https://orcid.org/0000-0002-2115-3060
Abstract
Hydroxylated chalcones are phytochemicals which are biosynthetic precursors of flavonoids and their 1,3-diaryl-prop-2-en-1-one structure is used as a scaffold for drug development. In this study, the structure-dependent activation of aryl hydrocarbon receptor (AhR)-responsive CYP1A1, CYP1B1, and UGT1A1 genes was investigated in Caco2 colon cancer cells and in non-transformed young adult mouse colonocytes (YAMC) cells. The effects of a series of di- and trihydroxychalcones as AhR agonists was structure dependent with maximal induction of CYP1A1, CYP1B1, and UGT1A1 in Caco2 cells observed for compounds containing 2,2′-dihydroxy substituents and this included 2,2′-dihydroxy-, 2,2′,4′-trihydroxy-, and 2,2′,5′-trihydroxychalcones. In contrast, 2′,4,5′-, 2′3′,4′-, 2′,4,4′-trihydroxy, and 2′,3-, 2′,4-, 2′,4′-, and 2′,5-dihydroxychalcones exhibited low to non-detectable AhR activity in Caco2 cells. In addition, all of the hydroxychalcones exhibited minimal to non-detectable activity in YAMC cells, whereas 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) induced CYP1A1, CYP1B1, and UGT1A1 in Caco2 and YAMC cells. The activity of AhR-active chalcones was confirmed by determining their effects in AhR-deficient Caco2 cells. In addition, 2,2′-dihydroxychalcone induced CYP1A1 protein and formation of an AhR-DNA complex in an in vitro assay. Simulation and modeling studies of hydroxylated chalcones confirmed their interactions with the AhR ligand-binding domain and were consistent with their structure-dependent activity as AhR ligands. Thus, this study identifies hydroxylated chalcones as AhR agonists with potential for these phytochemicals to impact AhR-mediated colonic pathways.
1,3-Diaryl-prop-2-en-1-ones are chalcone natural products belonging to the flavonoid family and are precursors in the biosynthesis of flavones and isoflavones (Aoki et al., 2000; Kozlowski et al., 2007). Chalcones are also broadly classified as phenolics and part of a large family of structurally diverse phenolic phytochemicals many of which are used in traditional medicines and are available as nutraceuticals (Akihisa et al., 2006; de Mello et al., 2018; Mirossay et al., 2017; Nowakowska, 2007; Rani et al., 2019; Singh et al., 2014; Zhuang et al., 2017). Chalcones uniquely contain a conjugated enone function in the propane bridge between two aryl rings and there is evidence that this moiety plays an important role in the biological activity of chalcones (Singh et al., 2014). In addition, chalcones have been extensively used as a scaffold for the synthesis of structurally diverse analogs that have activity against multiple diseases including cancer (Rani et al., 2019; Singh et al., 2014; Zhuang et al., 2017).
Isoliquiritigenin (ISL) or 2′,4,4′-trihydroxychalcone is one of the bioactive components found in the roots of licorice plants and ISL has also been identified in multiple food products and is available as a nutraceutical (Peng et al., 2015). The pharmacological activity of ISL has been extensively investigated and ISL exhibits anticancer activities in multiple tumor types through modulating diverse pathways. ISL also protects against liver and cardiac injury, exhibits antidiabetic and antimicrobial activity, protects agonist neurological damage and modulates immune responses (Peng et al., 2015). ISL and other components of licorice extracts promote induction of Treg cells and attenuate many of the pro-inflammatory responses induced by dextran sodium sulfate (DSS) in a mouse model of inflammatory bowel disease (Guo et al., 2015; Wu et al., 2016). The molecular mechanisms associated with many of these responses are not well defined but, in some studies, intracellular targets have been identified. For example, ISL inhibits arachidonic acid formation from phospholipids and its metabolism into PGE2 in breast cancer cells and the proposed mechanisms involve direct interaction of ISL with multiple genes including phospholipase A2 and cyclooxygenase (Li et al., 2013). In addition, different chalcones including ISL, 2,2′,5′- and 2,2′,4′-trihydroxy-chalcone exhibit different neuroprotective effects (Jiwrajka et al., 2016; Lee et al., 2012; Zhu et al., 2010) and there is evidence that structurally diverse chalcones induce a similar pattern of responses and this has also been observed for flavonoid compounds (Kim et al., 2004; Kumar and Pandey, 2013; Ross and Kasum, 2002; Salaritabar et al., 2017).
