at the ﬁrst catalytic step (substrate deprotonation) we

were expecting to shift the rate-limiting reaction step and

hence, observe a saturation of the reaction rate at high O₂

concentrations. Instead, the apparent k_cat/K_mvalue is ten times

higher than the corresponding k_cat/K_mOvalue, which strongly

2

suggests that the O₂-dependent steps are rate limiting the

reaction rate in this variant similar to what is observed for wt

HOD. Then, why does this initial step have a major impact on

K

DOI: 10.1021/jacs.5b03836

J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society

Article

proteases is much shorter (2.7 Å).³¹This structural diﬀerence

suggests that weakening the interaction between these two

conserved residues, the serine is no longer able to act as

nucleophile. Hence, this loss of function in the serine is

promoting its role hydrogen bonding the reaction intermedi-

ates.

Additionally, we generate two diﬀerent variants (W160F/

W160A) to study the Trp₁₆₀involvement in catalysis. The

phenylalanine substitution only partially aﬀects the overall

activity while the substitution by a less bulky residue (alanine)

decreases the activity dramatically. The corresponding energy

proﬁles for the W160F/W160A models were shifted with

3

respect to wt HOD, and show increased barriers for TS₁and

¹TS₃, and thus aﬀect the O₂activation mechanism. The crystal

structure coordinates locate the Trp₁₆₀residue in the cap

domain and contributes to the shape of the active site

pocket.^11bThis, together with the fact that Trp₁₆₀does not

directly interact with the substrate, suggest that its function

might be restricting the position of the substrate favoring its

orientation toward catalytically relevant residues (His₂₅₁). As

HOD and QDO perform chemically similar reactions, it is

interesting to note that another tryptophan residue (Trp₁₅₃

)

occupies a similar position and orientation in QDO, but little is

known of its function. Noteworthy, additional active site

residues, such as His₁₀₂in HOD and His₉₆in QDO, have been

found to contribute to the overall catalysis through mutagenesis

studies reported previously.^11b,12bAbsence of His₁₀₂(as in

Models B) appears to increase the barrier for endoperoxide

Figure 8. Valence bond curve crossing diagram for the formation of a

quintet spin peroxo complex from reactants. Dots represents relevant

valence electrons and lines between dots are bonds.

1

breakdown, i.e., TS₃, by 8 kcal mol⁻¹.

Why Does HOD Not Need a Cofactor? Finally, we

answer the question, how the reaction mechanism could be

inﬂuenced by the use of a cofactor in HOD, e.g., iron? In

particular, we hypothesize how the catalytic mechanism of

HOD would be aﬀected by a bound cofactor. In order to

understand the details of the oxygen activation in HOD and to

clarify why an iron cofactor is not needed, we set up another

VB curve crossing diagram (Figure 8) for a hypothetical system

that uses an iron(III)-superoxo cofactor. The above reﬂects the

active species as appeared in intradiol dioxygenases, which

incidentally catalyze a very similar reaction as to that described

in Scheme 1. Note that iron(II)-superoxo will give rise to the

same conclusions. Thus, the iron(III)-superoxo complex (⁵R′)

will covalently bind the dioxygen as an end-on superoxo

structure and most likely will bind the substrate as a bidentate

ligand via atoms O³and O⁴. Calculations of iron(III)-superoxo

complexes of nonheme iron dioxygenases resulted in a high-

spin ground state with three unpaired electrons on the metal

and one on the superoxo group.³²Typically calculations on

nonheme iron(III)-superoxo complexes give chemical systems

with considerably lesser radical character (and more anionic

Furthermore, the nitrogen atom (N¹) in the MHQ ring will

increase the electron density of the neighboring carbon atom

(C²) and make it more susceptible for C−O bond formation

with dioxygen. As such, the diﬀerence in substrate between

HOD and catechol dioxygenases will require lesser energy for

substrate dioxygenation, which can even be directly done with

³O₂.

As above, the height of the barrier will be a fraction of the

excitation energy (G_FeOO) for the transition from the ground

state to the excited state representing the product wave

function. An analysis of the VB structures for the ground and

excited states implicates that also G_FeOOis connected to the

breaking of the π-bond between C²and C³(E_π) as well as the

bond formation energy for the C−O interaction (E_CO), eq 7.

