Systematic Diversity-Oriented Synthesis of Reduced Flavones from γ-Pyrones to Probe Biological Performance Diversity

Article abstract of DOI:10.1021/acschembio.9b00294

Diversity-oriented synthesis (DOS) has historically focused on the development of small-molecule collections with considerable chemical diversity with the hypothesis that chemical diversity will lead to diverse biological activities. We took a systematic approach to DOS to develop a focused library of reduced flavones from ?-pyrones with diversity of appendage, stereochemistry, and chemical properties to determine which features of small molecules are most predictive of biological performance diversity. The effects of these systematic modifications on biodiversity were determined using Cell Painting and cytotoxicity assays to compare the results of multiple methods of assessment. We observed that a greater fraction of sp³ hybridized atoms (fsp³) does not always lead to enhanced biodiversity, that stereochemistry and appendage diversity both contribute to biodiversity, and that lipophilicity of the pyrone class of compounds correlates with biodiversity. These results will contribute to the development of a general algorithm to predict which chemical features should be considered during the synthesis of DOS libraries to create biological performance-diverse collections of small molecules for probe and drug discovery.

Full text of DOI:10.1021/acschembio.9b00294

Articles
pubs.acs.org/acschemicalbiology
Cite This: ACS Chem. Biol. XXXX, XXX, XXX−XXX
Systematic Diversity-Oriented Synthesis of Reduced Flavones from
γ‑Pyrones to Probe Biological Performance Diversity
Erica M. Gerlach,‡ Melissa A. Korkmaz,‡ Ivan Pavlinov, Qiwen Gao, and Leslie N. Aldrich*
Department of Chemistry, University of Illinois at Chicago, 845 West Taylor Street, Chicago, Illinois 60607, United States
S
* Supporting Information
ABSTRACT: Diversity-oriented synthesis (DOS) has historically focused on the development of small-molecule collections
with considerable chemical diversity with the hypothesis that chemical diversity will lead to diverse biological activities. We took
a systematic approach to DOS to develop a focused library of reduced flavones from γ-pyrones with diversity of appendage,
stereochemistry, and chemical properties to determine which features of small molecules are most predictive of biological
performance diversity. The effects of these systematic modifications on biodiversity were determined using Cell Painting and
cytotoxicity assays to compare the results of multiple methods of assessment. We observed that a greater fraction of sp3
hybridized atoms (fsp3) does not always lead to enhanced biodiversity, that stereochemistry and appendage diversity both
contribute to biodiversity, and that lipophilicity of the pyrone class of compounds correlates with biodiversity. These results will
contribute to the development of a general algorithm to predict which chemical features should be considered during the
synthesis of DOS libraries to create biological performance-diverse collections of small molecules for probe and drug discovery.
Rapid and efficient biological performance diversity assays
have recently begun to be explored.2 One such assay, Cell
Painting, is a phenotypic screen that multiplexes six dyes to
stain eight cellular components and organelles.3 Cells are
imaged, and hundreds of features are assessed to generate
unique morphological profiles induced by each molecule that
can be compared based on similarity. It has previously been
shown that selection of compounds for diversity of Cell
Painting and gene expression profiles results in greater
biological performance diversity than random selection or
selection based on chemical diversity.4 In addition, clustering
of well-annotated, bioactive compounds based on Cell Painting
profiles successfully grouped compounds with similar bio-
logical activities, highlighting this as a useful method for the
determination of mechanism of action through comparison of
the similarity of profiles obtained in-house or through public
profiles. Cytotoxicity profiling has also been used historically as
a method to compare small molecules to determine the
mechanism of action through profile similarity.5−8 Methods
that enable the determination of mechanism similarity can also
be used to quantify the dissimilarity of cellular profiles to
determine performance diversity.
mall-molecule libraries constructed using diversity-ori-
Sented synthesis (DOS) approaches have enabled discovery
of novel probes for numerous biological pathways.1 DOS has
historically focused on the creation of libraries with
considerable chemical diversity; however, population of
chemical space does not guarantee that small molecules will
have diverse biological properties. Biological performance
diversity, the idea that a small-molecule library possesses the
widest range of possible targets with minimal redundancy,
must be assessed and considered for the development of
optimal libraries for probe and drug discovery. The lack of a
predictive algorithm to link chemical diversity to biological
activity diversity poses a major challenge for modern DOS. A
better understanding of the structural and physical features
that make the most significant contributions to diversity of
biological response is required to refine and optimize small-
molecule libraries. We hypothesize that systematic modifica-
tion of natural product-like precursors to provide diversity of
appendages, stereochemistry, scaffold, and physicochemical
parameters will enable the generation of a predictive algorithm
for library design. The trends identified through these studies
will facilitate the curation of smaller libraries with greater
biological performance diversity, thus lowering the cost and
improving the efficiency and success of high-throughput
screens (HTS) to discover novel probes and therapeutics.
Received: April 12, 2019
Accepted: June 11, 2019
Published: June 11, 2019
© XXXX American Chemical Society
A
DOI: 10.1021/acschembio.9b00294
ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology
Articles
Pyrones and flavones are commonly occurring biologically
active natural products.9−17 It is particularly enticing to
develop a natural-product inspired library of flavonoid
derivatives from γ-pyrones because novel entities from these
structural classes could possess unique and useful biological
properties (Figure 1). The reduced flavone scaffold is of great
Cell Painting and cytotoxicity assays to compare the results of
two methods for the evaluation of performance diversity and to
determine the importance of considering multiple methods of
assessment. We conclude that greater fsp3 does not always lead
to greater performance diversity, that stereochemistry and
appendage diversity are important for performance diversity,
and that lipophilicity of the pyrone analogues correlates with
performance diversity in the cytotoxicity assay but not in the
Cell Painting assay, highlighting how the two methods for
annotation provide complementary results and additional
information that would not have been obtained by only one
assay. The generality and predictability of these features will
continue to be explored to provide a better understanding of
the chemical diversity elements that should be considered
during the construction of DOS libraries.
