Accepted Manuscript Effects of conjugation metabolism on radical scavenging and transport properties of quercetin – In silico study Višnja Stepanić, Sara Matić, Ana Amić, Bono Lučić, Dejan Milenković, Zoran Marković PII:
S1093-3263(18)30535-7
DOI:
https://doi.org/10.1016/j.jmgm.2018.10.023
Reference:
JMG 7261
To appear in:
Journal of Molecular Graphics and Modelling
Received Date: 19 July 2018 Revised Date:
22 October 2018
Accepted Date: 24 October 2018
Please cite this article as: Viš. Stepanić, S. Matić, A. Amić, B. Lučić, D. Milenković, Z. Marković, Effects of conjugation metabolism on radical scavenging and transport properties of quercetin – In silico study, Journal of Molecular Graphics and Modelling (2018), doi: https://doi.org/10.1016/j.jmgm.2018.10.023. 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 proof before it is published in its final 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.
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT
Effects of conjugation metabolism on radical scavenging and transport
2
properties of quercetin – in silico study
3
Višnja Stepanić a*, Sara Matić,a Ana Amić,b Bono Lučić,a Dejan Milenković,c Zoran Markovićd
4
a
Ruđer Bošković Institute, Bijenička 54, 10000 Zagreb, Croatia
5
b
Department of Chemistry, Josip Juraj Strossmayer University of Osijek, Cara Hadrijana 8a,
6
31000 Osijek, Croatia
7
c
8
Kragujevac, Serbia
9
d
SC
Bioengineering Research and Development Center, Prvoslava Stojanovića 6, 34000
Department of Chemical-Technological Sciences, State University of Novi Pazar, Vuka
Karadžića bb, 36300 Novi Pazar, Serbia
11
M AN U
10
RI PT
1
Corresponding Author
13
*(V.S.) Phone: +385 1 457 1248. Fax: +385-1-4561-010.
14
E-mail:
[email protected]
15
ORCID
16
Višnja Stepanić: 0000-0001-9518-4153
18 19
EP
AC C
17
TE D
12
1
ACCEPTED MANUSCRIPT ABSTRACT
21
Quercetin (Q) is a natural polyphenol with high radical scavenging capacity, but low in vivo
22
bioavailability. It is extensively transformed by host phase II metabolism and microbiota.
23
Herein, effects of major in vitro and in vivo conjugation transformations of Q on its radical
24
scavenging capacity and human serum albumin (HSA) binding were studied by using
25
appropriate computational approaches, DFT (U)B3LYP/6-31+G(d,p) and molecular docking,
26
respectively. With regard to radical scavenging capacity of Q, conjugation transformations
27
generally reduce its antioxidant capacity including regeneration efficiency through
28
disproportionation of an intermediate radical species since these structural modifications
29
occur mainly at its radical scavenging –OH groups. They were also found to alter dominant
30
radical scavenging mechanism in a specific way dependent upon conjugation type and site.
31
Concerning distribution by HSA, binding to this main plasma transporter protein may not be
32
dominant transport mechanism for Q and its metabolites in vivo. Like Q aglycon, most of its
33
metabolites are bound non-specifically at multiple binding sites of HSA, with relatively weak
34
affinities. Only sulfo-conjugates including plasma abundant isomer Q-3’-O-SO3–, were
35
predicted to bind specifically in warfarin-like manner, but also with relatively low binding
36
affinity.
37
Keywords: quercetin; conjugation metabolism; radical scavenging; disproportionation,
38
serum albumin
SC
M AN U
TE D
EP
AC C
39
RI PT
20
40
Abbreviations: BDE, bond dissociation enthalpy; BS I, binding site I; ETE, electron transfer enthalpy;
41
FE, free energy; HAT, hydrogen atom transfer; HSA, human serum albumin; LGA, Lamarckian genetic
42
algorithm; Q, quercetin; QBM, quercetin binding mode; RS, radical scavenging; SA, serum albumin;
43
SPLET, sequential proton loss electron transfer; WBM, warfarin binding mode
2
ACCEPTED MANUSCRIPT 44
1. Introduction Natural compounds are notable sources of drugs and determine functional properties of
46
food and beverages. Quercetin (Q) is one of the most studied natural compounds. It is an
47
abundant plant polyphenol from the subgroup of flavonols with average dietary intake in EU
48
exceeding 20 mg/day [1]. It has been known mainly as an antioxidant with beneficial
49
cardioprotective and chemopreventive health effects [2].
50
In various fruits and vegetables, Q is present in the form of diverse glycosides. Glycosides
51
undergo moderately fast cleavage of sugar fragment(s) by the action of bacterial β-
52
glucosidase(s) and intestinal absorption (the absorption half-time t1/2 is ~ 1 h) [3].
53
Subsequently, the aglycon Q is extensively metabolised in two ways, through: (i) conjugation
54
by the host phase II metabolic enzymes mainly in intestine and liver and (ii) C-ring fission by
55
host enterobacteria [3]. Consequently, after oral consumption, the plasma level of Q is very
56
low, generally in sub-μM range and the free aglycon is present in the amount less than 10%
57
of the total plasma Q. After the extensive first-pass metabolism, Q and its metabolites are
58
slowly excreted with the elimination half-time t1/2 longer than 15 h and largely determined
59
by dietary source. Disposition of Q in human is complex and the great portion of Q (from
60
23.0–81.1% of a dose) is reported to be totally metabolized to carbon dioxide and exhaled
61
into air [4].
62
The main metabolites of Q found in human plasma are Q-3-O-β-D-glucuronide (Q-3-O-Glr–),
63
Q-3’-O-sulfate (Q-3’-O-SO3–) and a hybrid diconjugate 3’-O-methyl-3-O-glucuronide (3'-O-
64
Me-Q-3O-Glr–) [5]. While the conjugated metabolites circulate in plasma, the bacterial
65
catabolites - small molecular weight phenolic acids such as caffeic acid, hydrocaffeic acid,
66
homoprotocatechuic acid and homovanillic acid, are also present in urine and feces [6]. Q is
67
also metabolised in specific ways in vitro, in cellular cultures [7]. Upon 24 h incubation of Q
AC C
EP
TE D
M AN U
SC
RI PT
45
3
ACCEPTED MANUSCRIPT (50 μM and 100 μM) with human colon carcinoma cell line HT29 or rat hepatocellular
69
carcinoma cell line H4IIE, the major identified Q metabolites were Q-4’-O-Glr– (40%) and Q-
70
3-O-Glr– (27%) in HT29 cells, and Q-7-O-Glr– (67%) in H4IIE cells with the residual level of
71
aglycon present in the incubation media less than 1% [8]. In vitro cellular effects of Q are
72
greatly determined by its concentration and by cell type implying plausible impacts of
73
specific cellular metabolic transformation on its activity [9].
74
Fig. 1. Quercetin and its conjugated metabolites detected in vivo and/or in vitro and considered here in silico.
