Flavonoids as promoters of the (pseudo-)halogenating activity of lactoperoxidase and myeloperoxidase

Article abstract of DOI:10.1016/j.freeradbiomed.2016.06.026

In this study several flavonoids were tested for their potential to regenerate the (pseudo-)halogenating activity (hypothiocyanite formation) of the heme peroxidases lactoperoxidase (LPO) and myeloperoxidase (MPO) after hydrogen peroxide-mediated enzyme inactivation. Several flavonoid subclasses with varying hydroxylation patterns (especially of the flavonoid B-ring) were examined in order to identify structural properties of efficient enzyme regenerators. Kinetic parameters and second-order rate constants were determined. A 3′,4′-dihydroxylated B-ring together with C-ring saturation and hydroxylation were found to be important structural elements, which strongly influence the flavonoid binding and oxidizability by the LPO/MPO redox intermediates Compounds I and II. In combination with docking studies these results allow an understanding of the differences between flavonoids that promote the hypothiocyanite production by LPO and MPO and those that inhibit this enzymatic reaction.

Full text of DOI:10.1016/j.freeradbiomed.2016.06.026

Author’s Accepted Manuscript
Flavonoids as promoters of the (pseudo-
)halogenating activity of lactoperoxidase and
myeloperoxidase
Jana Gau, Paul G. Furtmüller, Christian Obinger,
Martine Prévost, Pierre Van Antwerpen, Jürgen
Arnhold, Jörg Flemmig
www.elsevier.com
PII:
S0891-5849(16)30314-8
http://dx.doi.org/10.1016/j.freeradbiomed.2016.06.026
FRB12917
DOI:
Reference:
To appear in:
Free Radical Biology and Medicine
Received date: 12 April 2016
Revised date: 23 June 2016
Accepted date: 23 June 2016
Cite this article as: Jana Gau, Paul G. Furtmüller, Christian Obinger, Martine
Prévost, Pierre Van Antwerpen, Jürgen Arnhold and Jörg Flemmig, Flavonoids
as promoters of the (pseudo-)halogenating activity of lactoperoxidase and
myeloperoxidase, Free
Radical
Biology
and
Medicine
http://dx.doi.org/10.1016/j.freeradbiomed.2016.06.026
This is a PDF file of an unedited manuscript that has been accepted fo
publication. As a service to our customers we are providing this early version o
the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting galley proof before it is published in its final citable form
Please note that during the production process errors may be discovered which
could affect the content, and all legal disclaimers that apply to the journal pertain

Flavonoids as promoters of the (pseudo-)halogenating activity of lactoperoxidase
and myeloperoxidase
Jana Gaua*, Paul G. Furtmüllerb, Christian Obingerb, Martine Prévostc, Pierre Van Antwerpend,
Jürgen Arnholda, Jörg Flemmiga
a Institute for Medical Physics and Biophysics, Medical Faculty, University of Leipzig, Härtelstraße 16 - 18,
04107 Leipzig, Germany
jana.gau@medizin.uni-leipzig.de, joerg.flemmig@medizin.uni-leipzig.de, juergen.arnhold@medizin.uni-
leipzig.de
b Department of Chemistry, Division of Biochemistry, University of Natural Resources and Life Sciences
(BOKU) Vienna, Muthgasse 18, 1190 Vienna, Austria
paul.furtmueller@boku.ac.at, christian.obinger@boku.ac.at
c Laboratory of Structure and Function of Biological Membranes, Université Libre de Bruxelles, Boulevard
du Triomphe, 1050 Brussels, Belgium
mprevost@ulb.ac.be
d Laboratory of Pharmaceutical Organic Chemistry , Institute of Pharmacy, Université Libre de Bruxelles,
Boulevard du Triomphe, 1050 Brussels, Belgium
Pierre.Van.Antwerpen@ulb.ac.be
Corresponding author:
Jana Gau
Phone: +49 341 9715773
Fax:
+49 341 9715709
E-mail: jana.gau@medizin.uni-leipzig.de
1

Abstract
In this study several flavonoids were tested for their potential to regenerate the (pseudo-)halogenating
activity (hypothiocyanite formation) of the heme peroxidases lactoperoxidase (LPO) and
myeloperoxidase (MPO) after hydrogen peroxide-mediated enzyme inactivation. Several flavonoid
subclasses with varying hydroxylation patterns (especially of the flavonoid B-ring) were examined in
order to identify structural properties of efficient enzyme regenerators. Kinetic parameters and second-
order rate constants were determined. A 3’,4’-dihydroxylated B-ring together with C-ring saturation and
hydroxylation were found to be important structural elements, which strongly influence the flavonoid
binding and oxidizability by the LPO/MPO redox intermediates Compounds I and II. In combination with
docking studies these results allow an understanding of the differences between flavonoids that
promote the hypothiocyanite production by LPO and MPO and those that inhibit this enzymatic reaction.
Keywords
polyphenols; flavonoids; lactoperoxidase; myeloperoxidase; hypothiocyanite; inflammation
Abbreviations
APF, aminophenyl fluorescein; DMSO, dimethyl sulfoxide; DTNB, 5,5’-dithiobis-(2-nitrobenzoic acid);
HC4, hydroxycinnamic acid; HOMO, highest occupied molecular orbital; HX1, 2-(3,5-
bistrifluoromethylbenzylamino)-6-oxo-1H-pyrimidine-5-carbohydroxamic acid; LPO, lactoperoxidase;
LUMO, lowest unoccupied molecular orbital; MPO, myeloperoxidase; OPLS, optimized potentials for
liquid simulations; RMSD, root-mean-square deviation; TNB, 5-thio-2-nitrobenzoic acid
2

Introduction
The heme containing enzymes lactoperoxidase (LPO) and myeloperoxidase (MPO) belong to the
1
immunological relevant chordata peroxidases . Lactoperoxidase is released from mucosal glands and
2,
can be found in secretions like saliva, milk or tears 3. Myeloperoxidase is secreted from invading
neutrophils at inflammatory sites 3, 4. In the buccal cavity, both LPO and MPO play an important role in
innate immune response due to their (pseudo-)halogenating activity 5-7. While MPO is able to oxidize
-
-
-
-
chloride (Cl ), bromide (Br ), iodide (I ) as well as thiocyanate (SCN ) to the corresponding hypo(pseudo-
)halous acid, LPO forms only hypoiodous acid (HOI) and hypothiocyanite ( OSCN) 7, 8. Because of its easy
-
8
-
oxidizability and high abundance in the oral cavity , SCN is the favored substrate both for MPO and
9-12
-
LPO. Thus, the formation of OSCN clearly dominates
since it is in equilibrium with the protonated
-
form hypothiocyanous acid (HOSCN). HOSCN/ OSCN exerts an antimicrobial activity by entering bacterial
biofilms and reacting with free thiol groups predominantly inside the microbes 9, 13
.
-
The formation of OSCN by LPO and MPO is initiated by the activation of the enzymes via two-electron
reduction of hydrogen peroxide (H2O2) (see Scheme 1) 14, 15
.
3

