Dioxybenzone triggers enhanced estrogenic effect via metabolic activation: in silico, in vitro and in vivo investigation

Article abstract of DOI:10.1016/j.envpol.2020.115766

Dioxybenzone is widely used in cosmetics and personal care products and frequently detected in multiple environmental media and human samples. However, the current understanding of the metabolic susceptibility of dioxybenzone and the potential endocrine disruption through its metabolites in mimicking human estrogens remains largely unclear. Here we investigated the in vitro metabolism of dioxybenzone, detected the residue of metabolites in rats, and determined the estrogenic disrupting effects of these metabolites toward estrogen receptor α (ERα). In vitro metabolism revealed two major metabolites from dioxybenzone, i.e., M1 through the demethylation of methoxy moiety and M2 through hydroxylation of aromatic carbon. M1 and M2 were both rapidly detected in rat plasma upon exposure to dioxybenzone, which were then distributed into organs of rats in the order of livers > kidneys > uteri > ovaries. The 100 ns molecular dynamics simulation revealed that M1 and M2 formed hydrogen bond to residue Leu387 and Glu353, respectively, on ERα ligand binding domain, leading to a reduced binding free energy. M1 and M2 also significantly induced estrogenic effect in comparison to dioxybenzone as validated by the recombinant ERα yeast two-hybrid assay and uterotrophic assay. Overall, our study revealed the potential of metabolic activation of dioxybenzone to induce estrogenic disrupting effects, suggesting the need for incorporating metabolic evaluation into the health risk assessment of benzophenones and their structurally similar analogs. Dioxybenzone was metabolized into two major metabolites via hydroxylation and demethylation, which exhibited higher binding affinity and agonistic activity toward ERα than dioxybenzone.

Full text of DOI:10.1016/j.envpol.2020.115766

Journal Pre-proof
Dioxybenzone Triggers Enhanced Estrogenic Effect via Metabolic Activation: in Silico
and in Vitro Investigation
Tingjie Zhan, Leili Zhang, Shixuan Cui, Weiping Liu, Ruhong Zhou, Shulin Zhuang
PII:
S0269-7491(20)36455-1
DOI:
https://doi.org/10.1016/j.envpol.2020.115766
ENPO 115766
Reference:
To appear in: Environmental Pollution
Received Date: 1 July 2020
Revised Date: 28 September 2020
Accepted Date: 29 September 2020
Please cite this article as: Zhan, T., Zhang, L., Cui, S., Liu, W., Zhou, R., Zhuang, S., Dioxybenzone
Triggers Enhanced Estrogenic Effect via Metabolic Activation: in Silico and in Vitro Investigation,
Environmental Pollution, https://doi.org/10.1016/j.envpol.2020.115766.
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition
of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of
record. This version will undergo additional copyediting, typesetting and review before it is published
in its final form, but we are providing this version to give early visibility of the article. 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.
© 2020 Elsevier Ltd. All rights reserved.

CRediT authorship contribution statement
Shulin Zhuang and Ruhong Zhou: Conceptualization, Writing - Reviewing and
Editing. Tingjie Zhan: Methodology, Investigation, Writing - original draft. Leili
Zhang: Investigation, Formal analysis, Writing - original draft. Shixuan Cui: Data
curation, Validation, Formal analysis. Weiping Liu: Supervision, Writing - Review &
Editing.

Graphical Abstract
1

1
2
Dioxybenzone Triggers Enhanced Estrogenic Effect via Metabolic Activation: in Silico and in
Vitro Investigation
3
4
Tingjie Zhana, Leili Zhangb, Shixuan Cuia, Weiping Liua, Ruhong Zhoub, c, Shulin Zhuanga, *
a
5
Key Laboratory of Environment Remediation and Ecological Health, Ministry of Education,
6
College of Environmental and Resource Sciences, Zhejiang University, 310058 Hangzhou, China.
b Computational Biological Center, IBM Thomas J. Watson Research Center, Yorktown Heights, NY,
United States.
7
8
9
c Department of Chemistry, Columbia University, New York, NY, United States.
10
11
12
* Correspondence authors: (S. Z.) shulin@zju.edu.cn
1

13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
Abstract
Dioxybenzone is widely used in cosmetics and personal care products and frequently detected in
multiple environmental media and human samples. However, the current understanding of the
metabolic susceptibility of dioxybenzone and the potential endocrine disruption through its
metabolites in mimicking human estrogens remains largely unclear. Here we investigated the in vitro
metabolism of dioxybenzone, detected the residue of metabolites in rats, and determined the
estrogenic disrupting effects of these metabolites toward estrogen receptor α (ERα). In vitro
metabolism revealed two major metabolites from dioxybenzone, i.e., M1 through the demethylation
of methoxy moiety and M2 through hydroxylation of aromatic carbon. M1 and M2 were both rapidly
detected in rat plasma upon exposure to dioxybenzone, which were then distributed into organs of
rats in the order of livers > kidneys > uteri > ovaries. The 100 ns molecular dynamics simulation
revealed that M1 and M2 formed hydrogen bond to residue Leu387 and Glu353, respectively, on
ERα ligand binding domain, leading to a reduced binding free energy. M1 and M2 also significantly
induced estrogenic effect in comparison to dioxybenzone as validated by the recombinant ERα yeast
two-hybrid assay and uterotrophic assay. Overall, our study revealed the potential of metabolic
activation of dioxybenzone to induce estrogenic disrupting effects, suggesting the need for
incorporating metabolic evaluation into the health risk assessment of benzophenones and its
structurally similar analogs.
30
31
2

32
33
34
35
36
37
Capsule Abstract
Dioxybenzone was metabolized into two major metabolites via hydroxylation and demethylation,
which exhibited higher binding affinity and agonistic activity toward ERα than dioxybenzone.
Keywords: dioxybenzone, metabolism, estrogenic effect, molecular modeling, mass spectrometry
3

