Reaction coordinate analysis for β-diketone cleavage by the non-heme Fe2+-dependent dioxygenase Dke1

Article abstract of DOI:10.1021/ja042313q

Acetylacetone dioxygenase from Acinetobacter johnsonii(Dke1) utilizes a non-heme Fe²⁺ cofactor to promote dioxygen-dependent conversion of 2,4-pentanedione (PD) into methylglyoxal and acetate. An oxidative carbon-carbon bond cleavage by Dke1 is triggered from a C-3 peroxidate intermediate that performs an intramolecular nucleophilic attack on the adjacent carbonyl group. But how does Dke1 bring about the initial reduction of dioxygen? To answer this question, we report here a reaction coordinate analysis for the part of the Dke1 catalytic cycle that involves O₂ chemistry. A weak visible absorption band (ε ≈ 0.2 mM^-1 cm^-1) that is characteristic of an enzyme-bound Fe²⁺-β-keto-enolate complex served as spectroscopic probe of substrate binding and internal catalytic steps. Transient and steady-state kinetic studies reveal that O₂-dependent conversion of the chromogenic binary complex is rate-limiting for the overall reaction. Linear free-energy relationship analysis, in which apparent turnover numbers (k_cat^app) for enzymatic bond cleavage of a series of substituted β-dicarbonyl substrates were correlated with calculated energies for the highest occupied molecular orbitals of the corresponding β-keto-enolate structures, demonstrates unambiguously that k _cat^app is governed by the electron-donating ability of the substrate. The case of 2′-hydroxyacetophenone (2′HAP), a completely inactive β-dicarbonyl analogue that has the enol double bond delocalized into the aromatic ring, indicates that dioxygen reduction and C-O bond formation cannot be decoupled and therefore take place in one single kinetic step.

Full text of DOI:10.1021/ja042313q

Published on Web 08/13/2005
Reaction Coordinate Analysis for â-Diketone Cleavage by the
Non-Heme Fe2+-Dependent Dioxygenase Dke1
Grit D. Straganz* and Bernd Nidetzky*
Contribution from the Institute of Biotechnology and Biochemical Engineering,
Graz UniVersity of Technology, Petersgasse 12, A-8010 Graz, Austria
Received December 21, 2004; Revised Manuscript Received July 15, 2005; E-mail: grit.straganz@TUGraz.at; bernd.nidetzky@TUGraz.at
Abstract: Acetylacetone dioxygenase from Acinetobacter johnsonii (Dke1) utilizes a non-heme Fe2+ cofactor
to promote dioxygen-dependent conversion of 2,4-pentanedione (PD) into methylglyoxal and acetate. An
oxidative carbon-carbon bond cleavage by Dke1 is triggered from a C-3 peroxidate intermediate that
performs an intramolecular nucleophilic attack on the adjacent carbonyl group. But how does Dke1 bring
about the initial reduction of dioxygen? To answer this question, we report here a reaction coordinate analysis
for the part of the Dke1 catalytic cycle that involves O2 chemistry. A weak visible absorption band (ꢀ ≈ 0.2
mM-1 cm-1) that is characteristic of an enzyme-bound Fe2+-â-keto-enolate complex served as
spectroscopic probe of substrate binding and internal catalytic steps. Transient and steady-state kinetic
studies reveal that O2-dependent conversion of the chromogenic binary complex is rate-limiting for the
app
cat
overall reaction. Linear free-energy relationship analysis, in which apparent turnover numbers (k ) for
enzymatic bond cleavage of a series of substituted â-dicarbonyl substrates were correlated with calculated
energies for the highest occupied molecular orbitals of the corresponding â-keto-enolate structures,
app
cat
demonstrates unambiguously that k is governed by the electron-donating ability of the substrate. The
case of 2′-hydroxyacetophenone (2′HAP), a completely inactive â-dicarbonyl analogue that has the enol
double bond delocalized into the aromatic ring, indicates that dioxygen reduction and C-O bond formation
cannot be decoupled and therefore take place in one single kinetic step.
of bound Fe2+. Only Fe2+ but not other divalent metal ions such
as Zn2+, Mn2+, Cu2+, Ni2+, or Co2+ can be employed to
reconstitute active Dke1 from the metal free apo-enzyme.1b
Selective use of Fe2+ for catalysis links Dke1 to prominent
NHME superfamilies such as the R-keto-acid-dependent dioxy-
genases and extradiol-catechol ring-cleaving dioxygenases. The
all-histidine metal-binding site in Dke1, which is constituted
by His-62, His-64, and His-104,3 however, differs from the
active site configuration in these NHMEs that use a facial triad
Introduction
Acetylacetone dioxygenase Dke1 from Acinetobacter
johnsonii catalyzes the O2-dependent conversion of 2,4-
pentanedione (PD) into methylglyoxal and acetate.1 Oxidative
carbon-carbon bond cleavage in the â-dicarbonyl substrate
proceeds with stoichiometric consumption of O2. Results of 18
O
labeling experiments support classification of Dke1 as a novel
dioxygenase (EC 1.13.11.50), showing that one atom of 18O is
incorporated into each reaction product upon the enzymatic bond
fission.1a Dke1 is a functional homotetramer composed of
subunits of 153 amino acids with a molecular mass of 16 607
Da. It belongs to a group of mononuclear non-heme metal-
dependent enzymes (NHMEs) that elegantly couples dioxygen
chemistry with a wide range of substrate transformations,
including C-C bond cleavage, decarboxylation, hydroxylation,
and carbon-heteroatom bond formation.2 Fe2+ appears to be
central to the enzymatic mechanism because the turnover
number of Dke1 is strictly correlated with the mole equivalent
of two histidines and one carboxylate to bind Fe2+ 2e
.
Dke1 can accept a range of â-dicarbonyl compounds that
represent a wide variety of substituents at both the central C-3
and the methyl groups of the natural PD substrate.1 The relaxed
substrate specificity of Dke1 suggested the possibility of probing
the enzymatic mechanism through quantitative structure-
reactivity correlations of kinetic substituent effects. Considering
possible pathways for enzymatic C-C bond cleavage, we
rationalized that variation of the substrate structure could affect
different points of the reaction coordinate, arguably leading to
changes in the catalytic rate (this work) and the cleavage
specificity of the enzyme.1a We showed in a recent article that
the oxygenative cleavage of asymmetric â-dicarbonyl com-
pounds is controlled by the electronic properties of the substit-
uents and takes place at the carbon adjacent to the most electron-
deficient carbonyl group. The linear free-energy relationship
(1) (a) Straganz, G. D.; Hofer, H.; Steiner, W.; Nidetzky, B. J. Am. Chem.
Soc. 2004, 126, 12202-12203. (b) Straganz, G. D.; Glieder, A.; Brecker,
L.; Ribbons, D. W.; Steiner, W. Biochem. J. 2003, 369, 573-581.
(2) For a review, see: (a) Costas, M.; Mehn, M. P.; Jensen, M. P.; Que, L., Jr.
Chem. ReV. 2004, 104, 939-986. (b) Bugg, T. D. Curr. Opin. Chem. Biol.
2001, 5, 550-555. (c) 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, 235-350. (d) Lange, S. J.; Que, L., Jr.
Curr. Opin. Chem. Biol. 1998, 2, 159-172. (e) Hegg, E. L.; Que, L., Jr.
