Quinone-induced protein modifications: Kinetic preference for reaction of 1,2-benzoquinones with thiol groups in proteins

Article abstract of DOI:10.1016/j.freeradbiomed.2016.05.019

Oxidation of polyphenols to quinones serves as an antioxidative mechanism, but the resulting quinones may induce damage to proteins as they react through a Michael addition with nucleophilic groups, such as thiols and amines to give protein adducts. In this study, rate constants for the reaction of 4-methylbenzoquinone (4MBQ) with proteins, thiol and amine compounds were determined under pseudo first-order conditions by UV-vis stopped-flow spectrophotometry. The chemical structures of the adducts were identified by LC-ESI-MS/MS. Proteins with free thiols were rapidly modified by 4MBQ with apparent second order rate constants, k₂ of (3.1±0.2)×10⁴ M^-1 s^-1 for bovine serum albumin (BSA) and (4.8±0.2)×10³ M^-1 s^-1 for human serum albumin at pH 7.0. These values are at least 12-fold greater than that for α-lactalbumin (4.0±0.2)×10² M^-1 s^-1, which does not contain any free thiols. Reaction of Cys-34 of BSA with N-ethylmaleimide reduced the thiol concentration by ~59%, which resulted in a decrease in k₂ by a similar percentage, consistent with rapid adduction at Cys-34. Reaction of 4MBQ with amines (Gly, Nα-acetyl-l-Lys, N?-acetyl-l-Lys and l-Lys) and the guanidine group of Nα-acetyl-l-Arg was at least 5×10⁵ slower than with low-molecular-mass thiols (l-Cys, Nα-acetyl-l-Cys, glutathione). The thiol-quinone interactions formed colorless thiol-phenol products via an intermediate adduct, while the amine-quinone interactions generated colored amine-quinone products that require oxygen involvement. These data provide strong evidence for rapid modification of protein thiols by quinone species which may be of considerable significance for biological and food systems.

Full text of DOI:10.1016/j.freeradbiomed.2016.05.019

Author’s Accepted Manuscript
Quinone-induced protein modifications: Kinetic
preference for reaction of 1,2-benzoquinones with
thiol groups in proteins
Yuting Li, Sisse Jongberg, Mogens L. Andersen,
Michael J. Davies, Marianne N. Lund
www.elsevier.com
PII:
S0891-5849(16)30255-6
http://dx.doi.org/10.1016/j.freeradbiomed.2016.05.019
FRB12882
DOI:
Reference:
To appear in: Free Radical Biology and Medicine
Received date: 10 March 2016
Revised date: 11 May 2016
Accepted date: 18 May 2016
Cite this article as: Yuting Li, Sisse Jongberg, Mogens L. Andersen, Michael J.
Davies and Marianne N. Lund, Quinone-induced protein modifications: Kinetic
preference for reaction of 1,2-benzoquinones with thiol groups in proteins, Free
Radical
Biology
and
Medicine,
http://dx.doi.org/10.1016/j.freeradbiomed.2016.05.019
This is a PDF file of an unedited manuscript that has been accepted for
publication. As a service to our customers we are providing this early version of
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.

Quinone-induced protein modifications: Kinetic preference for reaction of
,2-benzoquinones with thiol groups in proteins
1
a,b,c
b
b
d
b,d*
Yuting Li , Sisse Jongberg , Mogens L. Andersen , Michael J. Davies , Marianne N. Lund
a
School of Food Science and Technology, South China University of Technology, 381 Wushan
Road, Tianhe District, Guangzhou, China 510640.
b
Department of Food Science, Faculty of Science, University of Copenhagen, Rolighedsvej 30,
1
958 Frederiksberg, Denmark.
c
Guangdong Province Key Laboratory for Green Processing of Natural Products and Product
Safety, South China University of Technology, Guangzhou, China
d
Department of Biomedical Sciences, Faculty of Health and Medical Sciences, University of
Copenhagen, Blegdamsvej 3, 2200 Copenhagen, Denmark.
Email addresses: Yuting Li: yli.scut@gmail.com, Sisse Jongberg: Jongberg@food.ku.dk,
Mogens L. Andersen: mola@food.ku.dk, Michael J. Davies: davies@sund.ku.dk, Marianne N.
Lund: mnl@food.ku.dk (*corresponding author, tel.: +453533 3547).
1

ABSTRACT
Oxidation of polyphenols to quinones serves as an antioxidative mechanism, but the resulting
quinones may induce damage to proteins as they react through a Michael addition with
nucleophilic groups, such as thiols and amines to give protein adducts. In this study, rate
constants for the reaction of 4-methylbenzoquinone (4MBQ) with proteins, thiol and amine
compounds were determined under pseudo first-order conditions by UV-Vis stopped-flow
spectrophotometry. The chemical structures of the adducts were identified by LC-ESI-MS/MS.
Proteins with free thiols were rapidly modified by 4MBQ with apparent second order rate
4
-1 -1
3
-
constants, k of (3.1 ± 0.2) × 10 M s for bovine serum albumin (BSA) and (4.8 ± 0.2)×10 M
2
1
-1
s for human serum albumin at pH 7.0. These values are at least 12-fold greater than that for α-
2
-1 -1
lactalbumin (4.0 ± 0.2) × 10 M s , which does not contain any free thiols. Reaction of Cys-34
of BSA with N-ethylmaleimide reduced the thiol concentration by ~59%, which resulted in a
decrease in k by a similar percentage, consistent with rapid adduction at Cys-34. Reaction of
2
4
MBQ with amines (Gly, Nα-acetyl-L-Lys, Nε-acetyl-L-Lys and L-Lys) and the guanidine group
5
of Nα-acetyl-L-Arg was at least 5×10 slower than with low-molecular-mass thiols (L-Cys, Nα-
acetyl-L-Cys, glutathione). The thiol-quinone interactions formed colorless thiol-phenol products
via an intermediate adduct, while the amine-quinone interactions generated colored amine-
quinone products that require oxygen involvement. These data provide strong evidence for rapid
modification of protein thiols by quinone species which may be of considerable significance for
biological and food systems.
Keywords plant phenols, rate constants, thiols, quinone-protein adduct, protein modification.
2