38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ꢀ Introduction
Dioxybenzone, a commercial benzophenone-type ultraviolet light absorbent, is widely used in many
personal care products (PCPs) (Lu et al. 2018). It has also been added in polymers and surface
coatings for industrial products, such as plastics, food packaging and textiles, to limit
photodegradation (Tsui et al. 2017). Exposure to dioxybenzone occurred through direct skin contact,
ingestion of its residues in environmental water at the ng/L level, and inhalation of dust and soil
particles at the ng/g level (Tsui et al. 2015; Wang et al. 2013; Wu et al. 2013). Evidences show that
dioxybenzone has been detected in the urine of children 8-11 years old residing in China with the
detection frequency up to 78.3% (Lu et al. 2018), and in the urine, blood and breast milk of adults
with measurable levels at ꢀg/L (Gao et al. 2015; Rodriguez-Gomez et al. 2015; Tarazona et al. 2013).
It has also been of a great threat to aquatic organisms, as evidenced by its frequent detection in coral
tissues with mean concentration of 24.7 ng/g (Tsui et al. 2017).
Within the adverse outcome pathway (AOP) framework, the interaction between chemical
polluants and ERα is identified as one of the key molecular initiating events attributed to endocrine
disruption and adverse reproductive health effect, such as the abnormalities in cyclicity and fertility
(Becker et al. 2015; Suvorov and Waxman 2015). In vivo study observed that benzophenones elicited
ERα-dependent uterine weight increase in rats, and benzophenone-1 (Table S1) showed association
with the progression of endometriosis (Kunisue et al. 2012; Schlumpf et al. 2001). Several
benzophenones exerted estrogenic potency mainly by interfering with estrogen receptor (ER) α and
ERα-mediated signaling (Miller et al. 2001). The exposure to benzophenones with different side
chains resulted in the upregulation of estrogen target genes and the induction of proliferation in ERα
expressing breast cancer cells (Kerdivel et al. 2013; Meng et al. 2020). Dioxybenzone was reported
to exhibit weak estrogenic effects as revealed by an in vitro test using MCF-7 cells and decreased
4

61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
hatching rates of larvae in zebrafish (Kerdivel et al. 2013; Zheng et al. 2020). Additionally, a
growing body of evidence demonstrated that the metabolites of some benzophenones were detected
in human samples (Molins-Delgado et al. 2018). Benzophenone-2 (Table S1) was easily metabolized
into glucurono- and sulfo-conjugated metabolites in a hepatic cell line (Le Fol et al. 2015). In vitro
metabolism of benzophenone-3 (Table S1) resulted in the formation of its metabolites, including
dioxybenzone and benzophenone-1 with higher estrogenic potency, which were also formed in rats
and human after an in vivo exposure (Kim and Choi 2014; Tarazona et al. 2013; Yoko Watanabe et
al. 2015). However, the metabolic susceptibility of dioxybenzone remained unknown. The potential
metabolism was reported to add difficulty in distinguishing the real adverse outcome between the
parent compound and its metabolites, posing challenge to perform adequate in vitro-to-in vivo
extrapolations (Leist et al. 2017). Thus, the investigation on whether the metabolism of
dioxybenzone induced the deteriorated estrogenic disrupting effects towards ERα is highly needed.
In this study, we first analyzed the structure of dioxybenzone metabolites by
HPLC-QTOF-MSMS after in vitro incubation with liver microsome (Table S1). The major
metabolites of dioxybenzone were further detected in plasma and target organs of female rats after
oral exposure to dioxybenzone. We used the 100 ns molecular dynamics simulations to probe the
distinct binding characteristics of dioxybenzone and its major metabolites toward ERα ligand
binding domain (LBD). The estrogenic disrupting effects were further confirmed by the recombinant
ERα yeast two-hybrid assay and uterotrophic assay. Our findings here reinforce the important notion
that the full assessment of the estrogenic effect of the test chemicals, dioxybenzone included, should
consider the full spectrum of their active metabolites.
ꢀ Materials and Methods
5

83
84
Chemicals. Dioxybenzone (catalog D0586, > 98% purity) and benzophenone-6 (BP-6, catalog
D0575, > 90% purity, Table S1) were obtained from TCI Development Co., Ltd. (Shanghai, China).
17β-Estradiol (E2, catalog 140534, 99% purity) and formic acid (catalog 299272, 98% purity) were
obtained from J&K Scientific Ltd. (Beijing, China). Dimethyl sulfoxide (DMSO, catalog
60313ES60, 99.5% purity) was obtained from Yeasen Biotech Co., Ltd. (Shanghai, China).
β-Glucuronidase/arylsulfatase (catalog 10127698001), ethyl acetate (catalog 270989, 99.8% purity),
the acetonitrile (catalog 271004, 99.8% purity) and methanol (catalog 34860, 99.8% purity) of
chromatographic grade were obtained from Sigma-Aldrich (St. Louis, MO, USA).
85
86
87
88
89
90
91
In vitro metabolism. The in vitro metabolism of dioxybenzone mediated by human liver microsome
(HLM) was conducted following reported protocols (Lu and Wong 2011; Lv et al. 2017; Warner et
al. 2009). Dioxybenzone was incubated with pooled HLM (catalog 452161, Corning, Inc.) to
identify its potential metabolites. The substrate was obtained by diluting dioxybenzone to a final
concentration of 0.1 mΜ with phosphate buffer (PBS, 100 mM, pH = 7.4) at 37 . The incubation
mixture consisted of the substrate, HLM (final concentration: 0.09 mg/mL), and NADPH-generating
system (solution A, catalogue 451220; solution B, catalogue 451200; solution A: solution B = 5:1) in
a volume of 1 mL. The reaction mixture was incubated for 2.5 h at 37 and stopped by the ice-cold
methanol. Samples were centrifuged for 10 min at 12000 rpm before UPLC-QTOF-MSMS analysis.
Dioxybenzone incubated in the absence of HLM and PBS incubated with HLM were used as the
negative control. The assay was performed for three technical replicates.
92
93
94
95
96
97
98
99
100
101
102
103
104
105
Synthesis of the metabolites. M1 and M2 were synthesized through the demethylation of
dioxybenzone and BP-6, respectively. Dioxybenzone or BP-6 was firstly dissolved in
dichloromethane and then was added to a 100 mL round flask equipped with a nitrogen balloon and
magnetic stirrer. Boron tribromide was added with the mole ratio of 1:3 (dioxybenzone: boron
6

106
107
108
109
110
111
tribromide) and 1:1.5 (BP-6: boron tribromide). After 48 h reaction, the reaction mixture was
transferred to a separatory funnel containing 20 mL H2O and 20 mL ethyl acetate for extraction. The
organic layer was separated and dried by anhydrous sodium sulfate, and then concentrated by rotary
evaporation. The mixture was separated and purified via the column chromatography. The synthetic
products were redissolved in methanol for the UPLC-Q-TOF-MS/MS analysis, or in deuterated
chloroform for the 1H and 13C NMR analysis.
112
113
In vivo metabolism. Mature female Sprague-Dawley rats (6-8 weeks old, 180-200 g) were
purchased from Shanghai Slac Laboratory Animal CO., LTD (Shanghai, China) and maintained in
the animal facility at Hangzhou Normal University under a controlled environment with 60%
humidity and a 12/12 h light-dark cycle at 25 ± 2 . All rats received basic diet and water ad libitum.
Female rats were randomly assigned to the control and treatment groups (n = 4 mice per group).
Dioxybenzone was dissolved in corn oil. In vivo kinetic study was carried out in plasma of rats
treated with dioxybenzone at 100 mg kg-1 bw d-1 by a single oral gavage. Rats in control group were
treated with corn oil only. Blood was collected from the rats at different times (0, 5, 20, 40, 60, 90,
120, 180 min) into heparinized tubes to isolate plasma. After the exposure to dioxybenzone at 100
mg kg-1 bw d-1 via oral gavage for three days, rats were killed, and then the livers, kidneys, ovaries
and uteri were all collected for the analysis of metabolites. All samples were stored at -20ꢁ°C until
analysis. All animal studies were approved by the Committee on Animal Welfare of Hangzhou
114
115
116
117
118
119
120
121
122
123
124
7