Studies in this laboratory have investigated the structure-dependent activity of flavonoids as aryl hydrocarbon receptor (AhR) agonists and antagonists (Jin et al., 2018; Park et al., 2019,; Zhang et al., 2003) and there is some evidence linking the effects of flavonoids to their activities as AhR ligands (Allen et al., 2001; Xue et al., 2017); however, the AhR activity of chalcones has not been well defined. One study reported that chalcones including 2- and 4-hydroxychalcone, 2′,4,4′-trihydroxychalcone, and 2′,3,4,4′-tetrahydroxychalcone (buten) inhibited CYP1A1 and CYP1B1 activities (Wang et al., 2005). 2-Hydroxychalcone was particularly effective as an inhibitor of 7,12-dimethylbenz[a]anthracene (DMBA) induced CYP1A1 gene expression and both 2- and 2′-hydroxychalcone induced luciferase activity in MCF-7 cells transfected with an AhR-responsive construct (XRE-luc) containing a luciferase reporter gene (Wang et al., 2005). These results suggest that like flavones and isoflavones, chalcones may also be AhR ligands. Hence, in this study, we investigated the structure-dependent AhR activity of chalcones in colon cancer cells and show that only a subset of 2,2′-dihydroxy-substituted chalcones-activated AhR-responsive genes in colon cancer cells.
MATERIALS AND METHODS
Chemicals
The hydroxylated chalcones were purchased from Indofine Chemical Co (Hillsborough, New Jersey) and include 2,2′-dihydroxychalcone (98%), 2′,3-dihydroxychalcone (98%), 2′,4-dihydroxychalcone (98%), 2′,5′-dihydroxychalcone (98%), 2′,4,5′-trihydroxychalcone (97%), 2′,4,4′-trihydroxychalcone (97%), 2′,3′,4′-trihydroxychalcone (97%), 2,2′,5′-trihydroxychalcone (97%), and 2,2′,4′-trihydroxychalcone (97%). 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) (>98%) was synthesized in this laboratory.
Cell culture
The young adult mouse colonic (YAMC) cell line was described in our previous studies (Cheng et al., 2017; Jin et al., 2018; Park et al., 2019). YAMC cell line was maintained in RPMI 1640 medium with 5% fetal bovine serum, 5 units/ml mouse interferon-γ (IF005) (EMD Millipore, Massachusetts), 0.1% ITS (insulin, transferrin, selenium) (41400-045) (Life Technologies, Grand Island, New York) at 33°C and experiments were carried out at 37°C. Caco-2 human colon cancer cells were obtained from American Type Culture Collection (ATCC, Manassas, Virginia). Caco-2 cells were maintained in DMEM with nutrient mixture supplemented with 20% FBS and 10 m/l 100× MEM nonessential amino acid solution (Gibco) at 37°C in the presence of 5% CO2. Caco2-AhR knockout cells were obtained by CRISPR/Cas9 as described previously (Cheng et al., 2017; Jin et al., 2018; Park et al., 2019).
Quantitative real-time reverse transcriptase PCR
Total RNA was extracted from cells using an RNA isolation kit C according to the manufacturer’s protocol. cDNA synthesis was performed from the total RNA of cells using High Capacity RNA-to-cDNA Kit (Applied Biosystems, Foster City, California). Real-time PCR was carried out in triplicate using a Bio-Rad SYBR Green Universal premix for 1 min at 95°C for initial denaturing, followed by 40 cycles of 95°C for 15 s and 60°C for 1 min in the Bio-Rad iCycler (MyiQ2) real-time PCR system. Gene expression was analyzed using the comparative CT method and normalized to expression levels of TATA-binding protein (TBP). The sequences of the primers used for real-time PCR were as follows: CYP1A1 sense 5′-GAC CAC AAC CAC CAA GAA C-3′, antisense 5′-AGC GAA GAA TAG GGA TGA AG-3′; UGT1A1 sense 5′-GAA TCA ACT GCC TTC ACC AAA AT-3′, antisense 5′-AGA GAA AAC CAC AAT TCC ATG TTC T-3′; TBP sense 5′-GAT CAG AAC AAC AGC CTG CC-3′, antisense 5′-TTC TGA ATA GGC TGT GGG GT-3′.