Eur. J. Biochem. 1997, 250, 625-629. (f) Que, L., Jr.; Ho, R. Y. Chem.
ReV. 1996, 96, 2607-2624.
(3) Stranzl, G. D. Ph.D. Thesis, Karl Franzens University, Graz, Austria, 2002.
9
12306
J. AM. CHEM. SOC. 2005, 127, 12306-12314
10.1021/ja042313q CCC: $30.25 © 2005 American Chemical Society

â
-Diketone Cleavage by Fe2+-Dependent Dke1
A R T I C L E S
Scheme 1. Proposed Mechanism of Oxygenative C-C Bond Cleavage in PD Catalyzed by Dke1a
a The steps that involve dioxygen activation and reduction of O2 are hypothetical and adapted from related non-heme Fe2+ dioxygenases.2 Subsequent
reactions triggered from a peroxidate intermediate are supported by results of Straganz et al.1a
analysis for the cleavage specificity of Dke1 suggested a
mechanism of carbon-carbon bond fission that is summarized
in Scheme 1. The reaction of O2 with the PD substrate is
proposed to yield a C-3 peroxidate intermediate that performs
intramolecular nucleophilic attack on the juxtaposed carbonyl
group to generate a dioxetane that then decomposes into the
products. In this article we address the question of how Dke1
promotes the crucial first steps of the catalytic sequence that
involve reduction of O2 and formation of the peroxidate.
Reactions of singlet substrates and triplet dioxygen via
concerted two-electron transfer resulting in a singlet product
are formally spin-forbidden and therefore do not take place at
appreciable rates in the absence of a catalyst that alters the spin
state of dioxygen or substrate. The mechanism by which
NHMEs turn on their O2 reactions has thus aroused considerable
interest among biochemists.2,4 In consideration of the wide
variety of enzymatic transformations performed on totally
unrelated substrate structures and promoted by active sites
harnessing different metal ions, catalytic factors and their relative
contributions to rate acceleration are hardly uniform among the
Because one-electron reduction of dioxygen is thermodynami-
cally unfavorable,4b,6 it has been suggested that additional
electrons from bound ligand could be used to pull the reaction
forward, and this plausibly explains the increased reactivity of
Fe2+ in the presence of substrate. Another possibility is that
substrate binding lowers the redox potential of Fe2+, thereby
facilitating reduction of dioxygen through enhancement of the
stability of Fe3+. It consolidates the finding that inactivation of
dihydroxy-2,3-biphenyl dioxygenase as a result of Fe2+ oxida-
tion proceeds much more quickly during substrate turnover than
in the resting state. Time-dependent formation of Fe3+ in this
enzyme parallels the evolution of superoxide, suggesting that
one-electron reduction of dioxygen by ferrous iron is partially
decoupled from the subsequent catalytic cycle.7 In an alternative
scenario, the concept of substrate activation by the metal cofactor
as has been reported for Fe3+-dependent dioxygenases8 and
Cu2+-containing quercetin dioxygenase9 might also assist O2
reduction in Fe2+-dependent enzymes, as has been suggested
for the enzyme mechanism of Fe2+-dependent 2,3-dihydroxy-
biphenyl dioxygenase based on molecular orbital calculations
of the active site.4b Finally, a conformational rearrangement of
the active site, triggered by the event of substrate binding, may
prime the enzymatic reaction in one of several ways. R-Keto-
glutarate-dependent dioxygenases change the geometry of the
Fe2+ cofactor from six-coordinate to five-coordinate upon
formation of the ternary complex between enzyme, R-keto-
glutarate cosubstrate, and substrate and thus generate an open
NHMEs. A classical example is provided by the group of Fe3+
-
dependent catechol intradiol-cleaving dioxygenases which are
thought to activate their substrate (not dioxygen) via the metal
cofactor. NHMEs requiring ferrous iron for activity utilize a
different catalytic mechanism. A chemically convincing scenario
is that activation of dioxygen initiates the enzymatic reaction
and takes place through one-electron reduction of O2 mediated
by Fe2+, concomitant with formation of an Fe3+-superoxide
intermediate. For the first time, very recently an Fe3+-bound
superoxide intermediate could be spectroscopically characterized
during the oxygenation reaction of a non-heme diiron(II)
complex.5
binding position for dioxygen on the previously protected Fe2+
,
which will consequently react with O2.4a A more favorable
orientation of the molecular orbitals of Fe2+ and O2 might also
be induced upon substrate binding, thereby bringing about the
reaction.
A dioxygenase reaction that is promoted by the oxidation of
enzyme-bound Fe2+ poses the question of how the enzyme
prevents the conversion of Fe2+ to Fe3+, uncoupled from the
event of substrate transformation. A significant observation
reported in several articles is that dioxygen reactivity increases
in the substrate-bound enzyme by as much as ∼107-fold,
compared with the reference reaction of O2 and the resting Fe2+
enzyme in the absence of substrate or ligand. A number of
factors, summarized later in this article, have been suggested
to contribute to this large reactivity enhancement toward
dioxygen upon substrate binding.2,4 The lack of an observable
Fe3+ intermediate in NHMEs has called into question a distinctly
stepwise conversion of dioxygen to peroxide via superoxide.
Here we report a detailed reaction coordinate analysis for the
Dke1-catalyzed conversion of PD into the C-3 peroxidate
intermediate (Scheme 1) which decomposes into the products.
The results provide insights into kinetics and chemistry of
dioxygen reduction by Dke1.
Experimental Section
Chemicals and Enzyme. PD, 1,1,1-trifluoro-2,4-pentanedione (TFPD),
4,4-difluoro-1-phenyl-1,3-butanedione (DFPB), 4,4,4-trifluoro-1-phenyl-
1,3-butanedione (TFPB), 1-phenyl-1,3-butanedione (PBD), and 2′-
(6) (a) Bassan, A.; Borowski, T.; Siegbahn, P. E. M. J. Chem. Soc., Dalton
Trans. 2004, 20, 3153-3162. (b) Sawyer, D. T. In Oxygen Complexes and
Oxygen ActiVation by Transition Metals; Martell, A. E., Sawyer, D. T.,
Eds.; Plenum: New York, 1988; pp 131-148.
(7) Vaillancourt, F. H.; Labbe, G.; Drouin, N. M.; Fortin, P. D.; Eltis, L. D. J.
Biol. Chem. 2002, 277, 2019-2027.
(4) (a) Solomon, E. I.; Decker, A.; Lehnert, N. Proc. Natl. Acad. Sci. U.S.A.
2003, 100, 3589-3594. (b) Davis, M. I.; Wasinger, E. C.; Decker, A.;
Pau, M. Y. M.; Vaillancourt, F. H.; Bolin, J. T.; Eltis, L. D.; Hedman, B.;
Hodgson, K. O.; Solomon, E. I. J. Am. Chem. Soc. 2003, 125, 11214-
11227.
(8) (a) Que, L., Jr.; Lipscomb, J. D.; Munck, E.; Wood, J. M. Biochim. Biophys.
Acta 1977, 485, 60-74. (b) Hamilton, G. A. In Molecular Mechanisms of
Oxygen Actiuation; Hayaishi, O., Ed.; Academic Press: New York, 1974;
pp 443-445.
(5) Shan, X.; Que, L., Jr. Proc. Natl. Acad. Sci. U.S.A. 2005, 102, 5340-
(9) Steiner, R. A.; Kalk, K. H.; Dijkstra, B. W. Proc. Natl. Acad. Sci. U.S.A.