SC
75
RI PT
68
M AN U
76
The metabolic transformations of Q have considerable impact on its antioxidative
78
properties. They modify its direct radical scavenging (RS) capacity [10,11], its potency to
79
inhibit low-density lipoprotein (LDL) [12] or protein oxidation [13], as well as its inhibitory
80
activity to enzymes involved in endogenous free radical production such as xanthine oxidase,
81
lipoxygenase or NADPH oxidase activity of generating superoxide radical anion [14,15].
82
Phase II metabolism also alters various other biological activities of Q including interactions
83
with membrane transporters and effects on angiogenesis [16].
84
The metabolic conjugations of Q may alter its interactions with serum albumin (SA), the
85
main transporter protein of endogenous and exogenous substances in blood [17]. According
86
to in vitro studies [18], Q and most of its main metabolites are bound by human serum
87
albumin (HSA) what may impact their in vivo bioavailability and distribution [12,19].
AC C
EP
TE D
77
88 89
Scheme 1 Assumed RS mechanisms of Q and its metabolites: (a) hydrogen atom transfer (HAT) and (b)
90
sequential proton loss electron transfer (SPLET). In double HAT pathway (a), aryloxyl monoradical Q-O is an
91
intermediate and quinoid specie Q(=O)2 is a final product. (c) Disproportionation of intermediate radical Q-O
92
regenerates the parent reducing form Q-OH.
•
•
4
ACCEPTED MANUSCRIPT 93
The purpose of our study was to explore the effects of main metabolic transformations of Q,
95
their types and positions, on its RS capacity and HSA binding by using appropriate
96
computational methods. The RS computational models as well as molecular docking
97
calculations have already been applied for study of polyphenols [20]. The reliable and
98
insightful results obtained for polyphenol aglycons as well as similar (semi)synthetic
99
compounds, encouraged us to apply in silico approaches on quercetin metabolites. The
100
effects of conjugation transformations on biological properties of Q have not been studied
101
enough so far. In addition to the major in vivo metabolites Q-3-O-Glr−, Q-3’-O-SO3− and 3'-O-
102
Me-Q-3O-Glr– found in human plasma [5], the conjugated 3’-O- (Q-3’-O-Me, isorhamnetin)
103
and 4’-O- (Q-4’-O-Me, tamarixetin) methyl ethers, 3-O- (Q-3-O-SO3−) and 7-O- (Q-7-O-SO3−)
104
sulfates as well as 4’-O-β-D-glucuronide (Q-4’-O-Glr−) were also included into the studied set
105
of Q metabolites (Fig. 1). Some of these metabolites were detected in tissues of pigs and/or
106
rats [21] or in cellular cultures [7,22]. For some of the metabolites, in vitro data on RS activity
107
[11] and SA binding have been reported [19]. In this extensive study, quercetin and its
108
metabolites were considered in their neutral (Q-OH) and anionic (Q-O–) reduced states,
109
neutral aryloxyl monoradical intermediate state (Q-O•) as well as in doubly oxidized quinoid
110
states (Q(=O)2) (Scheme 1). In particular, effects of metabolic transformations on the double
111
sequential RS processes and on disproportionation of radical intermediate were investigated
112
to explore their contributions to the unusually high RS extent of Q [9]. The results obtained
113
by relatively inexpensive in silico methods, provide valuable mechanistic insights at
114
molecular level which can be further used for better understanding redox activity of
115
extensively metabolised polyphenol Q under in vitro and in vivo conditions. It is shown that
116
RS properties of most of conjugates are considerably changed as compared with the aglycon
AC C
EP
TE D
M AN U
SC
RI PT
94
5
ACCEPTED MANUSCRIPT 117
Q. They are determined by the type and site of conjugation. Regarding ways of
118
transportation through the body, conjugates of Q are expected to be distributed by other
119
ways than serum albumin transport.
120
2. Materials and methods
122
2.1 Quantum-chemical calculations of radical scavenging activities
123
The influence of methylation, sulfurylation and glucuronidation on RS efficiency and
124
mechanism of Q, was studied by applying quantum-mechanical density functional theory
125
(DFT) approach. The DFT model (U)B3LYP/6-31+G(d,p) was used in combination with
126
polarizable continuum model of solvation (IEFPCM) for the estimation of RS parameters
127
(Scheme 1) for Q and its major metabolites (Fig. 1) (Table 1) [23,24,25,26,27,28,29].
128
Recently, the similar analysis has been done for the set of Q catabolic ring fission phenolic
129
acid products (Supplementary Table S1) [6]. Sulfate (-OSO3–) and glucuronide (-OGlr–)
130
conjugates were considered with deprotonated, negatively charged -SO3– and -COO– groups,
131
respectively. Equilibrium structures in neutral, anionic and quinoid closed-shell ground
132
electronic states as well as in monoradical open-shell doublet ground electronic state, were
133
fully optimized in the gas phase. The minima were confirmed by no imaginary vibrational
134
frequencies. The expectation values of the
operator calculated by the (U)B3LYP/6-
135
31+G(d,p) model for the open-shall radical species FlO• were within the range 0.760-0.784.
136
The NBO spin density analysis was performed by using the NBO 6.0 software (Fig. S1) [30].
137
Calculated bond dissociation enthalpy (BDE) for hydrogen atom transfer (HAT) mechanism
138
[31] and proton affinity (PA) / electron transfer enthalpy (ETE) values for sequential proton
139
loss electron transfer (SPLET) pathway [32] in the gas phase as well as corresponding free
140
energies (BDFEaq and ∆Gbasicity (i.e. pKa) / ETFEaq, respectively) in aqueous phase, were used
AC C
EP
TE D
M AN U
SC
RI PT
121
6
ACCEPTED MANUSCRIPT as parameters for prediction of relative RS activities (Scheme 1) [27]. These thermodynamic
142
reaction parameters were estimated at temperature of 298.15 K, pressure of 1 atm for the
143
gas phase and standard state of 1 M for aqueous phase [27]. In the case of a set of flavonols,
144
significant linear correlation between aqueous parameters BDFEaq and ETFEaq was found,
145
demonstrating that BDFEaq can be used as a general parameter for estimating relative RS
146
capacities of phenolic compounds regardless underlying RS mechanism [27]. Hence, herein
147
we considered only double HAT to estimate relative capacity of a metabolite molecule to
148
scavenge two free radicals. DFT calculations were done by using Gaussian 03 software [33].
SC
RI PT
141
M AN U
149
2.2 Molecular docking into human serum albumin
151
Rigid, non-covalent molecular docking of Q and its conjugated metabolites including their
152
differently deprotonated species and doubly oxidized quinoid species, was performed into
153
HSA by using the program AutoDock 4.2 [34]. The crystal structure 2BXD of a complex of HSA
154
with R-warfarin was downloaded from Protein Data Bank [17]. The initial 3D conformations
155
of Q and its metabolites were taken from DFT calculations. The structure of a hybrid
156
conjugate 3'-O-Me-Q-3-O-Glr– and all reference ligands (Table S2) were generated by the
157
program Marvin 6.0.0 using Dreiding force field [35]. AutoDockTools 1.5.6 was applied for
158
preparation of compounds and HSA for molecular docking [36]: B-chain and ligands were
159
removed from the structure 2BXD; nonpolar hydrogens and lone pairs were merged and
160
Gasteiger partial atomic charges were assigned to HSA as well as to each compound
161
considered. The binding site was defined as the grid map centred at the point (4,-11,3), with
162
60x60x80 points spaced by 0.375 Å. Using the Lamarckian genetic algorithm (LGA), 100
163
docking calculations were performed per molecule with default values of other algorithm
164
parameters. Two clusters of binding conformations - the most stable cluster (E1) and most
AC C
EP
TE D
150
7
ACCEPTED MANUSCRIPT populated cluster (Emp) with the binding affinity window (Emp – E1) not greater than 2
166
kcal/mol, were analysed for finding representative binding mode(s) of a studied compound.