Western blot analysis
Cells (3 × 105 cells/well) were plated in 6-well plates and grown for 24 h. Cells were treated with different concentrations of the compounds. Cell lysates were prepared in lysis buffer (10 mM Tris-HCl [pH 7.4], 150 mM NaCl, 1 mM EDTA, 1% triton-X-100, 0.5% Na deoxycholate, 0.1% SDS) with protease and phosphatase inhibitor cocktail (10 µl/ml each; Thermo, Rockford, Illinois). Protein concentrations from whole-cell lysates were quantified using the Bio-Rad protein assay (Bio-Rad Laboratories, Hercules, California). Aliquots of cellular proteins were electrophoresed on 10% SDS-polyacrylamide gel electrophoresis and transferred to a polyvinylidene difluoride membrane (Bio-Rad Laboratories). The membrane was allowed to react with a specific antibody, and detection of specific proteins was carried out by enhanced chemiluminescence. Loading differences were normalized using GAPDH antibody.
DNA binding assay
Episeeker DNA-protein-binding assay kit (Abcam) was used to measure the dioxin response element (DRE) binding of AhR by following the manufacturer’s protocol. A biotinylated 25-bp double-stranded oligonucleotides of the DRE promoter region (wild type; 5-GATCTGGCTCTTCTCACGCAACTCCG-3′ and mutant type; 5-GATCTGGCTCTTCTGTCATCACTCCG-3′) containing AhR-binding consensus sequence was used as a capture probe, and 25-bp double-stranded unlabeled oligonucleotide containing the identical consensus sequences was used as a competitor. Whole cell extract from Caco-2 treated with different doses of 2,2′-dihydroxychalcone was used in this experiment.
Statistics
All of the experiments were repeated a minimum of 3 times. The data are expressed as the means ± SD. Statistical significance was analyzed using either unpaired-Student’s t test (two-tailed) or analysis of variance test. A p-value of <.05 was considered statistically significant.
Computational methods
We computationally investigated the binding of AhR-active 2,2′,4′-trihydroxychalcone and AhR-inactive 2′,4,4′-trihydroxychalcone to the human AhR (hAhR) using molecular docking followed by molecular dynamics (MD) simulations. Both 2,2′,4′-trihydroxychalcone and 2′,4,4′-trihydroxychalcone were independently initially docked to hAhR using AutoDock Vina (Trott and Olson, 2009). The initial hAhR structure corresponded to the structure (residues 247–406) used in our previous study (Jin et al., 2018). Briefly, the hAhR structure was built using the crystal structure of hAhR in complex with a tetrazole-containing antagonist (PDB ID: 4XT2 [Scheuermann et al., 2015]), and residues missing from the crystal structure were modeled using I-TASSER (Yang et al., 2015). The initial structures of 2,2′,4′-trihydroxychalcone and 2′,4,4′-trihydroxychalcone were obtained from the PubChem compound database (www.pubchem.ncbi.nlm.nih.gov) (Wang et al., 2009). Prior to docking, the energy of the ligands were minimized using the MMFF94 force field and conjugate gradients algorithms in 200 steps via OpenBabel embedded in PyRx 0.8 (Wang et al., 2009). In the molecular docking runs, 21 residues located in hAhR were selected and made flexible in our effort to provide a more realistic ligand-protein interaction environment compared with rigid docking without an unmanageable increase in computer processing time (Abreu et al., 2012): His291, Lys292, Leu293, Phe295, Gly321, Tyr322, Ile325, Cys333, Ser336, His337, Met340, Ser346, Gly347, Met348, Phe351, Ser365, Asn366, Ala367, Ile379, Val381, and Gln383. A search space of 18 Å × 24 Å × 24 Å was established such that all possible binding site interactions pertaining to each ligand would be within the space. From the molecular docking runs, 9 docked structures of 2,2′,4′-trihydroxychalcone in complex with hAhR