2002, 99, 16625-16630.
5345.
9
J. AM. CHEM. SOC. VOL. 127, NO. 35, 2005 12307

A R T I C L E S
Straganz and Nidetzky
hydroxyacetophenone (2′HAP) and all other chemicals were from Sigma
Aldrich (St. Louis, MO) except 1,1-difluoro-2,4-pentanedione (DFPD),
which was from Matrix Scientific (Columbia, SC). They were all of
the highest available quality and unless otherwise mentioned greater
than 97% pure. Recombinant Dke1 harboring a C-terminal decapeptide
fusion was produced and purified by affinity chromatography as
described elsewhere.10 The tagging has no effect on Dke1 activity and
specificity. The molar concentration of purified Dke1 was determined
from the absorption (at 280 nm) of the enzyme in the presence of 6 M
guanidine HCl, using an extinction coefficient of 2.1 mM-1 cm-1, which
was calculated from the tagged polypeptide sequence. Protein-bound
Fe2+ was measured using ferene S as metal chelator. The assay
employed 500 µL of an appropriately diluted protein sample that was
made anaerobically by repeated cycles of evacuation and nitrogen
flushing. The protein solution was mixed with 500 µL of a nitrogen-
purged solution of ferene S (40 mM; ∼1000-fold molar excess to the
concentration of Dke1 subunits) in 20 mM Tris/HCl, pH 7.5, and
incubated at 25 °C. The increase in absorbance at 592 nm was
monitored over time, typically up to 6 h, until it reached a constant
value. Enzyme-bound Fe2+ concentrations were calculated from the
difference of initial and final absorbance with an extinction coefficient
fluorescence titration. Among two new absorbance bands that appeared
in the spectrum of substrate-bound Dke1 compared with the spectrum
of free enzyme (see the Results section), the intensity of the band that
was lowest in energy (λ3) was chosen as reporter of the binding event.
Aliquots of ligand solution were added as described above until the
band intensity did not increase further. Because binding of most ligands
was tight relative to the enzyme concentration used in the experiment,
a significant portion of the total concentration of each binding partner
is in the binary complex at equilibrium. Therefore, eq 1 was used to
fit the data with the program Microcal Origin Pro 6.1 (OriginLab
Corporation, Northampton, MA):
A3 ) ꢀ3{(Kd + cDke1 + L) - [(Kd + cDke1 + L)2 - 4cDke1 L]0.5}/2
(1)
where A3 is the absorbance at λ3 with the corresponding absorption
coefficient ꢀ3, cDke1 is the concentration of enzyme active sites, and L
is the ligand concentration.
The fluorescence titrations were performed in the same way just
described except that cDke1 was 0.2 µM (95% Fe2+ occupancy) and the
temperature was 4 °C or 25 °C, as indicated. Experiments were
performed with a Hitachi F-4500 fluorescence spectrophotometer
(Hitachi High-Technologies, Tokyo, Japan) using excitation and
emission wavelengths of 290 and 330 nm, respectively. Constant slit
widths of 5 nm were used. The quenching of the intrinsic Dke1
tryptophan fluorescence was measured as a function of the ligand
concentration. After corrections for dilution and blank readings, data
were plotted as (I0 - I)/(I0 - I)max against L, where I0 and I are
fluorescence intensities in the absence and presence of ligand,
respectively. Equation 2 was used to fit the data:
of 35.5 mM-1 cm-1 for the stable complex between ferene S and Fe2+ 11
.
The control used 20 mM ferene S in 20 mM Tris/HCl, pH 7.5, in the
absence of Dke1.
Initial Rate Studies. Initial rates were recorded by measuring the
depletion of substrate or dioxygen. It was proven that both methods
yielded consistent results. Substrate conversion was monitored spec-
trophotometrically with a DU 800 UV-vis spectrophotometer (Beck-
mann Coulter, Inc., Fullerton, CA) at 25 °C. The wavelength of
maximum absorption for the enolate of each â-diketone at pH 7.5 (λ1)
was determined in separate experiments, as described in the Supporting
Information, and used to monitor the time course of the enzymatic
reaction. Assays were performed in a total volume of 1.5 mL. Tris/
HCl buffer, 20 mM pH 7.5, was filled into a sealed quartz cuvette and
brought to a defined concentration of dissolved dioxygen (0-1200 µM)
by flushing with mixtures of O2 and N2. Then Dke1 was added, and
I0 - I ) (I0 - I )maxL/(L + Kd)
(2)
where (I0 - I)max is the difference between I0 and I at saturating L.
Gel Filtration Analysis of the Dke1-PD Complex. A HiTrap
desalting column (5 mL, Amersham Biosciences, Uppsala, Sweden)
was used to fractionate high and low molecular weight complexes of
Fe2+ and PD. The column was equilibrated with degassed Tris/HCl
buffer (20 mM, pH 7.5) containing 5 mM PD which was added to
prevent bleaching of the chromogenic Fe2+ complex in the presence
of residual O2. The anaerobic sample (0.5 mL) contained 1 mM Dke1
active sites dissolved in the buffer supplemented with PD, and 500 µL
of it was applied to the column. It was eluted at a flow rate of 1 mL
min-1 with automatic collection of 0.5-mL fractions. Color associated
with certain fractions was clearly visible and confirmed by recording
a UV-vis spectrum. Dke1-containing fractions were identified using
the Bio-Rad protein dye binding assay (Bio-Rad, Hercules, CA).
Transient Kinetic Measurements. Stopped-flow measurements
were carried out on an Applied Photophysics instrument (model SX.18
MV, Applied Photophysics LTD, Leatherhead, U.K.) equipped with a
modular optical system. Data acquisition and analysis were done using
Applied Photophysics software. Detection was by absorbance, using a
wavelength characteristic of the free substrate enolate or the enzyme-
substrate complex (A3). Solutions of the reactants were prepared in air-
saturated 20 mM Tris/HCl buffer, pH 7.5, the cO2 being 260 µM if not
otherwise stated. From two separate syringes, 100 µL of each purified
recombinant enzyme (580 µM subunits, 75%, Fe2+ content) and
substrate (400 µM) was shot into a 20-µL flow cell having a 1-cm
path length. Alternatively, for slow reactions, a DU 800 UV-vis
spectrophotometer was used to monitor transient kinetics under
analogous conditions. All experiments were done at 25 °C in triplicate
and analyzed, and the resulting k values were averaged. Appropriate
controls were recorded in all cases to exclude the possibility of artifacts.
The observed rate constants for formation (k1) and decay (k2) of the
binary complex were obtained from nonlinear fits of the appropriate
parts of the experimental absorbance trace at A3 to eqs 3 and 4,
respectively. In eqs 3 and 4, A3,max is the theoretical maximum
the resulting dioxygen concentration (cO ) was recorded. The enzymatic
2
reactions were started by addition of substrate (3-30 µL) dissolved in
the above Tris/HCl buffer using a Hamilton syringe. It will be shown
in the Results section that Dke1 could not be saturated with dioxygen
app
cat
at the steady state. We thus define an apparent turnover number (k
)
that is the mole of substrate cleaved per mole of enzyme-bound iron
and second at the respective O2 concentration. Initial rates of dioxygen
consumption and all other cO measurements were performed with a
2
micro-optode O2 sensor (Microtox TX3-AOT, PreSens, Regensburg,
Germany) that was introduced directly into the sealed cuvette. The
sensor used noninvasive fluorescence quenching as a method of
detection.
