Natural-Product-Inspired Compounds as Countermeasures against the Liver Carcinogen Aflatoxin B1

Article abstract of DOI:10.1021/acs.jnatprod.9b00290

Aflatoxin B₁ (AfB₁) ranks among the most potent liver carcinogens known, and the accidental or intentional exposure of humans and livestock to this toxin remains a serious global threat. One protective measure that had been proposed is employing small-molecule therapeutics capable of mitigating the toxicity of AfB₁; however, to date, these efforts have had little clinical success. To identify molecular scaffolds that reduce the toxicity of AfB₁, we developed a cell-based high-throughput high-content imaging assay that enabled our team to test natural products (pure compounds, fractions, and extracts) for protection of monolayers and spheroids composed of HepG2 liver cells against AfB₁. The spheroid assay showed notable potential for further development, as it afforded greater sensitivity of HepG2 cells to AfB1, which is believed to better mimic the in vivo response of hepatocytes to the toxin. One of the most bioactive compounds to arise from this investigation was alternariol-9-methyl ether (1, purified from an Alternaria sp. isolate), which inspired the synthesis and testing of several structurally related molecules. Based on these findings, it is proposed that several types of natural and synthetic polyarene molecules that have undergone oxidative functionalization (e.g., compounds containing 3-methoxyphenol moieties) are promising starting points for the development of new agents that protect against AfB₁ toxicity.

Full text of DOI:10.1021/acs.jnatprod.9b00290

Article
pubs.acs.org/jnp
Cite This: J. Nat. Prod. XXXX, XXX, XXX−XXX
Natural-Product-Inspired Compounds as Countermeasures against
the Liver Carcinogen Aflatoxin B1
Adam C. Carter,† Jarrod B. King,† Allison O. Mattes,† Shengxin Cai,† Narender Singh,‡
and Robert H. Cichewicz*,†
†Natural Products Discovery Group, Institute for Natural Products Applications and Research Technologies, Department of
Chemistry & Biochemistry, Stephenson Life Science Research Center, University of Oklahoma, 101 Stephenson Parkway, Norman,
Oklahoma 73019, United States
‡CFD Research Corporation, 701 McMillian Way, Suite D, Huntsville, Alabama 35806, United States
S
* Supporting Information
ABSTRACT: Aflatoxin B1 (AfB1) ranks among the most potent liver
carcinogens known, and the accidental or intentional exposure of humans and
livestock to this toxin remains a serious global threat. One protective measure
that had been proposed is employing small-molecule therapeutics capable of
mitigating the toxicity of AfB1; however, to date, these efforts have had little
clinical success. To identify molecular scaffolds that reduce the toxicity of
AfB1, we developed a cell-based high-throughput high-content imaging assay
that enabled our team to test natural products (pure compounds, fractions,
and extracts) for protection of monolayers and spheroids composed of HepG2
liver cells against AfB1. The spheroid assay showed notable potential for
further development, as it afforded greater sensitivity of HepG2 cells to AfB1,
which is believed to better mimic the in vivo response of hepatocytes to the
toxin. One of the most bioactive compounds to arise from this investigation
was alternariol-9-methyl ether (1, purified from an Alternaria sp. isolate), which inspired the synthesis and testing of several
structurally related molecules. Based on these findings, it is proposed that several types of natural and synthetic polyarene
molecules that have undergone oxidative functionalization (e.g., compounds containing 3-methoxyphenol moieties) are
promising starting points for the development of new agents that protect against AfB1 toxicity.
continue to play major roles in the global health and safety of
humans and other animals.
flatoxins are highly potent carcinogenic metabolites
Aproduced by several types of fungi including Aspergillus
flavus and Aspergillus parasiticus.1,2 Since their discovery in the
1960s following the poisoning of over 100 000 turkeys and
other poultry in England,1,3 a great deal of research has been
carried out pertaining to this family of toxins with regard to
their impact on human and animal health. Among the
aflatoxins reported to date, aflatoxin B1 (AfB1) is widely
considered to be one of the most carcinogenic naturally
occurring compounds known to humankind.4 AfB1 and related
aflatoxins are believed to be responsible for elevated
hepatocellular carcinoma rates observed in many parts of the
world including Africa, eastern Asia, and parts of South
America.5 It is estimated that 4.5 billion people worldwide are
exposed to aflatoxins in their diet and that aflatoxins are
responsible for up to 28% of all hepatocellular carcinoma cases
globally.6 The potency and ease of access to AfB1 and related
analogues has sparked concerns that mycotoxins such as AfB1
may be employed or developed for use in biological warfare, as
was the case in Iraq during the 1980s.7 In addition to their
negative effects on human health, aflatoxin contamination has
led to the loss of millions of dollars of agricultural products in
the United States alone.8 Thus, AfB1 and its analogues
In adult humans, conversion of AfB1 to its chemically
reactive form, AfB1-8,9-epoxide, through oxidative metabolism
involving several isoforms of cytochrome P450 enzymes
(CYP) (Figure 1), leads to many of the detrimental cancer-
related effects associated with AfB1 exposure. Several CYP
isoforms have been implicated as responsible for the metabolic
Figure 1. Metabolic activation of aflatoxin B1 (AfB1) to chemoactive
aflatoxin B1-8,9-epoxides.