167
The docking procedure (binding site definition, use of LGA for conformational sampling and
168
choice of docking parameter values) was validated by docking warfarin and other known
169
strong and weak ligands of binding site I (BS I), and comparison of the obtained in silico
170
results with the wet HSA binding data available in literature (Table S2). The program PyMOL
171
was used for visualization and preparation of figures [37].
SC
172
RI PT
165
3. Results and discussion
174
3.1 Influence of phase II metabolic transformations on radical scavenging capacity of
175
quercetin
176
Q is a very efficient direct free radical scavenger with generally high stoichiometric
177
coefficient (∼5) [11,38]. The high free RS capacity of Q implies multi-step process and is
178
commonly ascribed to the presence of more RS OH groups in Q and to ability of Q to
179
regenerate the active reducing forms, the parent form or aryloxyl radical intermediate. The
180
redox active groups of Q, the catecholic group (3’-OH and 4’-OH) and 3-OH group are
181
coupled by the extended π electron conjugation involving the double C2=C3 bond, and
182
additionally stabilized by intramolecular H-bonds (Fig. 1 and Fig. S1). Structural modification
183
that impairs any of these structural features generally diminishes the RS capacity of Q.
184
Accordingly, conjugation reactions of phase II metabolism– glucuronidation, sulfation and
185
/or methylation, have been found to reduce the RS capacity of Q [10,12]. Herein, by means
186
of computations, the metabolic structural transformations were found to affect not only
187
capacity, but also dominant RS mechanism of Q, in great dependence upon the position of
188
conjugation (Table 1).
AC C
EP
TE D
M AN U
173
8
ACCEPTED MANUSCRIPT As a direct radical scavenger, Q may act through HAT [31] or SPLET [32] mechanism, in great
190
dependence upon the medium polarity and pH. In the gas phase, which may well
191
approximate effects of non-polar environment, HAT is a dominant RS mechanism of Q. The
192
4’-OH group is the 1st RS centre of Q considering that the aryloxyl radical Q-4’-O• is more
193
stable than the Q-3-O• one, with more delocalized spin density (Fig. S1) and accordingly
194
BDEg,1(4’-OH) of 74.2 kcal/mol lower than BDEg,1(3-OH) of 80.8 kcal/mol (Table 1). However,
195
in aqueous phase, the prevalent RS mechanism is SPLET since Q is deprotonated at its 4’-OH,
196
3-OH or 7-OH group [27]. In anionic state, Q is most prone to react as a radical scavenger at
197
3-O position. The formation of Q-3-O• radical was predicted to be favourable oxidation
198
process of Q in water: ETFEaq(3-OH) < ETFEaq (4’-OH) and BDFEaq,1(3-OH) < BDFEaq,1 (4’-OH)
199
(Table 1) [27]. In environment of either type, the intermediate aryloxyl radicals Q-4’-O• and
200
Q-3-O• even more readily release H-atom (BD(F)E2 < BD(F)E1, Table 1) and transform into the
201
corresponding quinoid form as a final product of double RS process (Scheme 1) [6].
202
It’s well known that any compound with the catechol moiety may exert RS activity [39]. High
203
RS activity of the catechol group is ascribed to the stabilisation of an aryloxyl radical through
204
donating H-bond by the adjacent OH group (Scheme 1) [31]. However, the RS mechanism of
205
the catechol fragment is highly dependent upon the rest of the structure. The acidity of the
206
4’-OH group is greater for quercetin (pKa ~ 7.5) than for the simple phenolic acids with the
207
catecholic moiety (pKa > 9), and thus SPLET is significant RS mechanism for Q, but not for its
208
catabolic products at the physiological pH (Table S1) [6]. Nonetheless, HAT parameter values
209
(BDEg, BDFEaq) of phenolic acidic catabolites (Table S1) were estimated to be comparable
210
with those of Q (Table 1). Greater RS efficiency of Q [11] may, hence, be at least partially
211
ascribed to the SPLET mechanism since deprotonation strongly accelerates the RS reaction in
212
polar solvents [32]. For most of Q conjugated metabolites, SPLET is also predicted to be a
AC C
EP
TE D
M AN U
SC
RI PT
189
9
ACCEPTED MANUSCRIPT dominant RS mechanism in polar environment (ETFEaaq < BDFEaaq,1, Table 1), what is in
214
accordance with the in vitro results showing that their RS efficiencies in the ABTS assay at pH
215
7.4 are generally greater than in the FRAP assay at the acidic pH 3.6 [10]. As for Q glycoside
216
isomers [28], it has also evidenced herein that for successful interpretation of RS activities of
217
Q metabolites, it is essential to take into account molecular acidity and solvent effects.
218
For the in vivo abundant metabolite Q-3-O-Glr–, BDFEaq,1 is significantly higher than ETFEaq
219
(Table 1) indicating that SPLET is its main RS mechanism in polar medium. Accordingly its RS
220
activity may be considerably influenced by medium polarity as it has already been observed
221
in vitro [10]. In difference, for its isomer Q-4’-O-Glr– the predicted RS mechanism is HAT
222
since 4’-O-glucuronidation considerably reduces the molecular acidity (Table 1) and
223
consequently the RS activity of this isomer should be quite independent upon medium
224
polarity and pH. Since for Q-4’-O-Glr– double RS and disproportionation reactions are also
225
cancelled by 4’-O substitution, this metabolite can act only as a single H-atom donor. The in
226
vitro RS activity of Q-4’-O-Glr– is considerably lower than for the Q-3-O-Glr– isomer what has
227
been illustrated by lag times of copper(II)-induced LDL oxidation decrease in the order Q-7-
228
O-Glr– ≈ Q > Q-3-O-Glr– ≈ Q-3-glucoside >> Q-4’-O-Glr– (at concentrations of 2 μM) [12].
229
Analogous trend in RS activities was observed for 3-O- and 4’-O glycoside isomers of Q [11].