Schematic representation of the –OSCN active cycle and the –OSCN inactive cycle for LPO and MPO.
Upon reaction with hydrogen peroxide both enzymes undergoes a two-electron oxidation to form Compound I. Subsequently
–
SCN– reduces Compound I directly to native enzyme by a two-electron process thereby forming OSCN (–OSCN active cycle).
Alternatively the regeneration of native enzyme from Compound I can occur via two consecutive one-electron reactions
including the formation of Compound II (–OSCN inactive cycle). Furthermore, for LPO the spontaneous transition from
Compound I to Compound I* (oxoiron(IV) combined with radical site in the protein matrix) can also lead to the formation of
Compound II. In the presence of excess H2O2 Compound II can further react to Compound III that either decays to the ferric
state (thereby releasing superoxide) or to ferrous state (thereby releasing dioxygen). The tested flavonoids that reactivate the
(pseudo-)halogenating activity of LPO and MPO are supposed to react as one-electron donors with both Compounds I and II
(indicated by the 2-phenylchromane basic structure) and thus regenerate the native enzyme by resolving accumulated
Compound II.
Thin arrows indicate spontaneous transitions while standard arrows represent the reaction of LPO/MPO redox intermediates
-
either with H2O2 or SCN . Supposed influenced reactions by flavonoids within the catalytic cycles are marked with thick and
hatched arrows. Indicated rate constants for LPO and MPO are marked by normal or italic letters, respectively. Rate constants
were taken from the literature: 12, 16, 17 (* reaction conditions: pH 7.0; 15 °C); 14, 18 (** reaction conditions: pH 7.0; 25 °C).
4

+•
In this reaction the native enzyme is oxidized to Compound I (i.e. oxoiron(IV)porphyryl radical, Por-
3
Fe(IV)=O) . Subsequently, Compound I can be reduced back to the ferric enzyme by two-electronic
-
-
oxidation of SCN to OSCN. Alternatively, Compound I can also be reduced via two consecutive one-
electron steps in the peroxidase cycle 19, whereby the resolution of the enzyme redox intermediate
Compound II (i.e. oxoiron(IV), Por-Fe(IV)-OH) is the rate determining step 6, 20. Compound II does not
contribute to the (pseudo-)halogenating activity of LPO and MPO and accumulates in the absence of
21
suitable one-electron donors . Additionally in the presence of H2O2 excess Compound II can further
•-
react to Compound III (i.e. oxyperoxidase, Por-Fe(III)ꢀO2 /Por-Fe(II)ꢀO2) followed by the slow conversion
to the ferric state (accompanied by the release of superoxide) or spontaneous formation of ferrous
enzyme that can be oxidized to Compound II by H2O2 18. For LPO the (pseudo-)halogenating activity can
also be disrupted by the formation of Compound I* that is characterized by a shift of the radical site from
8
the porphyrin ring to the protein moiety . Compound III and ferrous MPO/LPO as well as Compound I*
22
-
are not able to contribute to the formation of OSCN . Yet an efficient halogenating activity of LPO and
MPO is important to inhibit bacterial growth and metabolism in the buccal cavity and upper airway fluids
23-25. Furthermore, H2O2 that is secreted by oral pathogens like Streptococci and Lactobacilli is rapidly
13
.
consumed by LPO (and MPO), thus protecting mammalian cells from the cytotoxic effect of H2O2
Oxidative destruction of LPO and MPO as well as enzyme inactivation by Compound II accumulation or
Compound I* formation may be associated with recurrent bacterial damage and inflammatory processes
like gingivitis, stomatitis or caries 26-28
.
Flavonoids are ubiquitously occurring secondary plant components with well-known health promoting
properties 29, 30. Although the enhancing effect of several flavonoids on the OSCN production by LPO has
-
19, 31,
been demonstrated
32, there is a lack of data concerning the relationships between structural
compositions of flavonoids, arrangement of flavonoids in the substrate channel of LPO, and the (pseudo-
)halogenating activity of LPO. For MPO, only few data exist about the enhancement of the halogenating
5

22, 33,
activity by (-)-epicatechin and some derivatives
34, but other flavonoids were not tested
systematically. In this study we addressed the questions how variations in the C-ring composition and
-
the B-ring hydroxylation pattern of flavonoids promote the OSCN generation in both LPO and MPO.
-
Kinetic data for the OSCN formation obtained with 5,5'-dithiobis-(2-nitrobenzoic acid) (TNB) were
compared with second-order rate constants for flavonoid oxidation by Compounds I and II as well as with
docking studies.
Experimental Section
Materials
Lactoperoxidase from bovine milk (about 80 % purity according to the A412/A280 ratio) was obtained from
Sigma-Aldrich, Steinheim, Germany as a lyophilized powder (ꢀ 200 U/mg). The high homology of bovine
6
LPO compared to other species makes the results well transferable to the human enzyme . Human
myeloperoxidase (99.5 % purity according to the A430/A280 ratio) was purchased from Planta Natural
Products, Vienna, Austria. Stock aliquots (5 µM) of the enzymes were prepared in phosphate buffered
saline (PBS; 10 mM; pH 7.4) for LPO or in phosphate buffer (10 mM; pH 7.4) for MPO, respectively, and
stored at -25 °C until usage. Unless otherwise stated the final enzyme concentration for both enzymes
was 5 nM during the steady-state kinetic measurements.
-
The formation of OSCN by LPO and MPO was quantified by following the generation of Ellman’s reagent
35
(5,5'-dithiobis-(2-nitrobenzoic acid), DTNB) from 5-thio-2-nitrobenzoic acid (TNB) at 412 nm using a
multiplate reader (Tecan Infinite 200 Pro, Männerdorf, Switzerland). The final reaction mixture contained
50 µM TNB as well as 2 mM thiocyanate and 80 µM hydrogen peroxide. A chloride concentration of
140 mM and 50 mM was chosen, respectively, for the LPO and MPO assay. For all measurements the
hydrogen peroxide working solution was freshly prepared each day. H2O2, sodium chloride (NaCl, ꢀ 99 %)
and Ellman’s reagent (DTNB, ꢀ 98 %) for TNB preparation were obtained from Sigma-Aldrich,
6

Taufkirchen, Germany. Sodium phosphate monobasic monohydrate (ꢀ 99 %) and sodium phosphate
dibasic dihydrate (ꢀ 99 %) for buffer preparation were purchased from Merck, Darmstadt, Germany and
Sigma-Aldrich, respectively.
To quantify the formation of the MPO-derived HOCl the non-fluorescent dye aminophenylfluorescein
(APF, Biomol GmbH, Hamburg, Germany) was used. The final reaction mixture contained 5 nM MPO,
-
-
50 mM Cl , 10 µM APF, 80 µM H2O2 and 0 - 2 mM SCN .
Apigenin (ꢀ 95 %), (+)-catechin hydrate (ꢀ 96 %), chrysin (ꢀ 98 %), (+)-dihydromyricetin (ꢀ 98 %), (-)-
epicatechin (ꢀ 90 %), (-)-epigallocatechin (ꢀ 95 %), (ꢁ)-eriodictyol (ꢀ 95 %), (-)-hesperetin (ꢀ 98 %),
homovanillic acid (ꢀ 97 %), (ꢁ)-naringenin (ꢀ 95 %), (+)-pinobanksin (ꢀ 95 %), (+)-pinocembrin (95 %) as
well as (+)-taxifolin (ꢀ 90 %) were obtained from Sigma-Aldrich. (+)-Afzelechin (> 95 %), (+)-aromadendrin
(> 95 %), (-)-butin (> 98 %) and kaempferol (ꢀ 98 %) were purchased from ChemFaces, Wuhan, China.
Luteolin (ꢀ 95 %) and quercetin (ꢀ 95 %) were bought from Biopurify Phytochemicals Ltd., Chengdu,
China. (-)-Homoeriodictyol (ꢀ 99 %) was obtained from Cymit Quimica S.L., Barcelona, Spain.
For all tested flavonoids a 5 mM stock solution was prepared in dimethyl sulfoxide (DMSO; Sigma-
Aldrich) and further diluted with buffer. The final flavonoid concentrations ranged between 0.01 and
20 µM. The final DMSO concentration did not exceed 1 % (v/v). Control measurements did not show an
effect of DMSO concentration up to 1 % on the enzyme activity.
Methods
Determination of enzyme kinetic parameters
-
The kinetic measurements with TNB (detection of OSCN formation) were performed in PBS (for LPO) or
chloride free phosphate buffer (for MPO) at pH 7.4 and 37 °C. Kinetic parameters (Michaelis-Menten
constant (KM), maximal rate (Vmax), turnover number (kcat) and specificity constant (KM/kcat)) were
calculated via Lineweaver-Burk plots according to previous studies 31. For flavonoids with inhibitory
effects the half maximal inhibitory concentration (IC50) was calculated by using exponential curve fitting.
7