INTRODUCTION
,2-Benzoquinones are highly reactive electrophiles, that are formed by oxidation of catechols or
1
phenols [1, 2]. Catechol structures are found in many naturally occuring polyphenols, and the
formation of 1,2-benzoquinones in biological systems can thus be expected to accompany the
antioxidative action of phenolic compounds, which takes place by a sacrificial oxidation. 1,2-
Benzoquinones undergo Michael addition reactions with nucleophilic groups such as thiols and
amines on free amino acids, peptides or proteins [3] potentially resulting in modified protein
structures and functions. Such modifications may be of importance in physiological and
pathological disorders and change the composition and properties of plant based foods and
beverages which often have high natural levels of phenolic compounds [4].
Polyphenol metabolites are absorbed into the circulatory system as a result of dietary
uptake. There is considerable epidemiological evidence indicating that the consumption of diets
rich in fruit and vegetables containing polyphenols is associated with a reduction in the risk of a
number of chronic pathologies, such as cardiovascular and neurodegenerative diseases and some
specific cancers [5]. Other effects of phenolic and quinoic compounds have also been reported in
biological systems, such as inhibition of protein glycation, antiallergic properties, upregulation of
protective enzymes via interaction with cellular and response elements and signaling pathways,
and cytotoxic effects [6-8]. However, the mechanisms by which polyphenols afford protection
are still incompletely characterized. In addition to dietary uptake of polyphenols and subsequent
metabolism, quinones can also be produced from enzymatic metabolism and metal-ion-induced
oxidation of a wide range of environmental contaminants [9] such as benzene [6], toluene [10],
cigarette smoke [11], polychlorinated biphenyls [12] and other chlorinated aromatics [13]. In
addition, a significant number of cellular metabolites (e.g. estrogens, cathecholamines [6, 14])
3

and commercial drugs (e.g. adriamycin, daunorubicin, daunomycin and mitomycin C) contain, or
can be oxidized to, quinones [15, 16]. Thus there are multiple routes by which quinones might
interact with proteins in biological and food systems and potentially result in protein
modification.
The electrophilic nature of 1,2-benzoquinones can lead to modification of nucleophilic
protein side chains. The interaction of quinones with the Cys residue of glutathione (GSH) in
bone marrow has been suggested to cause myelotoxicity and hematological disorders [17], and
quinone adduction to the active site thiol of calcium-ATPase can decrease enzyme activity in a
similar manner to that induced by the thiol blocking agent N-ethylmaleimide (NEM) [18]. Lens
proteins have been found to react with 1,2-naphthaquinone (from naphthalene metabolism) to
generate brown reaction products [19], which may contribute to the lens cataract formation
induced by naphthalene. Cys or GSH is added to wine to prevent quinone polymerization and
thereby discoloration as they form a colorless product upon reaction with quinones [20].
Quinones have also been shown to form adducts with thiol groups on muscle proteins [21, 22],
which results in undesirable changes in texture by disruption of the disulfide-based protein
network [23]. Phenol adducts formed with amine groups on cytochrome c have been found to
interfere with its proapoptotic function by changing the protein structure [24, 25]. Blocking of
amine groups through reaction with quinones has been suggested to interfere and potentially
inhibit protein glycation [26, 27] as an alternative mechanism to trapping of reactive carbonyls
[
7].
In the present study, the reactions of a model 1,2-quinone, 4-methylbenzoquinone (4MBQ),
with amino acid and protein nucleophiles were studied in order to examine the role and
quantitative importance of the reactions of benzoquinones with thiols, versus amine and
4

guanidine nucleophiles, present on free amino acids, peptides and proteins. Rate constants for
reaction of 4MBQ with three proteins (bovine serum albumin, BSA; human serum albumin,
HSA; α-lactalbumin) as well as thiol- (GSH, L-Cys and Nα-acetyl-L-Cys), amine- (Gly, L-Lys,
Nα-acetyl-L-Lys, and Nε-acetyl-L-Lys), and guanidine- (Nα-acetyl-L-Arg) containing amino
acids and peptides were determined by stopped-flow spectrophotometry, and the nature of the
reaction products was identified by LC-ESI-MS/MS.
MATERIALS AND METHODS
Chemicals and reagents
BSA, HSA, α-lactalbumin, L-GSH, L-Cys, Nα-acetyl-L-Cys, L-Lys, Gly, 4-methylcatechol
(4MC), Nα-acetyl-L-Lys, Nε-acetyl-L-Lys, Nα-acetyl-L-Arg, and acetonitrile (HPLC grade)
were purchased from Sigma-Aldrich (St. Louis, MO, USA). All reagents were of analytical
grade or of highest possible purity. Distilled water was obtained through a Milli-Q purification
device (Millipore, Bedford, MA).
Generation and reduction of 4MBQ
On each day of analysis, a solution of 2 mM 4MC was dissolved in phosphate buffer solution
(0.2 M, pH 4.5) and de-oxygenated by purging with nitrogen for 10 min. Cyclic voltammograms
of this solution was obtained by using a voltammetry analyzer CV-50W (BAS Co., Ltd.) with a
glassy carbon working electrode (3 mm diameter, BAS Co., Ltd.), a platinum coil counter
electrode (5 mm in diameter and 5 cm in length), and an Ag/AgCl (KCl, c=3 M) reference
electrode (Metrohm, Switzerland). The bulk electrolysis was performed at an initial potential of
4
60 mV vs. Ag/AgCl under nitrogen by using the same CV-50W potentiostat, and the initial
5