230
Due to reduced molecular acidity, the main RS mechanism of 7-O-sulfo conjugate Q-7-O-SO3–
231
is also predicted to be HAT. The place of the 1st HAT in Q-7-O-SO3– depends upon the
232
medium polarity: in non-polar/polar medium it should happen at 4’-OH / 3-OH group
233
according to DFT calculations (Table 1). For Q-3-O-SO3–, the prevailing RS mechanism may
234
also be HAT, in difference to the corresponding glucuronide isomer 3-O-Glr–. In opposite to
235
the 3-O-Glr– group, the electron accepting 3-O-SO3– group, reduces electron donating
236
capacity of the active 4’-O site. The most abundant human plasma sulfo-conjugate Q-3’-O-
AC C
EP
TE D
M AN U
SC
RI PT
213
10
ACCEPTED MANUSCRIPT SO3– [5], is predicted to act as a single free radical scavenger at 3-OH group since its RS
238
activity at 4’-OH site and disproportionation propensity are considerably weaker than for
239
aglycon Q (Table 1) [10]. Its RS mechanism should also be determined by medium polarity
240
and free radical type, as well. The analogous, but much less reducing effect on the RS activity
241
of 4’-O position is observed in the case of neutral and relatively small methyl group at 3’-O-
242
position in isorhamnetin. For isorhamnetin, the RS process is also predicted to start at 3-OH
243
group and corresponding BDFEaq and ETFEaq values are comparable with those for Q (Table
244
1). However, the 2nd HAT process at 4’-OH position is predicted to be somewhat less efficient
245
than for Q, which may at least partially explain experimentally observed smaller RS capacity
246
of isorhamnetin than for the non-modified aglycon [11]. In comparison, the 4’-O-methylated
247
isomer tamarixetin has relatively weak RS activity because double reducing and
248
disproportionation properties are cancelled by 4’-O-conjugation. Since Q metabolites
249
modified at the 4’-OH group may not form quinone forms, this metabolic transformation
250
generally considerably attenuate not only RS, but also pro-oxidative and associated cytotoxic
251
effects of Q.
252
Q has ability to regenerate its reducing parent form [40]. The aryloxyl radical intermediate
253
Q-O• was postulated to quickly disproportionate to regenerate the parent Q and produce a
254
quinone form Q(=O)2 (Scheme 1). A produced quinoid tautomer may further bind
255
nucleophilic solvent molecule (e.g. methanol) resulting in regeneration of catechol at the
256
ring B which may again exert RS activity [40]. According to the values of reaction parameters
257
predicted for formation of aryloxyl radical and its disproportionation in gas and water (Table
258
1), recycling to the parent reducing form by disproportionation mechanism, is most
259
favourable for Q. In general, at least one of parameters calculated for metabolites (i.e.
260
BDFEaq, pKa, ETFEaq, or disproportionation reaction energy) is less favourable than for
AC C
EP
TE D
M AN U
SC
RI PT
237
11
ACCEPTED MANUSCRIPT aglycon Q (Table 1). Such in silico results are in accordance with considerably less in vitro RS
262
efficiency of metabolites as compared to their aglycon Q [11,38].
263
The disproportionation process favours formation of para-quinone methide in comparison
264
with ortho-quinone (Table 1). The four possible tautomers of the doubly oxidized Q may be
265
formed and herein we considered only two, ortho-quinone and para-quinone-methide which
266
can be formed directly by sequentially donating two H-atoms from the most active 4’- and 3-
267
hydroxyl groups (Scheme 1) [9]. The para-quinone-methide is estimated to be energetically
268
more favourable tautomer and thus it may be formed also by isomerization of ortho-quinone
269
tautomer. The significant role of para-quinone methide in prolongation of RS activity of Q,
270
may represent the additional significant contribution of 3-OH group and conjugated C2=C3
271
bond to the high RS capacity of Q as well as of Q metabolites with these structural
272
characteristics non-modified [41].
TE D
273
M AN U
SC
RI PT
261
3.2 Influence of phase II metabolic transformations on human serum albumin binding of
275
quercetin
276
Q and its metabolites were docked into the rigid structure of the extended Sudlow’s BS I
277
[17]. The docking space comprised BS I and the adjacent cleft located on another side of the
278
fluorescent residue Trp214, among sub-domains IIA, IIB and IIIA [42] (Fig. 2). The ligand
279
warfarin in the 2BXD structure of HSA is placed within BS I and its binding mode is denoted
280
as warfarin binding mode (WBM). In vitro displacement studies by fluorescence
281
spectroscopy, has shown that Q binds via non-covalent interactions to HSA in area close to
282
its amino acid residue Trp214, but in different region than warfarin [18,43,44]. The positively
283
charged hydrophilic cleft has been suggested to be a binding site of Q and Q-3-O-Glr– [43].
284
According to our results, in the predicted binding mode with the most favourable docking
AC C
EP
274
12
ACCEPTED MANUSCRIPT score E1 (Table 2), Q is also placed within the cleft with its catechol ring B interacting with
286
amino acid residues Ser202 and Trp214 of domain II, and its benzopyranone A/C making H-
287
bonds and van der Waals contacts with the amino acid residues Ser454, Leu457, Cys461,
288
Leu481, Arg484, Arg485 of the sub-domain IIIA (Fig. 2). However, in the most populated
289
binding mode (predicted by 66 of 100 LGA runs) with only somewhat less favourable binding
290
score Emp (Table 2), Q is located within BS I and its quercetin binding mode (QBM) partially
291
overlaps with WBM (Fig. 2). Thus, Q is expected to be bound by HSA in multiple ways. In
292
addition, the binding score Emp for QBM is quite higher than for the strong BS I ligand
293
warfarin (Table 2). Hence, Q is expected to be easily replaced by warfarin as it has already
294
been noticed in many in vitro studies [43]. The binding score of Q is also less favourable than
295
the binding scores for other strong BS I binders of comparable molecular weight:
296
indomethacin, piroxicam and ciprofloxacin (Table S2). The calculated binding scores for all
297
strong BS I ligands including warfarin are lower than -8.7 kcal/mol. For Q, its three anions as
298
well as for its two di-oxidized quinone forms considered (Fig. 1), the most populated modes
299
QBM have binding scores quite higher than -8.7 kcal/mol (Table 2).
SC
M AN U
TE D
EP
300
RI PT
285
Fig. 2. a) Electrostatic surface of HSA illustrating positively charged cleft and upper entrance into BS I. b) The
302
extended BS I of HSA (PDB: 2BXD) with binding poses of Q (QBM green; in the cleft cyan) and warfarin (WBM
303
red). Interacting amino acid residues and a few atom distances (---, Å) are marked.
304
AC C
301
305
Q and its metabolites are analogous compounds sharing the same flavone scaffold. Hence,
306
binding scores normalized by molecular weights were used for estimation of contribution of
307
metabolic substituents to binding affinities of HSA for metabolites in reference to the parent
308
aglycon Q (Table 2) [45]. HSA binds Q and its metabolites with generally similar affinity [12].
13
ACCEPTED MANUSCRIPT Deprotonation and methylation of OH groups are estimated to weakly strengthen binding of
310
Q to HSA in agreement with in vitro findings [16]. Among other metabolites, only for the
311
most abundant plasma sulfo-conjugate Q-3’-O-SO3–, somewhat stronger binding to BS I has
312
been predicted. The computed relative binding scores are consistent with the in vitro results
313
according to which the binding affinity of HSA is similar for 3-O- and 7-O-sulfo-conjugates
314
[19], and decreases in the order Q-3’-O- SO3–> Q-3- Glr–> Q-4’- Glr– [12,16].