Received: March 29, 2019
© XXXX American Chemical Society and
American Society of Pharmacognosy
A
DOI: 10.1021/acs.jnatprod.9b00290
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
activation of AfB1 including CYP3A4, CYP3A5, and CYP1A2;
however, some uncertainty remains as to which CYP
isoform(s) is the most critical for causing in vivo toxicity
when genetic and comorbidity factors are considered.9−11
Once AfB1 is converted to its reactive electrophilic form, the
compound reacts with cellular targets (e.g., DNA, RNA, and
proteins), forming covalent bonds. One of the most problem-
atic of these covalent modifications involves the reaction of
AfB1-8,9-epoxide with DNA, which ultimately gives rise to
mutations that increase the risk for tumor development,
especially in the liver.2,11
Because of its severe toxic effects, research has focused on
the development of protective measures to counteract the
carcinogenic properties of AfB1. These efforts have largely
followed two general strategies: (i) reducing exposure to the
AfB1 and (ii) blocking its metabolic transformation to the
activated epoxide. Among the described chemoprotective
agents, two of the most widely reported are chlorophyllin
and oltipraz (Figure 2), which appear to have limited efficacy
compounds that afforded in vitro cellular protection against the
toxic effects of AfB1.18,19 Further inspection showed that these
reports focused exclusively on using colorimetric assays to
measure cell metabolic activity (i.e., Alamar blue, MTT,
intracellular ATP) as end points for assessing cell viability. Our
preliminary studies indicated that these assay methods were
imperfect because of their poor abilities to detect low levels of
AfB1 toxicity, as well as their failure to assess overall cell health
(e.g., cell morphology). Specifically, our microscopy-based
observations of cells treated with AfB1 alone and with putative
protective agents revealed two problems: (i) cells treated with
even low-level (noncytotoxic) concentrations of AfB1 displayed
substantial morphological changes that were not immediately
reflected in their metabolic activity, and (ii) in some instances,
the application of “protective” compounds to AfB1-treated cells
resulted in viable, but severely altered hepatocytes that did not
resemble healthy control cells (e.g., reduced cell areas, cells
appeared angular as opposed to rounded, cells grew in isolation
rather than appearing as islands of cells, and more). These
observations alerted us that a different method was required to
identify protective molecules that could maintain cell viability,
while also sustaining healthy cell morphologies. Accordingly,
we determined that a high-content imaging approach using an
Operetta system (PerkinElmer) would provide a reasonable
solution. Specifically, high-content imaging afforded the
opportunity to simultaneously collect quantitative cell viability
data, as well as qualitatively assess cell morphologies.20
To determine an appropriate cell density for high-content
imaging studies, a series of experiments were performed
examining the impact of different cell-seeding strategies on
assay performance. It was determined that 2 × 104 HepG2 cells
per well in a 96-well microtiter plate offered ample numbers of
cells per field to reproducibly quantify cell viability, while it
also prevented overcrowding, so that the morphologies of most
individual cells were discernible. Under these assay conditions,
healthy HepG2 cells grew as islands of polygonal-shaped cells
(Figure 3). Using the Operetta image analysis software, we
automated the determination of cell shape and area, which
provided two important assay-design benefits. First, this
approach allowed us to detect minor perturbations in cell
shape (an indicator of overall cell health), and second, cell area
could now be used as a metric for estimating viability. The
correlation between cell area and viability was made known to
us during the data exploration phase of our assay development
process; we had observed that cell viability (as determined by
MTT and SRB assays) of a cell population following AfB1
exposure was strongly correlated with average cell size. Thus,
cell area was chosen as a novel surrogate metric for testing
purposes because of its relationship to cell viability.
Next, the concentration−response effects of AfB1 on HepG2
cell viability was investigated with the goal of determining the
EC50 value of the toxin at 48 h post-treatment. An EC50 value
of 30 μM was established for AfB1, which was in the range of
other reported in vitro assay systems.18,19 The Z-factor for the
assay was determined to be 0.82, indicating that the assay
design provided a high likelihood of detecting bioactive
samples under high-throughput screening conditions.
Testing Two Putative AfB1 Protective Compounds:
Oltipraz and N-Acetylcysteine. Previous reports for two
compounds, oltipraz and N-acetylcysteine (Figure 2),
suggested that these compounds offered protection from
AfB1-induced cellular toxicity.21,22 However, these results have
been called into question by a subsequent investigation12
Figure 2. Structures of previously studied putative chemoprotective
agents chlorophyllin, oltipraz, and N-acetylcysteine.
in humans at risk for dietary exposure to AfB1.12−16 While
these compounds have not seen marked successful use in
humans, they do serve as inspiration for the development of
new and more effective chemoprotective agents that are
capable of mitigating the effects of AfB1 toxicity.
While AfB1’s metabolism, disposition, and clinical toxic
manifestations vary among different animals, age groups, and
people of different genetic backgrounds,11 our study was
focused on addressing the constellation of acute and short-
term toxic effects imposed by AfB1 on human adults under the
threat of receiving a bolus dose of the toxin. To address a key
feature posed by AfB1 in the context of this threat (i.e., liver
toxicity), we developed a bioassay system to enable the
identification of new natural products that reduced the cellular
toxicity of AfB1 toward human liver-derived cells (HepG2).
Historically, the HepG2 cell line has been used as a proxy for
hepatocyte function and drug metabolism since it is known to
expresses several CYP enzymes, albeit at levels that are
markedly lower than those observed in intact human liver
tissues.17 In this report, we detail the development and testing
of our new assay system and describe the identification of
several natural and related synthetic molecules that afforded in
vitro protection against AfB1 toxicity in monolayers and
spheroids composed of HepG2 cells.
RESULTS AND DISCUSSION
■
Assay Development and Optimization. At the outset of
this study, our team identified a small number of published
peer-reviewed methods that described the identification of
B
DOI: 10.1021/acs.jnatprod.9b00290
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
Figure 3. Two of the top bioactive samples (hits) from screening fungal and bacterial extracts for substances affording protective effects against
AfB1 toxicity (30 μM) in monolayers of HepG2 cells. The cells simultaneously challenged with AfB1 and treated with the extract from plate 58, well
B1 (source of compounds 1−3) exhibited partially restored cell morphologies compared to the cells treated with AfB1 alone (note how the cells
treated only with AfB1 are smaller and more angular in appearance). In comparison, the cells treated with AfB1 and the extract from plate 59, well
E6 (source of compounds 4 and 5) exhibited increased cell area, but were morphologically dissimilar to the cells receiving no AfB1 (vehicle-only
treated controls).
Figure 4. Oltipraz and N-acetylcysteine provided no discernible protection from AfB1 toxicity in HepG2 cell monolayers. Error bars represent
1
SD calculated from a two replicate measurements of live cell area.
showing that administration of oltipraz causes adverse effects
in human patients including gastrointestinal symptoms,
phototoxicity, thermal sensitivity, and paresthesias. N-Acetyl-
cysteine administration can cause adverse effects as well,
ranging from nausea to death.23 In our assay system, neither
compound showed protective effects against AfB1 (Figure 4),
with oltipraz actually increasing the toxic effects of AfB1.
Interestingly, we observed that both compounds caused
increases in HepG2 cell area when administered alone, which
may have been a contributing confounding factor in prior tests.
Nevertheless, it is also possible that the protective effects of
oltipraz and N-acetylcysteine are strictly manifested in vivo
where other, system-level biological processes may be at play.
Screening of Natural Product Extracts. An exploratory
screen was carried out to test natural product extracts derived
from bacteria and fungi for the presence of compounds that
offered protection against AfB1 toxicity. For these tests, 276
bacterial extracts (tested at 33.4 μg/mL) and 644 fungal
extracts (tested at 16.7 μg/mL) were screened. These
concentrations were chosen as reasonable starting points
based on previous experience applying our natural product
extract library to screening programs conducted in the lab.