125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
Normal University, and accorded with the Guidelines for Animal Management of Zhejiang province,
which meet the ethical guidelines for experimental animals in China.
Identification of metabolites of dioxybenzone by UPLC-Q-TOF-MS/MS. The metabolites of
dioxybenzone were identified by the UPLC-Q-TOF-MS/MS equipped with an UV absorption
detector (Waters Corp., Milford, MA, USA) and coupled with an AB Triple TOF 5600 + 127 system
(AB SCIEX, Framingham, USA) using the Agilent ZORBAC SB-C18 column (2.1 × 100 mm,
particle size: 3.5 ꢀm). The mobile phase contained 0.1% formic acid-water (A) and 0.1% formic
acid- acetonitrile (B), following the linear gradient method: 0/5, 20/50, 30/95, 31/5 (min/B%) with a
constant flow rate of 0.6 mL/min. Analytes was detected with the wavelength of 254 nm and with
the maximum error of ± 5 ppm. The column temperature was 35 . The optimal MS condition was
set in the negative ion mode of atmospheric pressure chemical ionization (APCI) source. The
calibration solution was delivered automatically through APCI for the mass calibration using the
Automated Calibration Delivery System. The collision energy was set at 40 V ± 20 V. The m/z
values of precursor ion and product ion were scanned with the range of 100-1500 Da and 50-1500
Da, respectively. The mass spectral data were analyzed by PeakView Software v1.2 (AB SCIEX,
Framingham, USA).
Sample preparation. The plasma or homogenized tissue samples (livers, kidneys, ovaries and uteri)
of female rats were extracted with ethyl acetate/ n-hexane/ formic acid (1/1/0.1), and thereafter were
sonicated for 15 min. The supernatant of mixture was dried with nitrogen and redissolved with 100
ꢀL methanol. For hydrolysis, samples were firstly mixed with the same volume of ammonium
acetate buffer containing β-glucuronidase/arylsulfatase (glucuronidase, 1000 U/mL, sulfatase, 750
U/mL) overnight at 37 . Samples were then extracted using the same method mentioned above. All
extracts were stored at -20ꢁ°C for further analysis. Extracted recoveries in plasma and tissues
8

148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
samples range from 73% to 112%. The concentration of dioxybenzone, M1 and M2 was
simultaneously quantified by the well-established UPLC-MS/MS-based methods (Text 1).
Molecular dynamic (MD) simulations. The molecular modelling was performed following the
procedure of our previously reported work (Zhuang et al. 2016; Zhuang et al. 2018). Molecular
docking was firstly conducted using the X-Ray diffraction structure of ERα LBD (PDB ID: 4MGC,
resolution: 2.15 Å) in complex with dioxybenzone or its metabolites. Dioxybenzone and its
metabolites were docked into the active site of ERα LBD using AutoDock Vina. The constructed
tertiary structures were prepared for MD simulations using NAMD 2.9 (Acun et al. 2018).
CHARMM36 force fields were used to treat both proteins and ligands (Huang and MacKerell Jr
2013). The 100 ns conventional MD simulations were conducted to the top 2 docked structures in
isothermal-isobaric (NPT) ensemble with 1 bar and 310 K using NAMD 2.9. The treatment for van
der Waals interactions was performed with a cutoff distance at 12 Å and switching distance at 10 Å.
Particle mesh Ewald (PME) with a grid size of 1 Å was used to analyze the electrostatic interactions.
The binding free energy of final structures of dioxybenzone and its metabolites from MD
simulations was calculated using free energy perturbation (FEP) method. The free-energy changes in
thermodynamic cycle contain the formations of dioxybenzone or its metabolites in the ERα LBD
binding pocket (ꢂGbound) and in water (ꢂGfree). Ethane (C2H6) was used as the reference structure.
Thus, the binding free energy ꢂꢂGbinding, was calculated as ꢂꢂGbinding = ꢂGbound –ꢂGfree. Standard
deviations of the mean were calculated by averaging over 5 runs of FEP simulations with different
initial random seeds.
Yeast Two-hybrid Assay. The estrogenic effects of dioxybenzone and its metabolites were in vitro
determined by recombinant yeast two-hybrid kit (catalog YC-ER-201, Wuxi CASA Environmental
Technology Co., Ltd, China). The yeast cells were transformed with the human ERα and the reporter
9

171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
gene expressing β-galactosidase. The yeast cells were grown on synthetic
dextrose/leucine/tryptamine (catalog 4823-6, Mobitec Company) medium at 30 overnight.
Dioxybenzone, M1 and M2 with concentrations ranging from 0.005 ꢀM to 50 ꢀM were cultured
with yeast medium for 2 h at 30 . The treated yeast cells were lysed by chloroform. The
β-galactosidase reaction mixture was started by the addition of o-nitrophenyl β-D-galactopyranoside
(o-NPG, 13.3 mM). Sodium carbonate (1 M) was then added to terminate the reaction after 1 h
incubation. The absorbance at 420 nm of samples was measured using Infinite 200 PRO NanoQuant
Multimode Reader (Tecan Group Ltd., Switzerland) and the β-galactosidase activity was determined
following the procedure of a reported protocol (Ma et al. 2005). DMSO and E2 (final concentration:
2.5 × 10-4 ꢀM) were used as the negative and positive control, respectively. IC10 values were
obtained on the basis of dose-response curves. The assay was performed in biological triplicate for
three technical replicates.
Uterotrophic assay using mice. Immature female ICR mice (PND 17-18 d, 17-19 g) were obtained
from Shanghai Slac Laboratory Aniamal CO., LTD (Shanghai, China) and maintained in the animal
facility at Hangzhou Normal University under a controlled environment with 60% humidity and a
12/12 h light-dark cycle at 25 ± 2 . Five mice were housed in an individually ventilated cage,
provided with basic diet and water ad libitum. All chemicals were dissolved in corn oil. Mice were
randomly assigned to different groups (n = 10) and treated with corn oil only or one of the following
treatment: dioxybenzone, M1, M2 at doses of 0.8, 4, 20 or 100 mg kg-1 bw d-1or E2 at a dose of 500
ꢀg kg-1 bw d-1 via intraperitoneal injection for 3 days beginning on PND 21. Mice were weighed and
then killed 24 h after the final treatment, and uteri of mice were obtained, blotted and weighed.
Uterine organ coefficient was calculated by the ratio of uterus weight to final body weight.
10