RESULTS AND DISCUSSION
■
Our goal was to prepare structurally diverse, reduced flavones
through a versatile, efficient route and to incorporate scaffold,
stereochemical, and appendage diversity to build an intelligent,
focused library to assess performance diversity. The γ-pyrones
were accessed from acid chlorides (1) and enaminones (2) in
good to excellent yields for all incorporated appendages
(Scheme 1).19 The pyrones were reacted with trans-3-(tert-
butyldimethylsilyloxy)-1-methoxy-1,3-butadiene to obtain a
1.5:1 mixture of endo:exo isomers after 24 h at reflux in
toluene in 38% yield (Table S1).20,21 Alternative solvents did
not improve selectivity, so we decided to explore various Lewis
acids and their ability to promote this reaction. Catalytic InCl3
and FeCl3 produced the endo product exclusively with >20:1 dr
after only 1 h in excellent yield, 85% and 92%, respectively
(Table S1). Lower catalyst loading and additional time or
alternative solvents did not improve the observed yields. The
Diels−Alder reaction conditions tolerated all substitution
patterns of evaluated γ-pyrones and desired products were
immediately deprotected to provide the enones (Scheme 1).
Treatment with catalytic In(OTf)3 (10 mol %) at −20 °C in
THF provided the enone products in good yields over two
steps (Scheme 1, Conditions A). However, some substrates
were sensitive to these conditions, and the enone would
decompose to an undesired byproduct, phenol 6. Alternative
conditions were developed in which the reaction was
performed in dichloromethanewith the addition of silica gel.
After 20 min at RT, the solvent was evaporated, and the
product was immediately purified to provide the enone in good
yield (Scheme 1, Conditions B). To access enantiomerically
enriched alcohols following reduction of the enone, a Corey−
Bakshi−Shibata (CBS) reduction was performed to obtain two
major diastereomers as the desired products in excellent yield
(Scheme 1).22 Both enantiomers of the catalyst provided good
yields of the diastereomeric products. The resulting alcohols
were then esterified (Scheme 1) to provide a library of
compounds with scaffold, stereochemical, and appendage
diversity with diverse chemical properties and fsp3 (9a−h
and 10a−h). Following esterification, the two diastereomers
were separated for biological evaluation using reverse-phase
HPLC. Waiting until the last step for stereoisomer separation
is incredibly efficient for library preparation and minimizes the
number of transformations necessary to obtain enantioen-
riched products. In this manner, we were able to rapidly obtain
62 pyrones 3 and reduced flavone analogues 7−10, with
stereochemical and appendage diversity for biological evalua-
Figure 1. Structure of natural and reduced flavones. The reduced
flavone scaffold was obtained from pyrones in 19−67% yield over 3 to
4 steps. (A) The fsp3 was considerably increased in the alcohol and
ester reduced flavones compared to the natural flavones. (B) Pyrones
and flavones had diverse activities in the toxicity assay. (C) Pyrones
displayed significant diversity in the Cell Painting assay. Higher fsp3
does not predict greater performance diversity for these scaffolds, so it
is important to include diversity of this feature and other key
parameters to curate performance-diverse libraries.
interest due to the increase in the fraction of sp3 hybridized
atoms (fsp3) and stereochemical complexity of the flavone
core, especially because small-molecule collections with higher
ratios of fsp3 have been shown to have selectivity for specific
targets that is higher than that of a collection of commercial
compounds with lower fsp3.18 Apigenin, galangin, quercetin,
chrysin, and genistein, natural products that have displayed
polypharmacology, including antioxidant, anti-inflammatory,
pro-apoptotic, and autophagy modulation activities, have been
chosen for comparison.12−17 Herein, we describe the system-
atic preparation of reduced flavones from γ-pyrones to
generate a library that incorporates scaffold, appendage,
stereochemical, and physicochemical property diversity and
the biological evaluation of this natural product-like library.
The library was designed to establish correlations between
biological performance diversity data and chemical diversity
through systematic adjustment of these parameters. Synthesis
of flavonoid analogues also enables the direct comparison of
the effect of increasing the fraction of sp3 hybridized atoms in a
particular scaffold to the biological performance diversity of a
class of compounds. The effects of these systematic
modifications on biological diversity were determined using
B
DOI: 10.1021/acschembio.9b00294
ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology
Articles
a
Scheme 1. Synthesis of the Library of Reduced Flavones
a
Enaminones and various acid chlorides were used to form a variety of pyrones 3. The pyrones 3 underwent a cycloaddition with trans-3-(tert-
butyldimethylsilyloxy)-1-methoxy-1,3-butadiene to provide 4a−d. Subsequent deprotection and elimination provided enones 5a−d. A separate
method (B) was developed to prevent decomposition of the enone to phenol 6. Four enones were reduced to the alcohols through the
enantioselective CBS reduction followed by subsequent esterification using isovaleryl acid chloride or benzoyl chloride. Diastereomeric ratios were
calculated by NMR using 3 or 4 separate signals. Diastereomeric mixtures 9a−g and 10a−h were separated by reverse phase preparatory HPLC.
tion and with analogues 9a−g and 10a−h as separated
diastereomers.
cancer cells per well were plated in a 384-well plate and
incubated overnight at 37 °C. Cells were then treated with the
compound library in a 6-point dose (40 to 1.25 μM in DMSO,
2-fold dilution) with DMSO-treated wells serving as a control.