While none of the bacterial samples were found to be active
(even when the initial concentration was doubled from 33.4
μg/mL to 66.8 μg/mL), 21 of the fungal samples (3.3%)
showed protective activities (protection was defined as the
percent change in live cell area for cells treated with AfB1 and
test compounds/extracts versus cells treated with vehicle only;
compounds/extracts that afforded >50% changes in relative
live cell areas were considered protective). A second round of
testing was undertaken to confirm the protective effects of the
bioactive samples resulting in 17 extracts that showed
protective activities at 16.7 μg/mL (i.e., increased live cell
area and partially restored HepG2 cell morphology so that cells
C
DOI: 10.1021/acs.jnatprod.9b00290
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
Figure 5. Structures of fungal secondary metabolites 1−6 tested in this study.
appeared more like untreated-control samples) (refer to Figure
3 for examples of cellular phenotypes generated by active
samples).
advantage of the supplies of these compounds that were
already available in our pure compound library, we confirmed
that 1 and 2 were indeed highly active, affording 67% and 66%
maximum relative live cell areas (MRLCAs) following AfB1
exposure, while 3 did not protect the HepG2 cells (MRLCA =
15%) (Table 1). Further testing of molecules 1 and 2 in the
Seven of the bioactive fungal extracts that provided
improvements in HepG2 cell morphology (AfB1-treated cells
appeared similar to cells not exposed to AfB1) were selected,
and the source fungi were subject to small-scale fermentation.
Fungi were grown for 4 weeks in 250 mL flasks on monolayers
of Cheerios breakfast cereal, at which point the cultures were
extracted with EtOAc and subjected to partitioning against
H2O. The organic phases were retained, the resulting mixtures
of natural products were subjected to LC-MS analyses, and the
Table 1. Maximum Protection Afforded by Compounds 1−
14 from AfB1 Toxicity in HepG2 Monolayers, Expressed as
Relative Live Cell Areas 1 SD (Compared to Vehicle-Only
a
Treated Cells)
b
compound
MRLCA (%)
maximum concentration tested (μM)
natural-product-containing residues were tested in the assay to
generate concentration−response curves (1.25 to 50 μg/mL).
Three of the samples (laboratory codes 58B1, 55C11, and
59E6) showed good concentration−response effects; however,
LC-MS revealed that two of the samples (58B1 and 55C11)
contained nearly identical natural product profiles. This left
two samples, 58B1 and 59E6 (Figure 3), which were selected
for further study. Notably, sample 58B1 provided greater
retention of HepG2 cell morphology compared to the vehicle-
only controls, while sample 59E6, which afforded comparable
live cell areas, lacked the ability to restore normal HepG2 cell
morphology. Both fungal isolates were subjected to ITS
sequencing, revealing their likely identities as an Alternaria sp.
(source of extract 58B1) and an Aspergillus sp. (source of
extract 59E6).
1
2
3
4
5
6
7
8
67.0 1.1
65.7 4.7
15.2 5.8
64.9 6.8
20.5 2.0
68.3 14.5
37.0 0.1
61.6 6.1
53.9 1.5
60.2 8.2
41.0 3.4
63.9 0.7
72.0 4.6
33.5 0.3
50
50
0.5
c
cd
,
0.5
c
5
50
c
5
c
12.5
50
50
9
10
11
12
13
14
50
cd
,
5
cd
,
0.5
0.5
c
a
When monolayers of cells were exposed to 30 μM AfB1, it resulted in
Purification of Compounds 1−3 from Alternaria sp.
Isolate 58B1. The small-scale extract produced from the
Alternaria sp. (58B1) was subjected to HPLC chromatography
over C18, and the resulting fractions were deposited into a
deep-well 96-well plate. After removal of the solvent, the
organic residues were resuspended in DMSO, and aliquots
were transferred by a pin tool to fresh 96-well microtiter plates
for bioassay testing. Our team has employed this methodology
as a strategy for accelerating the process of purifying bioactive
natural products from microorganisms and plants, and details
are provided elsewhere.24 Three wells exhibited protection
from AfB1 toxicity, and their contents were analyzed by LC-
MS. The first two active wells contained mixtures of two
major-component molecules, while the third active well
contained one major component. We were able to quickly
identify all three compounds as alternariol-9-methyl ether (1),
altenuisol (2), and alternariol (3) (Figure 5) by dereplication
against our lab’s internal database of fungal metabolites. Taking
relative live cell areas that were ∼30% compared to vehicle-only
treated cells. Accordingly, MRLCA (maximum relative live cell areas)
values of >50% were implemented as the threshold for substances to
b
qualify as potentially protective against AfB1. MRLCA values were
obtained by taking the live cell area recorded for the concentration of
a test substance that afforded the greatest total live cell area and
dividing that value by the live cell area obtained for vehicle-only
treated cells [(live cell area for treated cells/live cell area for cells
treated with vehicle only) × 100%]; data are expressed as percentages
c
1 SD, n = 2. The live cell area did not increase at higher
d
concentrations tested. Toxicity was observed at 50 μM.
absence of AfB1 showed that compound 2 was mildly cytotoxic
to the HepG2 cell line, while 1 exhibited no overt cytotoxic
effects. Upon comparing the structures of compounds 1 and 2
versus 3, we observed that the foremost difference between
these metabolites was the presence of a methoxy group at the
9-position of 1 and 2. On the basis of these observations, we
D
DOI: 10.1021/acs.jnatprod.9b00290
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
Figure 6. Structures of synthetic and commercial compounds and 7−14 tested in this study.
hypothesized that the presence of the methoxy moiety was an
important contributor to the protective effects of 1 and 2.
Purification of Compounds 4 and 5 from Aspergillus
sp. Isolate 59E6. The extract produced from the small-scale
preparation of the probable Aspergillus sp. isolate (59E6) was
processed, and the fractions were tested for protection against
AfB1 toxicity as described for the Alternaria sp. isolate (vide
supra). Two fractions were identified as bioactive, and their
contents analyzed by LC-MS. These wells contained relatively
pure, single metabolites that were identified as rubrofusarin B
(4) and funalenone (5) (Figure 5) by LC-MS-guided
dereplication and comparisons with samples previously
deposited in our pure compound library. Using the available
purified material, we determined that 4 and 5 afforded 65%
and 21% MRLCA values, respectively, at test concentrations of
50 μM (Table 1).
Assessing the Protective Properties of Compounds
6−14. When examined collectively, metabolites 1−5 are
relatively small, planar, aromatic molecules that are character-
istic of the types of substrates reported as inhibitors of
CYP1A2.25 These types of compounds are thought to readily
fit in the relatively narrow binding pocket of this enzyme.26 On
the basis of this observation, we investigated a series of
structurally similar molecules as candidates for protecting
HepG2 cells against AfB1-induced toxicity. Based on the lack of
activity observed for compound 3, the methoxyphenol moiety
found in compounds 1 and 2 was thought to be necessary to
afford protective activity. Therefore, we proceeded to test
another secondary metabolite from our library, monomethyl
sulochrin (6) (Figure 5), which is characterized as a
benzophenone. Compound 6 exhibited an MRLCA of 68.3%
(50 μM), which was comparable to the results observed for the
other bioactive fungal metabolites identified in our testing.