and 9 docked structures of 2′,4,4′-trihydroxychalcone in complex with hAhR were produced (for a total of 18 docked structures). These 18 structures were subsequently used as initial structures for 20 ns explicit solvent MD simulation studies with the aim to refine, optimize intermolecular interactions, as well as to determine the structural stability and evaluate the energetic favorability of the binding modes obtained from AutoDock Vina1, as described (Cheng et al., 2017; Jinet al., 2018). MD simulations in conjunction with energy calculations were used to assess the energetic favorability of different binding modes. The MD simulations were performed using CHARMM (Brooks et al., 2009) and CHARMM36 (Best et al., 2012) topology and parameters as described in our previous studies (Cheng et al., 2017; Jin et al., 2018). Topology and parameters for both ligands were generated from CGENFF (Vanommeslaeghe et al., 2010) without high penalty scores indicating fair parametrization. Upon completion of the MD simulations, the AutoDock Vina scoring function (Trott and Olson, 2009) was used to calculate the binding free energy throughout the 20 ns MD simulations in 20 ps intervals (corresponding to 1000 snapshots per simulation trajectory scored). The simulations of the 2,2′,4′-trihydroxychalcone-hAhR and 2′,4,4′-trihydroxychalcone-hAhR binding modes with the most favorable average binding free energy were considered to be the binding modes representing the most likely naturally occurring binding conformations of the ligands bound to hAhR. These simulations were subsequently extracted for further investigation. It is worth noting that for both ligands, independently, the second most energetically favorable binding modes was structurally similar to the lowest binding free energy simulation. Thus, further analysis was focused on only the lowest energy binding mode for each ligand. Also, notably, the two ligands were computationally predicted to bind with high-affinity to the protein (−8.4 ± 0.4 for 2,2′,4′-trihydroxychalcone and −8.5 ± 0.4 for 2′,4,4′-trihydroxychalcone). The extracted lowest binding free energy simulations of each ligand-hAhR complex were subsequently inspected with emphasis on comparing their binding as a means to provide insights into their structural differences which were also correlated with the experimentally observed activity differences.
RESULTS
Initial studies focused on a series of trihydroxychalcones and their induction of three AhR-responsive genes in Caco2 cells, namely CYP1A1, CYP1B1, and UGT1A1. TCDD (positive control), 2,2′,4′- and 2,2′,5′-trihydroxychalcones were potent inducers of CYP1A1 (Figure 1A) and the maximal induction responses for the chalcones were >50% of that observed for TCDD. In contrast, 4,2′,5′-, 2′,3′,4′-, and 2′,4,4′-trihydroxychalcones induced minimal CYP1A1 mRNA. A similar structure-dependent pattern of induction of CYP1B1 (Figure 1B) and UGT1A1 (Figure 1C) was observed for the trihydroxychalcones in Caco2 cells and for induction of UGT1A1 the magnitude of the responses observed for 1 and 5 µM 2,2′,4′- and 2,2′,5′- trihydroxychalcones was ≥100% of that observed for 10 nM TCDD. Thus, at least one important structural element for the AhR activity of trihydroxychalcones in Caco2 cells was the 2,2′-dihydroxy substitution pattern. We also investigated the effects of this set of compounds in non-transformed YAMC cells and the trihydroxychalcones did not significantly induce CYP1A1 (Figure 2A), CYP1B1 (Figure 2B), and UGT1A1 (Figure 2C). The lack of AhR activity by hydroxychalcones in YAMC cells differed from that reported for some flavonoids which did not induce CYP1A1 (Jin et al., 2018,Park et al., 2019) but induced UGT1A1 and CYP1B1 expression.