After 72 h, cell viability was measured using the CellTiter-Glo
assay. Three independent experiments in duplicate were used
to calculate the area under the curve (AUC) values for the
concentration response curves of each compound. The natural
flavones and pyrones were the most cytotoxic compound
classes; however, the natural flavones were considerably more
cytotoxic in HeLa cells (Figure 2A), whereas the pyrones were
more potent in the OVCAR-8 cell line (Figure 2B). By
comparing the AUC values, the pyrones had the greatest
diversity of response in both cell lines, and the natural flavones
were considerably more diverse in the HeLa cell lines (Figure
2C). The reduced flavone alcohols and esters were relatively
nontoxic in both cell lines. The effect of different appendages
on the activity of the pyrones and flavones was assessed by
Before performing any cell-based experiments, we assessed
the stability of each intermediate and final product in pH 7
phosphate buffered saline (PBS) solution at 37 °C. After 24 h
at 37 °C, enone 5a had almost completely decomposed (less
than 20% remaining) to phenol 6 (Figure S1). Alcohol 8a and
ester 10a were significantly more stable and showed no
significant decomposition over 72 h. In light of these results,
we were confident that the alcohols and esters would give
reliable results in biological experiments, whereas the enones 5
would not be stable to evaluation conditions and were not
included in the library.
We next compared the biological performance of our
reduced flavones 7−10 to natural flavones and γ-pyrones 3
to determine which features make the most significant
contribution to biological performance diversity. For cytotox-
icity profiling, 1000 cervical (HeLa) or ovarian (OVCAR-8)
C
DOI: 10.1021/acschembio.9b00294
ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology
Articles
Figure 2. Appendage and stereochemical diversity lead to biological performance diversity across compound classes in the cytotoxicity assay. (A, B)
Dose−response curves represent 72 h cell toxicity data for pyrones and natural flavones in HeLa and OVCAR-8 cell lines, respectively. Compounds
treated in HeLa (A) vs OVCAR-8 (B) have different dose−response curves, signifying variation in activity between the two cancer cell lines. (C)
Box and whisker plot represents the wide range of AUC values among different compound classes in HeLa (upper plot) and OVCAR-8 (lower
plot) cell lines, specifically among the pyrone compound class. (D−F) Heat maps were prepared using p-values that were generated from AUC
comparisons between each pair of compounds in the cytotoxicity assay in HeLa cells. Student’s t-test was used to calculate p-values, and p < 0.05
denotes a significant difference in activity between two compounds. Heat maps show a greater variation in p-values of pyrones 3 (D) than natural
flavones (E) and reduced flavone esters 9 and 10 (F) in the HeLa cell line, which supports the importance of scaffold for activity. Heat map for the
esters 9 and 10 (F) displays p-values generated from the AUC averages of all four stereoisomers to consider the effect of appendages alone, which
does not provide significant activity diversity. However, the great diversity observed among the pyrones can be attributed to appendage diversity.
(G) Heat maps show the effect of appendages for each of the four stereoisomers in the HeLa cell line, highlighting the importance of both
appendages and stereochemistry for performance diversity. (H) Heat maps compare all four stereoisomers with each set of appendages and reveal
high diversity in p-values among each appendage set for several sets of ester stereoisomers in the HeLa cell line, further confirming the importance
of stereochemistry and appendages to provide performance diversity.
calculating p-values for comparison of the AUC of each
pyrone, natural flavone, or reduced flavone with each
appendage (Figures 2D−F). For the pyrones 3, variation of
the appendages creates significantly diverse responses (Figure
D
DOI: 10.1021/acschembio.9b00294
ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology
Articles
Figure 3. Comparison of biological performance diversity for each compound class using the Cell Painting assay and analysis of the effect of
chemical diversity by principal component analysis. (A) Scatter plot shows the distribution of each compound based on the first two principal
components from the PCA. Each point represents the average of four replicates at 80 μM, and points shown in gray are inactive as determined by
their mp-values (>0.05). (B) Dot plot on the left shows the overall variation in activity scores for the entire library relative to DMSO with the cutoff
for determination of active compounds represented as a dotted line. Out of the entire library, 1 of the natural flavones and 11 of the 14 pyrones are
active. The box and whisker plot on the right displays the distribution of the average activity scores for each compound class. Each point is the
average of the activity scores for one compound across all doses. (C) Heat map represents the effect of appendages on the mp-values for each
pyrone. mp-Values were calculated based on the Mahalanobis distance between a pair of pyrones. mp-Value < 0.05 denotes a significant difference
in the morphologies produced by that pair of compounds.
Figure 4. Correlation between biological activity and cLogP. (A) Scatter plots reveal the correlation between AUC of pyrones (left) and reduced
flavone esters (right) with cLogP in HeLa cells. (B) Scatter plots reveal the correlation between average activity scores of pyrones (left) and esters
(right) with cLogP in HeLa cells. In panels A and B, cLogP values were calculated using SwissADME.13 (C) Dot plot of the Pearson’s correlation
between cLogP and AUC or average activity scores for each class of compounds with grayed out points showing insignificant correlations (p-value
> 0.05). The pyrones show a significant correlation between biological activity and cLogP in the cytotoxicity assays.
2D), especially in comparison with the esters 9 and 10, where
appendage does not make as large of a difference alone (Figure
2F). For the esters 9 and 10, the combination of appendages
(Figure 2G) and stereochemistry (Figure 2H) appears to be
important for diversity of response. By comparing the
contribution of appendages for each stereoisomer, scaffold 9-
E
DOI: 10.1021/acschembio.9b00294
ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology
Articles
Figure 5. IC50 values from cytotoxicity assays and average activity scores from the Cell Painting assay for pyrones (A), natural flavones (B), and
selected values for reduced flavones (C). Additional data are reported in Table S2.
1 and 10-1 had the most significant activity differences due to
appendages (Figure 2G). When comparing sets of stereo-
isomers, compounds 9a-1, 9a-2, 10a-1, and 10a-2 have very
similar activities with very little variation among p-values,
whereas compounds 9g-1, 9g-2, 10g-1, and 10g-2 have a
greater variation of p-values and thus more diverse responses
(Figure 2H). These results highlight the fact that both
appendage and stereochemistry diversity are important for
biological performance diversity.