However, when testing the synthetic compound 3,5-dimethox-
yphenol (7) (Figure 6), which contains two m-methoxyphenol
moieties, we observed very little activity, as evidenced by a low
MRLCA value (37.0% at 5 μM with no increase at higher
concentrations), suggesting that compound 7 may be too small
and/or at least two aromatic rings might be required to provide
activity. Since benzophenones are known substrates of
CYP1A227,28 and are more readily accessible via synthesis
than the polycyclic systems comprising compounds 1 and 2,
we decided to explore the activities of additional benzophe-
nones containing 3-methoxyphenol moieties. For these
reasons, oxybenzone (8) and benzophenone-6 (9) (Figure
6) were tested using the HepG2 monolayers, resulting in
MRLCAs of 61.6% (12.5 μM) and 53.9% (50 μM),
respectively. This suggested that increasing the number of 3-
methoxyphenol units in a molecule did not necessarily lead to
further increases in protection. Importantly, these observations
provided evidence that benzophenones may serve an
alternative scaffold for disrupting the cellular toxicity caused
by AfB1.
Published studies exploring other classes of compounds
reported to afford protection against AfB1 (e.g., flavonoids and
stilbenoids) revealed that small-molecule-dependent inhibition
of CYP1A2 and functionally similar CYPs was highly reliant on
the positioning of the pendant hydroxy and alkoxy substituents
among these chemical scaffolds.29−31 To determine if an
alternative substitution pattern on a benzophenone would
increase or diminish protection, we synthesized and tested 4-
hydroxy-3-methoxybenzophenone (10) (Figure 6). This
compound exhibited an MRLCA of 60.2% (50 μM), indicating
that o-methoxyphenol units were acceptable for inclusion on
this scaffold. However, when ketone 10 was reduced to yield 4-
hydroxy-3-methoxybenzhydrol (11) (Figure 6), the MRLCA
resulting from this compound dropped to 41.0% (50 μM),
suggesting that the carbonyl group is an important contributor
to protective activity. Given the planarity of many reported
CYP inhibitors, as well as the planarity of compounds 1, 2, and
4, we hypothesized that testing other planar aromatic scaffolds
might prove valuable. Accordingly, we identified 1-hydroxy-3-
methoxyxanthone (12) as a candidate xanthone scaffold
(Figure 6) to probe because of its structural similarity and
increased planarity compared to 8. Compound 12 exhibited a
similar MRLCA (63.9% at 5 μM) to that observed for 8;
however, this compound also showed signs of toxicity, thus
diminishing further interest in this compound series.
While reviewing published accounts of small-molecule CYP
inhibitors, the compound ritonavir (13)32 (Figure 6) caught
E
DOI: 10.1021/acs.jnatprod.9b00290
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
our attention since it represented a significant departure from
the arene-containing systems we had investigated. Further
fueling our interest in this compound was the fact that 13 is
widely reported to inhibit CYP3A4, which has also been
implicated in the activation of AfB1 into its toxic epoxide
form.2,11 To test the hypothesis that 13 may protect cells from
AfB1, its activity was assessed in our assay system. This test
revealed that 13 afforded strong protection against AfB1
(72.0% relative live cell area at 5 μM), suggesting that
CYP3A4 and its functional homologues may also be viable
targets for limiting the toxicity of AfB1. These observations
were supportive of earlier studies.9 Combining this observation
with other reports citing nitrogen-containing heterocycles as
inhibitors of CYPs,33−36 we attempted to combine a
heterocycle system with an arene to determine if an
amalgamation of these moieties would offer protection from
AfB1-induced toxicity. Accordingly, the compound 2-hydroxy-
4-methoxy-N-(3-pyridinylmethyl)benzamide (14) (Figure 6)
was prepared and tested. However, this molecule afforded little
to no protection (33.5% MRLCA at 0.5 μM with no increase at
higher concentrations tested) to HepG2 cells. Since our
studies had already generated several other promising
compounds, we did not pursue this or other classes of
compounds further.
Testing a HepG2 Spheroid Model for Assessing AfB1
Protection. While our assay system proved capable of
detecting compounds that afforded protection to HepG2
cells that had been exposed to AfB1, we were aware that certain
limitations remained in our approach. Specifically, CYP
expression in HepG2 monolayers is known to be reduced in
comparison to CYP expression/activity levels under in vivo
conditions.17,19,37,38 Furthermore, methods to consistently and
uniformly upregulate CYPs in vitro remain challenging,38 and
our own experiments with several small-molecule promotors of
CYP activity proved unreliable or provided only small increases
in putative CYP activity (data not shown). Problems such as
these were not only apparent in our studies but were observed
in virtually every published account assessing the in vitro
toxicity of AfB1 (i.e., the amount of toxin required to cause cell
death was consistently higher than what was required to
manifest toxicity in vivo).37,39
Figure 7. Relative live cell areas of HepG2 spheroids treated with
AfB1 and test compounds (10 μM). Compounds were considered
protective if they enabled spheroids to achieve relative live cell areas
that were >50% compared to vehicle-only treated spheroids. Error
bars are expressed as 1 SD, n = 3.
scopic analysis of the cells contained in the treated spheroids
indicated that these cells retained morphologies that were
indistinguishable from the unchallenged controls (Figure 8).
Using our newly developed high-content assay, we screened
a library of bacterial and fungal crude extracts, resulting in the
identification of three fungus-derived natural products that
afforded protection against AfB1 toxicity. In support of this
assay, we found the images collected using the Operetta to be
invaluable; these data enabled us to concentrate on active
samples that preserved the natural cellular morphology of
healthy liver cells. Using natural products as a starting point,
we were able to identify other structurally similar fungal
secondary metabolites that also exhibited strong protective
activities. Additionally, we used synthetic methods to further
explore bioactive scaffolds and obtain valuable insights
concerning the structural requirements affecting activity and
identify paths toward further scaffold optimization. While we
did not confirm the mechanism(s) by which the compounds
afford protection to HepG2 liver cells, the increased sensitivity
of HepG2 spheroids to AfB1 and the increased protection
afforded by these compounds in these spheroid assays
suggested that some of the bioactive molecules may be acting
wholly or in part through inhibition of one or more CYPs. This
potential mechanism, as well as the potent protective activity
observed for 13, a known CYP3A4 inhibitor, suggested that
the use of selective or semiselective inhibitors of CYPs (e.g.,
CYP1A2, CYP3A4, and CYP3A5) may provide a viable route
toward the development of chemoprotective measures that
defend against AfB1 toxicity. Ideally, such compounds may be
used to prevent liver damage and reduce the risk of acute liver
toxicity and hepatocellular carcinoma in people and animals at
risk for exposure to bolus doses of AfB1 and its toxic analogues.