G_FeOO= E_π+ E_OC

(7)

Binding of MHQ to a cofactor, such as iron(III) is only going

to marginally aﬀect the π-bonding energy (E_π). Moreover, little

changes in C−O bond formation energy are expected either as

the step does not involve an electron transfer or any charge

reorganization. Therefore, binding of substrate to a cofactor is

not going to increase the reaction rate dramatically from an

electronic perspective. Furthermore, binding of the substrate as

a bidentate ligand to the iron will make the distance between

the terminal oxygen atom of the iron(III)-superoxo and carbon

atom C²large and will lead to an increase in barrier heights

along the chemical reaction due to electrostatic interactions.

Consequently, a cofactor is not beneﬁcial for the HOD reaction

process and therefore it has developed without it. Thus, the

mechanism of HOD is diﬀerent from, for instance, catechol

intradiol dioxygenases, where the catechol-bound iron(III)

complex binds molecular oxygen to form an iron(III)-

superoxo.^8,33The latter attacks the carbonyl carbon of the

3

charge) on the terminal oxygen atom than O₂, and as a result

the terminal oxygen atom of the iron(III)-superoxo group will

become more electrophilic. However, the terminal oxygen atom

is located at a relatively large distance from carbon atom C²

(approximately 4.2 Å), consequently, in the case of HOD, the

substrate activation will incur a large geometrical change and

therefore introducing a cofactor will not lower the barriers with

respect to those found without cofactor.

Thus, in HOD the substrate N-heteroaromatic ring is

chemically diﬀerent to the aromatic ring present in the

substrates of catechol dioxygenases. Then, breaking of the π-

bond in HOD will require lesser energy than the breaking of

the π-bond of the aromatic ring in catechol dioxygenases.

L

DOI: 10.1021/jacs.5b03836

J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society

Article

N. Chem. Rev. 2005, 105, 2227−2252. (g) Kryatov, S. V.; Rybak-

Akimova, E. V.; Schindler, S. Chem. Rev. 2005, 105, 2175−2226.

(h) Kovaleva, E. G.; Lipscomb, J. D. Nat. Chem. Biol. 2008, 4, 186−

193. (i) He, P.; Moran, G. R. Curr. Opin. Chem. Biol. 2009, 13, 443−

450. (j) Ray, K.; Pfaff, F. F.; Wang, B.; Nam, W. J. Am. Chem. Soc.

2014, 136, 13942−13958. (k) Solomon, E. I.; Heppner, D. E.;

Johnston, E. M.; Ginsbach, J. W.; Cirera, J.; Qayyum, M.; Kieber-

Emmons, M. T.; Kjaergaard, C. H.; Hadt, R. G.; Tian, L. Chem. Rev.

2014, 114, 3659−3853. (l) Shaik, S.; Hirao, H.; Kumar, D. Nat. Prod.

Rep. 2007, 24, 533−552.

substrate to form a ring-structure analogous to I₁reported

above. This step, however, is not rate determining and the

following dioxygen bond cleavage is higher in energy.^9bAs a

consequence, changes in the rate-determining step and the

nature of the substrate and cofactor, will drive the reaction

diﬀerently.

CONCLUSIONS

■

We present a combined experimental and computational study

into the catalytic mechanism of MHQ dioxygenation by a

cofactor-free dioxygenase. Using spectroscopic and kinetic

studies we investigated substrate and substrate analogues

conversion to products using wt and active site variants.

These studies established key roles of a number of active site

residues, namely, His₂₅₁, Asp₁₂₆, Ser₁₀₁and Trp₁₆₀. Thus, an

active site dyad of His₂₅₁/Asp₁₂₆deprotonates the substrate,

which make it susceptible for C−O bond formation and later

electron transfer. The substrate is further stabilized into the

binding pocket through hydrogen bonding interactions with a

number of residues that hold it in the ideal orientation and

assist with C−O bond formation and electron transfer

pathways. Finally, computational modeling establishes what

the rate-determining step in the mechanism is and how it is

aﬀected by changes to the substrate, i.e., replacing methyl by n-

butyl, ﬂuoride, etc. A valence bond model explains the obtained

mechanisms and predicts that the addition of a cofactor to the

active site will not beneﬁt the reaction mechanism and most

likely will slow the reaction down.

(2) (a) Platten, M.; von Knebel Doeberitz, N.; Oezen, I.; Wick, W.;

Ochs, K. Front. Immunol. 2014, 5, 673−673. (b) Prendergast, G. C.;

Smith, C.; Thomas, S.; Mandik-Nayak, L.; Laury-Kleintop, L.; Metz,

R.; Muller, A. J. Cancer Immunol. Immunother. 2014, 63, 721−735.

(c) Munn, D. H.; Mellor, A. L. Trends Immunol. 2013, 34, 137−143.