315
Q, Q-3’-O-SO3– and Q-3-O-Glr– have been demonstrated to show protective effects against
316
peroxynitrite-induced nitration of tyrosine amino acid residues in HSA [13]. Since simple
317
phenolic acids did not inhibit the nitrotyrosine formation, it was assumed that the inhibitory
318
properties of Q, Q-3’-O-SO3– and Q-3-O-Glr– imply their binding to HSA. Herein, it is
319
predicted that sulfo-conjugates and 3-O-glycuronide may bind within BS I (Table 2) what
320
indicates their tentative protective effects against peroxynitrite-induced nitration of tyrosine
321
amino acid residue Tyr150 in the BS I of HSA. Predicted binding modes of Q-3’-O-SO3– and Q-
322
3-O-Glr– within BS I are, however, different pointing to their different protective reaction
323
mechanism. Q-3-O-Glr– is bound within BS I like Q in QBM with the redox active catechol ring
324
B placed close to Tyr150, while Q-3’-O-SO3–is bound in WBM with the B ring positioned at
325
the open side entrance and its redox active 3-OH group placed close to Tyr150 (Fig. 3).
326
Difference in binding mode may also indicate different effects of their HSA binding on their
327
RS activities [12]. In general, binding modes and affinities of doubly oxidized quinone forms
328
of Q and its metabolites are predicted to be like those of the parent species. This implies
329
that RS properties of Q and its metabolites may be retained in the HSA bound state.
AC C
EP
TE D
M AN U
SC
RI PT
309
330
14
ACCEPTED MANUSCRIPT 331
Fig. 3. Left: Difference in non-covalent binding of the most abundant human plasma mono-conjugated Q
332
metabolites Q-3-O-Glr (pink) and Q-3’-O-SO3 (yellow) within BS I of HSA. Atom distances (---) are in Å. Right:
333
DFT conformations of the three flexible metabolites.
–
–
334
RI PT
335
In difference to Q and other metabolites, binding of the most of considered sulfo-conjugated
337
species to HSA is predicted to be only within BS I. Their top-ranked clusters with the most
338
favourable binding score are also the most populated ones. They bind in WBM in the way
339
that enables negatively charged sulfo-group to interact with positively charged lysine
340
(Lys195, Lys199) and arginine (Arg218, Arg222) residues at the side entrance to BS I (Fig. 3).
341
Since they are estimated to be bound with weaker or similar affinity than warfarin (Table 2),
342
warfarin and other strong binders are expected to displace them and, consequently, they
343
may be released in blood. Similarly, the negatively charged glucuronyl group of Q-3-O-Glr– is
344
placed at the same position as the –O-SO3– group in both QBM and WBM (Fig. 3). However,
345
the isomer 4’-O-Glr– is predicted to protrude outside BS I into adjacent cleft due to its
346
stretched shape (Fig. 3). Relatively weaker HSA binding of glucuronyl-conjugates is in
347
accordance with their distribution by blood cells. The metabolite 3-O-Glr–accumulates within
348
the activated macrophages. Macrophages transport it to the acidic inflammatory sites where
349
β-glucosidase releases the aglycon which exerts antioxidant and other biological effects [9].
M AN U
TE D
EP
AC C
350
SC
336
351
4. Conclusions
352
Direct radical scavenging capacities and HSA binding of major mono- and di-conjugated
353
metabolites of the abundant plant polyphenol Q (Fig. 1) were studied by using DFT
354
((U)B3LYP/6-31+G(d,p)) and molecular docking computational approaches, respectively
355
[20,27,28]. The valuable mechanistic insights in contributions of the conjugated metabolites
15
ACCEPTED MANUSCRIPT to in vitro and in vivo activities of Q have been gained and supported by wet experimental
357
results available in literature [10-13,19].
358
Since phase II conjugation transformations of Q occur mainly at its radical scavenging –OH
359
groups, their attenuation effect on RS capacity of Q is primarily due to a reduced number of
360
active –OH groups, that is H-atoms that may participate in RS. However, at least one of the
361
calculated RS reaction parameters: BDFEaq (HAT), pKa, ETFEaq (SPLET), or aryloxyl radical
362
disproportionation energy (Scheme 1), was also found to be less favourable for metabolites
363
as compared to their aglycon Q (Table 1) [27]. The aryloxyl radical disproportionation
364
reaction resulting in the formation of the para-quinone methide product, was revealed to
365
contribute considerably to the extent of RS for unmodified Q. Such a finding may contribute
366
to the understanding of the importance of the free 3-OH group and conjugated C2=C3 bond
367
for unusually high RS capacity of Q [41].
368
The impact of the metabolic transformations on RS capacity and HSA binding of Q was found
369
to depend upon both, conjugation type and position. The dominant RS mechanism was
370
found to change upon conjugation. For example, in aqueous medium at physiological pH, the
371
dominant RS mechanism of Q is SPLET, while for its metabolites Q-7-O-SO3– and Q-4’-O-Glr–,
372
due to decreased OH acidity, it is HAT, which implies that these species target different free
373
radicals.
374
Most of Q metabolites were predicted to be bound by HSA in the area close to its
375
fluorescent amino acid residue Trp214, non-specifically, at multiple sites with similar and
376
relatively low affinity. Only sulfo-conjugates with anchoring interactions of their –O-SO3–
377
group with positively charged lysine and arginine residues at the side entrance to BS I, were
378
found to bind only within the BS I, similarly to warfarin, but with lower binding affinity.
379
Accordingly, albumin binding is not expected to contribute considerably to body distribution
AC C
EP
TE D
M AN U
SC
RI PT
356
16
ACCEPTED MANUSCRIPT of Q and its metabolites. The other ways of distribution such as uptake by macrophages is
381
expected to be more important [9]. Since the double oxidized quinoid forms share the same
382
binding mode and have similar or higher binding affinities as corresponding parent reducing
383
species (Table 2), RS properties of Q and its metabolites is expected to be similar in their HSA
384
bound state.
RI PT
380
385
Supplementary material
387
The supplemental file includes radical scavenging data for phenolic acid catabolites, binding
388
parameters for known ligands of HSA binding site I and spin density distributions for neutral
389
monoradicals of Q and its metabolites. The Cartesian coordinates for DFT optimized
390
structures used for calculations of parameters given in Table 1, are provided in the sdf
391
format.
M AN U
SC
386
TE D
392
Conflicts of interest
394
There are no conflicts to declare.
395
EP
393
Acknowledgment The research was supported by the basic funding of the Croatian Ministry of
397
Science and Education, the program of The Scientific Centre of Excellence for Marine Bioprospecting
398
–BioProCro (Competitiveness and Cohesion Operational Program, European Regional Development
399
Fund KK.01.1.1.01), as well as by the projects of the Ministry of Science of the Republic of Serbia
400
(172015 and 174028). All authors are grateful to Jelena Đorović for participating in computations.
401
Croatian authors are thankful to Croatian Ministry of Science and Education for supporting the
402
computational cluster Isabella (http://www.srce.unizg.hr/en/usluge/isabella-cluster).