A Cell Painting assay was used to further assess biological
performance diversity. 1500 HeLa cells per well were plated in
a 384-well plate and grown overnight at 37 °C. Cells were then
treated with the compound library in dose (80 to 5 μM in
DMSO, 2-fold dilution) with DMSO-treated wells serving as a
F
DOI: 10.1021/acschembio.9b00294
ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology
Articles
control. After 24 h, the cells were fixed; permeabilized; stained
with dyes to visualize the nucleus, endoplasmic reticulum,
nucleoli and cytoplasmic RNA, F-actin cytoskeleton, Golgi,
plasma membrane, and mitochondria (Table S2); and imaged
using a high-throughput fluorescence microscope (IXM-XLS).
Two independent experiments were performed in duplicate.
Images were analyzed using CellProfiler to generate per-cell
profiles containing over 1700 features for each cell in the
image.25 Per-cell profiles were aggregated, standardized, and
averaged into per-well profiles with z-scores generated based
on the median profile relative to DMSO-treated wells. Thus,
per-well profiles represent the average phenotype produced by
compound treatment. The profiles for different compounds
were compared through principal component analysis (PCA)
(Figure 3A). Mahalonobis distances (activity scores) were
calculated to assess the differences between cell morphologies
induced by each compound relative to DMSO-treated cells and
mp-values calculated to determine the statistical significance of
the observed difference (Figure 3B).23 By this metric, the
natural flavones and the pyrones 3 were the only groups of
compounds containing “active” molecules in the Cell Painting
assay as determined by a statistically significant activity score
(mp-value < 0.05). For the active pyrone library, it is clear that
appendages greatly contribute to the diversity of activity
(Figure 3C). A comparison of the lipophilicity of the pyrone
analogues 3 using cLogP reveals a correlation between
calculated cLogP and toxicity (AUC) (Figure 4A and B).
Pearson correlation values were calculated for each library to
compare activity score, AUC values, and cLogP (Figure 4C).
Whereas the toxicity of the pyrones significantly correlates with
lipophilicity, the activity score of the pyrones did not correlate
with lipophilicity, highlighting the utility of using several assays
to reveal different features that provide the most significant
biological diversity. In this case, it appears that incorporation of
compounds with diverse physicochemical properties, specifi-
cally lipophilicity, provides performance diversity.
It is of interest to note that for the pyrones 3, the most active
compounds in Cell Painting are not always the most toxic
(Figure 5A). This observation is also true for the natural
flavones (Figure 5B). The reduced flavones did not show
significant activity in the Cell Painting assay (Figures 3A and B
and 5C, Table S2). For the reduced flavones, 10b-2 is more
cytotoxic in the OVCAR8 cell line than in the HeLa cell line
(Figure 5C, Table S2). 10b-2 is also more cytotoxic than its
stereoisomers 9b-1, 9b-2, and 10b-1. These results highlight
the importance of stereochemistry for biological activity and
selectivity. The difference between the potency of 10d-1 and
its stereoisomers 9d-1, 9d-2, and 10d-2 further supports this
observation. Appendages also make a significant contribution.
In general, the isovaleryl esters at R2 provided more potent
compounds in the cytotoxicity assays than the phenyl esters,
whereas variation at the other positions was better tolerated to
maintain cytotoxicity.
to performance diversity. We also observed that the sp3-
enriched alcohols 7 and 8 and esters 9 and 10 were inactive in
the Cell Painting assay; however, they might have activities in a
set of different assays with various functional read-outs. In fact,
we observed significant activity for several esters 9 and 10 in
the 72-h cytotoxicity assays that indicated stereoisomerism and
appendages as chemical diversity elements that affect perform-
ance diversity. These observations reveal the importance of
using multiple assays such as Cell Painting and 72-h
cytotoxicity to fully assess performance diversity. Furthermore,
while it may be important to include molecules with diverse
fsp3 values to obtain performance diverse libraries, it is critical
to determine the impact of increasing fsp3 character on the
activity of a particular scaffold early in the library generation
stage, potentially in multiple assays, so a scaffold that lacks an
increase in diverse activity will not be over-represented. It is
also important to note that it is likely that sp3-deficient
scaffolds should be included in addition to sp3-enriched
scaffolds to obtain a more performance-diverse collection of
small molecules.