(d) Sikalidis, A. K. Pathol. Oncol. Res. 2015, 21, 9−17.

(3) Prescott, A. G.; John, P. Annu. Rev. Plant. Physiol. Plant. Mol. Biol.

1996, 47, 245−271.

(4) (a) Gibson, D. T.; Parales, R. E. Curr. Opin. Biotechnol. 2000, 11,

236−243. (b) Karlsson, A.; Parales, J. V.; Parales, R. E.; Gibson, D. T.;

Eklund, H.; Ramaswamy, S. Science 2003, 299, 1039−1042.

(5) (a) Bugg, T. D. H. Tetrahedron 2003, 59, 7075−7101. (b) de

Visser, S. P., Kumar, D., Eds.; Iron-Containing Enzymes: Versatile

Catalysts of Hydroxylation Reactions in Nature; RCS Publishing:

Cambridge, U.K., 2011. (c) Diaz, E.; Jimenez, J. I.; Nogales, J. Curr.

Opin. Biotechnol. 2013, 24, 431−442. (d) Blomberg, M. R. A.;

Borowski, T.; Himo, F.; Liao, R.-Z.; Siegbahn, P. E. M. Chem. Rev.

2014, 114, 3601−3658.

(6) (a) Groce, S. L.; Lipscomb, J. D. Biochemistry 2005, 44, 7175−

7188. (b) Walsh, T. A.; Ballou, D. P.; Mayer, R.; Que, L., Jr. J. Biol.

Chem. 1983, 258, 4422−4427.

(7) (a) Vaillancourt, F. H.; Barbosa, C. J.; Spiro, T. G.; Bolin, J. T.;

Blades, M. W.; Turner, R. F. B.; Eltis, L. D. J. Am. Chem. Soc. 2002,

124, 2485−2496. (b) Davis, M. I.; Orville, A. M.; Neese, F.; Zaleski, J.

M.; Lipscomb, J. D.; Solomon, E. I. J. Am. Chem. Soc. 2002, 124, 602−

614.

(8) (a) Kovaleva, E. G.; Lipscomb, J. D. Science 2007, 316, 453−457.

(b) Knoot, C. J.; Purpero, V. M.; Lipscomb, J. D. Proc. Natl. Acad. Sci.

U. S. A. 2015, 112, 388−393.

(9) (a) Deeth, R. J.; Bugg, T. D. H. J. Biol. Inorg. Chem. 2003, 8, 409−

418. (b) Borowski, T.; Siegbahn, P. E. M. J. Am. Chem. Soc. 2006, 128,

12941−12953.

(10) Fetzner, S.; Steiner, R. A. Appl. Microbiol. Biotechnol. 2010, 86,

791−804.

(11) (a) Bauer, I.; Max, N.; Fetzner, S.; Lingens, F. Eur. J. Biochem.

1996, 240, 576−583. (b) Steiner, R. A.; Janssen, H. J.; Roversi, P.;

Oakley, A. J.; Fetzner, S. Proc. Natl. Acad. Sci. U. S. A. 2010, 107, 657−

662. (c) Thierbach, S.; Bui, N.; Zapp, J.; Chhabra, S. R.; Kappl, R.;

Fetzner, S. Chem. Biol. 2014, 20, 1−9.

(12) (a) Frerichs-Deeken, U.; Ranguelova, K.; Kappl, R.;

Huttermann, J.; Fetzner, S. Biochemistry 2004, 43, 14485−14499.

(b) Frerichs-Deeken, U.; Fetzner, S. Curr. Microbiol. 2005, 51, 344−

ASSOCIATED CONTENT

* Supporting Information

■

S

Table with kinetics and computational data, as well as Schemes

and Figures with kinetic, spectroscopic and computational

results. We also provide Cartesian coordinates of optimized

DFT structures. The Supporting Information is available free of

charge on the ACS Publications website at DOI: 10.1021/

jacs.5b03836.

AUTHOR INFORMATION

Corresponding Authors

*nigel.scrutton@manchester.ac.uk

*sam.devisser@manchester.ac.uk

Notes

■

The authors declare no competing ﬁnancial interest.

ACKNOWLEDGMENTS

■

This research was supported by the BBSRC, which is thanked

for support via grant number BB/I020543/1 (SdV & NSS),

BB/I020411/1 (RAS) as well as for a DTP studentship to

MGQ. The EPSRC is thanked for an Established Career

Fellowship to NSS. The National service of Computational

Chemistry Software (NSCCS) is thanked for additional CPU

time.