Spectroscopic Characterization of Dke1-Substrate Complexes.
An anaerobic solution of Dke1 (270 µM subunits and 75% Fe2+ content,
giving 200 µM active sites) in 20 mM Tris buffer, pH 7.5, was prepared
in an evacuable quartz cuvette sealed with a septum. Spectroscopic
titrations were carried out by adding aliquots of 0.5-5 µL from an
anaerobic stock solution of ligand with a Hamilton syringe. Absorbance
wavelength scans in the range 300-700 nm were carried out on the
free enzyme and after each addition of ligand, and the resulting spectra
were compared after appropriate corrections for dilution. Controls were
obtained by the same procedure without Dke1, and corrections for blank
values at the corresponding ligand concentrations were made in all
cases.
Equilibrium Binding Studies. Dissociation constants (Kd) of Dke1-
substrate complexes were determined using data from absorbance or
(10) Straganz, G. D.; Slavica, A.; Hofer, H.; Mandl, U.; Steiner, W.; Nidetzky,
B. Biocatal. Biotransform., in press.
(11) Johnson-Winters, K.; Purpero, V. M.; Kavana, M.; Nelson, T.; Moran, G.
R. Biochemistry 2003, 42, 2072-2080.
9
12308 J. AM. CHEM. SOC. VOL. 127, NO. 35, 2005

â
-Diketone Cleavage by Fe2+-Dependent Dke1
A R T I C L E S
Table 1. Steady-State and Transient Kinetic Analysis of the
Reaction of Dke1 with a Series of â-Dicarbonyl Compounds and
Corresponding Calculated Energies of Their Highest Occupied
Molecular Orbital (ꢀHOMO) for the Monoanionic Substrate
Structuresa
app
cat
1
1
1
substrate
k
(s-
)
Km
(
µM)
ꢀHOMOb (eV)
k
1 (s-
)
k )
2 (s-
PD
6.5
9.1 ( 1.5c
2.7 ( 0.7
2.5 ( 0.5
-3.20
-3.78
-4.15
-4.03
-4.40
-3.31
225
9.5
6.6
DFPD
TFPD
DFPB
TFPB
2′HAP
3.6 × 10-2
4.2 × 10-2
5.3 × 10-3
4.1 × 10-4
5.2 × 10-5
0
4.9 × 10-3
0.73
20
4.3 × 10-4 d 2.0 ( 1.0d
n.d.e
0
n.d.e
-
n.d.e
n.d.e
app
cat
a All experiments were performed at 25 °C and cO ) 260 µM. k and
2
observed rate constants of enzyme-substrate complex formation (k1) and
decay (k2) are within the experimental error of 10%. b The values of ꢀHOMO
were obtained with the program Spartan’02 (Wavefunction, Inc., CA) using
semiempirical calculation with AM1 basis set. c Determined previously.1b
d Assays with DFPB were performed under conditions of single turnover
(cDke1 > cDFPB). e n.d. ) not determined.
absorbance value under the conditions used. Because formation and
breakdown of the enzyme-substrate complex occur in well-separated
time domains (k1 > 10k2), it was possible to treat binding and reaction
as independent processes.
A3(buildup) ) A3,max[1 - exp(-k1t)]
A3(breakdown) ) A3,max exp(-k2t)
(3)
(4)
Figure 1. Apparent steady-state kinetic parameters of the Dke1-catalyzed
conversion of â-diketones at varying dissolved dioxygen concentrations at
Results
app
A recent study has reported electronic substituent effects on
the specificity of carbon-carbon bond cleavage catalyzed by
Dke1.1a The rate of â-diketone substrate conversion at the steady
state, however, shows an even more pronounced substituent
dependence, prompting us to perform a detailed kinetic analysis
for the reaction of Dke1 with a range of substituted â-dicarbonyl
compounds.
25 °C. (A) Correlation of k and cO for PD, DFPD, and TFPD. (B)
cat
2
app
Comparison of the dependences of k and the transient reaction rate
cat
constant k2 (see later) on c for the conversion of DFPD. The inset shows
O2
Km values for DFPD determined at different levels of cO
.
2
app
cat
Estimates of k /Km for O2 are obtained from the slopes of the
straight lines in Figure 1 (panel A).
app
cat
These k /Km values are consistently lower than the corre-
Substituent Effects on Steady-State Kinetic Parameters.
Initial rates of Dke1-catalyzed cleavage of different â-diketones
were recorded under conditions in which the substrate concen-
sponding second-order rate constants for the â-diketone sub-
strates by a factor of about 50, recorded at a constant cO of
2
260 µM. Figure 1 also shows that the transient reaction rate
constant (k2), obtained from single-turnover kinetic measure-
ments described below, has exactly the same dependence on
tration was varied and the cO was constant at 260 µM.
2
Spectrophotometric detection of substrate conversion at its
characteristic λ1 was employed (see the Supporting Information).
Kinetic parameters were obtained from nonlinear fits of a
rectangular hyperbola to the data and are summarized in Table
1.
cO as the steady-state turnover number. These results validate
2
an experimental approach in which the reactivities of different
substrates are compared at a precisely controlled and identical,
however nonsaturating, concentration of dioxygen.
In search of an interpretable structure-reactivity correlation
of the strong kinetic substituent effect on kacaptp, we performed
frontier molecular orbital calculations on the monoanions of the
chosen â-diketones and used the thus-obtained energies of their
respective highest occupied molecular orbital (ꢀHOMO) as nu-
The apparent turnover number (kacaptp) value changed more
than 105-fold upon alteration of the substituents on the core
â-diketone structure of PD. The apparent Michaelis constant
Km, however, showed very little variation between 2 and 9 µM,
across the same series of substrate analogues, suggesting that
the kinetic substituent effect is almost exclusively on the
cleophilic reactivity scale for the Dke1 substrates (Table 1). A
app
cat
catalytic rate and not on binding. The catalytic efficiency
plot of log k against ꢀHOMO (Figure 2) was linear with a
app
cat
coefficient of determination (r2) of 0.97 and a slope of 4.4 eV-1
.
(k /Km), which represents the extrapolation to a limiting
substrate concentration, therefore shows a similar correlation
The mechanistic implication of this free-energy relationship
is as follows. The absence of a break in the correlation in Figure
2 suggests that there is no transition between different rate-
limiting steps across the series of homologous substrates,
app
as k
.
cat
We investigated the influence of the O2 concentration on
kinetic parameters for the conversion of PD, DFPD, and TFPD.
As shown in Figure 1 (panel B), the value of Km did not change
although it spans 5 orders of magnitude in enzymatic activity.
app
cat
significantly ((20%) in response to an increase of cO from 20
The value of k
which then represents a common rate-
2
to 1200 µM. kapp showed a linear dependence on cO irrespec-
limiting step is completely governed by the nucleophilic
character of the monoanionic substrate or, in other words, its
aptitude to donate electrons.
2
cat
tive of the substrate used (Figure 1, panel A), revealing clearly
that Dke1 cannot be saturated with dioxygen at the steady state.