concentration of 4MC solution was 2 mM. The working electrode was a reticulated vitreous
carbon tube with a diameter of 3.5 cm and length of 4.5 cm (BAS Co., Ltd.). Electrolysis was
terminated when the amount of charge passed was equivalent to 2 electrons per molecule. The
4
MBQ yield was estimated to be ca. 80% (1.6 mM) by UV-Vis spectrophotometry (Cintra 40,
-
GBC Scientific Equipment Pty. Ltd., Australia) using an extinction coefficient ε395 of 1350 M
1
-1
cm [28]. The 4MBQ solution was used within 1 h of generation, with the extent of loss being
ca. 16% over this time as determined by UV-Vis spectrophotometry.
Blocking of BSA thiol groups by NEM
The free thiol group in BSA (0.8 mM) was blocked by reaction with 80 mM NEM in 0.2 M
phosphate buffer (pH 7.0) for 90 min at 25 C. Excess NEM was removed by eluting the protein
through a PD-10 column (GE healthcare). Dilution of the protein was quantified by its
absorbance at 280 nm [29].
Quantification of thiol groups
Thiols were quantified by reaction with 5,5’-dithiobis (2-nitrobenzoic acid) (DTNB) based on
the method of Jongberg et al. [23] with slight modifications. An aliquot of 600 μL dissolved
protein (BSA, HSA, or thiol-blocked BSA (BSA-NEM)) or a blank sample containing 0.1 M
Tris buffer (pH 8.0) was mixed with 2400 μL 0.1 M Tris buffer (pH 8.0). Aliquots of 600 μL
DTNB (10 mM DTNB in 0.1 M Tris buffer, pH 8.0) was added to each sample and incubated in
the dark for 30 min at 25 C, and the absorbance at 412 nm was recorded. Thiol concentrations
were calculated based on a standard curve of L-Cys (1.67-83.3 μM) and given as mol thiol/mol
protein.
6

Kinetic experiments
o
The kinetic measurements were carried out at 25 C by using a SX20 stopped-flow
spectrophotometer (Applied Photophysics, London, UK) in the UV/Vis range (200-750 nm).
Reactants were dissolved into 0.2 M phosphate buffer at appropriate pH values with 4MBQ
in one syringe and the proteins, peptides or free amino acids in the other syringe. Equal volumes
of each reactant were injected into the optical cell and UV-visible spectra recorded over time.
The proteins, peptides or free amino acids were kept in at least 5-fold excess of 4MBQ
concentration to obtain pseudo first-order kinetics.
Identification of the reaction product from the reaction between 4MBQ with amines
Adducts were prepared by the reaction of 4MBQ (0.2 mM) with the corresponding amine (25
mM) in 0.2 M phosphate buffer (pH 7.0). The mixture was stirred for 1 min at 22 C, and then
reduced by bulk electrolysis (initial potential: -1500 mV) for 30 min to obtain the reduced
product. UV-visible spectra were recorded before and after reduction of each adduct.
Identification of chemical structure of adducts by LC-ESI-MS/MS
Reaction products of the reaction between 4MBQ and thiol and amine compounds were analyzed
by LC-ESI-MS/MS according to the method of Yin et al. [26] with slight modifications. Thiols
and amines were dissolved in 0.2 M phosphate buffer. Reaction products derived from thiol
compounds (GSH, L-Cys and Nα-acetyl-L-Cys) were prepared by reacting thiols (10 mM) with
4
MBQ (1 mM) at pH 5.5. The reaction product derived from Gly was generated by the reaction
of Gly (50 mM) with 4MBQ (1 mM) at pH 6.5. For the rest of the amine compounds (L-Lys, Nα-
7

acetyl-L-Lys and Nε-acetyl-L-Lys), they were prepared by the reaction of 4MBQ (1 mM) with
amines (100 mM) at pH 8.0. All the reaction solutions were stirred for 1 min at 25 C and then
filtered through a 0.22 μm syringe filter.
Samples (10 μL) were injected into an LC system (HP 1100 series, Agilent Technologies,
Waldbronn, Germany) equipped with a ZIC-HILIC column (2.1 mm i.d. × 15.0 cm, 3.5 μm;
Merck KGaA, Darmstadt, Germany) and eluted at a flow rate of 0.1 mL/min with a gradient
prepared from (A) ammonium acetate buffer (6.5 mM, pH 5.5) in acetonitrile/water (90:10) and
(B) ammonium acetate buffer (6.5 mM, pH 5.5) in acetonitrile/water (60:40). The gradient
consisted of 30-60 % B in 20 min, followed by an increase to 100% over the next 2 min, which
was kept for 10 min, followed by a decrease to 0% B over the next 3 min, which was kept for 10
min and the column was then re-equilibrated for 2 min to 30% B. The column was operated at 25
o
C. The LC system was controlled by Agilent ChemStation (Agilent Technologies 2001-2005).
Mass spectra were obtained using a HP 1100MSD spectrometer (Agilent Technologies,
Waldbronn, Germany) equipped with an ESI source and MSD trap control software (Bruker
Daltonik GmbH 1998-2005). MS operating parameters were as follows: positive mode, drying
o
gas flow of 9.0 L/min, drying gas temperature of 200 C, nebulizer pressure of 50 psi, scan range
of m/z 50–750, MS/MS fragmentation amplitude of 1.00 V.
Statistical data analysis
Analysis of variance (ANOVA) was carried out by Microsoft Excel 2007. Significance was
defined as a p value <0.05.
8