To quantify the rate of HOCl formation by MPO the increase of the fluorescence intensity after APF
33
oxidation by HOCl (monitored at 525 nm after excitation at 488 nm ) was plotted against the time and
pseudo-first order rate constants (kobs) were calculated by exponential curve fitting.
All measurements were performed at least in triplicates. As KM and Vmax values were calculated via
Lineweaver-Burk plots, the indicated standard deviations correspond to the coefficient of determination
for the linear curve fit. Significant differences between the data were determined by using a Student’s
two tailed t-test. This statistic test has been chosen as all data were successfully tested for Gaussian
distribution (Shapiro-Wilk test). The indicated asterisks correspond to the level of significance (p < 0.05
(*), p < 0.01 (**), p < 0.001 (***)), respectively.
Docking calculations
The X-ray structures of human myeloperoxidase complexed to 2-(3,5-bistrifluoromethylbenzylamino)-6-
36
oxo-1H-pyrimidine-5-carbohydroxamic acid (HX1) (PDB ID: 4C1M)
and of bovine lactoperoxidase
complexed to hydroxycinnamic acid (HC4) (PDB ID: 3UBA) 37 were used for docking studies with selected
flavonoids. Only one monomer (chains A and C) of MPO was kept. The water and other ligand molecules
were removed from the X-ray structures. The charge of the iron was set to +3, which corresponds to the
native enzyme state 3. In both enzymes Asp94/Asp108 and Glu242/Glu258 were covalently linked to the
methyl groups of rings A and C of the heme (ester bonds). In MPO Met243 was also covalently linked to
the vinyl group of the heme C ring creating a sulfonium cation3, 6. Following a pKa calculation, Asp98 and
Asp112 were protonated in MPO and LPO, respectively. All structures were prepared by using the
Protein Preparation Wizard protocol in the Schrodinger software (www.schrodinger.com). The ligand
input files were prepared as following: The initial 3D structures of the ligands were generated using the
Ligprep module from Schrodinger and the ligand partial charges were ascribed using the OPLS (optimized
potentials for liquid simulations) force-field. The program Epik was used to predict the different
8

protonation states of all ligands. Docking was performed using the program Glide (version 6.6) with a
rigid receptor and the Glide XP docking protocol with scoring function. The remaining parameters were
set to their default values. The reliability of the docking process was first assessed docking HC4 and HX1
in the binding site of LPO (PDB code 3UBA) and MPO (PDB code 4C1M), respectively. Luteolin, (-)-
epicatechin (RR enantiomer) and (+)-pinocembrin were applied in the docking studies.
Stopped-flow kinetic measurements
Second-order rate constants for the reaction of selected flavonoids with LPO as well as MPO
Compounds I and II were determined by using an SX-18 MV apparatus from Applied Photophysics. All
measurements were performed in 10 mM phosphate buffer, pH 7.4 at 25 °C.
If not otherwise stated 0.5 µM enzyme were pre-mixed with a 2- (LPO) or 10-fold (MPO) excess of H2O2
to pre-form Compound I. After an incubation time of 100 ms (LPO) or 40 ms (MPO) the flavonoid was
injected and absorbance changes at the Soret maximum of Compound II (431 nm for LPO 16 and 456 nm
12
for MPO ) were followed. The tested flavonoid concentrations varied between 0 and 20 µM. Since the
direct determination of the rate constant for the reaction of the tested substances with MPO
Compound II was not possible, the enzyme was pre-mixed with 0.5 µM homovanillic acid and a 10-fold
38
excess of H2O2 in the aging loop for 30 s to pre-form Compound II . Afterwards, different amounts of
flavonoid (10 - 150 µM) were added. For each substrate concentration at least three independent
measurements were performed and the resulting kinetic curves were averaged.
All measurements were performed in triplicate. The given standard deviations correspond to the R2
value of the respective linear fit.
9

Results
Effect of flavonoids on the -OSCN production by LPO and MPO
Setup of the test system
-
For all measurements 2 mM SCN were used to simulate physiological conditions in the buccal cavity. For
-
MPO the Cl concentration was also adapted to the physiological situation in human saliva by using
50 mM. As shown by control measurements with APF (Suppl. Fig. 1) no MPO-mediated HOCl production
takes place under these conditions, reflecting the preference of MPO for SCN 12. For LPO, which is not
-
16
-
-
able to oxidize chloride , control measurements exhibited no effect of Cl up to 140 mM on the OSCN
formation (see Suppl. Fig. 2).
A H2O2 concentration of 80 µM was used to simulate a pro-inflammatory situation. At these
-
concentrations for both enzymes a significant decrease of the OSCN-producing enzyme activity was
observed as compared to the presence of 20 µM H2O2 (Fig. 1). Next the ability of flavonoids to
-
regenerate the OSCN production by LPO and MPO was tested.
10

Figure 1
3.0x10-3
A
2.0x10-3
1.0x10-3
**
***
0.0
0
70
140
210
Hydrogen peroxide, µM
3.0x10-3
2.0x10-3
B
***
1.0x10-3
***
0.0
0
70
140
210
Hydrogen peroxide, µM
H2O2 dependent LPO- and MPO-mediated –OSCN formation rate related to the H2O2 concentration. The LPO and MPO systems
are shown in A and B, respectively. The effect of increasing H2O2 concentrations was tested at 37 °C and pH 7.4 in the presence
-
-
of 5 nM enzyme, 2 mM SCN and 50/140 mM Cl (MPO/LPO). For both enzymes H2O2 concentrations over 20 µM led to a
–
-
significant decrease of the pseudo-halogenating enzyme activity as indicated by lower relative OSCN formation rates ( OSCN
-
formation divided by the applied H2O2 concentration). At H2O2 concentrations over 100 µM nearly no OSCN formation was
observed. Mean and standard deviation of n = 3 experiments are given. **: p < 0.01, ***: p < 0.001.
Effect of B ring substituents
The most extensive studies were performed on flavanones and flavanols as these subclass compounds
are available with nearly all relevant B ring hydroxylation patterns (Fig. 2E and F).
11