CONCLUSION
■
In summary, we optimized a method to rapidly access
enantiomerically enriched reduced flavone analogues in 19−
67% yield over 3−4 steps from γ-pyrones. Using Cell Painting,
we determined that increased fsp3 of the reduced flavones does
not significantly improve biological performance diversity of
these scaffolds compared to the natural flavones and pyrones
within the limits of the detectable phenotypes for this
particular morphological profiling assay. We discovered that
lipophilicity, through the incorporation of different appen-
dages, predicted scaffold activity and resulted in performance
diversity among the pyrone analogues in the cytotoxicity assay
but not in Cell Painting. By comparison of the pyrones 3,
natural flavones, and reduced flavone esters 9 and 10, it is clear
that the scaffold makes a significant contribution to perform-
ance diversity. Our results also reveal cytotoxicity evaluation as
an effective method for preliminary assessment of biological
diversity through comparison of AUC values. We were able to
conclude that the pyrone compound class (3) had the greatest
variability among library members with different appendages
and that the ester compound class (9, 10) contained fewer
significant differences among biological performances of the
library members, making the diversity of this library much less
than that of the pyrone library, which is further supported by
the results from Cell Painting. Although the ester class had less
diverse activities in the evaluated assays, it was determined that
both appendages and stereochemistry are important for the
observed diversity of the ester collection in the cytotoxicity
assay. This work reveals the importance of including various
assays for the measurement of performance diversity to
evaluate diversity early in the library construction phase to
decrease performance-redundant scaffolds. It also reveals the
significance of chemical properties such as lipophilicity for
performance diversity and the observation that compounds
with lower fsp3 can also possess diverse biological activities. In
conclusion, a diverse combination of several factors, including
scaffold, appendages, stereochemistry, fsp3, and physicochem-
ical properties, combined with early assessment of performance
diversity of pilot libraries, will contribute to the development of
performance-diverse, small-molecule libraries. Additional scaf-
folds will need to be explored to determine if the observed
trends are general and to obtain a better understanding of the
Recently, Cell Painting has been used to assess the biological
performance diversity of a library based on a different scaffold,
which led to the conclusion that stereoisomerism and
appendages both make important contributions to biological
performance diversity on this scaffold.26 This is also the effect
that we have observed with both the cytotoxicity profiling and
Cell Painting approaches. These results highlight the
importance of systematic library preparation to explore the
contributions of as many chemical features as possible to
evaluate the generality of the observed trends that contribute
G
DOI: 10.1021/acschembio.9b00294
ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology
Articles
384 well plates. After attachment overnight, test compounds (6-point
dose, 2-fold dilution, 40 to 1.25 μM) and control (DMSO) were pin-
transferred into assay plates using a 96-well pin tool (V&P Scientific)
and a liquid handler, Biomek NXP Lab Automation Workstation
(Beckman Coulter). Assay plates were incubated at 37 °C in 5% CO2
for 72 h. Plates and CellTiter Glo reagent (Promega G9241) were
then equilibrated at RT for 30 min. CellTiter Glo reagent was added
directly to the wells (25 μL/well) and shaken on a rotating platform
shaker at 150 rpm for 10 min. Plates were removed from the shaker
and equilibrated on the benchtop for 10 min before reading
luminescence (SpectraMax i3x, Molecular Devices).
Computational Methods. For the Cell Painting assay,
CellProfiler (version 3.0) was used to process the raw images and
obtain per-cell morphological profiles following the pipeline described
by Bray et al.3,25 For our analyses, we used 1753 morphological
features relating to size, texture, intensity, local density, and radial
distribution of nuclei, cytoplasm, and entire cells. Per-well compound
profiles were generated from these features averaged across all cells in
a particular well and then normalized by calculating robust z-scores
based on the population of individual vehicle-treated cells from the
same assay plate. All postprocessing numerical analyses were
performed in R (version 3.4.4). Calculations of Mahalanobis distances
(activity scores) and multidimensional perturbation values were
performed as outlined by Hutz et al.12 Principal component analysis
was performed using the stats package (version 3.4.4). Pearson
correlations were calculated from cLogP and activity scores of each
compound using GraphPad Prism (version 7.04). For each
compound, cLogP was calculated using SwissADME.24 For the cell
viability assay, luminescence signal was normalized to DMSO
controls, and dose curves were generated from the mean standard
error of the mean across three independent experiments in duplicate.
Area under the curve values were calculated, and statistical
significance was determined by t-test using GraphPad Prism (version
7.04). IC50 values of pyrones and reduced flavones were obtained by
GraphPad Prism (version 7.04) using log(inhibitor) vs response−
variable slope (four parameters) analysis with top and bottom
constraints at 100 and 0, respectively. All plots and graphs were
generated using GraphPad Prism (version 7.04).
structural and physical properties of small molecules that are
most predictive of biological performance diversity to create
efficient, diverse compound libraries for probe and drug
discovery programs.
METHODS
■
Compounds. Flavone-like compounds were synthesized as
described. Purchased compounds included apigenin (Frontier
Scientific #JK470737), quercetin (Sigma #Q4951), genistein (LC
Laboratories #G-6055), chrysin (ACROS #AC110320050), and
galangin (Sigma #282200). All compounds were suspended in
dimethyl sulfoxide (DMSO), and final vehicle concentration in all
experiments did not exceed 0.1%.
Cell Culture. Ovarian cancer cell line, OVCAR8, was a gift from
Joanna Burdette at University of Illinois at Chicago and cervical
cancer cell line, HeLa, was purchased from Sigma-Aldrich
(#93021013). OVCAR-8 and HeLa cell lines were cultured and
assayed in DMEM (Corning, #15-013-CV) with 8.8% FBS (Sigma-
Aldrich, #12306C), 3.6 mM L-glutamine (Corning, #25-005-CI), and
1× penicillin−streptomycin (Corning, 30-002-CI). Cultured cells
were maintained in a humidified incubator at 37 °C in 5% CO2.
Cell Painting Assay. This protocol closely followed Bray et al.,3
and Melillo et al.26 1500 HeLa cells were seeded in 50 μL of media
per well in 384-well clear bottom, black, tissue culture treated imaging
plates (PerkinElmer, #6057308). After incubating for 24 h at 37 °C,
compounds (6-point dose, 2-fold dilution, 80 to 5 μM) and control
(DMSO, Corning, 25-950-CQC) were pin-transferred into assay
plates using a 96-well pin tool (V&P Scientific) and a liquid handler,
Biomek NXP Lab Automation Workstation (Beckman Coulter).
Following transfer, cells were incubated for 24 h at 37 °C in 5% CO2.