AhR responsiveness of trihydroxychalcones in Caco2 cells. Cells were treated with DMSO and 10 nM TCDD and different concentrations of chalcones for 18 h and effects on CYP1A1 (A), CYP1B1 (B), and UGT1A1 (C) gene expression were determined by real-time PCR as outlined in the Materials and Methods section. Results of gene expression studies (Figs. 1–5) are expressed as means ± SD for at least 3 separate experiments and significant (p < .05) induction is indicated.
AhR responsiveness of trihydroxychalcones in YAMC cells. Cells were treated with DMSO or 10 nM TCDD and different concentrations of chalcones for 18 h and effects on CYP1A1 (A), CYP1B1 (B), and UGT1A1 (C) gene expression were determined by real-time PCR as outlined in the Materials and Methods section.
The importance of the 2,2′-dihydroxy substitution pattern was further investigated using a series of dihydroxychalcone analogs. 2,2′-Dihydroxychalcone was a potent inducer of CYP1A1 (Figure 3A), CYP1B1 (Figure 3B), and UGT1A1 (Figure 3C) in Caco2 cells, and the magnitude of the induction was ≥100% of that observed for induction of these drug-metabolizing enzymes by 10 nM TCDD. 2′,5′-Dihydroxychalcone induced minimal (but significant) expression of CYP1A1, CYP1B1, and UGT1A1 mRNA. Some induction was also observed for 2′,3-dihydroxychalcone whereas minimal effects were observed for 2′,4- and 2′,4′-dihydroxychalcone. The dihydroxyflavones minimally induced CYP1A1 in YAMC cells (Figure 4A), most of these compounds induced CYP1B1 (Figure 4B) by 1.5–3.0 fold but were less active than TCDD (8-fold induction). The dihydroxychalcones did not induce UGT1A1 in YAMC cells (Figure 4C) and the major differences between tri- and dihydroxychalcones in YAMC cells was the minimal (but significant) induction of CYP1B1 by the latter group of compounds. Confirmation of the role of the AhR in mediating the induction responses by hydroxychalcones was investigated in wild-type Caco2 cells and Caco2-AhRKO cells where the AhR was silenced by CRISPR/Cas9 (Cheng et al., 2017). The AhR-active 2,2′-dihydroxy-, 2,2′,4′-, and 2,2′,5′-trihydroxychalcones induced CYP1A1 (Figure 5A), CYP1B1 (Figure 5B), and UGT1A1 (Figure 5C) in Caco2 cells and with few exceptions these induction responses were not observed in Caco2-AhRKO cells. In Caco2-AhRKO, the hydroxychalcones induced UGT1A1 (<2.5-fold) suggesting that induction of this gene was, in part, AhR independent.
AhR responsiveness of dihydroxychalcones in Caco2 cells. Cells were treated with DMSO or 10 nM TCDD and different concentrations of chalcones for 18 h and effects on CYP1A1 (A), CYP1B1 (B) and UGT1A1 (C) gene expression were determined by real time PCR as outlined in the Materials and Methods section.
AhR-responsiveness of dihydroxychalcones in YAMC cells. Cells were treated with DMSO or 10 nM TCDD and different concentrations of chalcones for 18 h and effects on CYP1A1 (A), CYP1B1 (B), and UGT1A1 (C) gene expression were determined by real-time PCR as outlined in the Materials and Methods section.
AhR responsiveness of 2,2′-dihydroxychalcone analogs in Caco2 cells expressing AhR and in Caco2-AhRKO cells where the AhR is silenced. Cells were treated with DMSO or 10 nM TCDD and different concentrations of chalcones for 18 h and effects on CYP1A1 (A), CYP1B1 (B), and UGT1A1 (C) gene expression were determined by real-time PCR as outlined in the Materials and Methods section.