Considering these problems, we were intrigued by reports
that CYP expression is increased in HepG2 cells when they are
grown as spheroids.40,41 Spheroids, which are a form of 3D cell
culture in which cells are allowed to aggregate free of a
noncellular substrate,42 have been used in a number of assay
systems because of their abilities to better mimic in vivo cell
behaviors (e.g., gene expression, drug susceptibility, drug
penetration, drug metabolism, morphology, and more).40−44
As such, we proceeded with further experiments utilizing
HepG2 spheroids in place of the HepG2 monolayers.
Our preliminary experiments creating and testing HepG2
spheroids revealed straightaway a remarkable 15× increase in
the sensitivity of these cells to AfB1 (LC50 of 2 μM in spheroids
versus 30 μM in monolayers). The increased sensitivity to AfB1
treatment appeared to be in agreement with the reported
increase in CYP; however, CYP expression and activity levels
were not directly quantified. Next, we proceeded to test several
of the top-performing monolayer-protecting compounds
(Table 1) in the spheroid model (Figure 7). Two of the
topmost performing compounds, 1 and 12, afforded strong
protection (95% and 72% relative live cell areas, respectively, at
10 μM) from AfB1 exposure (Figure 7). Moreover, micro-
EXPERIMENTAL SECTION
■
General Experimental Procedures. NMR data were collected
on Varian 400 and 500 MHz NMR spectrometers. Accurate mass
(HRESIMS) data were collected on a Waters SYNAPT G2-Si mass
spectrometer. LC-MS data were obtained on a Shimadzu LC-MS
2020 system (ESI quadrupole) coupled to a photodiode array
detector, with a Phenomenex Kinetex column (2.6 μm C18 column,
100 Å, 75 × 3.0 mm). The preparative HPLC system utilized SCL-
10A VP pumps and a system controller with a Luna 5 μm C18 column
(110 Å, 250 × 21.2 mm, 10 mL/min), and the analytical and
semipreparative HPLC system utilized Waters 1525 binary pumps
with Waters 2998 photodiode array detectors and Luna 5 μm C18
F
DOI: 10.1021/acs.jnatprod.9b00290
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
Figure 8. Protection afforded by compounds 1 and 12 from AfB1-induced toxicity in HepG2 spheroids. Spheroids were treated with calcein AM
and Hoechst prior to imaging. In live spheroids, calcein AM was converted to the green-fluorescent calcein product after hydrolysis by intracellular
esterases, whereas, in dead cells, calcein was not produced and only the Hoechst-stained (blue) DNA was observed.
columns (110 Å, 250 × 4.6 mm, 1 mL/min and 110 Å, 250 × 10 mm,
4 mL/min). All solvents were of ACS grade or better.
diluted at a ratio of 1:5 in phosphate-buffered saline (PBS), and 5 μL
of the diluted solution was added to each well of the microtiter plate.
Plates were incubated for 30 min, and the contents of each well were
imaged using Operetta. Harmony software was used to calculate the
live cell areas by finding all Hoechst-labeled nuclei and assigning live
or dead status to each cell based on a threshold of green fluorescence
(set to 350 fluorescent units). This assessment was made possible
because live cells contain esterases that can cleave the acetoxymethyl
(AM) ester group of calcein AM, resulting in the dye’s characteristic
bright green glow inside the cells. Vehicle-only treated controls
(DMSO) were used to calculate 100% culture viability. The
protection afforded by each compound was expressed as the live
cell area of the treated cells relative to the live cell area of the vehicle-
only controls and expressed as a percentage. A compound was
considered protective if the relative live cell area was >50% since
∼30% live cell area was typically observed for the AfB1-treated
samples.
The spheroid cell assay was carried out in a similar manner with the
major difference being the assays were conducted using ultralow
attachment plates (Corning 7007). Aliquots of 1000 cells were seeded
into each well containing 100 μL of EMEM, and the cells were grown
for 4 days until a tight spheroid of cells had formed. Spheroids were
treated with test compounds and 100 μL of complete EMEM with
AfB1 added. Spheroids were grown for three additional days, followed
by viability and live spheroid size determination using the Operetta
high-content analysis system.
Chemicals. Thionyl chloride and 2-hydroxy-4-methoxybenzoic
acid were purchased from Alfa Aesar. Benzophenone-6 (9), 3-
picolylamine, 2-(trimethylsilyl)phenyl trifluoromethanesulfonate, and
2,4,6-trihydroxybenzoic acid were purchased from TCI America.
Cesium fluoride, vanillic acid, and oxybenzone (8) were purchased
from Beantown Chemicals. Oltipraz, dimethyl sulfate, vanillin, and
phenylmagnesium bromide were purchased from Sigma-Aldrich.
Sodium azide was purchased from Fluka, and potassium carbonate
was purchased from Fisher Scientific. Ritonavir (14) was purchased
from Selleck Chemicals. N-Acetylcysteine was purchased from VWR.
3,5-Dimethoxyphenol (7) was purchased from Chem Impex Interna-
tional.
Aflatoxin B1 Protection Assay. The human hepatocellular
carcinoma cell line, HepG2 (HB-8065), was purchased from
ATCC. The cells were maintained in Eagle’s minimum essential
medium (EMEM) (Gibco) with 5% FetalClone III (Hyclone) and
penicillin/streptomycin (50 U/mL, 50 μg/mL) in a humidified
incubator in a 5% CO2 atmosphere. Monolayer-based cell assays were
performed in 96-well plates by seeding 20 000 cells per well in 100 μL
of medium and allowing the cells to attach overnight. Test
compounds were added the next day to the wells from freshly
prepared DMSO stock solutions. Immediately after compound
addition, 100 μL aliquots of a stock solution made from complete
EMEM and AfB1 in DMSO were added to the wells, bringing the total
well volumes to 200 μL. The cells were incubated for 48 h, at which
time viability was determined using a calcein AM and Hoechst 33342
live cell area assay on an Operetta high-content analysis system. To
perform this assay, a stock solution of 40 μM calcein AM and 160 μM
Hoechst 33342 was prepared in DMSO. The stock solution was
Screening Natural Product Extracts. A collection of 644 fungal
extracts dissolved in DMSO was screened at 16.7 μg/mL. Test
samples were transferred to test plates using a 1 μL pin tool. The
protective activities of the extracts were assessed as described above
for the AfB1 protection assay (vide supra). Additionally, a collection of
G
DOI: 10.1021/acs.jnatprod.9b00290
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
276 mammalian microbiome-derived bacterial extracts45 (solubilized
in DMSO) was screened at 33.4 and 66.8 μg/mL using a 2 μL pin
tool. All samples were screened in duplicate. The Operetta system was
used to collect two fields-of-view per well, and the images were saved
for analysis. Active samples were defined as extracts that afforded
protection, resulting in cells that most closely approximated the
vehicle-only treated controls (i.e., extracts and compounds that
provided >75% relative live cell area).