All following aspirating and washing steps were performed using a
Biotek Multiflo FX and plates were protected from light as much as
possible. A 1 mM DMSO solution of MitoTracker Deep Red
(Thermo Fisher, #M22426) was prepared and diluted in prewarmed
media to a concentration of 500 nM. Then, 40 μL of media were
removed from each well of the assay plates, and MitoTracker solution
was added to each well (30 μL/well). After incubating for 30 min at
37 °C, aqueous paraformaldehyde (4.0% in PBS) (Electron
Microscopy Services, #15710, methanol-free) was added to each
well (10 μL/well) to fix the cells. After incubating for 20 min, cells
were washed once with 70 μL 1× HBSS (Corning, 21-023-CV), and a
0.1% HBSS solution of Triton X-100 (VWR, #0694-1L) was added to
each well (30 μL/well) to permeabilize the cells. After incubating for
20 min at room temperature, wells were washed twice with 70 μL 1×
HBSS. A 1 mg mL−1 dH2O solution of wheat germ agglutinin
(WGA), Alexa Fluor 555 conjugate (Thermo Fisher, #W32464); 1
mg mL−1 solution in 0.1 M aqueous NaHCO3 of Concanavalin A,
Alexa Fluor 488 conjugate (Thermo Fisher, #C11252); and a 1.5 mL/
vial MeOH solution of Phalloidin, Alexa Fluor 568 conjugate
(Thermo Fisher, #A12380) were combined with a 10 mg mL−1
aqueous solution of Hoechst 33342 (Invitrogen) and a 5 mM DMSO
solution of SYTO 14 Green Fluorescent Nucleic Acid Stain (Thermo
Fisher, #S7576) in 1× HBSS supplemented with 1% bovine serum
albumin to afford a staining solution of 1.5 μg/mL WGA, 100 μg/mL
concanavalin A, 5 μL/mL phalloidin, 5 μg/mL Hoechst 33342, and 3
μM SYTO 14. This solution was added to each well (30 μL/well),
and the cells were incubated at RT for 30 min. Wells were then
washed three times with 70 μL of 1× HBSS with no final aspiration,
and the plates were sealed with foil (Axygen, #HS-200). Images were
captured on an ImageXpress Micro XLS Widefield High-Content
Analysis System (Molecular Devices) using five filter cubes (Table
S2): DAPI (Hoechst), FITC (concanavalin A), TRITC (SYTO14),
Texas Red (phalloidin and WGA), and Deep Red (MitoTracker).
Photobleaching of low-intensity dyes is avoided by imaging in the
order of: MitoTracker, WGA, phalloidin, SYTO 14, Concanavalin A,
and Hoechst 33342. Nine sites were imaged per well in a 3 × 3 array
with laser-based autofocus on the first site per well.
PBS Stability. Enone (5a), alcohol (8a), and ester (10a) were
suspended in DMSO to make 20 mM solutions of each compound.
From the 20 mM solution of 5a, 40 μL was added to 0.5 mL of PBS
and 0.5 mL of MeOH in an LC/MS vial, affording a 0.8 mM final
concentration. From the 20 mM solution of 8a, 20 μL was added to 1
mL of PBS in an LC/MS vial, affording a 0.4 mM final concentration.
From the 20 mM solution of 10a, 10 μL was added to 1 mL of PBS in
an LC/MS vial, affording a 0.2 mM final concentration. Vials were
then heated to 37 °C or RT, and the percent remaining calculated
using an LC/MS (Agilent Technologies 1260 Infinity) by comparing
the percent conversion from product to major degradation product.
Samples were measured at 0, 0.5, 1, 1.5, 2, 3, 4, 6, 8, 18, 24, 36, 48, 60,
and 72 h. UV readings of the compound remaining over time at UV
254 and UV 280 were recorded and plotted in GraphPad Prism
(version 7.04).
General Synthetic Methods. All chemicals were purchased from
Sigma-Aldrich, Alfa-Aesar, Acros Organics, Matrix Scientific, TCI
America, or Oakwood Chemicals and were used without further
purification unless otherwise noted. Reaction progress was monitored
by TLC (Merck KGaA Silica gel 60 F254 glass backed plates) and LC/
MS (Agilent 1260 Series automated chromatographic system outfitted
with a Thermo Scientific Accucore column (2.1 × 50 mm, 2.6 μm
particle size)) utilizing a gradient elution mobile phase of 25%
ACN:H2O to 95% ACN:H2O over 2 min and then holding at 95%
ACN:H2O for 2 min (0.800 mL/min flow rate, 30 °C column
compartment, detection modes: wavelengths of 254 and 280 nm and
an Agilent 6120 quadrapole MS). NMR data were collected on a
Bruker AV 500 MHz spectrometer outfitted with a Bruker 5 mm
1H−19F/BBO S2 Z-gradient probe or a Bruker DPX 400 MHz
1
spectrometer outfitted with a Bruker 5 mm H−19F/BBO Z-gradient
Cell Viability Assay. OVCAR-8 cells or HeLa cells were seeded at
a density of 1000 cells per well in 50 μL of media per well in white
probe, and data were processed utilizing Mestrenova (Mestrelab
Research). Data were recorded at ambient temperature and are
H
DOI: 10.1021/acschembio.9b00294
ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology
Articles
reported as chemical shift (ppm) relative to solvent peak (1H NMR:
CDCl3 = 7.26 ppm, DMSO = 2.50 ppm; 13C NMR: CDCl3 = 77.16
ppm, DMSO = 39.52 ppm). Multiplicity is reported as follows: s =
singlet, d = doublet, t = triplet, q = quartet, bs = broad singlet, dd =
doublet of doublets, dt = doublet of triplets, td = triplet of doublets, tt
= triplet of triplets, ddd = doublet of doublet of doublets, ddt =
doublet of doublet of triplets, dtd = doublet of triplet of doublets,
quint = quintet, sext = sextet, hept = heptet. High resolution mass
spectrometry data were collected on a Waters Synapt G2-Si ESI MS, a
Micromass 70-VSE equipped with an EI/CI source, or a Waters Q-
TOF Ultima ESI and were collected at the University of Illinois at
Urbana−Champaign School of Chemical Sciences. Optical rotation
data were collected on a Rudolph Research Analytical AutoPol IV
Automatic Polarimeter outfitted with a TempTrolType 40T Polar-
imeter Sample Cell(5.0 × 50 mm). Chiral HPLC analysis was
performed on an Agilent 1100 series automated chromatographic
system outfitted with a ChiralPak IA-3 column (4.6 mm × 250 mm, 3
μm particle size) utilizing an isocratic mobile phase specific to each
compound. IR data were collected on a ThermoScientific Nicolet IS5
spectrometer outfitted with a ThermoFisher Scientific iD5 ATR.