The effect of 2,2′-dihydroxychalcone on ligand-induced binding of the AhR to a consensus DRE derived from the CYP1A1 gene was determined using an Episeeker DNA-protein binding assay (Figure 6A). Treatment with DMSO did not induce binding whereas TCDD bound the oligonucleotide containing wild-type (intact) but not mutant DRE and the binding was competitively decreased after incubation with excess (10-fold) free DNA which contained wild-type DRE. Similar results were observed after incubation with 0.1 or 1.0 µM 2,2′-dihydroxychalcone (Figure 6B), thus confirming that like TCDD, 2,2′-dihydroxychalcone induces AhR binding to an intact DRE but not to a mutant DRE. We also show that in Caco2 cells treated with 10 nM TCDD or 0.1–2.5 µM 2,2′-dihydroxychalcone, CYP1A1 protein expression was induced by both compounds. 2,2′-Dihydroxychalcone (≥0.5 µM) induced a response similar to that observed for 10 nM TCDD and this correlated with the gene expression results (Figure 3A).
2,2′-Dihydroxychalcone activation of AhR transformation and induction of CYP1A1 in Caco2 cells. A, Wild-type and mutant oligonucleotides derived from the human AhR (hAhR) promoter were incubated with whole cell lysates from Caco2 cells treated with TCDD (10 nM) and 2,2′-dihydroxychalcone (0.1 and 1 μM) and binding was determined in a colorimetric Episeeker DNA-protein assay as outlined in the Materials and Methods section. B, Caco2 cells were treated with DMSO (NC), 10 nM TCDD, or 2,2′-dihydroxychalcone for 24 h and whole cell lysates were analyzed by Western blots as outlined in the Materials and Methods section.
Because differences in the activity of AhR-active vs. AhR-inactive 2,2′,4′- and 2′,4,4′-trihydroxychalcones, respectively, may be related to their interactions within the AhR ligand-binding domain, we compared their modes of binding computationally using molecular docking followed by MD simulations as described (Jin et al., 2018; Park et al., 2019). Within the simulations of the lowest binding free energy modes, the two ligands share the same binding pocket, primarily comprising residues Thr289, His291, Phe295, Pro297, Leu315, Gly321, Tyr322, Phe324, Ile325, Cys333, Ser336, His337, Met340, Ser346, Met348, Phe351, Leu353, Ser365, Ala367, I379, Val381, and Gln383 (Figure 7). Interestingly, the two ligands also share nearly identical orientation in their complex structures (Figure 7). Both ligands form predominantly hydrophobic and aromatic interactions with protein residues, a portion of which His291, Phe295, Phe324, Cys333, Met348, and Phe351 are shown in Figure 7. In general, the hydrophobic and aromatic interactions with all residues in the binding pocket appear to contribute significantly to the stabilization of their binding, as well as binding with low predicted binding free energies. Although the carbonyl group of both ligands participates in a hydrogen bond with the side chain amide of Gln383 and the hydroxyl group of Ser365 (Figure 7), the main difference of the binding modes for the two ligands was associated with the fact that the hydroxyl groups of 2,2′,4′-trihydroxychalcone are frequently engaged in hydrogen bonds with surrounding residues, whereas this was not observed for the hydroxyl groups of 2′,4,4′-trihydroxychalcone. Specifically, the presence of the single hydroxyl group at position 2 in 2,2′,4′-trihydroxychalcone allowed the formation of a bifurcated hydrogen bond with the backbone carboxyl atoms of Gly321 and Phe295. The presence of the hydroxyl group at position 2 also correlated with the ability of the two other hydroxyl groups of 2,2′,4′-trihydroxychalcone to be engaged in hydrogen bonding with the side chain hydroxyl groups of Ser336 and Ser365 as well as in polar attractions (which can be viewed as longer distance hydrogen bonds) with the side chain amide group of Gln383 and side chain hydroxyl group of Ser346. It is worth noting that the polar attraction involving the side chain amide group of Gln383 was initially formed at approximately the last 5 ns of the simulation trajectory.