RESULTS
Generation of 4MBQ
4
MBQ was generated by constant potential anodic oxidation of 4MC [30]. The cyclic
voltammogram of 4MC at pH 4.5 shows a reversible couple with an anodic peak at 281 mV and
a corresponding cathodic peak at 252 mV and with a mid-point potential (E ) of 267 mV vs
1
/2
Ag/AgCl. The separation of the peak potentials was 29 mV close to the theoretical value (28.5
mV) for reversible processes (Figure 1, supplementary materials) [31]. Moreover, during
repetitive cycling the peakheights of 4MC redox couple were constant (data not shown)
suggesting that both the oxidized and reduced species were stable [32]. In buffer solutions with
pH above 5.0, the cyclic voltammogram of 4MC became irreversible (data not shown). For the
kinetic studies 4MBQ was generated electrochemically at pH 4.5 using a constant potential of
4
60 mV vs Ag/AgCl. The electrolyzed solution from the electrochemical cell was then mixed
with phosphate buffer solutions of higher pH in the stopped-flow apparatus in order to undertake
the kinetic characterisation of the reactions of 4MBQ with thiols and amines in the range of pH
values from 4.5 to 7.0. The stability of the electrochemically generated 4MBQ was investigated
by UV-Vis spectrophotometry (Figure 1). 4MC had a characteristic absorption at 280 nm, while
4
MBQ generated by electrolysis from 4MC had a characteristic absorption at 401 nm [21]. When
subjecting the solution of 4MBQ to constant potential cathodic reduction the UV-Vis spectrum
of the resulting solution demonstrated around 85 % conversion back to 4MC (Figure 1).
9

3
2
2
1
1
0
0
.0
.5
.0
.5
.0
.5
.0
4
4
4
MBQ
MBQ reduced by bulk electrolysis
MC
2
50
300
350
400
450
500
550
/
nm
Figure 1 UV-Visible spectra of 1 mM 4MC at 22 C, pH 4.5, 4MC oxidized by bulk electrolysis
yielding 4MBQ, and 4MBQ reduced by reverse bulk electrolysis regenerating 4MC.
Reaction between 4MBQ and proteins
Stopped flow kinetic experiments showed that the characteristic optical absorption of 4MBQ
(
0.04 mM) at 401 nm significantly decreased within 10 s when mixed with excess HSA at pH 7.0
Figure 2a). The absorption changes could be fitted to exponential decay curves allowing the
(
calculation of pseudo first-order rate constants (k ) (Figure 2b).
obs
0
0
0
.50
.25
.00
0.08
b
a
0
0
.06
.04
3
00
400
500
0.0
2.5
5.0
7.5
10.0
Time (s)
/
nm
Figure 2 a) Time dependent changes in the absorption spectra detected on reaction of 0.2 mM
o
HSA with 0.04 mM 4MBQ in aqueous phosphate buffer (0.2M, 25.0 C, pH 7.0) over a period of
1
0

1
0 s. The arrows indicate the decrease of 4MBQ concentration at 401 nm and the formation of an
adduct species with < 300 nm. b) Absorption changes recorded at 401 nm fitted to a single
max
exponential decay curve.
A linear correlation was obtained between kobs and the protein concentrations (Figure 3),
3
-1 -1
allowing the determination of an apparent second-order rate constant, k (4.8 ± 0.2) × 10 M s
2
for the reaction between 4MBQ and HSA.
1
1

3
2
1
0
-4
-4
-4
0.0
2.0x10
4.0x10
6.0x10
[
HSA] / M
T
Figure 3 Pseudo first-order plot of kobs versus protein concentration for the reaction of 4MBQ
o
(
0.04 mM) with HSA (0.2 mM - 0.5 mM) in aqueous phosphate buffer (0.2M, 25 C, pH 7.0).
The non-zero intercept on the vertical axis is due to the slow spontaneous decay of 4MBQ
determined in the absence of added protein.
4
-1 -1
Similar kinetic experiments with BSA gave k (3.1 ± 0.2) × 10 M s , while α-
2
2
-1 -1
lactalbumin, a protein without free Cys residues, gave k (4.0 ± 0.2) × 10 M s (Figure 4 and
2
Table 1).
1
2

0
0
0
0
0
0
0
0
0
.040
.035
.030
.025
.020
.015
.010
.005
.000
BSA
NEM-BSA
-lactalbumin
0
1
2
3
4
5
Time (s)
Figure 4 Time dependent changes in 4MBQ concentration (0.04 mM initial value) on reaction
with BSA (0.28 mM), NEM-BSA (0.28 mM, ~59% thiols blocked) or α-lactalbumin (0.28 mM)
o
in phosphate buffer (0.2M, pH 7.0, 25 C).
The low value for α-lactalbumin is consistent with reaction predominantly at Cys-34 of
HSA and BSA. This observation is in agreement with a previous study where the adduction of
4
MBQ to the free Cys residue of BSA was confirmed by LC-MS analysis of hydrolyzed samples
[
21]. As the Cys content of HSA and BSA is typically significantly less than the theoretical value
of 1 due to modification in vivo or during processing [33], the actual thiol concentration of the
proteins was determined using DTNB. Such experiments gave values of 0.21 mol/mol protein for
HSA and 0.38 mol/mol protein for BSA (Table 1). Based on these concentrations, corrected
second-order rate constants for reaction with the Cys residues, k2,SH, at pH 7.0 were calculated
4
4
-1 -1
for BSA and HSA as (8.3 ± 0.5) × 10 and (2.3 ± 0.1) × 10 M s , respectively.
1
3