solvent. The resulting residue was redissolved in 80 mL of EtOAc and
partitioned 3× with 35 mL of H2O. The organic phase was removed
under vacuum to yield 1.0 g of crude vanillonitrile. To a dry 20 mL
glass vial was added 2 mL of 3 M phenylmagnesium bromide in
diethyl ether (6 mM) (diluted with 1 mL of dry THF). The
phenylmagnesium bromide was cooled to 0 °C, and 100 mg of crude
vanillonitrile in 0.6 mL of dry THF was added dropwise over a 5 min
period. The reaction mixture was removed from the ice bath and
stirred at room temperature for 25 h. The reaction was heated to
reflux for 21 h. Following reflux, the vial was removed from the heat
source, the reaction quenched with 3 mL of saturated sodium
bicarbonate, and the reaction mixture partitioned against EtOAc (3×
10 mL). The solvent was removed under vacuum and subjected to
silica gel vacuum liquid chromatography using a stepwise gradient
(10:1 hexanes/EtOAc, 1:1 DCM/hexanes, DCM, 10:1 DCM/MeOH,
and MeOH; 300 mL each). The DCM fraction was retained, and the
solvent removed under vacuum. The resulting residue was subjected
to C18 semipreparative HPLC under isocratic conditions with 45%
ACN in H2O to yield 10 (9.2 mg, 6.7% overall yield). The identity of
Purification of Compounds 1−3. The fungal isolate was
identified as an Alternaria sp. (internal designation 58B1) based on
ITS sequence analysis (GenBank accession number MK038723) and
comparison to the BLAST database. The isolate had been obtained
via an oral swab taken from a roadkill opossum.45 The fungal isolate
was grown for 4 weeks on Cheerios breakfast cereal supplemented
with a 0.3% sucrose solution with 0.005% chloramphenicol in a 250
mL flask, at which point it was extracted with EtOAc. The EtOAc
extract (55 mg) was prepared at a concentration of 100 mg/mL, and 5
mg (50 μL injection volume) was subjected to semipreparative C18
HPLC (250 mm × 10 mm, 5 μm) with a MeOH/H2O gradient from
10% to 100% MeOH in 30 min with a 10 min MeOH wash. Fractions
were collected in a deep-well 96-well plate, and the solvent was
removed under vacuum. Each well of the 96-well plate was tested for
activity, resulting in three wells that were bioactive in the AfB1
protection assay. Analysis of the PDA chromatogram revealed that
each active well corresponded to a well-delineated peak(s). LC-MS
analysis of the contents of these wells provided strong matches (i.e.,
retention times, UV-absorption profiles, and MS data) to compounds
already cataloged in our laboratory’s pure compound library
(compounds 1−3). To obtain additional material for compound 1,
the fungal isolate was grown for 4 weeks on Cheerios breakfast cereal
supplemented with a 0.3% sucrose solution with 0.005% chlor-
amphenicol in three large mycobags (Unicorn Bags, Plano, TX, USA).
The resulting fungal biomass was homogenized and extracted with
EtOAc. The crude extract (28.3 g) was subjected to silica gel vacuum
liquid chromatography with elution steps of 1:1 hexanes/DCM,
DCM, 10:1 DCM/MeOH, and MeOH, yielding four fractions. The
MeOH fraction (7.3 g) was subjected to HP20SS vacuum liquid
chromatography and eluted with a step gradient of MeOH in H2O
(30%, 50%, 70%, 90%, 100%) and 1:1 DCM/MeOH, yielding a total
of six fractions. Upon storage in MeOH at −20 °C, a precipitate was
observed in the 100% MeOH fraction from HP20SS VLC, which was
removed by centrifugation. The precipitate was subjected to
semipreparative HPLC using isocratic elution conditions (60:40
acetonitrile (ACN)/H2O) to yield compound 1 (126 mg), which was
10 was verified by comparison of its HRESIMS, as well as 1H and 13
C
NMR spectra to published values.48
Synthesis of 4-Hydroxy-3-methoxybenzhydrol (11). Com-
pound 11 was prepared as described in a previous report with some
slight modifications.49 Specifically, to a 20 mL glass vial were added
290 mg of vanillin (1.25 mM) and 1.5 mL of dry tetrahydrofuran
(THF), and the mixture was cooled to 0 °C. To the cooled vanillin
solution, 1.9 mL of 3 M phenylmagnesium bromide in diethyl ether
(5.7 mM) was added, and the mixture was further diluted in 1 mL of
dry THF (added dropwise over a 5 min period with stirring). The
mixture was stirred at 0 °C for an additional 10 min. The vial was
removed from the ice bath and held at room temperature for an
additional 21 h. The reaction was quenched with the slow addition of
2 mL of MeOH, and 11 (14.9 mg, 3.4% yield) was purified by C18
semipreparative HPLC. The identity of 11 was verified through
comparisons of its HRESIMS, as well as 1H and 13C NMR spectra, to
published values.49
Synthesis of 1-Hydroxy-3-methoxyxanthone (12). The
preparation of compound 12 was initiated by adding 160 mg of
2,4,6-trihydroxybenzoic acid (0.94 mM), 240 mg of potassium
carbonate (1.7 mM), and 5 mL of acetone to a 20 mL glass vial.
To this mixture was added 0.16 mL of dimethyl sulfate (1.7 mM), and
the mixture was stirred at room temperature for 19 h. To the flask was
added 10 mL of H2O, and the reaction mixture partitioned against
EtOAc (3 × 10 mL). The organic phases were combined, and the
solvent was removed under vacuum. The crude product mixture was
purified by C18 semipreparative HPLC under isocratic conditions
(40% ACN in H2O) to yield methyl 2,6-dihydroxy-4-methoxyben-
zoate (88 mg, 47% yield), whose identity was verified by comparison
1
identified based on comparison of its HRESIMS, as well as H and
13C NMR data, to values reported in the literature.46
Purification of Compounds 4 and 5. The fungal isolate was
identified as an Aspergillus sp. (internal designation 59E6) based on
ITS sequence analysis (GenBank accession number MK038724) and
comparison to the BLAST database. The fungus was obtained from a
soil sample obtained from Malibu, California, as part of the University
of Oklahoma, Citizen Science Soil Collection Program.47 The fungal
isolate was grown and extracted in the same manner as above to yield
108 mg of fungal extract. This material was subjected to semi-
preparative HPLC as described above to give fractions that were
collected in a 96-well plate. Each well of the 96-well plate was tested
for activity, producing two wells that were bioactive in the AfB1
protection assay. Analysis of the PDA chromatograms for the contents
of both wells revealed that each active well corresponded to a well-
delineated peak. LC-MS analysis of the contents of these wells
provided strong matches (i.e., retention times, UV-absorption profiles,
and MS data) to compounds already cataloged in our laboratory’s
pure compound library (compounds 4 and 5).