193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
Statistical Analysis. Data were presented as mean ± standard deviation (SD) or mean ± standard
error of mean (SEM). Data were analyzed by one-way analysis of variance (ANOVA) followed by
post-hoc test for multiple comparison using IBM SPSS Statistics 20 (SPSS Inc., Chicago, IL).
Difference with a two-sided P < 0.05 was considered to be statistically significant and marked with
an asterisk.
ꢀ Results
Dioxybenzone was metabolized into demethylated and hydroxylated products.
To explore the biotransformation susceptibility of dioxybenzone, we investigated both in vitro and in
vivo metabolism of dioxybenzone. Dioxybenzone and its metabolites (Table 1) were analyzed by
UPLC-Q-TOF-MS/MS method. Dioxybenzone at the retention time (RT) of 16.4 min was detected
with the intensity of 7.7 × 104 (Figure 1A). The intensity of dioxybenzone was significantly
decreased by 59.4% after the incubation with HLM for 2.5 h (Figure 1B), indicating the metabolism
of dioxybenzone. Two metabolites M1 and M2 were formed at the corresponding retention time of
12.3 and 14.1 min, respectively. The intensity of M1 is higher than that of M2, indicating a higher
yield of M1 than that of M2. Compared with dioxybenzone, the shorter retention time of M1 and M2
in reversed phase chromatography suggests the higher hydrophilicity of two metabolites. The
structure of M1 and M2 were further identified from the MS spectrum (Table S2). The m/z of M1 is
229.0516 Da, 14 Da lower than that of dioxybenzone (243.0672 Da), indicating the removal of one
methyl moiety from the structure of dioxybenzone (Figure 1C, 1D). The m/z of M2 is 259.0612 Da,
16 Da higher than that of dioxybenzone, suggesting the insertion of an O atom into the structure of
dioxybenzone (Figure 1E). Dioxybenzone shows main fragment ions of 93.0348 Da, 108.0203 Da
and 123.0446 Da in MS/MS spectrum. The m/z of 93.0348 Da and 123.0446 Da refer to the phenolic
ring without and with methoxy moiety of dioxybenzone, respectively. The fragment ion of 108.0203
11

216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
Da was obtained from the further breakage of methyl moiety from methoxylated phenolic moiety
(123.0446 Da). The m/z 108.0203 Da and 123.0446 Da were also detected as the main fragment ions
of M2, suggesting the intact phenolic ring with methoxy moiety. Another main fragment ion of M2
is 109.0299 Da, 16 Da higher than that of the phenolic ring without methoxy moiety (93.0348 Da),
indicating one hydroxyl moiety bound to the phenolic ring without methoxy moiety. In comparison
of the smaller fragment ions of M1, fragment ions of 67 Da and 91 Da were also detected in M2,
suggesting that a hydroxy moiety might be bound to the para-position of carbanyl moiety of phenolic
ring in the sructure of M2. M1 and M2 have higher hydrophilicity with LogKow of 2.51 and 2.65,
respectively, than that of dioxybenzone with LogKow of 3.15 (Table S1), in line with the observation
of their shorter retention times. To validate the molecular structure, we synthesized M1 and M2 with
purity > 95% via the demethylation of dioxybenzone and BP-6, respectively, and compared their
retention times and MS/MS spectra (Figure S1). The 13C NMR and 1H NMR spectra show the
molecular structure for synthesized M1 and M2, respectively (Figure S2, S3). In 1H NMR of
synthesized M1 and M2, it ends up with 7 and 9 hydrogen atoms, respectively, in line with the
number of hydrogen atoms in M1 and M2 (except for the hydrogen from hydroxyl). The chemical
shift of hydrogen atoms ranging from 3.0 - 4.0 refers to the hydrogens on methyl, indicating the
existence of methyl moiety of synthesized M2 and absence of methyl moiety of synthesized M1. In
13C NMR spectra, the number of peaks for both synthesized M1 and M2 is 13 and 14, respectively,
same as their number of carbon atoms. The retention times and MS/MS spectra are the same
between the synthesized M1 and M2 and the corresponding metabolites obtained from the in vitro
metabolism study.
237
The residue of dioxybenzone and its metabolites in female rats.
12

238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
We further investigated the in vivo metabolism of dioxybenzone using rat model. Since rats have
relatively adequate blood supply for continuous blood collection, rats have been frequently used for
the metabolic kinetics study (Mi et al. 2017; Mutlu et al. 2019). After the treatment of female rats
with dioxybenzone at a dose of 100 mg kg-1 body weight (bw) per day by a single oral gavage, the
rat plasma was collected at different times. Using our established UPLC-MS/MS-based method,
dioxybenzone, M1 and M2 were simultaneously analyzed in rat samples (Table S3, S4). The
residues of dioxybenzone, M1 and M2 from phase I metabolism were all detected in rat plasma.
Figure 2A shows the the residue concentration changes over time. The concentration of
dioxybenzone was 294.0 ± 132.5 ꢀg/L at 5 min, and the concentration ratio of metabolites to parent
compound (M/P ratio) was 0.35and 0.02 for M1 and M2, respectively. M1 and M2 at their detectable
concentrations indicates the rapid metabolism of dioxybenzone mainly through the demethylation in
phase I metabolic process, followed by the hydroxylation. The concentrations of M1 and M2 were
decreased after the exposure to dioxybenzone from 5 min to 180 min. The M/P ratio was 0.25 and
0.01 for M1 and M2 at 180 min, respectively. Dioxybenzone, M1 and M2 can be further conjugated
with glucuronic acid or sulfuric acid through phase II metabolism. At 5 min, the concentration of
conjugated dioxybenzone, M1 and M2 were 16.5, 6.5 and 10.5 times of that of the unconjugated
forms (Figure 2B).
The concentrations of dioxybenzone, M1 and M2 and phase II conjugated forms in tissues of
female rats, including liver, kidney, uterus and ovary were further measured after the exposure to
dioxybenzone at a dose of 100 mg kg-1 bw d-1 for three days. The concentration of dioxybenzone in
liver, kidney, uterus and ovary was 447.9 ± 129.8, 237.2 ± 122.4, 153.3 ± 82.1 and 39.5 ± 25.9 ꢀg/kg,
respectively (Figure 2C). In these tissues, M/P ratio of M1 was 0.84, 0.39, 0.15 and 0.30, while that
of M2 was 0.03, 0.01, 0.004 and 0.01, respectively. The measured concentrations of dioxybenzone,
13