Figure 2
180
1.5
1.2
0.9
0.6
0.3
0.0
180
120
60
A
B
***
***
120
60
0
***
***
***
0
n
n
n
n
hi
n
n
n
ol
ty
i
i
ol
ty
hi
hi
hi
c
c
ec
el
But
)-
(-
atec
c
atec
embri
)-
(-
peretin
odi
odi
Catec
oc
l
)-
Afz
)-
Eri
)-
noc
Naringen
Hes
Epi
)-
(-
(+
)-
gal
Pi
(+
(
)-
(
Epi
)-
(-
Homoeri (+
)-
(-
9000
6000
3000
0
10
8
9000
D
C
*
*
**
6000
3000
0
6
***
*
**
4
2
0
n
n
n
n
hi
n
n
n
ol
ty
i
i
ol
ty
hi
hi
hi
c
c
ec
el
But
atec
c
atec
)-
(-
embri
peretin
odi
odi
Catec
oc
l
)-
Afz
)-
Eri
)-
noc
Naringen
Hes
)-
(-
Epi
)-
(-
(+
)-
gal
Pi
)-
(+
(
(
Epi
Homoeri (+
)-
)-
(-
(-
R3'
B
R3'
B
E
F
R4'
R5'
R4'
R5'
HO
O
HO
O
C
A
C
A
OH
R5
O
R5
R3‘
H
R4‘
R5‘ R5
R3‘ R4‘ R5‘ R5
OH
OH
H
H
H
H
H
H
OH
OH
OH
H
(+)-Afzelechin
(+)-Catechin
H
OH
H
H
H
OH
OH
OH
(ꢁ)-Naringenin
(ꢁ)-Eriodictyol
OH
OH OH
OH OH
(-)-Homoeriodictyol OCH3 OH
(-)-Epicatechin
(-)-Butin
OH
H
OH
H
OH OH OH OH
(-)-Epigallocatechin
(+)-Pinocembrin
(-)-Hesperetin
OH
OH
OH
OCH3
Specificity constants (kcat/KM) and IC50 values for the reaction of different flavanones and flavanols with LPO and MPO. A and B
correspond to measurements with the LPO system, C and D indicate the data obtained for the MPO system. Specificity constants
are shown as black bars, IC50 values are illustrated as belted bars. Chemical structure properties of the tested flavonoids are
given in E and F. In the group of the flavanones (A and C) for both enzymes 3’,4’-dihydroxylated (ꢁ)-eriodictyol exhibits the
12

highest kcat/KM value. An unhydroxylated B-ring ((+)-pinocembrin) as well as a 4’-methoxylation ((-)-hesperetin) leads to an
inhibition of the enzyme activity. Concerning flavanols (B and D) again 3’,4’-dihydroxylated compounds are found to be efficient
substances to increase the enzymatic activity. The cis-trans isomers (-)-epicatechin and (+)-catechin showed differences in their
reactivation potential of LPO and MPO. Calculated data of n = 3 - 6 experiments. *: p < 0.05, **: p < 0.01, ***: p < 0.001.
The specificity constant (kcat/KM, black bars in Fig. 2) was used to quantify the regenerating effect of the
tested flavonoids while for those substances, which emerged as inhibitors of the (pseudo-)halogenating
enzyme activity, the IC50 value was determined (Fig. 2, belted bars). For the underlying KM and kcat values
see Tabs. 1 and 2.
Table 1
Determined kinetic parameters for LPO.
Substance
class
KM,
kcat
kcat/KM
IC50
µM
R3’
R4’
R5’ R5 Compound
µM
s-1
µM-1 s-1
H
H
H
H
OH (+)-Pinocembrin
-
-
-
0.37 0.10
H
OH
OH
OCH3
OH
OH
0.19 0.00
0.36 0.01
-
(ꢁ)-Naringenin
(ꢁ)-Eriodictyol
19.02 ꢁ 0.02
102.5 ꢁ 0.3
OH
OH
H
H
H
H
52.40 ꢁ 0.68
-
146.4 ꢁ 3.8
-
OH (-)-Hesperetin
0.99 0.24
OCH3 OH
OH (-)-Homoeriodictyol 1.01 0.04
20.39 ꢁ 0.87
24.15 ꢁ 0.94
20.2 ꢁ 1.7
145.4 ꢁ 11.3
OH
OH
H
(-)-Butin
0.17 0.01
H
H
H
H
H
OH Chrysin
OH Apigenin
OH Luteolin
-
-
-
0.70 0.30
H
OH
OH
0.11 0.01
0.19 0.02
13.07 ꢁ 0.67
91.59 ꢁ 8.61
122.3 ꢁ 12.55
472.1 ꢁ 88.8
OH
H
H
H
H
H
OH (+)-Pinobanksin
OH (+)-Aromadendrin
OH (+)-Taxifolin
-
-
-
-
H
OH
OH
OH
3.44 0.55
0.77 0.02
23.30 ꢁ 3.72
48.17 ꢁ 1.32
30.04 ꢁ 1.57
6.8 ꢁ 2.2
63.0 ꢁ 3.4
23.6 ꢁ 2.5
OH
OH
OH OH (+)-Dihydromyricetin 1.27 0.07
H
OH
H
OH Kaempferol
0.54 0.05
50.10 ꢁ 1.44
92.6 ꢁ 5.3
13

OH
OH
H
OH Quercetin
1.68 0.18
101.95 ꢁ 10.98 60.7 ꢁ 13.1
H
OH
OH
OH
OH
H
H
H
OH (+)-Afzelechin
0.19 0.020
0.67 0.00
0.61 0.02
0.81 0.01
8.80 ꢁ 0.95
66.82 ꢁ 0.25
41.40 ꢁ 1.52
43.57 ꢁ 0.48
47.0 ꢁ 10.1
100.3 ꢁ 0.7
68.4 ꢁ 5.0
53.7 ꢁ 1.2
OH
OH
OH
OH (+)-Catechin
OH
(-)-Epicatechin
OH OH (-)-Epigallocatechin
Table 2
Determined kinetic parameters for MPO.
Substance
class
KM,
kcat
kcat/KM
IC50
µM
R3’
R4’
R5’ R5 Compound
µM
s-1
µM-1 s-1
H
H
H
H
OH (+)-Pinocembrin
-
-
-
0.77 0.26
H
OH
OH
OCH3
OH
OH
0.05 0.00
0.03 0.00
-
(ꢁ)-Naringenin
(ꢁ)-Eriodictyol
202.27 ꢁ 3.88
4405.3 ꢁ 169.2
OH
OH
H
H
H
H
181.34 ꢁ 8.79
-
5586.6 ꢁ 541.9
-
OH (-)-Hesperetin
6.53 1.50
OCH3 OH
OH (-)-Homoeriodictyol 0.01 0.00
18.20 ꢁ 0.17
39.49 ꢁ 4.16
3338.9 ꢁ 64.1
705.0 ꢁ 148.6
OH
OH
H
(-)-Butin
0.06 0.01
H
H
H
H
H
OH Chrysin
OH Apigenin
OH Luteolin
-
-
-
0.19 0.08
H
OH
OH
0.05 0.00
0.05 0.00
43.00 ꢁ 1.16
131.52 ꢁ 2.53
818.3 ꢁ 44.2
OH
2656.0 ꢁ 102.0
H
H
H
H
H
OH (+)-Pinobanksin
OH (+)-Aromadendrin
OH (+)-Taxifolin
-
-
-
2.11 0.55
H
OH
OH
OH
0.02 0.00
0.05 0.00
79.13 ꢁ 2.25
131.18 ꢁ 2.52
60.78 ꢁ 8.23
3268.0 ꢁ 185.6
2777.8 ꢁ 106.7
1763.7 ꢁ 477.6
OH
OH
OH OH (+)-Dihydromyricetin 0.03 0.01
H
OH
OH
H
H
OH Kaempferol
0.17 0.03
40.62 ꢁ 6.13
200.48 ꢁ 3.85
235.7 ꢁ 71.1
OH
OH Quercetin
0.05 0.00
4048.6 ꢁ 155.5
H
OH
OH
H
H
OH (+)-Afzelechin
0.03 0.00
168.62 ꢁ 4.03
5509.6 ꢁ 263.4
OH
OH (+)-Catechin
0.04 0.00
198.08 ꢁ 10.34 5102.0 ꢁ 532.7
14