9
J. AM. CHEM. SOC. VOL. 127, NO. 35, 2005 12309

A R T I C L E S
Straganz and Nidetzky
Electron-withdrawing substituents lead to lower energy
transitions, suggesting that the observed absorption bands can
be assigned to metal-to-ligand charge transfer (MLCT) transi-
tions in the Fe2+-â-keto-enolate complex of Dke1 and PD
(Scheme 1). The spectral properties of the Dke1 complexes
apparently resemble those of R-ketoacid-dependent non-heme
Fe2+ dioxygenases bound to a bidentate R-keto carboxylate
ligand.11,13 Evolution of band intensity of λ3 in response to the
presence of increasing amounts of dicarbonyl compound was
used to titrate the enzyme, and nonlinear fits of the data to eq
1 yielded estimates for the respective dissociation constants (Kd).
Dke1 bound to the inactive ligand 2′HAP also showed MLCT
transitions, the Kd value of this complex being ∼22 µM (see
Table 2). Results of fluorescence titrations confirm the Kd value
estimations that are based on measurements of MLCT band
intensity (Table 2).
Figure 2. Linear free-energy relationship analysis for Dke1-catalyzed
app
conversion of different â-diketones at 25 °C. k values were determined
cat
in this study or, in the cases of 2-acetylcyclohexanone (ACH), 3,5-
heptanedione (HD), 2,4-octanedione (OD), and 3-methylpentanedione
(MPD), were reported previously.1b For TFPD the transient rate constant
is given.
Single Turnvover Transient Kinetics. Rapid-mixing stopped-
flow experiments were performed under conditions in which
the substrate concentration limited the reaction to less than a
single turnover of the total active site concentration present.
The kinetic transients were recorded with the aim of character-
izing microscopic steps leading to and from the Dke1-â-keto-
enolate complex (Scheme 1). Figure 4 shows representative
single absorbance traces for the reaction of Dke1 with four
selected â-diketones in the presence of an air-saturated initial
Spectral Properties of the Enzyme-Substrate Complex.
Upon addition of PD or substituted â-dicarbonyl substrate
analogues to a concentrated Dke1 solution under anaerobic
conditions, two new bands (at λ2 and λ3) appeared in the UV-
visible region of the protein absorbance spectrum. Because
TFPB is turned over by the enzyme at a very slow rate, the
spectrum of the anaerobic Dke1-TFPB complex could be
compared with the spectrum of the same complex in an air-
saturated buffer (Figure 3, panel A), and superimposable
absorbance traces were obtained. If, however, Dke1-TFPB was
incubated in the presence of dioxygen for long (>2 h), the pink
color bleached gradually over time, as shown in Figure 3 (panel
A). The time-dependent loss in band intensity at λ2 and λ3
paralleled the time course of substrate depletion recorded at λ1
(Figure 3, panel B).
cO . Equations 3 and 4 were used to fit the appropriate part of
2
the time course from each stopped-flow experiment, as described
in the Experimental Section. Table 1 summarizes pseudo-first-
order rate constants of formation (k1) and disappearance (k2) of
the MLCT band at λ3 and compares them to the corresponding
steady-state kinetic parameters. Note that all results are from
measurements at a single dioxygen concentration (260 µM),
which is not saturating in the steady state. However, as already
shown in Figure 1, the dependence of k2 on cO for the
2
app
conversion of DFPD was exactly the same as that of k for
In concentrated solutions of fully active Dke1 (g100 µM,
Fe2+ g 90%) we found that a significant portion, up to 10% of
the total Fe2+, is dissociated, reflecting the binding equilibrium
for the metal cofactor. The â-dicarbonyl substrates of Dke1 are
reported to form stable complexes with Fe2+ in solution,12 and
the spectral properties thereof are comparable to those seen for
the Dke1-substrate complexes. We did not detect formation
of colored Fe2+-diketone complexes in the concentration range
relevant for our experiments (e10 µM free Fe2+; e50 µM
ligand). However, it was necessary to rigorously eliminate the
possibility that a solution complex of Fe2+ interferes with the
spectroscopic characterization of the enzyme-substrate adduct.
Results of gel filtration analysis revealed that the color
developed upon binding of PD or TFPB to fully active Dke1
eluted quantitatively in the protein-containing fractions, clearly
indicating that it was completely associated with the enzyme.
A control experiment used 200 µM Fe2+ in place of Dke1, and
the metal complex with 5 mM PD was washed off the column
wholly in the salt fraction. We conclude therefore that absorption
transitions described in Figure 3 and Table 2 distinctly reflect
noncovalent enzyme-substrate interactions at the level of the
binary complex. Table 2 summarizes the values of λ2 and λ3
and the corresponding molar extinction coefficients for a series
of complexes of Dke1 with substituted â-diketones.
cat
the same reaction. The results in Table 1 reveal that k1 and k2
were strongly influenced by the substrate structure, whereby
electron-withdrawing substituents slowed the formation and even
more dramatically the decay of the chromophoric complex. k2
app
cat
and k are congruent, and consequently their correlations
with ꢀHOMO are almost superimposable, which implies that the
intrinsic nucleophilic reactivity of the substrate governs k2 in
analogy to kacaptp. By contrast, log k1 cannot be so clearly
correlated against the chosen reactivity scale, indicating that
factors other than the electronic ones also influence binding.
The analysis of kinetic transients recorded at wavelengths
diagnostic of selected free and Dke1-bound substrates, namely
TFPD, DFPB, and TFPB, also showed that within the limits of
experimental error enzymatic cleavage of the â-keto-enolate
and loss of the MLCT band occurred at the same pre-steady-
state rate (data not shown).
The plot of k2 against cO in Figure 1 thus contains additional
2
information of mechanistic relevance. k2 is a pseudo-first-order
rate constant, and for the simple reaction, E-S + O2 T E-P,
where E-S is the Dke1-PD complex and E-P is the Dke1-
product complex, it is described by the relationship k2 ) k+2[O2]
+ k-2 where k+2 and k-2 are rate constants in the forward and
(13) (a) Ryle, M. J.; Padmakumar, R.; Hausinger, R. P. Biochemistry 1999, 38,
15278-15286. (b) Hegg, E. L.; Whiting, A. K.; Saari, R. E.; McCracken,
J.; Hausinger, R. P.; Que, L., Jr. Biochemistry 1999, 38, 16714-16726.
(12) Eng, S. J.; Motekaitis, R. J.; Martell, A. E. Inorg. Chim. Acta 1998, 278,
170-177.
9
12310 J. AM. CHEM. SOC. VOL. 127, NO. 35, 2005

â
-Diketone Cleavage by Fe2+-Dependent Dke1
A R T I C L E S
Figure 3. Spectra of a Dke1-TFPB complex and time-resolved analysis of band decay at λ1 - λ3 during enzymatic substrate conversion in the presence
of 260 µM dioxygen at 25 °C. The complex was formed by mixing Dke1 (175 µM active sites) and TFPB (200 µM), each dissolved in an air-saturated
Tris/HCl buffer (20 mM, pH 7.5). (A) Wavelength scans performed at incubation times of 0, 150, 300, 500, 750, and 1000 min show depletion of the
TFPB-enolate band (λ1) and the bands at λ2 and λ3 that are lacking in the free enzyme but appear in the Dke1-TFPB complex (enlarged inset). (B)
Superimposition of the time courses of band decay at λ1 (black) and λ2,3 (pink).