AC C
396
403 404
17
ACCEPTED MANUSCRIPT References
[1] A. Vogiatzoglou, A.A. Mulligan, M.A. Lentjes, R.N. Luben, J.P. Spencer, H. Schroeter, K.T. Khaw, G.G. Kuhnle, Flavonoid intake in European adults (18 to 64 years), PLoS One, 10 (2015) e0128132 [2] D. Del Rio, A. Rodriguez-Mateos, J.P.E. Spencer, M. Tognolini, G. Borges, A. Crozier, Dietary
chronic diseases, Antioxid. Redox Signal. 18 (2013) 1818-1892.
RI PT
(poly)phenolics in human health: Structures, bioavailability,and evidence of protective effects against
[3] J. Lee, A.E. Mitchell, Pharmacokinetics of quercetin absorption from apples and onions in healthy
humans, J. Agric. Food Chem. 60 (2012) 3874-3881.
[4] T. Walle, U.K. Walle, P.V. Halushka, Carbon dioxide is the major metabolite of quercetin in
SC
humans, J. Nutr. 131 (2001) 2648–2652.
[5] A.J.Day, F. Mellon, D. Barron, G. Sarrazin, M.R. Morgan, G. Williamson, Human metabolism of
M AN U
dietary flavonoids: Identification of plasma metabolites of quercetin, Free Radic. Res. 35, (2001) 941952.
[6] A. Amić, B. Lučić, V. Stepanić, Z. Marković, S. Marković, J.M. Dimitrić Marković, D. Amić, Free
radical scavenging potency of quercetin catecholic colonic metabolites: Thermodynamics of 2H+/2eprocesses, Food Chem. 218 (2017) 144-151.
[7] H. van der Woude, M.G. Boersma, J. Vervoort, I.M. Rietjens, Identification of 14 quercetin phase II
TE D
mono- and mixed conjugates and their formation by rat and human phase II in vitro model systems, Chem. Res. Toxicol. 17 (2004) 1520-1530.
[8] Y.Y. Lee-Hilz, M. Stolaki, W.J. van Berkel, J.M. Aarts, I.M. Rietjens, Activation of EpRE-mediated gene transcription by quercetin glucuronides depends on their deconjugation, Food Chem. Toxicol.
EP
46 (2008) 2128-2134.
[9] V. Stepanić, A. Čipak Gašparović, K. Gall Trošelj, D. Amić, N. Žarković, Selected attributes of polyphenols in targeting oxidative stress in cancer, Curr. Top. Med. Chem. 15 (2015) 496-509.
AC C
405
[10] M. Dueñas, F. Surco-Laos, S. Gonzalez-Manzano, A.M. Gonzalez-Paramas, C. Santos-Buelga, Antioxidant properties of major metabolites of quercetin, Eur. Food Res. Technol. 232 (2011) 103111.
[11] P. Goupy, C. Dufour, M. Loonis, O. Dangles, Quantitative kinetic analysis of hydrogen transfer reactions from dietary polyphenols to the DPPH radical, J. Agric. Food Chem. 51 (2003) 615-622. [12] K.M. Janisch, G. Williamson, P. Needs, G.W. Plumb, Properties of quercetin conjugates: modulation of LDL oxidation and binding to human serum albumin, Free Radic. Res. 38 (2004) 877884.
18
ACCEPTED MANUSCRIPT [13] A. Yokoyama, H. Sakakibara, A. Crozier, Y. Kawai, A. Matsui, J. Terao, S. Kumazawa, K. Shimoi, Quercetin metabolites and protection against peroxynitrite-induced oxidative hepatic injury in rats, Free Radic. Res. 43 (2009) 913-921. [14] A.J. Day, Y. Bao, M.R. Morgan, G. Williamson, Conjugation position of quercetin glucuronides and effect on biological activity, Free Radic. Biol. Med. 29 (2000) 1234-1243.
RI PT
[15] Y. Steffen, C. Gruber, T. Schewe, H. Sies, Mono-O-methylated flavanols and other flavonoids as inhibitors of endothelial NADPH oxidase, Arch. Biochem. Biophys. 469 (2008) 209-219.
[16] K. Beekmann, L. Actis-Goretta, P.J. van Bladeren, F. Dionisi, F. Destaillats, I.M. Rietjens, A stateof-the-art overview of the effect of metabolic conjugation on the biological activity of flavonoids,
SC
Food Funct. 3 (2012) 1008-1018.
[17] J. Ghuman, P.A. Zunszain, I. Petitpas, A.A. Bhattacharya, M. Otagiri, S. Curry, Structural basis of the drug-binding specificity of human serum albumin, J. Mol. Biol. 353 (2005) 38-52.
M AN U
[18] D.W. Boulton, K.U. Walle, T. Walle, Extensive binding of the bioflavonoid quercetin to human plasma proteins, J. Pharm. Pharmacol. 50 (1998) 243-249.
[19] C. Dufour, O. Dangles, Flavonoid-serum albumin complexation: determination of binding constants and binding sites by fluorescence spectroscopy, Biochim. Biophys. Acta. 1721 (2005) 164173.
TE D
[20] M. Skrt, E. Benedik, C. Podlipnik, N.P. Ulrih, Interactions of different polyphenols with bovine serum albumin using fluorescence quenching and molecular docking. Food Chem. 135 (2012) 24182424.
[21] V.C. de Boer, A.A. Dihal, H. van der Woude, I.C. Arts, S. Wolffram, G.M. Alink, I.M. Rietjens, J.
EP
Keijer, P.C. Hollman, Tissue distribution of quercetin in rats and pigs, J. Nutr. 135 (2005) 1718-1725. [22] D.J. Jones, J.H. Lamb, R.D. Verschoyle, L.M. Howells, M. Butterworth, C.K. Lim, D. Ferry, P.B. Farmer, A.J. Gescher, Characterisation of metabolites of the putative cancer chemopreventive agent
AC C
quercetin and their effect on cyclo-oxygenase activity, Br. J. Cancer. 91 (2004) 1213-1219. [23] A.D. Becke, Density-functional thermochemistry. III. The role of exact exchange, J. Chem. Phys. 98 (1993) 5648-5652.
[24] S.H. Vosko, L. Wilk, M. Nusair, Accurate spin-dependent electron liquid correlation energies for local spin density calculations: A critical analysis, Can. J. Phys. 58 (1980) 1200-1211. [25] P.C. Hariharan, J.A. Pople, The influence of polarization functions on molecular orbital hydrogenation energies, Theo. Chim. Acta. 28 (1973) 213-222. [26] M.J. Frisch, J.A. Pople, J.S. Binkley, Self-consistent molecular orbital methods. 25. Supplementary functions for Gaussian basis sets, J. Chem. Phys. 80 (1984) 3265-3269.
19
ACCEPTED MANUSCRIPT [27] V. Stepanić, K. Gall Trošelj, B. Lučić, Z. Marković, D. Amić, Bond dissociation free energy as a general parameter for flavonoid radical scavenging activity, Food Chem. 141 (2013) 1562-1570. [28] L. Lespade, S. Bercion, Theoretical investigation of the effect of sugar substitution on the antioxidant properties of flavonoids, Free Radic. Res. 46 (2012) 346-358. [29] M.T. Cancès, B. Mennucci, J. Tomasi, New integral equation formalism for the polarizable
RI PT
continuum model: Theoretical background and applications to isotropic and anisotropic dielectrics, J. Chem. Phys. 107 (1997) 3032-3304.