261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
M1 and M2 in four tissues ranked as liver > kidney > uterus > ovary. The concentration of M1 was
higher than that of M2 in four tissues, in line with the order of concentrations in rat plasma.
Conjugated dioxybenzone, M1 and M2 were also detected in these tissues (Figure 2D). The
conjugated concentrations of M1 and M2 in uterus and ovary were at least 10 times of that of
unconjugated forms, indicating the distribution of M1 and M2 into the organs of reproductive system
via phase II conjugated forms.
Different binding characteristics of dioxybenzone and its metabolites.
To probe ERα-related MIEs and evaluate the estrogenic effect of dioxybenzone, M1 and M2, we
investigated their binding interactions with ERα ligand binding domain (LBD) by 100 ns MD
simulations before in vitro validation of their agonistic effects toward ERα. All three ligands bind
with the same pocket of ERα LBD, however, the binding modes were different (Figure 3). The
binding pocket is mostly hydrophobic, featuring residues such as L346, L349, A350, L387, M388,
L391, L424, L524 and L525, which are mostly the same across three ligands. Dioxybenzone showed
no hydrogen bonds formed with ERα LBD (Figure 3A); its metabolite M1 formed weak hydrogen
bond between phenol-OH group and the backbone of L387 (by breaking the internal
carbonyl-phenol-OH hydrogen bond) (Figure 3B) whereas M2 formed a unique hydrogen bond
between its additional phenol-OH group and E353 (Figure 3C). Dioxybenzone as well as M1- and
M2-bound ERα LBD displayed only subtle differences in the tertiary structures, mostly in the loop
regions (Figure S4). Based on the binding structures, we further performed free energy perturbation
(FEP) calculations to obtain the binding free energies (ꢂꢂGbinding) of dioxybenzone, M1 and M2 to
ERα LBD. In comparison with dioxybenzone, M1 and M2 showed a relative lower binding free
energy of -0.74 (0.38) and -6.10 (0.46) kcal/mol, and exhibited stronger binding affinity to ERα
LBD. To further elucidate hydrophilic/hydrophobic contributions to the binding modes, we reported
14

284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
energy decomposition in Table 1. The hydrogen bond between M2 and ERα LBD significantly
lowered the electrostatic penalties in the binding mode, leading to a much stronger binding affinity
of M2 than dioxybenzone and M1. The much higher binding potency between M2 and ERα LBD
could induce a stronger estrogen disrupting effect than dioxybenzone.
Enhanced estrogenic effect by dioxybenzone metabolites toward ERα.
The estrogenic potency of dioxybenzone, M1 and M2 was validated by the well-established
recombinant human ERα gene yeast two-hybrid assay. E2, as the agonist of ERα, induced the
β-galactosidase maximum activity at 0.25 nM (Figure S5). Figure 4A showed the estrogenic potency
of dioxybenzone, M1 and M2 by the calculated ratio of their β-galactosidase activity at various
concentrations to that of E2 at 0.25 nM. Dioxybenzone showed no significant estrogenic effect at the
concentration ranging from 0.05 - 50 ꢀM. The dose-response curve showed that M1 and M2
significantly induced the estrogenic effect. M1 showed the estrogenic effect at the concentration of
1.67 ꢀM and gradually increased the activity up to 17.9 ± 0.9% with the increasing concentration.
M2 induced a rapid increase of estrogenic effects to 34.6 ± 2.0% at 5 ꢀM, despite that M2 showed
no significant estrogenic effect at the concentration below 1.67 ꢀM. In comparison with
dioxybenzone, M1 and M2 showed enhanced estrogenic effect with the EC20 of 42.56 ꢀM and 4.04
ꢀM, respectively. The estrogenic disrupting effect ranks as M2 > M1 > dioxybenzone, in line with
their relative binding affinity to ERα LBD.
To probe whether the enhanced estrogency of dioxybenzone metabolites would alter uterine
growth, we conducted the uterotrophic assay of dioxybenzone, M1 and M2 in mice. Mice model has
the advantage for developmental and reproductive toxicity studies of various chemicals (Markey et
al. 2001). The mice were intraperitoneally injected with dioxybenzone, M1 and M2 at doses of 0.8, 4,
20, 100 mg kg-1 bw d-1 for three days, using E2 as the positive control. Compared with mice treated
15

307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
with corn oil, no significant changes were observed in uterine weight of mice after the treatment
with dioxybenzone, M1 and M2 at doses of 0.8 and 4 mg kg-1 bw d-1. Dioxybenzone slightly
increased the uterus weight at doses of 20 and 100 mg kg-1 bw d-1. M1 and M2 at 20 mg kg-1 bw d-1
significantly increased the uterine weight and showed no effect on body weight, further confirming
their enhanced estrogenic effects in vivo (Figure 4B, Figure S6). However, the uterine weight was
decreased after the exposure to M1 and M2 at 100 mg kg-1 bw d-1. In mice, dioxybenzone at doses of
166-5000 mg kg-1 bw was previously reported to induce a decreased activity, piloerection, and
exophthalmos (Liebert 1983), whereas our results revealed its estrogenic effects in uterus via
metabolic activation after lower dose and short term exposure. Thus our in vivo results revealed the
higher estrogenic disrupting effects of the metabolites compared to dioxybenzone, in line with the
results obtained from in vitro and in silico assays.
ꢀ Discussion
Current health assessments of many estrogenic chemicals with metabolic susceptibility are hindered
by insufficient data which requires the characterization and testing of the impact of metabolic
products beyond their parent compounds. A subtle difference in the structure of chemicals during
metabolism could markedly alter the ADMET (absorption, distribution, metabolism, excretion, and
toxicity) properties (Miao and Scott Obach 2010), leading to the metabolic activation or metabolic
detoxication. One rising health concern on dioxybenzone is its usage and frequent exposure
accompanied with insufficient evaluation of its metabolism. The impact of metabolism on health risk
after exposure to dioxybenzone has been largely ignored, which motivated us to investigate the
estrogenic disrupting effect jointed with its metabolic products. Our study provided abundant
evidence of the metabolic susceptibility of dioxybenzone and the critical insight into the role of
metabolic evaluation in health risk assessment.
16