Table 1 Apparent second-order rate constant (k ) for reaction of 4MBQ with HSA, BSA, NEM-
2
BSA, and -lactalbumin at 25 C, thiol concentration determined for each protein, and the
second-order rate constants (k2,SH) corrected for reaction with Cys residues (n.d.; not determined).
-
1 -1
Protein
Thiol concentration
mol/mol protein)
pH
7.0
7.0
k (M s )
2
(
3
HSA
0.21 ± 0.00
(4.8 ± 0.2) x 10
4
4
k2,SH (2.3 ± 0.1) x 10
4
BSA
0.38 ± 0.00
(3.1 ± 0.2) x 10
k2,SH (8.3 ± 0.5) x 10
4
NEM-BSA
0.16 ± 0.01
n.d.
7.0
7.0
6.5
(1.0 ± 0.1) x 10
2
-lactalbumin
(4.0 ± 0.2) x 10
3
BSA
0.38 ± 0.00
(7.7 ± 0.2) x 10
In order to confirm the key role of the Cys-34 residue in these reactions, similar
experiments were carried out with BSA pre-treated with NEM in order to block the Cys-34
residue. The thiol concentration of the NEM-BSA preparation was found to be 0.16 ± 0.01
mol/mol protein, corresponding to a thiol-blocking of ~59%. The k of the reaction of 4MBQ
2
with NEM-BSA at pH 7.0 was decreased by a similar percentage (~67%) compared to non-
blocked. These observations are consistent with Cys-34 being the major target for 4MBQ.
Reaction between 4MBQ and low molecular mass thiols
In order to assess the potential role of protein structure in modulating the reaction rates of
quinones with thiols, the kinetics of reaction of 4MBQ with N -acetyl-L-Cys, L-Cys and GSH
were examined. 4MBQ reacted with all three thiols with rate constants that were highly
1
4

dependent on pH. The characteristic absorption of 4MBQ at 401 nm disappeared within 1 s in
the presence of excess concentrations of these thiols, accompanied by the generation of thiol-
quinone adducts with absorption maximum at 294 nm (Figure 5a and below). Exponential fits of
the kinetic traces of absorbance changes at 401 nm and 294 nm gave pseudo first-order rate
constants (k ) for loss of 4MBQ and adduct formation (Figure 5 and Table 2-7, supplementary
obs
material). The rate constants obtained at both wavelengths were similar indicating that reaction
intermediates do not build up to any significant extent over the time scales examined.
1
5

0
0
0
0
0
.4
.3
.2
.1
.0
a
3
00
400
500
/
nm
0
0
0
0
.06
.04
.02
.00
b
0
.00
0.25
0.50
0.75
1.00
Time (s)
0
.14
.12
.10
.08
.06
c
0
0
0
0
0
.00
0.25
0.50
0.75
1.00
Time (s)
Figure 5 a) Time dependent changes in the absorption spectra of reaction mixtures containing
o
Nα-acetyl-L-Cys (0.5 mM) and 4MBQ (0.04 mM) in aqueous phosphate buffer (0.2 M, 25.0 C,
pH 5.0). Arrows indicate the decrease of 4MBQ concentration at 401 nm and the adduct
formation at 294 nm over a period of 1 s; b) Fit of the exponential decay curve to absorption
recorded at 401 nm for 4MBQ decay over time; c) Fit of exponential curve to absorption
recorded at 294 nm.
1
6

Plots of kobs against the thiol concentrations did not give linear correlations. Similar non-
linear correlations have been observed by Jameson et al. [30] for the reaction of dopaminoquione
with Cys; these data were therefore analysed using equation 1, where [R-SH] is the pseudo first-
T
order concentration of thiol.
(1)
Equation 1 is based on a mechanism that involves an initial fast reversible formation of a
tetrahedral intermediate between the quinone and the thiol, characterized by an equilibrium
constant, K (see [30]). This tetrahedral intermediate decays by a first-order reaction with rate
constant, k, to give the thiol-quinone adduct (Scheme 1).
Scheme 1 Proposed mechanism for reaction between 4MBQ and thiols (in the case of L-Cys as
described by Jameson et al [30]).
K and k can be obtained from the A and B values in equation 1 and the corresponding pH
value as described by Jameson et al. [30] and equations 2 and 3:
(
2)
3)
(
1
7

4
01nm
The dependence of the observed rate constants determined at 401 nm and 294 nm, kobs and
2
94nm
k
obs , on the concentration of Nα-acetyl-L-Cys followed the non-linear relation predicted by
equation (1), with the A and B values obtained by fitting to equation 1 (Figure 6).
8
6
4
2
0
0
0
0
0
a
b
1
00
5
0
0
0.0
-3
-3
-3
-3
-3
-3
0.0
2.0x10
4.0x10
6.0x10
2.0x10
4.0x10
6.0x10
[
N -acetyl-L-Cys] / M
[N -acetyl-L-Cys] / M
T
T
Figure 6 Variation of observed first-order rate constant kobs with different Nα-acetyl-L-Cys
concentrations at pH 5.0. a) Following the loss of 4MBQ at 401 nm. b) Following formation of
the adduct species at 294 nm. The lines are fits to equation 1.
Similarly, kobs and the corresponding A and B values were determined for L-Cys and GSH
(Tables 2-7 in the supplementary materials). The A and B values vary with the reaction pH
allowing K and k to be determined by using equations 2 and 3 (Table 2 and Figure 7).
1
8

5
5
5
3
3
3
2
6
4
2
.0x10
.0x10
.0x10
2.0x10
a
1
1
5
.5x10
.0x10
.0x10
0.0
0.0
6
6
6
0
1x10
2x10
3x10
+
-1
1
/[H ] / M
5
3
8
6
4
2
.0x10
.0x10
.0x10
.0x10
2.0x10
1.5x10
1.0x10
5.0x10
0.0
b
5
5
5
3
3
2
0
.0
6
6
6
0
1x10
2x10
3x10
+
-1
1
/[H ] / M
+
Figure 7 Plots of A( ) and B(o) as a function of 1/ [H ] for the reaction of 4MBQ and Nα-acetyl-
L-Cys. a) Following the loss of 4MBQ at 401 nm; b) Following adduct formation at 294 nm.
Analyses of the consumption of 4MBQ and formation of products for each thiol gave a
similar set of values for K and k. The equilibrium constants K for the formation of the tetrahedral
-4
-4
intermediates with 4MBQ varied from 1.6 x 10 for Nα-acetyl-L-Cys to 4.2 x 10 for GSH
(Table 2). The corresponding values of k for decay of the intermediate into products varied from
3
-1
3 -1
1
.0 x 10 s for Nα-acetyl-L-Cys to 2.2 x 10 s for L-Cys.
1
9