[30] E.D. Glendening, C.R. Landis, F. Weinhold, NBO 6.0: Natural bond orbital analysis program, J. Comput. Chem. 34 (2013) 1429-1437.
SC
[31] J.S. Wright, E.R. Johnson, G.A. Di Labio, Predicting the activity of phenolic antioxidants:
theoretical method, analysis of substituent effects, and application to major families of antioxidants, J. Am. Chem. Soc. 123 (2001) 1173-1183.
M AN U
[32] G. Litwinienko, K.U. Ingold, Solvent effects on the rates and mechanisms of reaction of phenols
with free radicals, Acc. Chem. Res. 40 (2007) 222-230.
[33] M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria, M.A. Robb, J.R. Cheeseman, et al. (2004) Gaussian 03, Revision E.01. Wallingford CT: Gaussian, Inc.
[34] G.M. Morris, D.S. Goodsell, R.S. Halliday, R. Huey, W.E. Hart, R.K. Belew, A.J. Olson, Automated
Chem. 19 (1998) 1639-1662.
TE D
docking using a Lamarckian genetic algorithm and empirical binding free energy function, J. Comput.
[35] Marvin 6.0.0. http://www.chemaxon.com Accessed 1 December 2017. [36] G.M. Morris, R. Huey, W. Lindstrom, M.F. Sanner, R.K. Belew, D.S. Goodsell, A.J. Olson,
EP
AutoDock4 and AutoDockTools4: automated docking with selective receptor flexiblity, J. Comput. Chem. 16 (2009) 2785-2791.
[37] The PyMOL Molecular Graphics System, Version 1.8 Schrödinger, LLC
AC C
[38] D. Villaño, M.S. Fernández-Pachón, M.L. Moyá, A.M. Troncoso, M.C. García-Parrilla, Radical scavenging ability of polyphenolic compounds towards DPPH free radical, Talanta 71 (2007) 230-235. [39] J. Baell, M.A. Walters, Chemistry: Chemical con artists foil drug discovery, Nature 513 (2014) 481-483.
[40] O. Dangles, G. Fargeix, C. Dufour, One-electron oxidation of quercetin and quercetin derivatives in protic and non protic media, J. Chem. Soc., Perkin Trans. 2, 12 (1999) 1387-1395. [41] P. Trouillas, P. Marsal, D. Siri, R. Lazzaroni, J.L. Duroux, A DFT study of the reactivity of OH groups in quercetin and taxifolin antioxidants: the specificity of the 3-OH site, Food Chem. 97 (2006) 679688.
20
ACCEPTED MANUSCRIPT [42] A. Bujacz, K. Zielinski, B. Sekula, Structural studies of bovine, equine, and leporine serum albumin complexes with naproxen, Proteins 82 (2014) 2199-2208. [43] H. Rimac, C. Dufour, Ž. Debeljak, B. Zorc, M. Bojić, Warfarin and flavonoids do not share the same binding region in binding to the IIA subdomain of human serum albumin, Molecules 22 (2017) 1153.
RI PT
[44] B. Sengupta, P.K. Sengupta, The interaction of quercetin with human serum albumin: a fluorescence spectroscopic study, Biochem. Biophys. Res. Commun. 299 (2002) 400-403.
[45] C. Abad-Zapatero, Ligand efficiency indices for effective drug discovery, Expert Opin. Drug
EP
TE D
M AN U
SC
Discov. 2 (2007) 469-488.
AC C
Fig. 1. Quercetin and its conjugated metabolites detected in vivo and/or in vitro and considered here in silico.
21
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
Scheme 1 Assumed RS mechanisms of Q and its metabolites: (a) hydrogen atom transfer (HAT) and (b) •
sequential proton loss electron transfer (SPLET). In double HAT pathway (a), aryloxyl monoradical Q-O is an intermediate and quinoid specie Q(=O)2 is a final product. (c) Disproportionation of intermediate radical Q-O
AC C
EP
TE D
regenerates the parent reducing form Q-OH.
•
22
SC
RI PT
ACCEPTED MANUSCRIPT
Fig. 2. a) Electrostatic surface of HSA illustrating positively charged cleft and upper entrance into BS I. b) The
M AN U
extended BS I of HSA (PDB: 2BXD) with binding poses of Q (QBM green; in the cleft cyan) and warfarin (WBM
AC C
EP
TE D
red). Interacting amino acid residues and a few atom distances (---, Å) are marked.
Fig. 3. Left: Difference in non-covalent binding of the most abundant human plasma mono-conjugated Q –
–
metabolites Q-3-O-Glr (pink) and Q-3’-O-SO3 (yellow) within BS I of HSA. Atom distances (---) are in Å. Right: DFT conformations of the three flexible metabolites.
23
ACCEPTED MANUSCRIPT
Table 1 In silico parameters (kcal/mol) for estimating relative RS capacities for the most reactive 4’-OH and 3-OH groups of Q and its major
a
b
BDEg,2
Q
4’
74.2
76.2/72.9
Q-3’-OCH3
4’
81.8
Q-3’-O-SO3
4’
Q-3-O-SO3
−
− Q-7-O-SO3
b
c
∆GDISPRO,aq
In vitro RS activities expressed as TEAC values
2.1/-1.2
-4.8/-6.7
408.3 ; 7.52// 3.07
-8.4
-13.2
2.13// 1.82
-24.4
-36.7
93.0/ 452.7 ; 1.83 //1.42
6.5
-0.06
Basicityaq (pKa)
71.6
66.8/64.9
8.1 (6.0)
70.8
73.4
76.7
63.6
11.8 (8.6)
72.3
93.9
69.5
89.8
53.0
7.3 (5.3)
89.8
4’
73.7
69.6
68.7
68.0
2.7 (2.0)
73.3
4’
69.2
74.3/69.8
73.3
66.8/65.6
12.8 (9.3)
67.8
5.1
-6.4
4’
81.6
71.7
72.3
68.3
8.0 (5.9)
68.2
6.2
-0.1
331.7/ 131.1 ; 3.76 //0.72
Q
3
80.8
66.3
70.3
66.2
9.5 (7.0)
68.2
-14.5
-4.1
7.52// 3.07
Q-3’-O- Me
3
80.5
74.7
70.2
70.1
8.9 (6.6)
68.6
-5.8
-0.1
2.13// 1.82
Q-4’-O- Me
3
82.1
/
69.1
/
9.1 (6.7)
67.3
/
/
3.39// 1.65
Q-3’-O- SO3
3
78.4
85.0
71.1
71.7
9.6 (7.1)
68.8
0.8
-1.6
1.83 //1.42
−
3
77.9
61.1
71.3
67.6
12.5 (9.2)
66.1
-16.8
-3.7
Q-3-O-Glr
−
−
Q-7-O- SO3
−
d
ETFEaq
∆HDISPRO,g
BDFEaq,2
−
BDFEaq,1
SC
BDEg,1
M AN U
Active OH
TE D
MOLECULE
RI PT
metabolites in gas (g, non-polar) and aqueous (aq, polar) phases as well as in vitro measured RS parameters.