330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
Previous studies have focused more on the detection of the chemical itself in vivo (Calafat et al.
2008), whereas their metabolites might be potential toxicants. In our work, dioxybenzone mainly
underwent metabolism and excretion in female rat livers and kidneys which were detected with
higher concentrations of metabolites (Figure 1, 2). Kinetic studies in rat plasma showed the rapid
generation of major phase I metabolites after the short term exposure to dioxybenzone, with 47% of
M1 and 32% of M2 still remained in rat plasma after 3 h exposure (as compared to those at 5 min).
Many estrogenic compounds were also found to be distributed in the reproductive tract of mice, and
the concentration in uterus was found to be greater than in other tissues (Pollock and deCatanzaro
2014). Similarly, a significantly enhanced residue of dioxybenzone, M1, M2 and their conjugated
forms was detected in hormone-sensitive organs, including uterus and ovary. These findings imply
that different metabolites could be deposited in the target organs through blood circulation following
the biotransformation of xenobiotics.
Biochemical and genetic studies suggest that ERα is implicated in the adverse developmental
and reproductive outcomes induced by estrogenic xenobiotics (Hewitt and Korach 2011; Li et al.
2018a; Pillon et al. 2005). Distinct structures of dioxybenzone and its major phase I metabolites,
including demethylated product M1 and hydroxylated product M2, implicates the variation of
estrogenic potency in triggering distinct binding characteristics with ERα (Figure 3). Although the
concentration of M2 was relatively low in comparison with that of M1 after the metabolism of
dioxybenzone, M2 showed 10.5 times higher estrogenic disrupting potency than that of M1 (EC20)
(Figure 4).This discrepancy was elucidated from our MD simulations: (1) the hydrogen bond with
E353 regarded as the key E2-binding residue (Li et al. 2018b); (2) the interactions with several
hydrophobic residues of L346, L349, A350, L387, M388, L391, L424, L524 and L525. The
interaction between many other environmental estrogens and ERα was reported to affect the
17

353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
transcription response that was primarily associated with steroid synthesis and the female
reproductive health (Hewitt and Korach 2011; Milesi et al. 2012). Especially, the formation of
metabolites with active hydroxyl moiety contributed to the increased health risks (Ashrap et al. 2017;
Y. Watanabe et al. 2015). Besides M1 and M2, a trace but detectable amount of another
hydroxylated metabolite M3 with the RT of 11.4 min was also noted (Figure S7). On the basis of the
MSMS spectra of M3 and site of metabolism (SOM) predicted by SMARTCyp server, it is deduced
that the hydroxy moiety of M3 was bound to the phenolic ring with methoxy moiety (Figure S7A,
S7B and S7C). The proposed M3 also showed higher binding affinity to ERα LBD in comparison
with that of dioxybenzone (Figure S7D, S7E). Since the binding affinity of M3 is between that of the
two predominant metabolites (M2 and M1) based on in silico simulations, M3 was not synthesized
for further in vivo evaluation. Further study is surely warranted to acquire more data on structural
confirmation and estrogenic effect of M3. Combining with our computational simulations and yeast
two-hybrid assay, we concluded that the metabolic activation of dioxybenzone enhances estrogenic
disrupting effect toward ERα.
ꢀ Conclusion
Our in vitro and in vivo study showed the metabolic susceptibility of dioxybenzone, which leads to
the formation of hydroxylated and demethylated metabolites. Our in silico, in vitro and in vivo study
further revealed that metabolic activation of dioxybenzone induced distinct binding mode to ERα
LBD and exerted higher binding affinity and agonistic activity toward ERα, hence the potential of
higher health risks. On the basis of the differential effects of dioxybenzone and its metabolites,
future work on the investigation of reproductive health risks of dioxybenzone and its metabolites is
needed for a thorough risk assessment of this emerging contaminant. It is also cautioned that a
dermal exposure of dioxybenzone was not administrated in this study, which could become essential
18

376
377
378
379
considering its primary exposure route upon benzophenone exposure from the environment.
Nevertheless, our study provides new evidence and mechanistic insight into our current
understanding of the enhanced estrogenic disrupting effects likely triggered by the metabolism of
dioxybenzone and its structurally related chemicals with metabolic susceptibility.
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
ꢀ Associated Content
Surpporting Information
Text 1. Quantification of dioxybenzone, M1 and M2 by UPLC-MS/MS.
Table S1. Structures of dioxybenzone and its metabolites.
Table S2. The product ions of dioxybenzone, M1 and M2 observed from UPLC-Q-TOF-MS/MS.
Table S3. MS parameters of dioxybenzone, M1 and M2 for UPLC-MS/MS.
Table S4. Analytical recoveries of dioxybenzone, M1 and M2 in rat plasma and organs. Data are
presented as mean ± SD (in biological triplicate).
Figure S1. The extracted ion chromatograms of synthesized M1 and M2 (A), the MS spectrum of
synthesized M1 (B) and M2 (C), and the MS/MS spectrum of synthesized M1 (D) and M2 (E).
Figure S2. 1H and 13C NMR spectrum of synthesized M1.
Figure S3. 1H and 13C NMR spectrum of synthesized M2.
Figure S4. The superimposition of structures of ERα bound with dioxybenzone, M1 and M2. Purple
for the dioxybenzone binding mode, grey for the M1 binding mode and orange for the M2 binding
mode.
Figure S5. The standard dose-response curve for E2 by human recombinant ERα yeast two-hybrid
assay. Error bars show the standard deviation of triplicate assays.
19

397
398
399
Figure S6. Body weight of female mice treated with dioxybenzone, M1 and M2 at doses of 20 and
100 mg kg-1 bw d-1
Figure S7. The identification of metabolite M3 and its binding characteristics to ERα LBD.
400
ꢀ AUTHOR INFORMATION
401
402
403
404
405
406
Corresponding Author
* E-mail: (S. Z.) shulin@zju.edu.cn
ORCID
Shulin Zhuang: 0000-0002-7774-7239
Notes
The authors declare no competing financial interest.
407
408
409
410
411
ꢀ ACKNOWLEDGEMENTS
We thank the financial support from the National Natural Science Foundation of China (No.
21876153) and Fundamental Research Funds for the Central Universities (No. 2020FZZX001-06,
2019XZZX003-11). R.Z. acknowledges the financial support from the IBM Blue Gene Science
Program.
412
ꢀ References
413
414
415
416
417
418
419
420
421
Acun, B., Hardy, D. J., Kale, L. V., Li, K., Phillips, J. C., Stone, J. E. 2018. Scalable molecular
dynamics with NAMD on the summit system. IBM J. Res. Dev. 62, 4:1-4:9.
Ashrap, P., Zheng, G., Wan, Y., Li, T., Hu, W., Li, W., et al. 2017. Discovery of a widespread
metabolic pathway within and among phenolic xenobiotics. Proc. Natl. Acad. Sci. USA 114,
6062-6067.
Becker, R. A., Ankley, G. T., Edwards, S. W., Kennedy, S. W., Linkov, I., Meek, B., et al. 2015.
Increasing scientific confidence in adverse outcome pathways: application of tailored
Bradford-Hill considerations for evaluating weight of evidence. Regul. Toxicol. Pharmacol. 72,
514-537.
20