Table 2 Equilibrium constant (K) and rate constant (k) for reaction of 4MBQ with Nα-acetyl-L-
Cys, L-Cys and GSH. K401 nm and k401 nm were measured by following the loss of 4MBQ at 401
nm, K294 nm and k294 nm were measured by following the adduct formation at 294 nm.
-
1
-1
Thiols
Nα-Acetyl-L-Cys
L-Cys
K
401 nm
k
401 nm (s )
K
294 nm
k
294 nm(s )
-4
-4
-4
3
3
3
-4
-4
-4
3
3
3
(1.6 ± 0.1)×10
(3.3 ± 0.9)×10
(4.2 ± 1.1)×10
(1.0 ± 0.7)×10
(2.2 ± 0.8)×10
(1.3 ± 0.4)×10
(1.8 ± 0.2)×10
(3.2 ± 0.4)×10
(5.9 ± 0.4)×10
(1.2 ± 0.2)×10
(2.7 ± 0.4)×10
(1.3 ± 1.0)×10
GSH
When the term B[RSH)]T < 1 (see equation 1) the reaction approximates to pseudo first-
order behavior with a linear dependence between kobs and the thiol concentrations. This occurs
with low concentrations of thiols, or when the value of B is low due to low pH, or with less
favorable formation of the initial tetrahedral intermediate due to a low value of K. In such
situations, A will be equal to an apparent second order rate constant. This appears to be the case
for thiol-containing proteins and 4MBQ, with the apparent second order behavior consistent with
a low value of B, and small equilibrium constants for the formation of an initial tetrahedral
5
-1 -1
intermediate. The high values of A for Nα-acetyl-L-Cys, L-Cys and GSH (5.2 x 10 M s (pH
5
-1 -1
5
-1 -1
6
.5), 7.0 x 10 M s (pH 6.0) and 5.4 x 10 M s (pH 6.0), respectively. (Tables 2, 4, and 6,
supplementary material) indicate that these low-molecular-mass thiols react overall much faster
~68 fold or higher) with 4MBQ than the thiol containing proteins, where k for BSA at pH 6.5
(
2
3
-1 -1
was 7.7 x 10 M s (Table 1).
Reaction between 4MBQ and low molecular mass nitrogen nucleophiles
As amine and guanidine groups are highly abundant on many proteins (e.g. 59 Lys and 23
Arg residues versus 1 Cys residue for BSA; Uniprot database), these are potential alternative
2
0

nucleophilic reaction sites for quinones. Consequently, rate constants were determined for the
reaction of 4MBQ with the guanidine group of Nα-acetyl-L-Arg and a range of low molecular
mass amines (Gly, L-Lys, Nε-acetyl-L-Lys and Nα-acetyl-L-Lys) under pseudo first-order
conditions. The reaction of 4MBQ with the guanidine group of Nα-acetyl-L-Arg was too slow to
be determined accurately under the conditions employed. At pH 6.5, kobs for the decay of 4MBQ
-2
-1
(
0.04 mM) in the presence of 50 mM Nα-acetyl-L-Arg was (7.5 ± 0.2) × 10 s , which was not
-2
significantly different from the value of kobs for spontaneous decay of 4MBQ ((7.9 ± 0.2) × 10
-1
s , p > 0.05). Similar data were obtained at pH 7.0 (data not shown). These data suggest that
reaction of 4MBQ with Arg residues is very slow, which is in agreement with the study by
Pierpoint [34].
The apparent second-order rate constant, k
2
, for loss of 4MBQ in the presence of Gly at pH
-1
-1
5
6
.5 was 0.7 ± 0.1 M s (Table 8, supplementary material), which is 7 × 10 times lower than for
5
-1 -1
Nα-acetyl-L-Cys (5.2 x 10 M s ) at the same pH. At pH 7.0, k
2
for Gly increased to 2.0 ± 0.3
-1
-1
-1 -1
M s . The reaction of 4MBQ with L-Lys at pH 7.0 (k 8.4 ± 0.5 M s ) was faster than for Gly,
2
but still significantly slower than that for any of the thiols (Table 9, supplementary material). The
reactivity of the two amino groups in L-Lys was investigated by analysis of the rates of
reactions of Nε-acetyl-L-Lys and Nα-acetyl-L-Lys, where the side chain and -carbon amine
group are blocked, respectively. At pH 7.0, the α-carbon amine group reacted faster with 4MBQ
than the side chain amine with the order of reactivity being Lys > Nε-acetyl-L-Lys > Nα-acetyl-
L-Lys (Table 3). The increased reactivity of the -carbon amine over the side chain amine is
consistent with its lower pK value in agreement with previous studies [34].
a
2
1