Table 2. Characterization of â-Diketone Binding to Dke1 Using
whereas individually the â-dicarbonyl compounds and Dke1 are
essentially inert? There is evidence from the studies of heme
proteins14 that the redox potentials of their respective iron centers
UV-Visible and Fluorescence Spectroscopy
a
a
a
b
ꢀ1
(
λ1
)
ꢀ2
(
λ2
)
ꢀ3
(
λ3
)
Kd
S
mM-1 cm-1 (nm)
mM-1 cm-1 (nm)
mM-1 cm-1 (nm)
µM
(E0) may be altered in response to substrate binding. The log
app
cat
PD
2.2 (280)
10 (290)
14 (290)
13 (324)
16 (322)
3.3 (322)
0.27 (355)
0.22 (380)
0.28 (382)
0.40 (510)
0.34 (505)
0.25 (472)
0.25 (420)
0.28 (450)
0.27 (450)
0.37 (530)
0.35 (540)
0.25 (500)
0.75
5.8
6.7
4.1
2.2
k
- ꢀHOMO correlation for Dke1 implies the possibility that
DFPD
TFPD
DFPB
TFPB
2′HAP
E0 of the Fe2+ cofactor is lowered upon formation of the binary
complex, thereby enhancing the stability of Fe3+ and thus
favoring one-electron reduction of dioxygen by Fe2+ in the rate-
determining step (Scheme 1). We plotted literature values of
E0 for mononuclear iron complexes of the form Fe3+L3, where
L is the enolate of PD or a substituted derivative thereof,15
against the corresponding ꢀHOMO values of L. A good linear
correlation was obtained (r2 > 0.90; Figure 5) in which the
midpoint potential decreased with increasing ꢀHOMO from +0.07
V for Fe(TFPB)3 to 0.68 V for Fe(PD)3. By way of comparison,
the standard redox potential of the unligated Fe2+/Fe3+ couple
is +0.77 V, suggesting that substrate-derived donor ligands to
the Fe2+ cofactor of Dke1 could indeed make one-electron
reduction of O2 more favorable.
22
a The wavelengths of maximum absorption (λ1) for the free cis-â-keto-
enolate form of the substrate (S) and the anaerobic Dke1-ligand complex
(λ2, λ3) are given together with the corresponding molar extinction
coefficients (ꢀ). Note that the molarity of Fe2+-containing active sites is
the basis for the calculation of ꢀ. b The Kd values are from fits of eq 2 to
data from fluorescence titrations carried out at 4 °C and have a standard
deviation of (25%. Data obtained at 25 °C yield similar Kd values with,
however, greater standard deviation.
reverse direction of the reaction with O2, respectively. The
finding that this plot passes through the origin means that k-2
≈ 0, or in other words, MLCT band decay in the presence of
O2 is not detectably reversible. Therefore, this implies that the
O2-dependent conversion of Fe2+-bound diketonate is the first
irreversible step of the catalytic cycle of Dke1 and is rate-
determining overall.
2′HAP can be seen as a cis-â-keto-enol structure with the
enol double bond fixed in the aromatic ring. It is totally
unreactive toward oxygenative carbon-carbon bond cleavage
by Dke1 and a potentially useful probe through which the steps
involved in dioxygen reduction by the enzyme might be
dissected kinetically. Its ꢀHOMO is similar to that of PD, and its
E0 can be compared to that of PD and 2-acetylcyclohexanone
(ACH), as shown in Figure 5. Except for a partial delocalization
in the aromatic ring, the HOMO of 2′HAP is localized at the
same positions as that of ACH and â-diketonate structures in
general, that is, at the central carbon atom and the nearby oxygen
atoms (see the Supporting Information). Formation of a Dke1-
2′HAP complex gives MLCT bands at 472 and 500 nm (Table
Addition of 200 µM TFPB (=20 × Kd) to air-saturated
solutions of 300 µM Dke1 did not cause O2 depletion at a rate
app
cat
appreciably faster than that expected from k for the conver-
sion of TFPB. This result implies that formation of the Dke1-
TFPB complex does not promote O2 binding to the active site,
which should lead to an observable, stepwise decrease of cO at
2
high enzyme concentrations, nor does it induce significant O2
reduction prior to the rate-limiting step, which should be
mirrored by an increased dioxygen consumption rate under the
described single-turnover conditions. On the other hand, MLCT
band decay was detected only in the presence of O2. These
results together with the fact that the enzyme is not saturable
with dioxygen in the steady state (Figure 1) strongly indicate
that it is the chemical reaction of the binary complex with O2
that is rate-determining.
(14) (a) Denisov, I. G.; Makris, T. M.; Sligar, S. G.; Schlichting, I. Chem. ReV.
2005, 105, 2253-2277. (b) Ost, T. W.; Clark, J.; Mowat, C. G.; Miles, C.
S.; Walkinshaw, M. D.; Reid, G. A.; Chapman, S. K.; Daff, S. J. Am. Chem.
Soc. 2003, 125, 15010-15020. (c) Miles, C. S.; Ost, T. W.; Noble, M. A.;
Munro, A. W.; Chapman, S. K. Biochim. Biophys. Acta 2000, 1543, 383-
407. (d) Daff, S. N.; Chapman, S. K.; Turner, K. L.; Holt, R. A.; Govindaraj,
S.; Poulos, T. L.; Munro, A. W. Biochemistry 1997, 36, 13816-13823.
(e) Sligar, S. G. Biochemistry 1976, 15, 5399-5406.
The Redox Potential of the Cofactor. Why is it that Dke1-
bound substrates are so efficiently primed for reaction with O2,
(15) Endo, K.; Furukawa, M.; Yamatera, H.; Sano, H. Bull. Chem. Soc. Jpn.
1980, 53, 407-410.
9
J. AM. CHEM. SOC. VOL. 127, NO. 35, 2005 12311

A R T I C L E S
Straganz and Nidetzky
Figure 4. Representative absorbance traces at λ3 in single-turnover stopped-flow reactions of Dke1 (220 µM active sites) and PD (A), DFPD (B), TFPD
(C), or DFPB (D), each at a concentration of 200 µM. cO was 260 µM and constant, and all experiments were performed in 20 mM Tris/HCl buffer, pH
2
7.5, at 25 °C. The insets of panels A-D show the part of the time course enlarged in which binary complex formation occurs. Solid pink lines are the
experimental absorbance traces, and the broken lines are nonlinear fits to eq 4 or 3 (inset). Gray lines show the control containing substrate in air-saturated
buffer but lacking Dke1.
conversion concomitant with the evolution of superoxide but
uncoupled from C-O bond formation, which in the case of
2′HAP is greatly disfavored because it would destroy ring
aromaticity. Accelerated autoxidation of the 2′HAP-bound Dke1,
compared with free enzyme, would result, leading to rapid
inactivation. Increased autoxidation of NHMEs has previously
been causally related to a ligand-induced oxidation of Fe2+ 7
.
The observed O2 consumption rate of 8 × 10-6 s-1 for the
Dke1-2′HAP complex (0.2 mM), however, corresponds to the
basal autoxidation rate of unliganded Dke1 in the presence of
O2 (7.5 × 10-6 s-1), which is low but clearly significant. The
finding essentially rules out that E0 of the Fe2+ cofactor is
markedly influenced upon binding of 2′HAP.16
Discussion
Figure 5. Correlation of E0 values of Fe3+-â-diketonate complexes with
the ꢀHOMO values of the corresponding free ligands. The midpoint potentials
are from the literature15 except the E0 value of 2′HAP, which was determined
experimentally (see the Supporting Information). ꢀHOMO values are from
Table 1. Newly introduced abbreviations are: HFPD, 1,1,1,5,5,5-hexa-
fluoropentanedione; TMH, 2,2,6,6-tetramethyl-3,5-heptanedione; DPP, 1,3-
diphenylpropanedione.