Molecular graphics images of representative snapshots extracted from the lowest free energy binding modes of the human AhR with the (A) inactive 2′,4,4′-trihydroxychalcone and (B) the active 2,2′,4′-trihydroxychalcone. In both panels, the ligand is shown in thick licorice representation. AhR is shown in transparent cyan new cartoon representation, and a portion of the interacting residues is shown in thin licorice representation. Hydrogen bonds (and polar attractions) are shown in dotted lines. Both compounds form hydrogen bonds or polar interactions with Ser365 and however 2,2′,4′- and 2′,4,4′- chalcone form two simultaneous hydrogen bonds and a single hydrogen bond, respectively, with Gln383. Comparable modeling analysis of the interactions of 2,2′,4′- and 2′,4,4′-chalcone with the mouse AhR gave similar interactions to those observed with human AhR (data not shown).
DISCUSSION
Some flavones and isoflavones are AhR-active compounds (Ciolino et al., 1998, 1999; Van der Heiden et al., 2009; Walle and Walle, 2002; Zhang et al., 2003) and recent studies show that they induce CYP1A1, CYP1B1, and UGT1A1 gene expression in colon-derived cell lines (Jin et al., 2018; Park et al., 2019). However, their effects are structure, response, and cell context dependent. Induction of CYP1A1, a well characterized AhR-responsive gene by flavones and flavanones was structure dependent and the order of activity was pentahydroxyflavones > hexahydroxyflavones > tetrahydroxyflavones in Caco2 cells (Jin et al., 2018). Quercetin was among the most active of these compounds over a range of doses (10–100 µM) and the maximum induction response (CYP1A1) was approximately >80% of that observed for 10 nM TCDD. The structure-activity induction of CYP1A1 gene expression in Caco2 cells by chalcones demonstrated that for the di- and trihydroxychalcone analogs maximal induction required the 2,2′-dihydroxyl groups and any changes in this substitution pattern resulted in a dramatic loss of activity (Figs. 1 and 3). Using 2,2′-dihydroxychalcone as a model we also demonstrated that this compound induced AhR interactions with a DRE-containing oligonucleotide using an Episeeker DNA–protein binding assay (Figure 6A). We also observed that 2,2′-dihydroxychalcone induced CYP1A1 protein in Caco2 cells (Figure 6B) and the most active hydroxylated chalcones (2,2′-, 2,2′,4′-, and 2,2′,5′-) did not induce CYP1A1 in Caco2-AhRKO cells in which the AhR was silenced by CRISP/Cas9 (Jin et al., 2018,Park et al., 2019) (Figure 5). Moreover, compared with TCDD the chalcones induced minimal to non-detectable CYP1A1 mRNA in YAMC cells (Figs. 2 and 4) and similar results were observed for the hydroxylated flavones (unpublished results). Thus, cell context was important for the activity of hydroxylated flavones and chalcones as CYP1A1 inducers. However, the latter compounds were more dependent on the 2,2′-dihydroxy substitution patterns, whereas both the substitution pattern and number of hydroxyl groups were required for induction of CYP1A1 by flavones (Jin et al., 2018; Park et al., 2019). Differences in chalcone-induced CYP1A1 in human Caco2 (high) and mouse YAMC (low) cells was not due to differences in the binding of the chalcones to the human vs. mouse AhR (mAhR). Additional simulations indicate that the interactions of 2,2′,4′- and 2′,4,4′- trihydroxychalcones with the mAhR were similar to those observed for the hAhR illustrated in Figure 7.
The AhR-active 2,2′-di and 2,2′,4′- and 2,2′,5′-trihydroxychalcones were also potent inducers of CYP1B1 and UGT1A1 in Caco2 cells whereas, induction of these genes in YAMC cells was minimal. In contrast, the tetrahydroxyflavones such as luteolin and apigenin which did not induce CYP1A1 in Caco2 cells and were partial antagonists of induction of CYP1A1 by TCDD induced CYP1B1 in Caco2 cells to levels similar to that observed for TCDD (Jin et al., 2018). In contrast, the AhR-active microbial metabolite 1,4-dihydroxy-2-naphthoic acid induced levels of CYP1A1 mRNA in both Caco2 and YAMC cells that were comparable to that observed for TCDD (Cheng et al., 2017). Thus, the different AhR agonist activities of hydroxylated chalcones, 1,4-dihydroxy-2-naphthoic acid, and hydroxylated flavones were variable in the human Caco2 and mouse YAMC cells and this is consistent with these compounds being selective AhR modulators (SAhRMs). This selectivity observed for chalcones and flavonoids is consistent with their AhR-dependent health promoting effects compared with TCDD and other toxic halogenated aromatics which not only induce CYP1A1 but also a pattern of toxic responses not observed for AhR-active phytochemicals.