OH
OH
OH
OH
H
OH
0.03 0.00
(-)-Epicatechin
183.52 ꢁ 7.14
41.68 ꢁ 7.86
6756.8 ꢁ 525.7
2229.9 ꢁ 841.1
OH OH (-)-Epigallocatechin
0.02 0.00
Concerning LPO in case of flavanones (Fig. 2A) the 3’,4’-dihydroxylated compounds (ꢁ)-eriodictyol
(kcat/KM: 146.4 ꢁ 3.8 µM-1 s-1) and (-)-butin (kcat/KM: 145.4 ꢁ 11.3 µM-1s-1) exhibited the strongest effect
while the kcat and KM values obtained for (-)-butin were about 50 % lower (see Tab. 1). Compared to (ꢁ)-
eriodictyol the C4’ monohydroxylated derivate (ꢁ)-naringenin exhibited a 30 % lower kcat/KM value
(105.5 ꢁ 0.26 µM-1 s-1). (-)-Homoeriodictyol (C3’ methoxy group; C4’ hydroxyl group) showed a
considerably higher KM value (1.01 0.04) while its kcat value (20.39 ꢁ 0.87) was comparable to that of
(ꢁ)-naringenin. (-)-Hesperetin (C3’ hydroxyl group; C4’ methoxy group) inhibited the (pseudo-
)halogenating LPO activity with an IC50 value of 0.99 ꢁ 0.24 µM. Likewise, (+)-pinocembrin (no
hydroxylation at the B ring) also acted as an inhibitor of the enzyme with an IC50 value of 0.37 ꢁ 0.10 µM.
With regard to flavanols (Fig. 2B) again the 3’,4’-dihydroxylated compounds exhibited the strongest
effects on the (pseudo-)halogenating LPO activity. The highest kcat/KM value was found for (+)-catechin
(100.3 ꢁ 0.7 µM-1s-1) followed by the cis-isomer (-)-epicatechin (68.4 ꢁ 5.0 µM-1s-1). Again the C4’
monohydroxylated compound (+)-afzelechin exhibited a significantly reduced kcat/KM value (47.0 ꢁ
10.1 µM-1s-1) compared to the 3’,4’-dihydroxylated derivatives as a result of a strongly reduced kcat value
(see Tab 1). In case of the 3’,4’,5’-trihydroxylated compound (-)-epigallocatechin a considerably
increased KM value (the kcat value was in the range of that observed for (-)-epicatechin) also led to a
lower kcat/KM ratio (53.7 1.2 µM-1s-1) as compared to the 3’,4’-dihydroxylated flavanols.
These flavonoids were also tested for their regeneration potential on the MPO system (Fig. 2C and 2D).
The general trend of the compounds, either as promoters or inhibitors of the (pseudo-)halogenating
activity, was comparable to the results observed with LPO. Yet, several significant differences were found
as well: For flavanones (Fig. 2C) the 3’,4’-dihydroxylated compound (ꢁ)-eriodictyol again exhibited the
highest regeneration potential (kcat/KM: 5586.6 ꢁ 541.9 µM-1 s-1). In contrast, the other 3’,4’-
15

-
dihydroxylated flavanone, namely (-)-butin, had nearly no promoting effect on the OSCN production by
MPO. This behavior is in striking contrast to that obtained for LPO. The 4’-monohydroxylated compound
(ꢁ)-naringenin (see Tab 2) exhibited an about 21 % lower kcat/KM value (4405.3 ꢁ 169.2 µM-1s-1) compared
to (ꢁ)-eriodictyol, which is quite comparable to the decrease observed in LPO. (-)-Homoeriodictyol (C3’
methoxy group; C4’ hydroxyl group) exhibited a very low KM value of about 0.01 µM as well as a
significantly reduced kcat value (about 91 % lower compared to (ꢁ)-naringenin). This is in total contrast to
the results obtained for LPO. Like in LPO inhibitory effects were observed for MPO, i.e. for (+)-
pinocembrin (no hydroxylation at the B-ring) and (-)-hesperetin (C3’ hydroxyl group, C4’ methoxy
group). (+)-Pinocembrin emerged as the most efficient inhibitor (IC50: 0.77 ꢁ 0.26 µM) while considerably
higher concentrations of (-)-hesperetin (IC50: 6.53 ꢁ 1.50 µM) were needed.
In the flavanol subclass (Fig. 2D) again the 3’,4’-dihydroxylated compounds (-)-epicatechin and (+)-
catechin emerged as the most efficient MPO substrates, which is in line with the results obtained for
LPO. However, in contrast to LPO in MPO the cis-isomer (-)-epicatechin had a 32 % higher kcat/KM value
(6756.8 ꢁ 525.7 µM-1s-1) as compared to the trans-isomer (+)-catechin (5102.0 ꢁ 532.7 µM-1s-1). For the
C4’ monohydroxylated compound (+)-afzelechin a kcat/KM value of 5509.6 ꢁ 236.4 µM-1s-1 was calculated
to be in between those of (-)-epicatechin and (+)-catechin, which is in line with the results obtained for
(ꢁ)-naringenin and (ꢁ)-eriodictyol (see Fig. 2C). In the presence of an additional C5’ hydroxylation next to
a 3’,4’-dihydroxylation, as in (-)-epigallocatechin, again a considerably decreased kcat/KM ratio
(2229.6 ꢁ 841.1 µM-1s-1) was obtained, resulting from a strongly reduced kcat value.
Effect of A and C ring modifications
As 3’,4’-dihydroxylated flavonoids appear to be the best reactivators of LPO and MPO, we next
systematically studied compounds from different flavonoid subclasses with a 3’,4’-dihydroxylated B-ring.
The calculated specificity constants are summarized in Fig. 3A for LPO and in Fig. 3B for MPO,
respectively.
16

Figure 3
600
OH
B
A
OH
400
**
R =
**
**
**
**
**
200
HO
O
C
R
HO
O
C
R
A
A
0
OH
O
OH
O
n
n
n
i
fol
n
n
hi
i
ol
ty
ti
hi
c
i
ax
atec
c
odi
Luteol
(
uerce
T
)-
(+
Catec
(-
Q
Eri
)-
Epi
)-
)-
(+
Luteolin
(ꢁ)-Eriodictyol
HO
O
R
HO
O
C
R
8000
B
A
C
A
OH
OH
6000
4000
2000
OH
O
OH
O
***
**
***
*
(+)-Taxifolin
Quercetin
HO
O
R
HO
O
C
R
A
C
A
OH
OH
n
n
n
i
fol
n
n
hi
i
ol
ty
ti
OH
OH
hi
c
i
ax
atec
c
odi
Luteol
(
uerce
T
)-
Catec
(-
Q
Eri
)-
)-
(+
Epi
)-
(+
(+)-Catechin
(-)-Epicatechin
Specificity constants (kcat/KM) for the reaction of 3’,4’-dihydroxylated compounds from different flavonoid subclasses with LPO
and MPO. Flavones are marked in black, flavanones in dark gray, flavonols in gray, flavanonols in light gray and flavanols are
indicated by white bars. Results for LPO are shown in A and for MPO in B, respectively. Appropriate chemical structures are
given in C. For LPO the flavone luteolin shows the highest kcat/KM value followed by its non-planar pendant (ꢁ)-eriodictyol. The
lowest effect on the enzyme activity was observed for the flavanonol (+)-taxifolin. Regarding MPO the highest enzyme
reactivation potential was calculated for the flavanol (-)-epicatechin and the flavanone (ꢁ)-eriodictyol. For luteolin, the most
promising candidate for LPO, the lowest effect was determined for the MPO system. Calculated data of n = 3 - 6 experiments.
*: p < 0.05, **: p < 0.01, ***: p < 0.001.
In Fig. 3 flavones are indicated by black bars while flavanones are shown in dark gray, flavonols in gray,
flavanonols in light gray and flavanols as white bars. Regarding the flavanol subclass, both the trans- ((+)-
catechin) and the cis-isomer ((-)-epicatechin) were tested. The flavanone (-)-butin (no hydroxylation at
the C5 position of the A ring) was omitted from the comparison as no compounds with such a
17