Synthesis of 4-Hydroxy-3-methoxybenzophenone (10). To
a 250 mL round-bottom flask, 1.0 g of vanillin (4.3 mM), 800 mg of
sodium azide (12.3 mM), and 15 mL of ACN was added. The mixture
was stirred for 5 min, followed by the slow addition of 2 mL of
trifluoromethanesulfonic acid (22.6 mM) over a 5 min period. The
mixture was stirred for 30 min; then 15 mL of ACN was added, and
the crude reaction mixture was placed under vacuum to remove the
50
1
of its HRESIMS data and H NMR spectrum to published values.
To a 20 mL glass vial were added 9 mg of methyl 2,6-dihydroxy-4-
methoxybenzoate (0.045 mM), 110 mg of cesium fluoride (0.73
mM), 25 μL of 2-(trimethylsilyl)phenyl trifluoromethanesulfonate
(0.10 mM), and 1 mL of THF. The mixture was refluxed with stirring
for 24 h. After cooling, MeOH and dichloromethane (DCM) were
added, and the solvent was removed under vacuum. The residue was
redissolved in diethyl ether and washed with H2O (3×). The organic
phase was removed under vacuum, and the product was purified from
the reaction mixture by C18 semipreparative HPLC (gradient 10−
100% ACN in H2O with 0.1% formic acid) to yield 12 (1.1 mg, 10%
yield). The identity of 12 was confirmed by comparison of its
51
1
HRESIMS data and H NMR spectrum to published values.
Synthesis of 2-Hydroxy-4-methoxy-N-(3-pyridinylmethyl)-
benzamide (14). To a 2 mL glass vial were added 14 mg of 2-
hydroxy-4-methoxybenzoic acid (0.08 mM), 20 μL of 3-picolylamine
(0.20 mM), 80 μL of triethylamine (0.57 mM), 20 μL of thionyl
chloride (0.27 mM), and 400 μL of DCM. The mixture was stirred at
room temperature for 23 h; then the solvent was removed under
vacuum. The resulting material was redissolved in a mixture
composed of equal volumes of DCM and 1 M aqueous hydrochloric
acid (3×). The organic solvent was removed under vacuum, and the
H
DOI: 10.1021/acs.jnatprod.9b00290
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
residue subjected to C18 semipreparative HPLC (gradient 10−100%
ACN in H2O with 0.1% formic acid) to yield 14 (1.4 mg, 7% yield).
The identity of 14 was verified through comparisons of its HRESIMS
(13) Mycotoxin control in low- and middle-income countries (IARC
Working Group Reports, No. 9); International Agency for Research on
Cancer: Lyon (FR), 2015.
(14) Deng, J.; Zhao, L.; Zhang, N.; Karrow, N.; Krumm, C.; Qi, D.;
Sun, L. Mutat. Res., Rev. Mutat. Res. 2018, 778, 79−89.
(15) Kensler, T.; He, X.; Otieno, M.; Egner, P.; Jacobson, L.; Chen,
B.; Wang, J.; Zhu, Y.; Zhang, B.; Wang, J.; Wu, Y.; Zhang, Q.; Qian,
G.; Kuang, S.; Fang, X.; Li, Y.; Yu, L.; Prochaska, H.; Davidson, N.;
52
1
data and H NMR spectrum to published values.
ASSOCIATED CONTENT
■
S
* Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.jnat-
prod.9b00290.
Gordon, G.; Gorman, M.; Zarba, A.; Enger, C.; Mun
Helzlsouer, K. Biomarkers Prev 1998, 7, 127−134.
̃
oz, A.;
(16) Egner, P.; Wang, J.; Zhu, Y.; Zhang, B.; Wu, Y.; Zhang, Q.;
Qian, G.; Kuang, S.; Gange, S.; Jacobson, L.; Helzlsouer, K.; Bailey,
G.; Groopman, J.; Kensler, T. Proc. Natl. Acad. Sci. U. S. A. 2001, 98,
14601−14606.
NMR spectra for compounds 10−12 and 14 (PDF)
(17) Gerets, H.; Tilmant, K.; Gerin, B.; Chanteux, H.; Depelchin, B.;
Dhalluin, S.; Atienzar, F. Cell Biol. Toxicol. 2012, 28, 69−87.
(18) Chen, X.; Murdoch, R.; Shafer, D.; Ajuwon, K.; Applegate, T. J.
Appl. Toxicol. 2016, 36, 1437−1445.
AUTHOR INFORMATION
■
Corresponding Author
*E-mail: rhcichewicz@ou.edu. Tel: 405-325-6969. Fax: 405-
325-6111.
(19) Yokoyama, Y.; Sasaki, Y.; Terasaki, N.; Kawataki, T.; Takekawa,
K.; Iwase, Y.; Shimizu, T.; Sanoh, S.; Ohta, S. Biol. Pharm. Bull. 2018,
41, 722−732.
ORCID
Robert H. Cichewicz: 0000-0003-0744-4117
Notes
The authors declare no competing financial interest.
(20) Ang, M.; Pethe, K. Cytometry, Part A 2016, 89, 755−760.
(21) Bammler, T.; Slone, D.; Eaton, D. Toxicol. Sci. 2000, 54, 30−
41.
́
(22) Valdivia, A.; Martínez, A.; Damian, F.; Quezada, T.; Ortíz, R.;
Martínez, C.; Llamas, J.; Rodríguez, M.; Yamamoto, L.; Jaramillo, F.;
ACKNOWLEDGMENTS
■
Loarca-Pina, M.; Reyes, J. Poult. Sci. 2001, 80, 727−734.
̃
This material is based upon work supported by the Army
Contracting Command, Aberdeen Proving Ground - Natick
Contracting Division (ACC-APG NCD) under Contract No.
W911QY-17-C-0008. Any opinions, findings, and conclusions
or recommendations expressed in this material are those of the
author(s) and do not necessarily reflect the views of the ACC-
APG NCD. The Shimadzu LC-ESIMS instrument used for this
project was provided in part by a Challenge Grant from the
Office of the Vice President for Research, University of
Oklahoma, Norman Campus, and an award through the
Shimadzu Equipment Grant Program. The Waters LC-ESIMS
instrument was obtained in part through an NSF MRI
equipment grant (1626372). We are grateful for the
contribution of the soil sample from citizen scientist L.
Brewster, from which the Aspergillus sp. isolate was obtained.
We appreciate the helpful comments and advice provided by J.
Phillips during the design and performance of these experi-
ments.
(23) Sandilands, E.; Bateman, D. Clin. Toxicol. 2009, 47, 81−88.
(24) Cai, S.; Risinger, A.; Nair, S.; Peng, J.; Anderson, T.; Du, L.;
Powell, D.; Mooberry, S.; Cichewicz, R. J. Nat. Prod. 2016, 79, 490−
498.
(25) Sansen, S.; Yano, J.; Reynald, R.; Schoch, G.; Griffin, K.; Stout,
C.; Johnson, E. J. Biol. Chem. 2007, 282, 14348−14355.