́

352. (c) Hernandez-Ortega, A.; Quesne, M. G.; Bui, S.; Heuts, D.;

Steiner, R. A.; Heyes, D. J.; de Visser, S. P.; Scrutton, N. S. J. Biol.

Chem. 2014, 289, 8620−8632.

(13) (a) de Visser, S. P.; Quesne, M. G.; Martin, B.; Comba, P.; Ryde,

U. Chem. Commun. 2014, 50, 262−282. (b) Kumar, S.; Faponle, A. S.;

Barman, P.; Vardhaman, A. K.; Sastri, C. V.; Kumar, D.; de Visser, S. P.

J. Am. Chem. Soc. 2014, 136, 17102−17115. (c) Vardhaman, A. K.;

Barman, P.; Kumar, S.; Sastri, C. V.; Kumar, D.; de Visser, S. P. Angew.

Chem., Int. Ed. 2013, 52, 12288−12292.

REFERENCES

■

(1) (a) Solomon, E. I.; Brunold, T. C.; Davis, M. I.; Kemsley, J. N.;

Lee, S. K.; Lehnert, N.; Neese, F.; Skulan, A. J.; Yang, Y. S.; Zhou, J.

Chem. Rev. 2000, 100, 235−349. (b) Bugg, T. D. H. Curr. Opin. Chem.

Biol. 2001, 5, 550−555. (c) Costas, M.; Mehn, M. P.; Jensen, M. P.;

Que, L., Jr. Chem. Rev. 2004, 104, 939−986. (d) Bruijnincx, P. C. A.;

van Koten, G.; Klein Gebbink, R. J. M. Chem. Soc. Rev. 2008, 37,

2716−2744. (e) Bugg, T. D. H.; Ramaswamy, S. Curr. Opin. Chem.

Biol. 2008, 12, 134−140. (f) Abu-Omar, M. M.; Loaiza, A.; Hontzeas,

(14) Frisch, M. J.; et al. Gaussian 09, Revision C.01; Gaussian, Inc:

Wallingford, CT. 2010.

(15) (a) Becke, A. D. J. Chem. Phys. 1993, 98, 1372−1377. (b) Lee,

C. T.; Yang, W. T.; Parr, R. G. Phys. Rev. B: Condens. Matter Mater.

Phys. 1988, 37, 785−789. (c) Ditchfield, R.; Hehre, W. J.; Pople, J. A. J.

Chem. Phys. 1971, 54, 724−728. (d) McLean, A. D.; Chandler, G. S. J.

Chem. Phys. 1980, 72, 5639−5648.

M

DOI: 10.1021/jacs.5b03836

J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society

Article

(16) (a) Vardhaman, A. K.; Sastri, C. V.; Kumar, D.; de Visser, S. P.

Chem. Commun. 2011, 47, 11044−11046. (b) Heyes, D. J.; Sakuma,

M.; de Visser, S. P.; Scrutton, N. S. J. Biol. Chem. 2009, 284, 3762−

3767.

(17) Linstrom, P. J.; Mallard, W. G. J. Chem. Eng. Data 2001, 46,

1059−1063.

(18) (a) de Visser, S. P. Chem. Commun. 2007, 171−173. (b) Quesne,

M. G.; Latifi, R.; Gonzalez-Ovalle, L. E.; Kumar, D.; de Visser, S. P.

Chem.Eur. J. 2014, 19, 435−446.

(19) Hammett, L. P. J. Am. Chem. Soc. 1937, 59, 96−103.

(20) Torres Pazmino, D. E.; Winkler, M.; Glieder, A.; Fraaije, M. W.

J. Biotechnol. 2010, 146, 9−24.

(21) Veldhuizen, E. J. A.; Vaillancourt, F. H.; Whiting, C. J.; Hsiao,

M. M. Y.; Gingras, G.; Xiao, Y. F.; Tanguay, R. M.; Boukouvalas, J.;

Eltis, L. D. Biochem. J. 2005, 386, 305−314.

(22) Vaillancourt, F. H.; Han, S.; Fortin, P. D.; Bolin, J. T.; Eltis, L.

D. J. Biol. Chem. 1998, 273, 34887−34895.

(23) Entsch, B.; Ballou, D. P. Biochim. Biophys. Acta 1989, 999, 313−

322.

(24) Bull, C.; Ballou, D. P.; Otsuka, S. J. Biol. Chem. 1981, 256,

2681−2686.