Reaction mixtures were purified on a Biotage Isolera One automated
chromatography system outfitted with silica gel columns. Reversed-
phase high-performance liquid chromatography (HPLC) was
performed using Gilson model GX-241 Liquid Handler with Verity
4020 syringe pump, 159 UV−vis detector, and 321 solvent pump. A
Kinetex EVO C18 100 Å LC column (50 × 21.2 mm, 5 μm particle
size) was used, and the line was fitted with a Security Guard PREP
cartridge holder 21.2 mm with core−shell EVO C18 15 × 21.2 mm
cartridge and HPLC PREP column in-line filter (2.0 μm porosity,
21.2 mm). The mobile phase (acetonitrile:purified water) was
pumped at a flow rate of 15 mL/min for an initial hold of 1 min,
gradient for 4 min, and end hold for 1 min. Absorbance was measured
with a 159 UV−vis Gilson detector set at 320 nm to control fraction
collection. The fractions were concentrated using a Techne sample
concentrator.
REFERENCES
■
(1) Gerry, C. J., and Schreiber, S. L. (2018) Chemical probes and
drug leads from advances in synthetic planning and methodology.
Nat. Rev. Drug Discovery 17, 333−352.
(2) Pavlinov, I., Gerlach, E. M., and Aldrich, L. N. (2019) Next
generation diversity-oriented synthesis: a paradigm shift from
chemical diversity to biological diversity. Org. Biomol. Chem. 17,
1608−1623.
(3) Bray, M.-A., Singh, S., Han, H., Davis, C. T., Borgeson, B.,
Hartland, C., Kost-Alimova, M., Gustafsdottir, S. M., Gibson, C. C.,
and Carpenter, A. E. (2016) Cell Painting, a high-content image-
based assay for morphological profiling using multiplexed fluorescent
dyes. Nat. Protoc. 11, 1757−1774.
(4) Wawer, M. J., Li, K., Gustafsdottir, S. M., Ljosa, V., Bodycombe,
N. E., Marton, M. A., Sokolnicki, K. L., Bray, M.-A., Kemp, M. M.,
Winchester, E., Taylor, B., Grant, G. B., Hon, C. S.-Y., Duvall, J. R.,
̌
Wilson, J. A., Bittker, J. A., Dancík, V., Narayan, R., Subramanian, A.,
Winckler, W., Golub, T. R., Carpenter, A. E., Shamji, A. F., Schreiber,
S. L., and Clemons, P. A. (2014) Toward performance-diverse small-
molecule libraries for cell-based phenotypic screening using multi-
plexed high-dimensional profiling. Proc. Natl. Acad. Sci. U. S. A. 111,
10911−10916.
(5) Bates, S. E., Fojo, A. T., Weinstein, J. N., Myers, T. G., Alvarez,
M., Pauli, K. D., and Chabner, B. A. (1995) Molecular targets in the
National Cancer Institute drug screen. J. Cancer Res. Clin. Oncol. 121,
495−500.
(6) Schreiber, S. L., Shamji, A. F., Clemons, P. A., Hon, C., Koehler,
A. N., Munoz, B., Palmer, M., Stern, A. M., Wagner, B. K., Powers, S.,
Lowe, S. W., Guo, X., Krasnitz, A., Sawey, E. T., Sordella, R., Stein, L.,
Trotman, L. C., Califano, A., Dalla-Favera, R., Ferrando, A., Iavarone,
A., Pasqualucci, L., Silva, J., Stockwell, B. R., Hahn, W. C., Chin, L.,
DePinho, R. A., Boehm, J. S., Gopal, S., Huang, A., Root, D. E., Weir,
B. A., Gerhard, D. S., Zenklusen, J. C., Roth, M. G., White, M. A.,
Minna, J. D., MacMillan, J. B., and Posner, B. A. (2010) Towards
patient-based cancer therapeutics. Nat. Biotechnol. 28, 904−906.
(7) Woo, J. H., Shimoni, Y., Yang, W. S., Subramaniam, P., Iyer, A.,
ASSOCIATED CONTENT
* Supporting Information
■
́
Nicoletti, P., Rodríguez Martínez, M., Lopez, G., Mattioli, M.,
S
Realubit, R., Karan, C., Stockwell, B. R., Bansal, M., and Califano, A.
(2015) Elucidating Compound Mechanism of Action by Network
Perturbation Analysis. Cell 162, 441−451.
(8) Schulze, C. J., Bray, W. M., Woerhmann, M. H., Stuart, J., Lokey,
R. S., and Linington, R. G. (2013) Function-First” Lead Discovery:
Mode of Action Profiling of Natural Product Libraries Using Image-
Based Screening. Chem. Biol. 20, 285−295.
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acschem-
bio.9b00294.
Crystallographic information file for Compound 9i-2
(CIF)
(9) Wang, C.-Y., Wang, B.-G., Brauers, G., Guan, H.-S., Proksch, P.,
and Ebel, R. (2002) Microsphaerones A and B, Two Novel γ-Pyrone
Derivatives from the Sponge-Derived Fungus Microsphaeropsis sp. J.
Nat. Prod. 65, 772−775.
AUTHOR INFORMATION
Corresponding Author
*E-mail: aldrich@uic.edu.
