(76d) Gut Bacterial Product of Flavonoid Metabolism Exhibits Enhanced Aryl Hydrocarbon Receptor Agonist Activity in Caco2 Human Intestinal Epithelial Cells

Flavonoids, with over 6,000 unique structures occurring in nature [4] are the largest class of polyphenolic phytochemicals. Abundant in plants considered health-promoting, flavanoids’ health benefits are associated with their antioxidant and anti-inflammatory properties [1], the potency of which vary widely even among flavonoids of the same subclass. Further, there is no consensus regarding the molecular mechanisms underlying flavonoid bioactivity.

In the intestine, one potential molecular target is the aryl hydrocarbon receptor (AhR), a nuclear receptor capable of binding both endogenous metabolites and xenobiotic chemicals to regulate inflammatory pathways [3]. Recent studies from our lab and others have shown that flavonoids can exhibit agonist or antagonist activities to induce or inhibit AhR-responsive genes [5]. Some flavonoids, however, show minimal activity, which is consistent with findings that the immunomodulatory effects of flavonoids are structure dependent.

As dietary compounds that are only partially absorbed, flavonoids are available for metabolism by gut microbiota, leading to diverse products. In addition to hydrolyzing glycosidic bonds to release flavonoids from their glycoside forms, gut bacteria can perform a variety of reactions, e.g., ring fission, that cannot be formed through host metabolism alone. In this regard, exploiting the bioactivities of flavonoids requires a thorough understanding of their interactions with gut bacterial enzymes, resultant structural modifications, and the effects of these modifications on activation of host cellular receptors such as the AhR. The present study aims to computationally predict the microbial metabolites of dietary flavonoids, identify the enzymes that catalyze these transformations, and investigate their immunomodulatory effects as AhR agonists.

Based on the genomes of operational taxonomic units (OTUs) detected by 16S rRNA sequencing in anaerobic batch cultures of murine cecum contents, we assembled a model of metabolic reactions catalyzed by bacterial enzymes in these cultures. We curated the OTU table by eliminating species that are unlikely to be found in the murine intestine. The remaining species were searched against KEGG, UniProt and GenBank to retrieve their annotated genome sequences and match gene orthologs with corresponding enzyme commission (EC) numbers. Using a previously developed algorithm [6], the reactant-product pair data associated with the model’s EC numbers were used to learn a set of biotransformation operators for the murine cecum culture. Excluding reactions with molecular oxygen, more than 12,000 unique operators were derived. These operators were applied to 19 dietary flavonoids representative of the major subclasses that have been shown to exhibit anti-inflammatory activities. Our prediction results showed that unique microbial flavonoid metabolites were mostly a result of C-glycoside hydrolysis, de/methylation, de/hydroxylation and ring fission. The predictions regarding microbial metabolites overlapped with recently published in vitro experimental results. However, the enzymatic mechanisms have not been investigated in great detail. Our correlation analysis on predicted metabolites and corresponding microbial enzymes indicates that transferases, rather than oxidoreductases, likely catalyze the de/hydroxylation reactions in the presence of appropriate co-factors. Whereas the transferases were distributed widely across the bacterial families in our model, ring fission enzymes for specific flavonoid subclasses were found in only a few strains. One particularly interesting finding was the ring fission of naringenin to naringenin chalcone, which, according to reaction definitions in KEGG, requires an enzyme known to occur only in plants.

To validate the predictions, varying doses of the flavanone naringenin were added to the murine cecum contents culture and samples collected at different times were analyzed for the predicted metabolic products. Untargeted LC-MS experiments detected a dose- and time-dependent increase in one of the predicted metabolites, naringenin chalcone (Figure 1). Gene expression assays for induction of CYP1A1, CYP1B1, and UGT1A1 in human epithelial (Caco-2) cells showed that the chalcone product is an AhR ligand. Moreover, naringenin chalcone exhibited higher induction of UGT1A1 gene expression. Further structure activity relationship studies showed that other chalcone isomers, including 2,2’-dihydroxychalcone, 2,2’,4’-trihydroxychalcone and 2,2’,5’-trihydroxychalcone, also show significant AhR activity in Caco-2 cells (Figure 2).

