Shuwei Zhng, Christin Ohlnd, Christin Join,c, Shengmin Sng,*

a Laboratory for Functional Foods and Human Health, Center for Excellence in Post-Harvest Technologies, North Carolina Agricultural and Technical State University, North Carolina Research Campus, Kannapolis, North Carolina 28081, USA

b Department of Medicine, University of Florida, Gainesville, Florida 32611, USA

c Department of Infectious Diseases and Immunology, University of Florida, Gainesville, Florida 32611, USA

Keywords:

Thea flavin

Black tea

C-ring cleavage

Anaerobic incubation

Germ-free mice

A B S T R A C T

Increasing evidence has shown that gut microbiota plays important roles in metabolizing large molecular polyphenols to bioavailable and bioactive microbial metabolites.Thea flavin (TF) is one of the major color compounds in black tea and has demonstrated anti-inflammation, antioxidant, and anticancer effects properties.However, little is known about the metabolism of TF by gut microbiota in vivo.In this study, following the administration of TF to mice, the C-ring cleavage metabolites, dihydro- and tetrahydro-thea flavin (DH-TF and TH-TF) were detected in mouse feces by LC-MS and validated by authentic standards from in situ chemical reaction.The observation of the C-ring cleavage metabolites in TF-treated conventionalized mice but not in germ-free (GF) mice confirmed the role of gut microbiota in cleaving the C-rings of TF.The detection of DH-TF from the anaerobic incubation of TF with catechin-converting gut bacteria, Eggerthella lenta (Eggerth),suggested that the microbes with the capacity to cleave the C-ring of catechins were able to metabolize TF following the same mechanism.Additionally, three small phenolic metabolites were detected in mouse feces, and one of them was primarily detected in SPF mice not GF mice, which revealed that TF, subsequent to the cleaved C-ring, can be further metabolized into smaller phenolic metabolites by gut microbiota.Dose-dependent production of these metabolites were observed from the administration of 100 mg/kg to 400 mg/kg body weight of TF.In conclusion, gut microbiota can metabolize TF to the open-ring metabolites and the phenolic metabolites through the C-ring cleavage in mice.

1.Introduction

Thea flavins (TFs), along with thearubigins (TRs), are two major color components in black tea.These compounds are the condensed oxidant products of flavan-3-ols by polyphenol peroxidase and oxidase during fermentation of black tea from tea leaves [1].Due to their antioxidant properties, TFs have shown protective effects against inflammation, cardiovascular diseases, and certain type of cancers [2-4].Due to the inability of TFs to be absorbed into the circulatory system because of their large molecular sizes [1,5],majority of TFs remain in the intestinal tract.We have reported that TF mono-gallate and TF di-gallate were hydrolyzed by gut microbiota to release gallic acid and generate TF [6,7].While a significant number of studies have been conducted on TF biological activities, little is known about TF metabolism.A previous report suggested that TF’s skeletal structure is broken down to 3-(4’-hydroxyphenyl) propionic acid duringin vitrofeeding [5].However, it should be noted that impure TFs were used in this assay and that 3-(4’-hydroxyphenyl) propionic acid is an endogenous metabolite.Therefore, it is unclear if the increased amount of 3-(4’-hydroxyphenyl) propionic acid is produced directly from a broken-down of TF or from the agonist action after feeding TFs.A very recent publication reported that after incubation with human fecal microbiota, a number of hydroxylated phenylcarboxylic acids were formed with low concentrations from TF di-gallate [8].However, there is noin vivostudy to investigate the metabolism of TF by gut microbiota.

In this study, various concentrations of pure TF were administrated to mice to prevent any interference from other compounds that are generally present in black tea.The microbial metabolites were characterized through comparing LC/MS spectra with the authentic standards from in situ chemical reactions.The formation of specific microbial metabolites of TF was further confirmed by comparing profile in fecal samples collected from TF-treated germ-free (GF)mice and specific pathogen free (SPF) mice and by incubating TF with specific bacterium that can cleave catechins.

2.Materials and methods

2.1 Materials

TF from epicatechin and epigallocatechin and 3,4,6-trihydroxy-5H-benzocycohepten-5-one from catechol and pyrogallol were synthesized in our lab using the peroxidase and H2O2system according to our previous procedure [9].Palladium on carbon (Pa/C)and sodium formate were purchased from Sigma (St.Louis, MO).Other common chemicals and solvents were purchased from VWR International (Radnor, PA).LC-MS-grade solvents were obtained from Thermo Fisher Scientific (Waltham, MA).

2.2 Animal experiments

2.2.1 Animal experimental design of CD-1 mice

The animal experiment was conducted according to a protocol approved by the Institutional Animal Care and Use Committee of the North Carolina Research Campus (No.16-016).CD-1 mice were purchased from Charles River (Wilmington, MA) and acclimated for at least one week before being randomly assigned to different experimental groups.The mice were housed (5 mice/cage) and maintained in air-conditioned quarters with a room temperature of(20 ± 2) °C, relative humidity of (50 ± 10)%, and a light-dark cycle of 12 h:12 h (7 am to 7 pm).The mice were allowed free access to water and were fed a normal diet.