Table 3 Second-order rate constants (k ) for the reaction of 4MBQ with L-Lys, Nε-acetyl-L-Lys,
2
o
401nm
2
and Nα-acetyl-L-Lys at 25 C. k
was determined by following the loss of 4MBQ at 401 nm; k
adduct
2
was determined by following the formation of adduct at 499 nm for Nα-acetyl-L-Lys, and at
5
08 nm for Nε-acetyl-L-Lys and L-Lys.
pH 7.0
pH 8.0
pH 8.0
Amines
4
2
01nm
-1 -1
401nm
2
-1 -1
adduct
2
-1 -1
k
(M s )
k
(M s )
k
(M s )
L-Lys
8.4 ± 0.5
4.0 ± 0.4
0.9 ± 0.1
45.7 ± 3.5
25.8 ± 0.8
24.8 ± 2.0
16.3 ± 0.9
9.8 ± 0.8
4.7 ± 0.5
Nε-acetyl-L-Lys
Nα-acetyl-L-Lys
Reaction of 4MBQ with L-Lys at pH 8.0 resulted in additional changes in the UV-visible
spectra to those observed at lower pH values. An increase in absorbance at ~500 nm was
observed in addition to the loss of absorbance at 401 nm and increase at 293 nm (Figure 8).
Similar additional absorbance changes, which give rise to colored solutions, were observed with
Nε-acetyl-L-Lys, Nα-acetyl-L-Lys and Gly at this pH, although the
for these new
max
absorbance bands varied with the substrate (Figure 9). The time scale over which this additional
absorbance band increased was significantly longer than that for loss of 4MBQ (as monitored at
4
01 nm) and occurs with a noticeable lag time (Figure 8b, insert). This indicates that the
compounds that give rise to the absorption at longer wavelengths arises from secondary species
which are formed at a slower rate than for the initial reaction of 4MBQ with the amine.
2
2

0
0
0
0
0
0
.5
.4
.3
.2
.1
.0
a
b
0
0
0
.03
.02
.01
0
.008
0.007
0.006
0
.005
0
.0 0.2 0.4 0.6 0.8 1.0
Time (s)
0.00
3
00
400
500
0
5
10
Time (s)
15
20
/
nm
Figure 8 a) Time dependent changes in the absorption spectra of reaction mixtures containing L-
o
Lys (2.5 mM) and 4MBQ (0.04 mM) in aqueous phosphate buffer (0.2 M, 25.0 C, pH 8.0) over
a period of 20 s. Arrows indicate the decrease of 4MBQ at 401 nm and adduct formation at 508
nm; b) Absorption values recorded at 508 nm over time fitted to an exponential curve. The insert
shows the lag phase occurring within the first 0.5 s.
The
for the species derived from Nα-acetyl-L-Lys (~499 nm) was at a shorter
max
wavelength than those for Gly and Nε-acetyl-L-Lys (504-508 nm) (Figure 9). The absorption
spectra of L-Lys appeared to have features arising from a combination of the Nα-acetyl-L-Lys
and N -acetyl-L-Lys adducts, though with the -carbon amine product predominating. This is
consistent with reaction at the side chain amine in Nα-acetyl-L-Lys and the -carbon amine for
Gly and Nε-acetyl-L-Lys, while for L-Lys both amines may react.
2
3

0
0
0
0
0
0
.5
.4
.3
.2
.1
.0
Gly-quinone adduct
L-Lys-quinone adduct
N- -acetyl-L-lys-quinone adduct
N- -acetyl-L-Lys-quinone adduct
4
00
500
600
/
nm
Figure 9 The UV-visible absorption spectra of amine-quinone adducts obtained using 25 mM
amine and 0.2 mM 4MBQ in aqueous phosphate buffer (0.2 M, pH 7.0 and 22 C).
Identification of adducts by LC-ESI-MS/MS
The adducts generated from reaction of 4MBQ with the thiol and amino compounds were
identified by LC-ESI-MS/MS (Table 4). The adduct formed between L-Cys and 4MBQ gave an
ion with m/z 244.0, consistent with the presence of a protonated L-Cys-phenol adduct.
Analogous ions consistent with the presence of thiol-phenol adducts were detected with the other
thiols examined.
2
4

Table 4 ESI (positive ion mode) MS data for thiol-phenol and amine-quinone adducts.
Molecular
Mass
Retention
time (min)
12.7
m/z
(relative intensity, %)
244.0
Proposed ion
structure
Compound
+
L-Cys-phenol adduct
Nα-acetyl-L-Cys-phenol adduct
GSH-phenol adduct
243.2
[M+H]
+
285.3
5.5
286.0
[M+H]
+
429.4
6.8
430.1
[M+H]
+
Gly-quinone adduct
195.2
6.3
196.0 (100)
[M+H]
+
218.0 (10)
[M+Na]
+
L-Lys-quinone adduct
266.3
308.3
7.2
9.7
267.1
[M+H]
+
Nα-acetyl-L-Lys-quinone adduct
309.1 (100)
31.1 (15)
309.1 (100)
31.1 (58)
[M+H]
+
3
[M+Na]
+
Nε-acetyl-L-Lys-quinone adduct
308.3
7.9
[M+H]
+
3
[M+Na]
The products arising from the amines all gave rise to ions that were assigned to amine-
quinone species (Table 4), which absorb at ~500 nm [35] in agreement with the current data. The
nature of these adducts was examined further by carrying out bulk electrolysis of the products,
which resulted in a significant decrease in coloration and loss of the long wavelength absorbance
bands (data not shown). This is consistent with reduction of quinone-derived compounds to
amine-phenol species. After electrolysis, a fast re-coloration was observed in the presence of O2
consistent with re-oxidation of amine-phenol to amine-quinone species (Scheme 2).
2
5

Scheme 2 Proposed mechanism for reaction between 4MBQ and amines.
The fragmentation ions detected from each adduct are given in Table 5. The most abundant
ion from the Cys-phenol adduct had m/z 226.9, which was 17 less than the parent ion suggesting
the loss of a hydroxyl group. The adduct derived from Nα-acetyl-L-Cys gave two major
fragmentation ions at m/z 268.0 and 161.9, where the former was consistent with loss of a
hydroxyl group, and 161.9 representing the Nα-acetyl-L-Cys residue. The main fragmentation
ion of the GSH-phenol adduct (m/z 301.0) suggested a loss of the Glu residue.
Table 5 ESI MS/MS (positive ion mode) data for thiol-phenol and amine-quinone adduct.
Precursor ion
m/z
Product ions, m/z
(relative intensity, %)
Compound
L-Cys-phenol adduct
244.0
226.9 (100), 155.0 (53)
Nα-acetyl-L-Cys- phenol adduct
286.0
268.0 (100), 244.0 (87), 227.0 (19),
198.0 (16), 181.0 (64), 161.9 (99)
GSH-phenol adduct
Gly-quinone adduct
430.1
196.0
267.1
309.1
309.1
355.0 (4), 301.0 (100), 284.0 (46),130.0 (34)
150.0 (100), 122.0 (20)
L-Lys-quinone adduct
250.0 (6), 205.8 (4), 128.0 (100)
204.1 (6), 170.0 (10), 128.0 (100)
265.0 (8), 170.0 (4), 128.0 (100)
Nα-acetyl-L-Lys-quinone adduct
Nε-acetyl-L-Lys-quinone adduct
2
6