In plants, flavanone ring fission is catalyzed by chalcone isomerase (CHI). However, none of the species in the murine cecum microbiota model have a gene for CHI (or a homolog). An alternative enzyme is chalcone synthase (CHS), which reacts 4-coumaroyl-CoA with malonyl-CoA to form naringenin chalcone. This enzyme has been shown to also bind naringenin as a substrate at an active site that is highly conserved across species [2]. To explore CHS catalysis as an alternative reaction for chalcone formation in the cecum culture, we created a phylogenic tree with all bacterial strains in the cecum culture model that have genes homologous to CHS from Medicago sativa (a legume expressing CHS with known naringenin binding activity) with a cutoff value of 80%. We found two species, Paenibacillus lactis and Bacillus subtilis, encoding CHS homologs with a high degree of similarity to M. sativa CHS. Enzyme docking simulations showed that CHS from M. sativa and P. lactis have similar binding scores (-7 and -4, respectively) when naringenin was used as the ligand, suggesting that bacterial CHS could catalyze ring fission of flavanones to form chalcone products that appear to be more potent AhR agonists than the parent compounds.

Taken together, these findings suggest that ring fission of flavanones can be catalyzed by microbial CHS and the resulting molecules can exhibit enhanced AhR agonist activity. In ongoing work, we are investigating if monocultures of Paenibacillus lactis and Bacillus subtilis can catalyze the ring fission of naringenin under anaerobic conditions. Prospectively, the prediction and validation tools developed in this work could facilitate the identification of bioactive metabolites of dietary flavonoids and their dependence on the enzymatic functions of the gut microbiota.

Figure 1. A schematic representation of the experimental design (a) Microbial metabolite predictions of selected flavonoid compounds. (b) In vitro validation of predictions using untargeted metabolomics (c) Phylogenic tree of the CHS sequences of Medicago sativa and modeled gut microbiome community members (cutoff >80%) (d) Enzyme docking simulation of CHS from Paenibacillus lactis with naringenin as the ligand

Figure 2. Induction of UGT1A1 expression by naringenin and naringenin chalcone in Caco-2 cells. Results are expressed as relative to UGT1A1 expression as a response to DMSO treatment. TCDD: 2,3,7,8-tetrachlorodibenzo-p-dioxin. DMSO: Dimethyl sulfoxide

1. 1. Chirag M Lakhani1 2, Arjun K Manrai1, 3, Jian Yang4, 5, Peter M Visscher#4, 5,*, and Chirag J Patel#1, 1Department, B.T.T., “乳鼠心肌提取 HHS Public Access,” Physiology & behavior, 176 (3), pp. 139–148 (2019).
2. 2. Ferrer, J.L., J.M. Jez, M.E. Bowman, R.A. Dixon, and J.P. Noel, “Structure of chalcone synthase and the molecular basis of plant polyketide biosynthesis,” Nature Structural Biology, 6 (8), pp. 775–784 (1999).
3. 3. Neavin, D.R., D. Liu, B. Ray, and R.M. Weinshilboum, “The role of the aryl hydrocarbon receptor (AHR) in immune and inflammatory diseases,” International Journal of Molecular Sciences, 19 (12), (2018).
4. 4. Panche, A.N., A.D. Diwan, and S.R. Chandra, “Flavonoids: An overview,” Journal of Nutritional Science, 5, (2016).
5. 5. Park, H., U. Jin, A.A. Orr, et al., “HHS Public Access,” 32 (11), pp. 2353–2364 (2020).
6. 6. Yousofshahi, M., S. Manteiga, C. Wu, K. Lee, and S. Hassoun, “PROXIMAL: A method for prediction of xenobiotic metabolism,” BMC Systems Biology, 9 (1), pp. 1–17 (2015).