modification were available for other flavonoid subclasses. The corresponding kcat and KM values are
given in Tabs. 1 and 2.
For LPO (Fig. 3A) the flavone luteolin (planar, no C3 hydroxylation) showed the highest kcat/KM ratio
(472.1 ꢁ 88.8 µM-1s-1). A 69 % lower kcat/KM ratio (146.4 3.8 µM-1s-1) was found for the flavanone (ꢁ)-
eriodictyol (non-planar, no C3 hydroxylation). The flavonol quercetin (planar, C3 hydroxylation) showed
an even lower kcat/KM value (60.7 13.1 µM-1s-1) than luteolin as a result of a much higher KM value (see
Tab. 1). The flavanonol (+)-taxifolin (63.0 3.4 µM-1s-1) and the flavanols (+)-catechin (100.3 ꢁ 0.7 µM-1s-
1) and (-)-epicatechin (68.4 ꢁ 5.0 µM-1s-1) (all non-planar, C3 hydroxylation) exhibited comparably low
specificity constants. Thus, the influence of an additional C4 oxo group (like in flavanones, flavonols,
flavones and flavanonols) is unclear but has apparently no crucial influence on the reactivating potential
of the tested flavonoids.
Considerably different results were observed for MPO (Fig. 3B): The highest kcat/KM values were
calculated for the non-planar flavanols (+)-catechin (5102.0 ꢁ 532.7 µM-1s-1) and (-)-epicatechin
(6756.8 ꢁ 525.7 µM-1s-1) (C3 hydroxylation) as well as for the non-planar flavanone (ꢁ)-eriodictyol
(5586.6 ꢁ 541.9 µM-1s-1) (no C3 hydroxylation). For the planar flavone luteolin (no C3 hydroxylation) a
considerably lower kcat/KM value was found (2656.0 ꢁ 102.0 µM-1s-1) while for the planar flavonol
quercetin (C3 hydroxylation) a higher value was obtained (4048.6 ꢁ 155.5 µM-1s-1). In this respect the low
effect of (+)-taxifolin (2777.8 ꢁ 106.7 µM-1s-1) compared to quercetin is surprising as this non-planar
flavanonol also contains a C3 hydroxylation and thus resembles the flavanols (+)-catechin and (-)-
epicatechin.
Docking studies
Based on former kinetic measurements docking studies were performed with the most promising
substances (luteolin for LPO; (-)-epicatechin for MPO) and with the LPO/MPO inhibiting compound (+)-
18

pinocembrin. For the best docked poses of these flavonoids at the binding site of LPO (PDB code 3UBA)
and MPO (PDB code 4C1M) 3D structures are shown in Fig. 4A for LPO and 4B for MPO.
Figure 4
A
B
His109
I
I
His95
Gln105
Gln91
Glu102
Glu116
Glu258
His109
II
II
His95
Gln91
Arg255
Glu102
Glu116
His109
III
III
Gln91
Arg255
Arg239
Best poses for docking calculations with LPO (shown in (A); PDB code 3UBA) and MPO (shown in (B); PDB code 4C1M)
generated by Glide?. The docked ligands are illustrated as green sticks and numbered in case of LPO and MPO, respectively as
following: I) luteolin; II) (-)-epicatechin and III) (+)-pinocembrin. The enzymatic heme group is depicted as gray sticks with the
central iron ion shown as a light blue sphere. Hydrogen bonding amino acids are named and illustrated as blue sticks. The
ligand amino acid interactions are marked by black dashed lines. The images were generated using PyMol version, 0.99rc6.
19

The homologous and conserved amino acids glutamine (LPO: Gln105; MPO: Gln91), histidine (LPO:
His109; MPO: His95), glutamic acid (LPO: Glu116; MPO: Glu102) and arginine (LPO: Arg255; MPO:
Arg239) were found to form non-covalent interactions with the flavonoids. In case of LPO, the binding of
luteolin and (-)-epicatechin presents notable differences: The B-ring of luteolin gets slightly deeper in
the distal heme pocket, which allows a better stacking with the pyrrole ring D. The C3’ hydroxyl group of
this flavonoid forms three hydrogen bonds with His109, Gln105 and Glu258 while the C4’ hydroxyl group
only forms one hydrogen bond to His109. In case of (-)-epicatechin Arg255, which is not involved in
interactions of LPO with luteolin, forms a hydrogen bond to the C3’ hydroxyl moiety while the C4’
hydroxyl also interacts with His109.
For MPO the binding of luteolin and (-)-epicatechin shows strong similarities: The flavonoid B-ring of
both compounds stacked onto the pyrrole ring D inside the distal heme cavity. In both cases the C3’
hydroxyl group interacts with Glu91/His95 while the C4’ hydroxyl moiety hydrogen bonds to His95. One
difference lies in the flipping of the B-ring in (-)-epicatechin relative to luteolin that makes the C3’
hydroxyl group of one compound almost superimposed onto the C4’ hydroxyl moiety of the other.
As revealed by the docking studies, the C7 hydroxyl moiety of the flavonoid A-ring also forms hydrogen
bonds with amino acids extending further away from the heme, namely Glu116 for LPO and Glu102 for
MPO 39.
The inhibitory compound (+)-pinocembrin exhibits completely different binding properties at the active
site of the two enzymes compared to the reactivating substances luteolin and (-)-epicatechin: The
phenyl moiety of the chroman part stacks onto the heme pyrrole ring D, and the C5 and C7 hydroxyl
oxygens are involved in the formation of hydrogen bonds with His109/Arg255 in LPO and Gln91/Arg239
in MPO.
The predicted values of the free energy (DG) (see Tab. 3 A for LPO and B for MPO) show a score ranking
as following: luteolin (DG = - 8.2 kcal/mol) > (-)-epicatechin (DG = - 7.5 kcal/mol) > (+)-pinocembrin (DG
20

= - 5.2 kcal/mol) for LPO and luteolin (DG = - 6.6 kcal/mol) ꢂ (-)-epicatechin (DG = - 6.9 kcal/mol) > (+)-
pinocembrin (DG = - 5.6 kcal/mol) for MPO.
Table 3
Free binding energy (DG) from docking calculations with LPO and MPO predicted by Glide score?
LPO (PDB code 3UBA)
MPO (PDB code 4C1M)
Compound
DG (kcal/mol)
HC4*
-5.7
--
--
HX1*
-7.9
-6.6
-6.9
-5.6
Luteolin
(-)-Epicatechin
-8.2
-7.5
(+)-Pinocembrin -5.2
*indicate ligands used during the validation procedure
Determination of apparent second-order rate constants for selected flavonoids
Flavonoids containing a 3’,4’-dihydroxy moiety were tested by multi-mixing stopped-flow kinetic
measurements for their potential to resolve accumulated Compound II of LPO and MPO. Fig. 5 depicts
representative spectral changes for LPO (Fig. 5A and C) and MPO (Fig. 5B and D) in the presence of H2O2
only (Fig. 5A and B) and after the injection of 4 µM (-)-epicatechin (Fig. 5C and D).
21