We further investigated interactions of hydroxylated chalcones with the hAhR using a modeling approach as described previously (Cheng et al., 2017; Jin et al., 2018; Park et al., 2019) and compared the AhR binding of 2,2′,4′-trihydroxychalcone, a potent AhR agonist with the binding of the AhR-inactive 2′,4,4′-trihydroxychalcone. Our docking and simulation studies suggest that both 2,2′,4′-trihydroxychalcone and 2′,4,4′-trihydroxychalcone share nearly the same binding pocket and orientation upon binding, as well as similar binding free energies according to the AutoDock Vina scoring function (Trott and Olson, 2009). However, our computational studies suggest that the two ligands differ with regard to their overall hydrogen bonding or polar attraction capabilities. Our previous computational study (Jin et al., 2018) on AhR-active quercetin and apigenin suggests that both molecules have hydrogen bonding capabilities with Ser365 and Gly321, whereas only quercetin has hydrogen bonding capabilities with Gln383, Ser336, and Ser346. In this study, both the 2,2′,4′- and the 2′,4,4′-trihydroxychalcones participate in hydrogen bonding or polar interactions with Gln383 and Ser365, whereas the 2,2′,4′-trihydroxychalcone also interacts with Gly321, Phe295, Ser336, Ser346, and forms two simultaneous hydrogen bonds with Gln383. In contrast, 2′,4,4′-trihydroxychalcone forms a single hydrogen bond with Gln383. Thus, it is possible that some hydrogen bonds or combination of hydrogen bonds and polar attractions (ie, with a portion of residues Gly321, Ser365, Ser336, Ser346, and Gln383) could serve as “activation” keys that determine AhR activity of 2,2′,4′- and 2′,4,4′-trihydroxychalcones. Our previous study on TCDD binding to mAhR (mAhR) (Cheng et al., 2017) suggested that TCDD also forms polar interactions with Ser336/330 (residue numbers correspond to hAhR/mAhR numbering) and hydrogen bonds with Gln383/377, which could serve as an indication that polar attractions in general involving these residues and agonist molecules are important for signaling properties. Furthermore, a comparison between the three agonists (2,2′,4′-trihydroxychalcone, quercetin (Jin et al., 2018), and TCDD (Cheng et al., 2017)) could also indicate that hydrophobic interactions with at least a portion of residues Ile325/319, Cys333/327, Met348/342, and Phe351/345 are important for signaling. It should also be noted that these interactions for the hydroxylated chalcones are consistent with their structure-dependent induction of CYP1A1, CYP1B1, and UGT1A1 in Caco2 cells. However, these same interactions are not sufficient for activating expression of AhR-responsive genes in YAMC cells whereas TCDD induces expression of the AhR-responsive genes in both cell lines. These results demonstrate that hydroxylated chalcones are AhR ligands and maximal activities are observed for chalcones containing 2,2′-dihydroxy substituents. The structure-dependent differences in the AhR activities of chalcones and flavonoids in Caco2 and YAMC cells indicate that these phytochemicals are SAhRMs and this has been observed for other classes of AhR ligands (Murray et al., 2010a,b). Ongoing studies in the laboratory are investigating the colonic effects of AhR-active and AhR-inactive chalcones, such as ISL in mouse models which express wild-type and inactivated AhR to determine the role of the AhR in mediating chalcone-induced protective responses in the colon (Nowakowska, 2007; Wu et al., 2016).
FUNDING
Texas AgriLife Research; the Sid Kyle Chair Endowment; the Allen Endowed Chair in Nutrition and Chronic Disease Prevention; and the National Institutes of Health (R01-ES025713, R01-AT010282, R35-CA197707, and P30-ES023512).
DECLARATION OF CONFLICTING INTERESTS
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Comments