ORCID
Melissa A. Korkmaz: 0000-0003-4234-8702
Leslie N. Aldrich: 0000-0001-8406-720X
■
(10) Stierle, A. A., Stierle, D. B., and Patacini, B. (2008) The
Berkeleyamides, Amides from the Acid Lake Fungus Penicillum
rubrum. J. Nat. Prod. 71, 856−860.
(11) Henrikson, J. C., Ellis, T. K., King, J. B., and Cichewicz, R. H.
(2011) Reappraising the Structures and Distribution of Metabolites
from Black Aspergilli Containing Uncommon 2-Benzyl-4H-pyran-4-
one and 2-Benzylpyridin-4(1H)-one Systems. J. Nat. Prod. 74, 1959−
1964.
(12) Wang, T.-Y., Li, Q., and Bi, K.-S. (2018) Bioactive flavonoids in
medicinal plants: Structure, activity and biological fate. Asian J. Pharm.
Sci. 13, 12−23.
(13) Raffa, D., Maggio, B., Raimondi, M. V., Plescia, F., and
Daidone, G. (2017) Recent discoveries of anticancer flavonoids. Eur.
J. Med. Chem. 142, 213−228.
(14) Shin, S. Y., Woo, Y., Hyun, J., Yong, Y., Koh, D., Lee, Y. H., and
Lim, Y. (2011) Relationship between the structures of flavonoids and
their NF-κB-dependent transcriptional activities. Bioorg. Med. Chem.
Lett. 21, 6036−6041.
(15) Ravishankar, D., Rajora, A. K., Greco, F., and Osborn, H. M. I.
(2013) Flavonoids as prospective compounds for anti-cancer therapy.
Int. J. Biochem. Cell Biol. 45, 2821−2831.
Author Contributions
‡E.M.G. and M.A.K. contributed equally.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
Funding for these studies was provided by the UIC
Department of Chemistry. The authors acknowledge S.
DiMagno (UIC) for helpful discussions and L. Anderson
(UIC) for helpful discussions and usage of analytical
equipment. The authors thank D. Wink (UIC) for providing
X-ray crystallography data and analysis and F. Song (UIUC)
and H. Yao (UIUC) for providing high-resolution mass
spectrometry data.
I
DOI: 10.1021/acschembio.9b00294
ACS Chem. Biol. XXXX, XXX, XXX−XXX

ACS Chemical Biology
Articles
(16) Spagnuolo, C., Moccia, S., and Russo, G. L. (2018) Anti-
inflammatory effects of flavonoids in neurodegenerative disorders.
Eur. J. Med. Chem. 153, 105−115.
(17) Sung, B., Chung, H. Y., and Kim, N. D. (2016) Role of
Apigenin in Cancer Prevention via the Induction of Apoptosis and
Autophagy. J. Cancer Prev. 21, 216−226.
(18) Clemons, P. A., Bodycombe, N. E., Carrinski, H. A., Wilson, J.
A., Shamji, A. F., Wagner, B. K., Koehler, A. N., and Schreiber, S. L.
(2010) Small molecules of different origins have distinct distributions
of structural complexity that correlate with protein-binding profiles.
Proc. Natl. Acad. Sci. U. S. A. 107, 18787−18792.
(19) McCombie, S. W., Metz, W. A., Nazareno, D., Shankar, B. B.,
and Tagat, J. (1991) Generation and in Situ Acylation of Enaminone
Anions: A Convenient Synthesis of 3-Carbethoxy-4(1H)-pyridinones
and −4-pyrones and Related Compounds. J. Org. Chem. 56, 4963−
4967.
(20) Groundwater, W., Hibbs, E., Hursthouse, B., and Nyerges, M.
(1997) Synthesis and reactions of reduced flavones. J. Chem. Soc.,
Perkin Trans. 1 1, 163−169.
(21) Seth, P. P., and Totah, N. I. (1999) On the Stereochemistry of
the Dihydropyrone Diels-Alder Reaction. Org. Lett. 1, 1411−1414.
(22) Corey, E. J., and Helal, C. J. (1998) Reduction of Carbonyl
Compounds with Chiral Oxazaborolidine Catalysts: A New Paradigm
for Enantioselective Catalysis and a Powerful New Synthetic Method.
Angew. Chem., Int. Ed. 37, 1986−2012.
(23) Hutz, J. E., Nelson, T., Wu, H., McAllister, G., Moutsatsos, I.,
Jaeger, S. A., Bandyopadhyay, S., Nigsch, F., Cornett, B., Jenkins, J. L.,
and Selinger, D. W. (2013) The Multidimensional Perturbation
Value: A Single Metric to Measure Similarity and Activity of
Treatments in High-Throughput Multidimensional Screens. J. Biomol.
Screening 18, 367−377.
(24) Daina, A., Michielin, O., and Zoete, V. (2017) Swiss ADME: a
free web tool to evaluate pharmacokinetics, drug-likeness and
medicinal chemistry friendliness of small molecules. Sci. Rep. 7, 42717.
(25) McQuin, C., Goodman, A., Chernyshev, V., Kamentsky, L.,
Cimini, B. A., Karhohs, K. W., Doan, M., Ding, L., Rafelski, S. M.,
Thirstrup, D., Wiegraebe, W., Singh, S., Becker, T., Caicedo, J. C., and
Carpenter, A. E. (2018) Cell Profiler 3.0: Next-Generation Image
Processing for Biology. PLOS Biol. 16, No. e2005970.
(26) Melillo, B., Zoller, J., Hua, B. K., Verho, O., Borghs, J. C.,
Nelson, S. D., Maetani, M., Wawer, M. J., Clemons, P. A., and
Schreiber, S. L. (2018) Synergistic Effects of Stereochemistry and
Appendages on the Performance Diversity of a Collection of Synthetic
Compounds. J. Am. Chem. Soc. 140, 11784−11790.
J
DOI: 10.1021/acschembio.9b00294
ACS Chem. Biol. XXXX, XXX, XXX−XXX