Figure 5
0.12
0.09
0.06
0.03
0.00
0.015
0.010
0.005
0.000
0.015
0.010
0.005
0.000
A
B
b
b
0.08
0.04
0.00
a
a
450
500
550
600
650
500
550
600
650
700
300
400
500
600
700
300
400
500
600
700
Wavelength, nm
Wavelength, nm
0.12
0.08
0.04
0.00
0.09
0.06
0.03
0.00
0.015
0.015
C
D
c
0.010
0.005
0.010
0.005
c
a/b
0.000
450
0.000
500
500
550
600
650
550
600
650
700
a/b
300
400
500
600
700
300
400
500
600
700
Wavelength, nm
Wavelength, nm
Spectral changes of ferric LPO and MPO upon addition of H2O2 in the absence or presence of (-)-epicatechin. While in A the
spectral changes after the pre-incubation of 1 µM LPO with 2 µM H2O2 are shown in C spectral transitions in the additional
presence of 4 µM (-)-epicatechin is depicted. In the presence of H2O2 only (shown in A) a slow transition of Compound I [a] at
412 nm to Compound II [b] (measured 10 s after mixing) at 431 nm takes place. In the presence of (-)-epicatechin in C the
transition of Compound I to Compound II is too fast to be clearly measured. The first spectrum (measured after 2 ms)
represents a mixture of Compounds I and II [a/b]. Within 5 s Compound II is rapidly reduced to ferric LPO [c]. Similar reactions
are shown for 0.5 µM MPO in (B) and (D). In (B) a relative slow formation of Compound II [b] with a Soret peak at 456 nm
(measured 5 s after mixing) from Compound I [a] with a Soret maximum at 431 nm was observed in the sole presence of 5 µM
H2O2. The additional presence of 4 µM (-)-epicatechin in (D) leads to a fast transition to a mixture containing MPO
Compounds I and II [a/b] with two maxima at 431 and 456 nm (measured within 445 ms) followed by the reconstitution of
ferric MPO [c] at 431 nm within the next 100 s.
In the absence of the flavonoid, the Soret maximum of LPO and MPO slowly shifted from 412 to 431 nm
(Fig. 5A) and from 431 to 456 nm (Fig. 5C), respectively, reflecting the formation of Compound II starting
22

from Compound I. Further characteristic spectral changes for this conversion are the formation of new
maxima at 536 and 568 nm for LPO (insert of Fig. 5A) and a new maximum at 638 nm as well as a
shoulder around 600 nm for MPO (insert of Fig. 5B) 40.
In contrast, in the presence of (-)-epicatechin (Fig. 5C and D) for both enzymes the observed spectral
changes indicate a faster Compound II formation followed by a regeneration of ferric native enzyme. For
LPO (Fig. 5C) the first transition could not be measured due to the high reaction rate. Yet, the slower
regeneration of ferric LPO was traceable with an absorbance increase at 412 nm and characteristic
isosbestic points at 423, 459 and 517 nm 16. In addition, the maxima at 500, 542, 590 and 630 nm (insert
3
of Fig. 5C) clearly indicate the formation of native enzyme . In contrast, for MPO a quick formation of
Compound II starting from Compound I within the first 50 ms was detectable (Fig. 5D). In accordance
18
with the literature data, this transition was accompanied by isosbestic points at 413, 462 and 545 nm
3
and new maxima at 570, 620 and 690 nm (insert of Fig 5D) . Subsequently again a slower reduction of
MPO Compound II back to native enzyme took place as indicated by characteristic isosbestic points at
18
442 and 586 nm . For both enzymes there was no regeneration of native enzyme after H2O2-induced
Compound II formation in the absence of the tested flavonoids.
In order to determine second-order rate constants for the reaction of Compounds I and II, time courses
at 431 nm (LPO) and 456 nm (MPO) were fitted to single exponential equations (see Fig. 6A and B). The
obtained kobs values were re-plotted against the substrate concentration allowing the estimation of the
second-order rate constants from the linear slope for the tested substrate (see Fig. 6C and D).
23

Figure 6
0.020
0.016
0.012
0.021
0.014
0.007
A
B
0.008
75
0.000
7.5
0.1
0.2
0.3
0.4
0.5
1
2
3
4
5
Time, s
Time, s
C
D
50
25
0
5.0
2.5
0.0
1
2
3
4
5
1
2
3
4
5
6
Concentration, µM
Concentration, µM
Time traces and single exponential fits to determine second-order rate constants for the reaction of (-)-epicatechin with LPO
Compounds I and II. In A and B typical time traces for the reaction of 0.5 µM LPO with 4 µM (-)-epicatechin at 431 nm are
shown. Both for the transition of LPO Compound I to Compound II (increase in absorbance; A) and the subsequent reduction of
Compound II to ferric enzyme (decrease in absorbance; B), kobs values were calculated by fitting the changes in the absorbance
to a single exponential functions (grey lines). In C and D corresponding kobs values were plotted against the (-)-epicatechin
concentration to determine the second-order rate constant for both reactions from the slope of the linear fits (grey lines).
Mean and standard deviation of n = 3 are given.
The rates determined for all tested flavonoids are summarized in Tab. 4. The highest potential to resolve
LPO Compound II (see Tab. 4A) was found for quercetin (59.5 × 106 M-1 s-1) and luteolin (35.1 × 106 M-1 s-
1). The rates obtained for (-)-epicatechin (0.8 × 106 M-1 s-1) and (+)-taxifolin (0.2 × 106 M-1 s-1) were about
100-fold lower.
24

Table 4
Apparent second-order rate constants for the reaction of several flavonoids with LPO/MPO Compounds I
and II
A
Reaction rate with LPO Reaction rate with LPO
Substance
Ratio
Compound I (M-1 s-1)
Compound II (M-1 s-1)
(35.1 0.7) × 106
(3.0 0.1) × 106
(0.8 0.0) × 106
(0.2 0.0) × 106
(59.5 0.2) × 106
Luteolin
--*
--
--*
--
(ꢁ)-Eriodictyol
(-)-Epicatechin
(+)-Taxifolin
Quercetin
(14.3 0.5) × 106
(6.8 0.2) × 106
--*
16
29
--
B
Reaction rate with MPO Reaction rate with MPO
Substance
Ratio
Compound I (M-1 s-1)
(5.8 0.0) × 106
(6.2 0.0) × 106
(19.9 0.8) × 106
(3.9 0.4) × 106
--**
Compound II (M-1 s-1)
(0.5 0.0) × 106
(1.3 0.0) × 106
(3.5 0.0) × 106
(0.8 0.0) × 106
(7.0 0.4) × 106
Luteolin
10
5
(ꢁ)-Eriodictyol
(-)-Epicatechin
(+)-Taxifolin
Quercetin
6
5
--
* Rates could not be determined as the reaction was too fast for measuring.
** The rate could not be determined due to the distinctive spectral properties of quercetin.
The reaction rates of LPO Compound I with luteolin, (ꢁ)-eriodictyol and quercetin were too high for their
measurement and can be assumed to be in the range of 107 - 108 M-1 s-1. For (-)-epicatechin and (+)-
taxifolin rate constants about 14.3 × 106 M-1 s-1 and 6.8 × 106 M-1 s-1 were determined. A low ratio
between the second-order rate constants for the reaction of the substrate with Compounds I and II (see
Tab. 4) indicated efficient Compound II-resolving substances.31.
25