The Rate-Limiting Step of the Dke1 Catalytic Cycle.
Results of steady-state kinetic analysis for the reactions of Dke1
with dioxygen in the presence of a â-diketone substrate and in
(16) The rate of dioxygen consumption and concomitant cO -dependent release
2
of metal cofactor from the enzyme into solution as Fe3+ constitutes the
major mechanism of enzyme inactivation in Dke1 (Straganz, G. D.;
Nidetzky, B., Graz University of Technology, Graz, Austria. Unpublished
results, 2005). The causal relationship of steps involved in Dke1 autoxi-
dation is that one-electron reduction of dioxygen by the Fe2+ cofactor
precedes the dissociation of Fe3+, which is released into solution because
in its oxidized form the metal has no detectable affinity for the active site.
Details of the Dke1 inactivation pathway will be reported elsewhere.
2), suggesting that enzyme-bound 2′HAP and ACH share
considerable electronic similarities. Therefore, if one-electron
reduction of O2 by substrate-bound Fe2+ was the rate-limiting
step of the enzymatic reaction, one would expect that formation
of the Dke1-2′HAP complex might trigger Fe2+ into Fe3+
9
12312 J. AM. CHEM. SOC. VOL. 127, NO. 35, 2005

â
-Diketone Cleavage by Fe2+-Dependent Dke1
A R T I C L E S
the absence thereof (referred to here as autoxidation) are
consistent with a ternary complex enzymatic mechanism that
is formally random. The apparent Km values for the substrate
complex in a precatalytic equilibrium at the steady state.
Transient kinetic analysis for the reaction of Dke1 with DFPD
leads to the same conclusion. A two-step mechanism in which
dioxygen binding to the Dke1-substrate complex precedes the
slow reaction would lead to a hyperbolic dependence of k2 on
are independent of cO , which indicates that binding of substrate
2
does not enhance the affinity of Dke1 for dioxygen and vice
versa. Considering catalytic efficiencies for substrate that are
cO . The observed linear correlation is thus consistent with
2
app
∼50 times the corresponding k /Km values for dioxygen, a
binding and reaction taking place in one kinetic step. A very
low binding affinity of O2 to ligand-bound Dke1 is also
emphasized by results of titration experiments showing that
saturation of Dke1 with a slowly reacting substrate (TFPB) or
an inactive analogue (2′HAP) does not cause significant
depletion of dioxygen (260 µM) from the bulk solution. At the
high concentration of enzyme employed (0.3 mM) in the
experiments, a fractional saturation of the Dke1-ligand complex
with O2 of less than 0.03 would have been detectable. Sum-
marizing, enzymatic reaction is thus suggested to take place
from an encounter complex with dioxygen at the rate determined
cat
kinetic pathway in which substrate binds before dioxygen is
clearly preferred under the experimental conditions used in this
study.
The evidence from steady-state and transient kinetic studies
of the Dke1-catalyzed conversion of â-diketone substrates
differing more than 105-fold in reactivity toward dioxygen in
the presence of the enzyme suggests a reaction coordinate in
which the chemical transformation of the enzyme-bound cis-
â-keto-enolate into an alkyl peroxidate intermediate is the rate-
limiting step. Linear free-energy relationship analysis has
app
cat
app
cat
22
revealed the pronounced increase of k (or the mechanisti-
by k
.
cally synonymous transient rate constant k2) in response to a
decrease of ꢀHOMO, which implies that the observable catalytic
rate is controlled by the nucleophilic participation of the bound
substrate in an irreversible chemical reaction with dioxygen.
There is limited precedent among NHMEs regarding the use of
structure-reactivity correlations of kinetic substituent effects
to address mechanistic questions. Electron-withdrawing sub-
The conversion of Dke1-bound â-diketone into the proposed
alkylperoxidate intermediate (Scheme 1) requires the net transfer
of two electrons from the substrate to dioxygen. Considering
that a one-step reduction of triplet dioxygen by the singlet
substrate giving the singlet diketone-peroxide is formally spin-
forbidden, we asked the question of the relative timing of
electron-transfer steps during the Dke1-catalyzed reduction of
dioxygen and the possible role of the Fe2+ cofactor in triggering
the reaction. Assuming that autoxidation of the Fe2+ cofactor
serves as a surrogate of one-electron transfer from Fe2+ to
dioxygen during the normal catalytic cycle of Dke1 as it has
been invoked previously,7 our results do not support a mech-
anism in which the reduction of dioxygen to a superoxide via
the Fe2+-substrate complex is rate-determining. If that was the
case, the E0 of the metal cofactor in the Dke1-2′HAP complex,
which according to studies of homologous metal-â-diketonate
complexes in solution should have a low lying value in
comparison with the Fe2+/Fe3+ couple in the free enzyme, would
be expected to stimulate strongly (that is, by several orders of
magnitude as seen for kacaptp) the rate of enzyme autoxidation.
Considering a nonfavorable overall equilibrium for one-electron
stituents were shown to slow the reactions catalyzed by the Fe3+
-
dependent enzymes protocatechuate dioxygenase17 and catechol-
1,2-dioxygenase,18 but because of kinetic complexity it remained
difficult to consolidate the results into a chemical mechanism
of the enzymatic reduction of dioxygen. The existence of a
spectroscopically accessible intermediate on the reaction profile
of Dke1 provided us with the powerful opportunity to examine
internal steps of the catalytic cycle that involve dioxygen
chemistry. These critical steps are, with the exception of
R-ketoacid-dependent dioxygenases11,13,19 and a recent example
of a homoprotocatechuate-2,3-dioxygenase that utilizes a colored
model substrate,20 spectroscopically inaccessible for mechanistic
study in NHMEs. This article describes for the first time a
comprehensive quantitative structure-activity relationship analy-
sis of the substituent effect on the kinetically unmasked chemical
step of the reaction catalyzed by a non-heme Fe2+-dependent
dioxygenase. The results can be parsed unambiguously into
substrate-derived catalytic factors used by Dke1 to achieve
efficient reduction of dioxygen coupled with C-O bond
formation.21
reduction of dioxygen by Fe2+ 6b
, we emphasize that autoxidation
of unliganded Dke1 took place at a finite rate, implying that an
incremental change of this rate in response to ligand binding
would not have escaped detection by the analytical methods
under the experimental conditions used. Note that, while it is
clear that a potential partial electron transfer from iron to
dioxygen does not determine the catalytic rate, our results do
not exclude the possibility that it has some role in activating
the O2 cosubstrate for reaction. However, putative formation
of the reactive Fe3+O2•- complex would then have to occur in
a precatalytic equilibrium that probably lies far on the side of
Fe2+O2.23 The enzymatic reaction would then take place from
the very small portion of activated complex, in our opinion
leading to an extremely inefficient enzyme.