(26) Dutkiewicz, Z.; Mikstacka, R. Bioinorg. Chem. Appl. 2018, 2018,
1−21.
(27) Watanabe, Y.; Kojima, H.; Takeuchi, S.; Uramaru, N.; Sanoh,
S.; Sugihara, K.; Kitamura, S.; Ohta, S. Toxicol. Appl. Pharmacol. 2015,
282, 119−128.
(28) Takemoto, K.; Yamazaki, H.; Nakajima, M.; Yokoi, T. Mutat.
Res., Genet. Toxicol. Environ. Mutagen. 2002, 519, 199−204.
(29) Shimada, T.; Tanaka, K.; Takenaka, S.; Murayama, N.; Martin,
M.; Foroozesh, M.; Yamazaki, H.; Guengerich, F.; Komori, M. Chem.
Res. Toxicol. 2010, 23, 1921−1935.
(30) Mikstacka, R.; Wierzchowski, M.; Dutkiewicz, Z.; Gielara-
́
́
Korzanska, A.; Korzanski, A.; Teubert, A.; Sobiak, S.; Baer-Dubowska,
W. MedChemComm 2014, 5, 496−501.
(31) Takemura, H.; Itoh, T.; Yamamoto, K.; Sakakibara, H.; Shimoi,
K. Bioorg. Med. Chem. 2010, 18, 6310−6315.
REFERENCES
(32) Zeldin, R.; Petruschke, R. J. Antimicrob. Chemother. 2003, 53,
4−9.
■
(1) Asao, T.; Bu
̈
chi, G.; Abdel-Kader, M.; Chang, S.; Wick, E.;
(33) Sevrioukova, I.; Poulos, T. Dalton Trans 2013, 42, 3116−3126.
(34) Riley, R.; Parker, A.; Trigg, S.; Manners, C. Pharm. Res. 2001,
18, 652−655.
Wogan, G. J. Am. Chem. Soc. 1965, 87, 882−886.
(2) Eaton, D.; Gallagher, E. Annu. Rev. Pharmacol. Toxicol. 1994, 34,
135−172.
(35) Kaur, P.; Chamberlin, A. R.; Poulos, T.; Sevrioukova, I. J. Med.
Chem. 2016, 59, 4210−4220.
(3) Monson, M.; Coulombe, R.; Reed, K. Agriculture 2015, 5, 742−
777.
(36) Halpert, J. Annu. Rev. Pharmacol. Toxicol. 1995, 35, 29−53.
(37) Nakamura, K.; Aizawa, K.; Aung, K.; Yamauchi, J.; Tanoue, A.
Sci. Rep. 2017, 7, 41093.
(4) Squire, R. Science 1981, 214, 877−880.
(5) Kew, M. J. Gastrointestin. Liver Dis. 2013, 22, 305−310.
(6) Liu, Y.; Wu, F. Environ. Health Perspect. 2010, 118, 818−824.
(7) Bennett, J.; Klich, M. Clin. Microbiol. Rev. 2003, 16, 497−516.
(8) Klich, M. Mol. Plant Pathol. 2007, 8, 713−722.
(9) Gallagher, E.; Kunze, K.; Stapleton, P.; Eaton, D. Toxicol. Appl.
Pharmacol. 1996, 141, 595−606.
(38) Herzog, N.; Katzenberger, N.; Martin, F.; Schmidtke, K. Journal
of Cellular Biotechnology 2015, 1, 15−26.
(39) Lauschke, V.; Hendriks, D.; Bell, C.; Anderson, T.; Ingelman-
Sundberg, M. Chem. Res. Toxicol. 2016, 29, 1936−1955.
(40) Terashima, J.; Goto, S.; Hattori, H.; Hoshi, S.; Ushirokawa, M.;
Kudo, K.; Habano, W.; Ozawa, S. Drug Metab. Pharmacokinet. 2015,
30, 434−440.
(41) Takahashi, Y.; Hori, Y.; Yamamoto, T.; Ohara, Y.; Tanaka, H.
Biosci. Rep. 2015, 35, No. e00208.
(42) Fennema, E.; Rivron, N.; Rouwkema, J.; van Blitterswijk, C.; de
Boer, J. Trends Biotechnol. 2013, 31, 108−115.
(10) Ramsdell, H.; Parkinson, A.; Eddy, A.; Eaton, D. Toxicol. Appl.
Pharmacol. 1991, 106, 436−447.
(11) World Health Organization & Joint FAO/WHO Expert
Committee on Food Additives (83rd, G., Switzerland); Safety
evaluation of certain contaminants in food; World Health Organization:
Malta, 2018.
(12) Sudakin, D. J. Toxicol., Clin. Toxicol. 2003, 41, 195−204.
I
DOI: 10.1021/acs.jnatprod.9b00290
J. Nat. Prod. XXXX, XXX, XXX−XXX

Journal of Natural Products
Article
(43) Sant, S.; Johnston, P. Drug Discovery Today: Technol. 2017, 23,
27−36.
(44) Shah, U.-K.; Mallia, J. d. O.; Singh, N.; Chapman, K. E.; Doak,
S. H.; Jenkins, G. J. S. Mutat. Res., Genet. Toxicol. Environ. Mutagen.
2018, 825, 51−58.
(45) Motley, J.; Stamps, B.; Mitchell, C.; Thompson, A.; Cross, J.;
You, J.; Powell, D.; Stevenson, B.; Cichewicz, R. J. Nat. Prod. 2017,
80, 598−608.
(46) Lou, J.; Yu, R.; Wang, X.; Mao, Z.; Fu, L.; Liu, Y.; Zhou, L.
Braz. J. Microbiol. 2016, 47, 96−101.
(47) Du, L.; Robles, A. J.; King, J. B.; Powell, D. R.; Miller, A. N.;
Mooberry, S. L.; Cichewicz, R. H. Angew. Chem., Int. Ed. 2014, 53,
804−809.
(48) Huang, J.; Guo, Y.; Jiang, J.; Liu, H.; Ji, Y. Res. Chem. Intermed.
2015, 41, 7115−7124.
(49) Cohrt, A.; Jensen, J.; Nielsen, T. Org. Lett. 2010, 12, 5414−
5417.
(50) Yuen, T.; Brimble, M. Org. Lett. 2012, 14, 5154−5157.
(51) Genoux-Bastide, E.; Lorendeau, D.; Nicolle, E.; Yahiaoui, S.;
Magnard, S.; Di Pietro, A.; Baubichon-Cortay, H.; Boumendjel, A.
ChemMedChem 2011, 6, 1478−1484.
(52) Petrlikova, E.; Waisser, K.; Palat, K.; Kunes, J.; Kaustova, J.
Chem. Pap. 2011, 65, 52−59.
J
DOI: 10.1021/acs.jnatprod.9b00290
J. Nat. Prod. XXXX, XXX, XXX−XXX