422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
Calafat, A. M., Wong, L. Y., Ye, X., Reidy, J. A., Needham, L. L. 2008. Concentrations of the
sunscreen agent benzophenone-3 in residents of the United States: national health and nutrition
examination survey 2003--2004. Environ. Health Perspect. 116, 893-897.
Gao, C. J., Liu, L. Y., Ma, W. L., Zhu, N. Z., Jiang, L., Li, Y. F., et al. 2015. Benzophenone-type
UV filters in urine of chinese young adults: concentration, source and exposure. Environ. Pollut.
203, 1-6.
Hewitt, S. C., Korach, K. S. 2011. Estrogenic activity of bisphenol A and
2,2-bis(p-hydroxyphenyl)-1,1,1-trichloroethane (HPTE) demonstrated in mouse uterine gene
profiles. Environ. Health Perspect. 119, 63-70.
Huang, J., MacKerell Jr, A. D. 2013. CHARMM36 all-atom additive protein force field: validation
based on comparison to NMR data. J. Comput. Chem. 34, 2135-2145.
Kerdivel, G., Le Guevel, R., Habauzit, D., Brion, F., Ait-Aissa, S., Pakdel, F. 2013. Estrogenic
potency of benzophenone UV filters in breast cancer cells: proliferative and transcriptional
activity substantiated by docking analysis. PLoS One 8, e60567.
Kim, S., Choi, K. 2014. Occurrences, toxicities, and ecological risks of benzophenone-3, a common
component of organic sunscreen products: a mini-review. Environ. Int. 70, 143-157.
Kunisue, T., Chen, Z., Louis, G. M. B., Sundaram, R., Hediger, M. L., Sun, L. P., et al. 2012.
Urinary concentrations of benzophenone-type UV filters in U.S. women and their association
with endometriosis. Environ. Sci. Technol. 46, 4624-4632.
Le Fol, V., Aït-Aïssa, S., Cabaton, N., Dolo, L., Grimaldi, M., Balaguer, P., et al. 2015. Cell-specific
biotransformation of benzophenone-2 and bisphenol-s in zebrafish and human in vitro models
used for toxicity and estrogenicity screening. Environ. Sci. Technol. 49, 3860-3868.
Leist, M., Ghallab, A., Graepel, R., Marchan, R., Hassan, R., Bennekou, S. H., et al. 2017. Adverse
outcome pathways: opportunities, limitations and open questions. Arch. Toxicol. 91,
3477-3505.
Li, Y., Hamilton, K. J., Wang, T., Coons, L. A., Jefferson, W. N., Li, R., et al. 2018a. DNA
methylation and transcriptome aberrations mediated by ERα in mouse seminal vesicles
following developmental DES exposure. Proc. Natl. Acad. Sci. USA 115, 4189-4198.
Li, Y., Perera, L., Coons Laurel, A., Burns Katherine, A., Tyler Ramsey, J., Pelch Katherine, E., et al.
2018b. Differential in vitro biological action, coregulator interactions, and molecular dynamic
analysis of bisphenol A (BPA), BPAF, and BPS ligand–ERα complexes. Environ. Health
Perspect. 126, 017012.
Liebert, M. A. 1983. Addendum to the final report on the safety assessment of benzophenones-1, -3,
-4, -5, -9, and -11 to include benzophenones-2, -6, and -8. Int. J. Toxicol. 2, 79-84.
Lu, S., Long, F., Lu, P., Lei, B., Jiang, Z., Liu, G., et al. 2018. Benzophenone-UV filters in personal
care products and urine of schoolchildren from Shenzhen, China: Exposure assessment and
possible source. Sci. Total Environ. 640-641, 1214-1220.
Lu, Z., Wong, C. S. 2011. Factors affecting phase I stereoselective biotransformation of chiral
polychlorinated biphenyls by rat cytochrome P-450 2B1 isozyme. Environ. Sci. Technol. 45,
8298-8305.
Lv, X., Pan, L., Wang, J., Lu, L., Yan, W., Zhu, Y., et al. 2017. Effects of triazole fungicides on
androgenic disruption and CYP3A4 enzyme activity. Environ. Pollut. 222, 504-512.
Ma, M., Li, J., Wang, Z. J. 2005. Assessing the detoxication efficiencies of wastewater treatment
processes using a battery of bioassays/biomarkers. Arch. Environ. Contam. Toxicol. 49,
480-487.
21

467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
Markey, C. M., Michaelson, C. L., Veson, E. C., Sonnenschein, C., Soto, A. M. 2001. The mouse
uterotrophic assay: a reevaluation of its validity in assessing the estrogenicity of bisphenol A.
Environ. Health Perspect. 109, 55-60.
Meng, Q., Yeung, K., Kwok, M. L., Chung, C. T., Hu, X. L., Chan, K. M. 2020. Toxic effects and
transcriptome analyses of zebrafish (Danio rerio) larvae exposed to benzophenones. Environ.
Pollut. 265, 114857.
Mi, X. B., Bao, L. J., Wu, C. C., Wong, C. S., Zeng, E. Y. 2017. Absorption, tissue distribution,
metabolism, and elimination of decabrominated diphenyl ether (BDE-209) in rats after
multi-dose oral exposure. Chemosphere 186, 749-756.
Miao, Z., Scott Obach, R. 2010. Chapter 10 Alcohols and Phenols: Absorption, Distribution,
Metabolism and Excretion. In: Metabolism, Pharmacokinetics and Toxicity of Functional
Groups: Impact of Chemical Building Blocks on ADMET:The Royal Society of Chemistry,
460-485.
Milesi, M. M., Varayoud, J., Bosquiazzo, V. L., Munoz-de-Toro, M., Luque, E. H. 2012. Neonatal
exposure to low doses of endosulfan disrupts the expression of proteins regulating uterine
development and differentiation. Reprod. Toxicol. 33, 85-93.
Miller, D., Wheals, B. B., Beresford, N., Sumpter, J. P. 2001. Estrogenic activity of phenolic
additives determined by an in vitro yeast bioassay. Environ. Health Perspect. 109, 133-138.
Molins-Delgado, D., Olmo-Campos, M. d. M., Valeta-Juan, G., Pleguezuelos-Hernández, V.,
Barceló, D., Díaz-Cruz, M. S. 2018. Determination of UV filters in human breast milk using
turbulent flow chromatography and babies’ daily intake estimation. Environ. Res. 161,
532-539.
Mutlu, E., Garner, C. E., Wegerski, C. J., McDonald, J. D., McIntyre, B. S., Doyle-Eisele, M., et al.
2019. Metabolism and disposition of 2-hydroxy-4-methoxybenzophenone, a sunscreen
ingredient, in Harlan Sprague Dawley rats and B6C3F1/N mice-a species and route comparison.
Xenobiotica 50, 1-16.
Pillon, A., Boussioux, A. M., Escande, A., Ait-Aissa, S., Gomez, E., Fenet, H., et al. 2005. Binding
of estrogenic compounds to recombinant estrogen receptor-alpha: application to environmental
analysis. Environ. Health Perspect. 113, 278-284.
Pollock, T., deCatanzaro, D. 2014. Presence and bioavailability of bisphenol A in the uterus of rats
and mice following single and repeated dietary administration at low doses. Reprod. Toxicol.
49, 145-154.
Rodriguez-Gomez, R., Zafra-Gomez, A., Dorival-Garcia, N., Ballesteros, O., Navalon, A. 2015.
Determination of benzophenone-UV filters in human milk samples using ultrasound-assisted
extraction and clean-up with dispersive sorbents followed by UHPLC-MS/MS analysis. Talanta
134, 657-664.
Schlumpf, M., Cotton, B., Conscience, M., Haller, V., Steinmann, B., Lichtensteiger, W. 2001. In
vitro and in vivo estrogenicity of UV screens. Environ. Health Perspect. 109, 239-244.
Suvorov, A., Waxman, D. J. 2015. Early programing of uterine tissue by bisphenol A: critical
evaluation of evidence from animal exposure studies. Reprod. Toxicol. 57, 59-72.
Tarazona, I., Chisvert, A., Salvador, A. 2013. Determination of benzophenone-3 and its main
metabolites in human serum by dispersive liquid–liquid microextraction followed by liquid
chromatography tandem mass spectrometry. Talanta 116, 388-395.
Tsui, M. M., Leung, H. W., Kwan, B. K., Ng, K. Y., Yamashita, N., Taniyasu, S., et al. 2015.
Occurrence, distribution and ecological risk assessment of multiple classes of UV filters in
marine sediments in Hong Kong and Japan. J. Hazard. Mater. 292, 180-187.
22