The Proposed Mechanism of Dioxygen Reduction. The
apparent turnover number of the Dke1-catalyzed conversion of
â-diketones is linearly dependent on the dioxygen concentration,
implying that, under the conditions used and within the limits
of detection by the analytical methods, O2 does not combine
with the substrate-bound enzyme to accumulate a ternary
(17) Walsh, T. A.; Ballou, D. P. J. Biol. Chem. 1983, 258, 14413-14421.
(18) (a) Ridder, L.; Briganti, F.; Boersma, M. G.; Boeren, S.; Vis, E. H.;
Scozzafava, A.; Veeger, C.; Rietjens, I. M. Eur. J. Biochem. 1998, 257,
92-100. (b) Walsh, T. A.; Ballou, D. P.; Mayer, R.; Que, L., Jr. J. Biol.
Chem. 1983, 258, 14422-14427.
(22) Obviously one could define a second-order rate constant kcat′ (µM-1 s-1
)
(19) Grzyska, P. K.; Ryle, M. J.; Monterosso, G. R.; Liu, J.; Ballou, D. P.;
Hausinger, R. P. Biochemistry 2005, 44, 3845-3855.
which incorporates the dependence of the apparent kcat on cO . However,
2
the results for three different â-diketones in Figure 1 show clearly that the
dependence of kcat′ on intrinsic nucleophilic reactivity of the substrate would
be exactly the same as it appears in a correlation that uses apparent kcat
values. The conclusion is also valid for the transient rate constants.
(23) For the metallocenter of various NHMEs the binding of Fe2+ to O2, which
is accompanied by some charge transfer to O2, has been estimated to be
endergonic by ∼10 kcal/mol.4b,6
(20) Groce, S. L.; Miller-Rodeberg, M. A.; Lipscomb, J. D. Biochemistry 2004,
43, 15141-15153.
(21) The clear electronic disposition of the Dke1-substrate complex for reaction
eliminates the possibility that active-site conformational factors, which have
been invoked with R-ketoglutarate-dependent non-heme Fe2+ enzymes,4a
primarily drive the process of O2 reduction by Dke1.
9
J. AM. CHEM. SOC. VOL. 127, NO. 35, 2005 12313

A R T I C L E S
Straganz and Nidetzky
Scheme 2. Suggested Mechanism of Concerted Dioxygen
Reduction and C-O Bond Formation by Dke1a
in Dke1 contributes to catalysis in several ways. First and
perhaps most importantly, it brings about coordination of the
substrate and may activate it for attack on dioxygen, as suggested
earlier in this article. Second, it may provide electrostatic
stabilization to the TS of dioxygen reduction. One possibility
is that reduction is facilitated by coordination of a transient
superoxide species to Fe2+, as it enables the formation of the
[Fe2+O2 T Fe3+O22-] resonance structure, which formally
•-
represents the thermodynamically more favorable two-electron
reduction of dioxygen.6 However, the direct interaction of O2
with iron is debatable. The emerging concept of substrate
a The structure in brackets depicts a hypothetical transition state in which
the first electron has been transferred to O2.
activation as one catalytic factor in the reactions of Fe2+
-
dependent NHMEs is supported by recent reports on quercetin
dioxygenases. In the crystal structure of an anaerobic substrate-
bound complex of Cu2+-containing quercetin dioxygenase the
ligand adopts a distorted conformation which was thought to
reflect some radical character of the bound substrate, induced
by charge transfer from the substrate to Cu2+.9 The proposed
mechanism is that the thus primed substrate attacks molecular
oxygen directly, or dioxygen is reduced by reoxidation of Cu+
and both substrate and superoxide radical recombine. Related
studies have revealed quercetin dioxygenases that can utilize
Fe2+ or Cu2+ as a cofactor,25 and remarkably, they do so with
comparable enzymatic activities. A detailed structural and
mechanistic characterization of Fe2+-dependent quercetin di-
oxygenase is presently not available. However, an interesting
result from electron spin resonance studies is that the Fe2+
cofactor is shielded from interactions with the O2 surrogate nitric
oxide by the enzyme-bound quercetin.26 It suggests the intrigu-
ing possibility that carbon-carbon bond-cleaving dioxygenase
reactions may proceed without any direct interaction of dioxygen
and Fe2+ in cases when substrate activation is involved.
We would like to therefore suggest an alternative mechanism
that involves the direct attack of dioxygen by the Fe2+-bound
substrate in the rate-controlling step, as outlined in Scheme 2.
Coordination of the substrate to Fe2+ such that an orbital mixing
of its HOMO with the HOMO of the metal can take place,
similar to what has been observed in 2,3-dihydroxybiphenyl-
1,2-dioxygenase,4b is proposed to prime the substrate for the
reaction with dioxygen in a now spin-allowed process. We
envision a cataytic reaction profile in which one-electron transfer
from the substrate to dioxygen is thermodynamically unfavorable
and requires coupling to a second, fast step of C-O bond
formation that pulls the reaction toward the virtually irreversible
peroxidate formation. The chemical steps of the Dke1-catalyzed
reaction therefore take place in a concerted manner. The imme-
diate implications of this are that dioxygen reduction and C-O
bond formation must occur in one single catalytic step and
electron transfer and recombination of the intermediates cannot
be kinetically resolved.24 The transition state (TS) of the reaction
is thus expected to resemble, structurally and charge-wise, the
product of the electron transfer from the substrate to dioxygen,
as illustrated in Scheme 2. The subsequent recombination to
the alkylperoxidate intermediate is kinetically and thermody-
namically advantageous and thus expected not to contribute to
TS energy. The lack of measurable activity with 2′HAP which
cannot form the peroxidate is thus understood in the context of
a thermodynamically unfavorable equilibrium for reduction of
dioxygen by the activated substrate that in the case of 2′HAP
cannot be overcome by subsequent C-O bond formation.
The Role of the Cofactor in the Reactions Catalyzed by
Dke1 and Related NHMEs. We think that the Fe2+ cofactor
Acknowledgment. We thank Dr. Mario Klimacek for as-
sistance with stopped-flow measurements and Dr. Claudia
Muresanu for redox potential measurements.
Supporting Information Available: Plots for determination
of the apparent ꢀ1 for the â-diketone substrates used in this study,
results of fluorescence titration, determination of E0 of 2′HAP,
and a graphical representation of the HOMOs of ACH and
2′HAP. This material is available free of charge via the Internet
at http://pubs.acs.org.
JA042313Q
(24) (a) The results point out a major difference between Dke1 and catechol-
cleaving dioxygenases,24b which otherwise share the requirement for
substrate-derived electrons to drive the catalytic reaction. Choice of an
appropriate substrate made it possible to uncouple superoxide production
from subsequent substrate transformations catalyzed by the latter enzymes.
Considering Scheme 2, similar approaches would seem to be elusive for
Dke1. (b) Mayer, R.; Widom, J.; Que, L., Jr. Biochem. Biophys. Res.
Commun. 1980, 92, 285-291.
(25) (a) Bowater, L.; Fairhurst, S. A.; Just, V. J.; Bornemann, S. FEBS Lett.
2004, 557, 45-48. (b) Barney, B. M.; Schaab, M. R.; LoBrutto, R.;
Francisco, W. A. Protein Expression Purif. 2004, 35, 131-141.
(26) Gopal, B.; Madan, L. L.; Betz, S. F.; Kossiakoff, A. A. Biochemistry 2005,
44, 193-201.
9
12314 J. AM. CHEM. SOC. VOL. 127, NO. 35, 2005