(25) (a) Shaik, S.; Kumar, D.; de Visser, S. P. J. Am. Chem. Soc. 2008,

130, 10128−10140. (b) Sainna, M. A.; Kumar, S.; Kumar, D.;

Fornarini, S.; Crestoni, M. E.; de Visser, S. P. Chem. Sci. 2015, 6,

1516−1529. (c) Kumar, D.; Sastry, G. N.; de Visser, S. P. Chem.Eur.

J. 2011, 17, 6196−6205. (d) Kumar, D.; Karamzadeh, B.; Sastry, G. N.;

de Visser, S. P. J. Am. Chem. Soc. 2010, 132, 7656−7667. (e) de Visser,

S. P. J. Am. Chem. Soc. 2010, 132, 1087−1097.

(26) Shaik, S. S.; Hiberty, P. C. A Chemist’s Guide to Valence Bond

Theory; John Wiley & Sons: Hoboken, NJ, 2008.

(27) (a) Nardini, M.; Dijkstra, B. W. Curr. Opin. Struct. Biol. 1999, 9,

732−737. (b) Holmquist, M. Curr. Protein Pept. Sci. 2000, 1, 209−235.

(28) (a) Roth, J. P.; Klinman, J. P. Proc. Natl. Acad. Sci. U. S. A. 2003,

100, 62−67. (b) Gadda, G. Biochemistry 2012, 51, 2662−2669.

(c) Zhao, G.; Bruckner, R. C.; Jorns, M. S. Biochemistry 2008, 47,

9124−9135. (d) Prabhakar, R.; Siegbahn, P. E. M.; Minaev, B. F.

Biochim. Biophys. Acta, Proteins Proteomics 2003, 1647, 173−178.

́

(e) Hernandez-Ortega, A.; Lucas, F.; Ferreira, P.; Medina, M.; Guallar,

V.; Martinez, A. T. Biochemistry 2012, 51, 6595−6608. (f) McDonald,

C. A.; Fagan, R. L.; Collard, F.; Monnier, V. M.; Palfey, B. A. J. Am.

Chem. Soc. 2011, 133, 16809−16811.

(29) Carter, P.; Wells, J. A. Nature 1988, 332, 564−568.

(30) (a) Dunn, G.; Montgomery, M. G.; Mohammed, F.; Coker, A.;

Cooper, J. B.; Robertson, T.; Garcia, J. L.; Bugg, T. D. H.; Wood, S. P.

J. Mol. Biol. 2005, 346, 253−265. (b) Li, J.-J.; Bugg, T. D. H. Org.

Biomol. Chem. 2007, 5, 507−513. (c) Li, C.; Montgomery, M. G.;

Mohammed, F.; Li, J. J.; Wood, S. P.; Bugg, T. D. H. J. Mol. Biol. 2005,

346, 241−251.

(31) Tsukada, H.; Blow, D. M. J. Mol. Biol. 1985, 184, 703−711.

(32) (a) Borowski, T.; Bassan, A.; Siegbahn, P. E. M. Chem.Eur. J.

2004, 10, 1031−1041. (b) Jastrzebski, R.; Quesne, M. G.;

Weckhuysen, B. M.; de Visser, S. P.; Bruijnincx, P. C. A. Chem.

Eur. J. 2014, 20, 15686−15691.

(33) (a) Vaillancourt, F. H.; Bolin, J. T.; Eltis, L. D. Crit. Rev.

Biochem. Mol. Biol. 2006, 41, 241−267. (b) Hitomi, Y.; Yoshida, M.;

Higuchi, M.; Minami, H.; Tanaka, T.; Funabiki, T. J. Inorg. Biochem.

2005, 99, 755−763. (c) Bruijnincx, P. C. A.; Lutz, M.; Spek, A. L.;

Hagen, W. R.; van Koten, G.; Klein Gebbink, R. J. M. Inorg. Chem.

2007, 46, 8391−8402. (d) Pau, M. Y. M.; Davis, M. I.; Orville, A. M.;

Lipscomb, J. D.; Solomon, E. I. J. Am. Chem. Soc. 2007, 129, 1944−

1958. (e) Siegbahn, P. E. M.; Haeffner, F. J. Am. Chem. Soc. 2004, 126,

8919−8932.

N

DOI: 10.1021/jacs.5b03836

J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Article Doi

Catalytic Mechanism of Cofactor-Free Dioxygenases and How They Circumvent Spin-Forbidden Oxygenation of Their Substrates

DOI: 10.1021/jacs.5b03836

Source and publish data:

Authors:

Article abstract of DOI:10.1021/jacs.5b03836

Full text of DOI:10.1021/jacs.5b03836

Products guided by the article

R&D Labs maybe for 70111-44-3

Relevant to this article

Hot Product