513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
Tsui, M. M. P., Lam, J. C. W., Ng, T. Y., Ang, P. O., Murphy, M. B., Lam, P. K. S. 2017.
Occurrence, distribution, and fate of organic uv filters in coral communities. Environ. Sci.
Technol. 51, 4182-4190.
Wang, L., Asimakopoulos, A. G., Moon, H.-B., Nakata, H., Kannan, K. 2013. Benzotriazole,
benzothiazole, and benzophenone compounds in indoor dust from the United States and East
Asian countries. Environ. Sci. Technol. 47, 4752-4759.
Warner, N. A., Martin, J. W., Wong, C. S. 2009. Chiral polychlorinated biphenyls are
biotransformed enantioselectively by mammalian cytochrome P-450 isozymes to form
hydroxylated metabolites. Environ. Sci. Technol. 43, 114-121.
Watanabe, Y., Kojima, H., Takeuchi, S., Uramaru, N., Sanoh, S., Sugihara, K., et al. 2015.
Metabolism of UV-filter benzophenone-3 by rat and human liver microsomes and its effect on
endocrine-disrupting activity. Toxicol. Appl. Pharmacol. 282, 119-128.
Watanabe, Y., Kojima, H., Takeuchi, S., Uramaru, N., Sanoh, S., Sugihara, K., et al. 2015.
Metabolism of UV-filter benzophenone-3 by rat and human liver microsomes and its effect on
endocrine-disrupting activity. Toxicol. Appl. Pharmacol. 282, 119-128.
Wu, J. W., Chen, H. C., Ding, W. H. 2013. Ultrasound-assisted dispersive liquid-liquid
microextraction plus simultaneous silylation for rapid determination of salicylate and
benzophenone-type ultraviolet filters in aqueous samples. J. Chromatogr. A 1302, 20-27.
Zheng, X., Ren, X.-M., Zhao, L., Guo, L.-H. 2020. Binding and activation of estrogen related
receptor γ as possible molecular initiating events of hydroxylated benzophenones endocrine
disruption toxicity. Environ. Pollut. 263, 114656.
Zhuang, S., Wang, H., Ding, K., Wang, J., Pan, L., Lu, Y., et al. 2016. Interactions of benzotriazole
UV stabilizers with human serum albumin: Atomic insights revealed by biosensors,
spectroscopies and molecular dynamics simulations. Chemosphere 144, 1050-1059.
Zhuang, S., Zhang, L., Zhan, T., Lu, L., Zhao, L., Wang, H., et al. 2018. Binding specificity
determines the cytochrome P450 3A4 mediated enantioselective metabolism of metconazole. J.
Phys. Chem. B 122, 1176-1184.
23

542
Table 1 Energy decomposition from the FEP calculations. Data are presented as mean ± SD.
*
Compound
ꢀꢀGbinding
ꢀꢀGElec
ꢀꢀGVDW ꢀꢀGcoupled
-4.31
(0.47)
-5.05
10.11
(0.60)
12.42
(0.19)
7.23
-10.85
(0.41)
-14.04
(0.20)
-13.13
(0.30)
-3.56
(0.31)
-3.44
(0.18)
-4.52
(0.26)
Dioxybenzone
M1
M2
(0.38)
-10.41
(0.46)
(0.24)
543
544
* ꢂꢂGcoupled is defined as ꢂꢂGbinding - ꢂꢂGElec - ꢂꢂGVDW. The unit of energies is kcal/mol.
24

545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
Figure Legends
Figure 1. The in vitro metabolism of dioxybenzone. The extracted ion chromatograms (XIC) of
dioxybenzone (A). The XIC of dioxybenzone after 2.5 h HLM-mediated metabolism (B). MS (left
panel) and MS/MS spectrum (right panel) of dioxybenzone (C), M1 (D) and M2 (E).
Figure 2. The in vivo metabolism of dioxybenzone. The concentrations of dioxybenzone, M1, M2
(A) and their conjugated forms (B) in plasma of rats treated with dioxybenzone at 100 mg kg-1 bw
d-1 by a single oral gavage (n=4). The concentrations of dioxybenzone, M1, M2 (C) and their
conjugated forms (D) in livers, kidneys, ovaries and uteri of rats treated with dioxybenzone at 100
mg kg-1 bw d-1 by oral gavage for three days (n=4). Data are shown as mean ± SEM.
Figure 3. The binding mode of dioxybenzone, M1 and M2 to human ERα LBD; the protein is
illustrated with “cartoon” representation from Visual Molecular Dynamics (VMD); dioxybenzone,
M1 and M2 and surrounding residues are illustrated in stick representation; the dashed lines indicate
the formed hydrogen bonds.
Figure 4. The estrogenic effects of dioxybenzone, M1 and M2 in comparison with that of E2 using
human recombinant ERα yeast two-hybrid assay (A). Error bars show the standard deviation of
triplicate assays. The uterotrophic effect of dioxybenzone, M1, M2 at doses of 0.8, 4, 20 and 100 mg
kg-1 bw d-1 (B). Data are presented as mean ± SD. *P < 0.05 (One-way ANOVA).
25

562
563
Figure 1
564
26

565
566
567
Figure 2
27