Regarding MPO (Tab. 4B) the highest rate for the reaction of a flavonoid with MPO Compound I was
found for (-)-epicatechin (19.9 × 106 M-1 s-1), which corresponds well with a former study by Spalteholz
et al. 34. About 3-5 times lower reaction rates were determined for the other tested flavonoids (luteolin:
5.8 × 106 M-1 s-1; (ꢁ)-eriodictyol: 6.2 × 106 M-1 s-1 and (+)-taxifolin: 3.9 × 106 M-1 s-1). Due to the distinctive
spectral properties of quercetin and its reaction products (Suppl. Fig. 5A) the apparent second-order rate
constant for the reaction with MPO Compound I could not be determined.
The MPO Compound II reaction rates were measured after pre-formation of Compound II in the
presence of homovanillic acid 38. This approach was necessary as it was impossible to differentiate clearly
between spectral transitions from Compound I to Compound II and Compound II to ferric enzyme. This is
reflected by low ratios (equal or lower 10) between the rate constants for the reaction of the substrates
with Compounds I and II (see Table 4B). In case of second order rate constants for the reaction with MPO
Compound II the highest rate was measured with quercetin (7.0 × 106 M-1 s-1) followed by (-)-epicatechin
(3.5 × 106 M-1 s-1) and (ꢁ)-eriodictyol (1.3 × 106 M-1 s-1).
Discussion
In the present study we have shown that a number of flavonoids can act as potent regenerators of the
(pseudo-)halogenating activity of LPO and MPO. However, their regeneration potential highly depends
on the chemical properties of the flavonoid B- and C-rings.
Quantifying the effect of flavonoids on the (pseudo-)halogenating activity of LPO and MPO
-
In order to quantify the effect of flavonoids on the OSCN formation of LPO and MPO, specificity
constants (kcat/KM) or IC50 values (in case of inhibitors) were calculated. Note that the kinetic parameters
reflect the binding properties of the flavonoids exclusively for the experimental settings in this study. The
complexity of the reaction system (e.g. different substrates and redox intermediates) complicates the
transferability to other methods.
26

The kcat/KM values observed for MPO are considerably higher than those determined for LPO, which
might suggest a better binding and reactivity of the flavonoids with MPO. Alternatively, these differences
can also be explained by the kinetic cycles of LPO and MPO under the chosen experimental conditions.
The restoration of LPO Compound I from other inactive intermediates seems to be less effective
compared to MPO. This might be related to the fact that LPO Compound I is known to be slowly
41, 42
converted to Compound I*
(see Scheme 1) which is accompanied by protein oxidation. MPO
Compound I reduction by the protein is suppressed by the fact that (besides flavonoids) also H2O2 can
mediate the reduction of Compound I to Compound II (therefore preventing Compound I* formation). In
any case, higher flavonoid concentrations are required to restore the (pseudo-)halogenating activity of
LPO, which leads to increased KM and decreased kcat values. Furthermore, the different architecture of
the substrate channel, which is narrow and cylindrical in LPO but more funnel-shaped in MPO, affects
the orientation and stabilization and thus the reactivity of the flavonoids in the substrate access channel
as well as at the heme cavity 43, 44
.
Structure related influence of flavonoids on the enzymatic activity of LPO and MPO
By comparing the determined specificity constants with the docking calculations it can be assumed that
the C4’ hydroxyl group of the flavonoid B-ring is essential for both binding and reactivity of the
flavonoids with the enzymes. In both LPO and MPO His109 (LPO) or His95 (MPO) are involved in the
interaction with the C4’ hydroxyl oxygen of luteolin and (-)-epicatechin. As a part of the catalytic triad
this histidine is located quite close to the heme iron 8, 45. The significance of the flavonoid C4’ hydroxyl
group is revealed by its easy oxidizability. It can be supposed that the C4’ hydroxyl hydrogen is
46, 47
abstracted by LPO/MPO Compounds I and II leading to the formation of a semiquinone
(see
Scheme 2 for the quercetin reaction). Subsequently, the radical product can disproportionate to an
ortho-quinone balanced with its para quinone methide 47.
27

Presumed oxidation mechanism of flavonoids by LPO and MPO shown for quercetin (as assumed by Galati et al. 47). After
abstracting the hydrogen atom of the C4’ hydroxyl moiety in the presence of LPO/MPO and H2O2 a semiquinone is formed that
subsequently disproportionates to the corresponding ortho-quinone balanced with its para-quinone methide.
The oxidizability of flavonoids is highly influenced by the C-ring pattern. The highest specificity constants
for LPO were found in the group of flavones containing a planar, unsaturated and non-hydroxylated C-
-
ring. For MPO the highest OSCN-regenerating effects were determined for flavanols showing a non-
planar, saturated and C3 hydroxylated C-ring. As no hydrogen bonds with the C3 hydroxyl or C4 oxo
group were observed, the enzyme-specific differences do not only result from their different binding
behavior but also from different redox properties of the flavonoids. Indeed, optimal oxidizing
requirements are only achieved by quercetin (C2-C3 double bound and C3 hydroxyl moiety), while a
saturated C-ring as well as the lack of a hydroxyl group at C3 position avoid an optimal electron
delocalization across the molecule47 (see Scheme 2). Furthermore, the active site of flavonoids can be
theoretically visualized by the frontier orbital distribution. For 3’4’-dihydroxylated flavonoids the HOMO
(highest occupied molecular orbital) frontier orbitals are distributed mainly in the B-ring while the LUMO
48-50
(lowest unoccupied molecular orbital) frontier orbitals are located predominantly in the C-ring
.
Several studies calculated quite high HOMO values for quercetin compared to other flavonoids like
28

48, 49, 51
52
luteolin or (ꢁ)-naringenin
indicating its strong capacity to donate electrons . These aspects are
nicely illustrated by the stopped-flow data. For both LPO and MPO, quercetin exhibited the highest
reaction rates with LPO/MPO Compound II. The rate for the reaction with LPO Compound I was too high
to be determined properly but can be estimated to be in the range of 107 - 108 M-1 s-1. Although
-
quercetin seems to be the flavonoid with the best redox properties, its specificity constant for the OSCN
formation by LPO is reduced compared to luteolin. Due to the high reaction rate with LPO, Compound I
-
quercetin could be able to compete efficiently with the natural substrate SCN for Compound I (see
-
reaction rates in Scheme 1). This promotes the peroxidase cycle and thus, may hinder the OSCN
production by LPO. For MPO quercetin was found as a quite potent reactivity regenerator of the
-
(pseudo-)halogenating activity suggesting that it does not compete with SCN for Compound I as strongly
as for LPO.
A missing or blocked C4’ hydroxylation leads to an inhibition of the (pseudo-)halogenating activity. For
(+)-pinocembrin (replacement of C4’ hydroxyl moiety by hydrogen) docking calculations indicated an
inverted orientation of the flavonoid in the active pocket as compared to the reactivating substances
luteolin and (-)-epicatechin. Both in LPO or MPO the C5 and C7 hydroxyl oxygens of (+)-pinocembrin
form hydrogen bonds to those amino acid residues near the heme center, which are part of the catalytic
triad (LPO His109/Arg255 and MPO Arg239/Gln91) 39. We can only speculate about the inhibitory
-
mechanism of (+)-pinocembrin. In the presence of higher SCN concentrations (≥ 10 mM), the inhibitory
effect of (+)-pinocembrin was reversible (data not shown). Thus it can be supposed that this flavonoid
acts as a competitive inhibitor that blocks the substrate access channel of LPO and MPO and hinders
further substrate binding. The calculation of the frontier orbital distribution for flavonoids without
hydroxylation at the B-ring showed that the LUMO is localized mainly at the B-ring 48, while the HOMO is
primarily located at the A-ring, having the highest capacity to donate electrons 48, 53. This could be one
possible reason for the inverted orientation of (+)-pinocembrin inside the enzyme substrate channel.
29