The adducts from the incubation of 4MBQ with L-Lys, Nα-acetyl-L-Lys and Nε-acetyl-L-
Lys yielded a common abundant fragmentation ion at m/z 128.0 which arises from loss of a
hydroxyl group from the Lys residue. The main fragmentation pathway of the Gly-quinone
adduct was the loss of the carboxyl group. The proposed structures and main fragmentation
pathways for each adduct are summarized in Figure 10.
Figure 10 Proposed structures and main fragmentation pathways during MS-analysis of L-Cys-
phenol adduct (a), Nα-acetyl-L-Cys-phenol adduct (b), Gly-quinone adduct (c), L-Lys-quinone
adduct (only -amine adduct is shown) (d), Nα-acetyl-L-Lys-quinone adduct (e), Nε-acetyl-L-
Lys-quinone adduct (f), and the GSH-phenol adduct (g).
DISCUSSION
Quinones are common species in biological systems found in a wide variety of drugs,
industrial chemicals and natural products. They are also generated on metabolism of many
exogenous compounds and as a result of oxidation of species such as polyphenols. It has been
2
7

demonstrated that quinones undergo multiple further reactions including one- or two- electron
reduction reactions, and reaction with nucleophiles. One-electron reduction reactions (e.g. by
superoxide radicals [36], or via enzymes such as NADPH-cytochrome P450 reductase, NADH
cytochrome b reductase and nitric oxide synthetases) result in the formation of semiquinones
5
that can undergo disproportionation, oligomerization, or reaction with ascorbate [37] and oxygen.
Two electron reduction can occur enzymatically (e.g. with NAD(P)H quinone oxidoreductases)
or chemically using powerful reducing agents (reviewed [38]). These reduction processes
compete with reaction with nucleophiles (such as thiols and amines), with products arising from
thiol addition (e.g. 5-S-cysteinyl conjugates to DOPA, dopamine, noradrenaline and adrenaline
[
39]) being readily detected in urine and plasma of healthy humans. Thiol adduction products
from oxidized natural products have also bee reported (e.g. 2-S-glutathionyl caffeoyl tartrate [40],
-S-glutathionyl quercetin and 8-S-glutathionyl quercetin [41]). The relative significance of these
6
competing pathways is currently unclear as the kinetics of these reactions are poorly understood;
the current study provides rate constant data for some of these reactions.
The reaction of quinones with proteins is known to alter protein structure and function [11,
1
8, 19, 23-25, 42], but the major sites of adduction, and particularly the rate constants for these
processes have not been established. This is of particular importance as amine (N-terminal and
Lys side chain) and guanidine (Arg side chain) groups are often present at much higher
concentrations than free Cys residues on peptides and proteins (e.g. serum albumin, the major
protein present in mammalian plasma, contains ~83 amine/guanidine groups and only a single
free Cys residue). The data obtained in the current study indicate that even with this large excess
of alternative nucleophiles, it is the Cys residues that are kinetically the most important targets
when proteins are in excess, as will be the case in most biological systems. The high and
2
8

preferential reactivity of the Cys residues is confirmed by the higher rate constants for HSA and
BSA compared to the non-thiol protein -lactalbumin, and the decrease in rate constant on
blocking of the Cys-34 residue on BSA (Figure 4).
Direct comparison of the rate constants for reaction at different sites has been obtained by
determining the rate constants for reaction of 4MBQ with a range of low-molecular-mass amino
acids and peptides. At pH 7.0, k for Nα-acetyl-L-Cys was too fast to be determined accurately,
2
5
but at pH 6.5 k (corresponding to the A value; Table 2, supplementary material) was 5.2 × 10
2
-1 -1
M s and since this value is expected to be higher at pH 7.0, the apparent second-order rate
5
constant is assumed to be at least 5 × 10 -fold greater than that for reaction with Nα-acetyl-L-Lys
-1 -1
(
k 0.9 ± 0.1 M s ). These data are consistent with previous qualitative studies [21, 43], which
2
have reported that the thiol group of Cys-34 in BSA is involved in quinone-protein reactions
when BSA is in excess.
Comparison of the k values for Nα-acetyl-L-Cys (i.e. the A value) and BSA at pH 6.5,
2
indicates that Nα-acetyl-L-Cys is more reactive (by ~68-fold) (Figure 11; Tables 1,2,
supplementary materials). This difference may be due to both steric and electronic interactions,
as well as differences in the pK of the thiols, as the thiolate anion is a better nucleophile than the
a
neutral species [44]. In the reaction of 4MBQ with different thiols, the thiol pK may influence
a
the equilibrium constant, K, to a greater extent than the rate constant, k (Scheme 1). The K values
determined for 4MBQ and Nα-acetyl-L-Cys were the smallest obtained in the current study,
which may be explained, at least in part, by the high pK of Nα-acetyl-L-Cys (thiol pK 9.5) [45].
a
a
However, the values of K determined for L-Cys (thiol pK 8.3) [45] were smaller than those of
a
GSH (thiol pK 8.8) [45]. The values for k for the quinone-thiol reaction appear to be mainly
a
determined by the stability of the product (Scheme 1). In particular it has been proposed that the
2
9