e
f
f e
f
e
f
f f f f
e
AC C
EP
111.7/258.6 Q-4’-O-Glr 3 78.2 / 69.3 / 24.2 (17.8) 75.5 / / a b Active group for the first HAT or deprotonation. The second consecutive HAT process is assumed to occur at the other of these two OH-groups (Scheme 1). BDEg and BDFEg values were calculated by using corresponding B3LYP/6-31+G** values for H-atom of -0.497912 a.u. and 0.510927 a. u., respectively. Index 1 or 2 indicates the first c θ • θ + θ − and second step of double HAT mechanism. Values pKa and ETFEaq correspond to the first SPLET process. For ΔG hyd(H ), ΔG hyd(H ) and ΔG hyd (e ), the experimental values d st nd of 27.8 kJ/mol, -1104.62 kJ/mol and -148.5 kJ/mol were used, respectively [23]. The 1 / 2 value correspond to the formation of ortho-quinone and para-quinone e − methide, respectively (Scheme 1). Lag time for LDL oxidation /min at 2.0 μM of free / HSA bound compound [12]. Values for free/ HSA bound Q-7-O-Glr and 3'-O-Me-Q-3– f O-Glr in min are 421.7 / 279.9 and 98.3 /330.9, respectively. Values determined by ABTS//FRAP assays [10]. For comparison ABTS/FRAP TEAC values for HCA are (2.83 / 0.08).
1
ACCEPTED MANUSCRIPT
metabolites. The most stable cluster C
E1
C
7
-8.09
–
C
13
Q-3- O
–
C
Q-7- O
–
-7.21
1.00
82
-7.37
1.03
83
-7.69
1.07
% conf.
1.00
QBM
66
-8.68
1.08
QBM
3
-8.12
1.01
QBM
C
5
-8.08
1.00
Q-o- quinone
C
1
-8.96
1.11
Q-p- quinone methide
QBM
68
-7.98
0.99
0.95
c
Q Q-4'-O
C
1
-8.06
–
C
1
-9.32
–
C
3
-9.36
–
Q-3'-O-Me-7- O
C
14
-8.76
Q-3'-O-Me-p-qm
QBM
65
-7.71
Q-4'-O-Me
Q-3'-O-Me-3- O
Q-3'-O-Me-4'- O
-7.63
1.06
QBM
74
-7.54
1.05
QBM
68
-7.98
1.11
QBM
57
-7.27
0.96
QBM
81
-7.83
1.04
1.11
QBM
72
-7.94
1.06
1.04
QBM
19
-7.58
1.01
0.92
QBM
65
-7.71
1.03
3
-8.55
1.01
QBM
27”
-7.43
0.98
C
2
-8.26
0.98
QBM
43
-7.94
1.06
–
QBM
31
-7.72
0.91
QBM
31
-7.72
1.03
WBM
81
-8.47
0.83
WBM
81
-8.47
0.93
–
C
13
-8.91
0.88
WBM
60
-8.5
0.94
–
WBM
88
-8.96
0.88
WBM
88
-8.96
0.99
C
5
-8.66
0.85
WBM
61
-8.24
0.91
Q-4'-O-Me-7- O
Sulfated metabolites −
Q-3-O-SO3 −
Q-3- O-SO3 4'- O O
− Q-3-O-SO3 -o-quinone
AC C
C –
Q-4'-O-Me-3- O
− Q-3-O-SO3 -7-
79
1.10
EP
Q-3'-O-Me
Emp
QBM
TE D
Methylated metabolites
2
Emp,rel
Site
Metabolite
E1,rel
b
SC
% conf.
The most populated cluster c
M AN U
Site
b
RI PT
Table 2 Predicted binding modesa and scores (kcal/mol)b as well as in vitro determined binding parameters for Q and its major conjugated
In vitro HSA binding parameters c d
5.73 ; 50
58
e
e
ACCEPTED MANUSCRIPT
inv
31
-8.24
0.81
WBM
WBM
inv
53
-8.65
0.85
WBM
WBM
inv
WBM
−
Q-7-O-SO3 -o-q − Q-7-O-SO3 -p-qm − Q-3'-O-SO3
39
-7.73
0.85
inv
53
-8.65
0.96
inv
67
-8.29
0.92
13
-8.75
0.86
WBM
WBM
47
-8.54
0.84
WBM
47
Q-3'-O-SO3 -3- O
–
WBM
40
-9.12
0.90
WBM
40
− Q-3'-O-SO3 -4'-
–
WBM
52
-9.25
0.91
WBM
52
WBM
45
-8.93
0.88
WBM
45
WBM
64
-8.76
0.85
WBM
WBM
15
-9.15
0.72
− Q-3'-O-SO3 -7
O
O
–
— Q-3'-O-SO3 -p-qm
Q-3-O-Glr
– –
QBM
50
-9.11
0.71
–
–
WBM
21
-9.47
0.74
WBM
27
-10.06
0.79
WBM
6
-6.84
0.54
QBM
36
-9.15
0.70
WBM
68
-8.83
Q-3-O-Glr -7- O –
Q-3-O-Glr -o-q Q-4'-O-Glr
–
3'-O-Me-Q-3O-Glr
–
R-Warfarin a
1.07
inv
WBM- warfarin binding mode; QBM – quercetin binding mode, C- cleft, WBM b
0.94
-9.12
1.01
-9.25
1.02
-8.93
0.98
-8.76
0.96
e
87
d
4.28
d
QBM
37
-8.11
0.71
QBM
50
-9.11
0.80
QBM
34
-8.51
0.75
QBM
29
-9.05
0.80
C
15
-6.33
0.56
20.05
QBM
36
-9.15
0.78
12.92
WBM
68
-8.83
1.20
TE D
–
Q-3-O-Glr -4' O
64
M AN U
Glucuronidated
-8.54
SC
−
inv
RI PT
−
Q-7-O-SO3
8.54
d d
– flipped WBM, with the ring B at the position of the coumarin core, and benzopyrone at
c
EP
the entrance side; E1 and Emp are the lowest binding energies in the first and most populated clusters of binding conformations, respectively. % conf. denotes the number of conformations in the corresponding cluster. The binding energies E1 and Emp normalized with molecular weights and expressed relative to corresponding normalized d
e
3
-1
AC C
values for Q, in order to estimate contributions to binding affinity per mass unit. Concentration required for 50% fluorescence quenching of Trp214 [12]. The ligand–HSA o
binding constants (10 M ) for compounds forming fluorescent intermolecular complexes (pH 7.4 phosphate–NaCl buffer, 25 C) [19].
3
ACCEPTED MANUSCRIPT Highlights Radical scavenging (RS) properties of quercetin (Q) are altered by metabolism.
-
Dominant RS mechanism of metabolites depends on conjugation type and site.
-
Disproportionation regeneration of scavenging form is most efficient for aglycon.
-
Only sulfoconjugates bind specifically to human serum albumin (HSA).
-
RS properties of Q and its metabolites may be retained in the HSA bound state.
AC C
EP
TE D
M AN U
SC
RI PT
-