Two groups of mice (5 mice/group) were fed a purified diet, AIN-93G diet, for 3 days to wash out the potential phenolic metabolites from chow diet.After fasting overnight, group 1 (G1) was treated with vehicle (50% DMSO in water) as control, and group 2 (G2)was administered TF (25 mg/mL in 50% DMSO) by oral gavage at the dose of 100 mg/kg body weight.After 3 days wash out, G2 was administered TF (50 mg/mL in 50% DMSO) by oral gavage at a dose of 200 mg/kg body weight.This procedure was repeated for the administration of a third dose at 400 mg/kg body weight.The urine and stool samples were collected for 24 h after the administration of each dose and stored at -80 °C before analysis.

2.2.2 Germ-free vs.special-pathogen-free mouse experiment

The experiment with GF and SPF mice was carried out according to the protocol approved by the Institutional Animal Care and Use Committee at the University of Florida (IACUC# 201609606).GF wild type (WT) 129/SvEv mice (10–14-week old), were gavaged with SPF microbiota and transferred to regular housing for 2 weeks(conventionalized mice).Both GF mice (n= 7 with 3 female and 4 male mice) and the conventionalized mice (n= 8 with 5 male and 3 female mice) were individually housed in metabolic cages and orally gavaged with 200 mg/kg TF, and then the urine and stool samples were collected for 24 h and stored at -80 °C before analysis.

2.2.3 Preparation of mouse fecal samples

The fecal samples were dried by compressed nitrogen.After being refined, mouse fecal samples (100 mg) were homogenized in 80% aqueous methanol (1 000 mL) containing 0.1% formic acid and extracted by ultrasonication for 30 min.The samples were then centrifuged at 16 100 ×gfor 15 min at 4 °C, and the supernatants were picked and analyzed by LC-MS directly.

2.3 Biotransformation of TF by Eggerthella lenta (Eggerth)

Eggerth (ATCC25559 =E.lentaJCM 9979) [10]was purchased from the American Type Culture Collection (ATCC, Manassas, VA).Gifu Anaerobic Medium (GAM) broth was obtained from Fisher Scientific (Hampton, NH).All culture and biotransformation were performed under anaerobic condition in an Anaerobic Chamber(Sheldon manufacturing, Inc., Cornelius, OR) at 37 °C in darkness.The anaerobic condition was established by using a biological atmosphere mixture consisting of 5.0% carbon dioxide, 5.0%hydrogen, and balance nitrogen.The bacterium was inoculated in GAM broth (5 mL) in a Hungate tube for 2 days.The culture (1 mL)was transferred to fresh GAM broth (4 mL) and pre-cultured for 2 days.Then TF stock (10 μL; 15 mg/mL in DMSO) was added into the pre-culture.The culture without adding TF and the incubation of TF with GAM broth in the absence of ATCC25559 were used as the controls.After incubating for 72 h, the cultures were harvested and extracted with ethyl acetate three times.The extract was reconstituted in methanol (1 mL) for the LC/MS/MS analysis.

2.4 Chemical reaction to synthesize C-ring cleavage products of TF

Sodium formate (HCOONa, 0.5 mmol, 34 mg) and Pd/C(0.5 mmol, 53 mg) were added into the solution of TF (0.25 mmol,141 mg) in isopropanol (10 mL) and formic acid (60 μL).The reaction was initiated by heating the solution to 90 °C, and the mixture was stirred at 90 °C for 5 h.To stop the reaction, Pd/C was removed from the solution by adding methanol followed by centrifugation.The supernatant was collected.After dried by rotary evaporator,the products were reconstituted in methanol (final concentration of 0.1 mg/mL) for LC-MS/MS analysis.

2.5 LC-MS/MS analysis

LC-MS analysis was performed with a same instrument and same conditions as those in our previous paper [11].Brie fly, a Q-Exactive Plus Orbitrap tandem mass spectrometer was coupled to a Thermo Scientific Vanquish UHPLC system via electrospray ionization interface (Thermo Scientific, Waltham, MA).A Gemini C18 110? column (50 mm × 2.0 mm, 3 μm; Phenomenex, Torrance, CA) at a lf ow rate of 0.2 mL/min was used to separate the metabolites.Mass spectrometer was calibrated before detection and the source was tuned with TF.Selected ion monitoring (SIM) mode was used to search the target metabolites.

3.Results

3.1 The formation of C-ring cleavage metabolite dihydrothea flavin (DH-TF) in mice

SIM mode was used to search the formation of DH-TF, the cleavage product of one of the C-rings of TF.As seen in Fig.1A, there were two new peaks in fecal samples collected from TF-treated mice but not from un-treated, control mice.Both peaks had a molecular ion atm/z565.135 1 [M-H]-(calculatedm/z565.134 6 [M-H]-),with a molecular formula of C29H26O12, which was two protons more than TF.Based on their MS2spectrum (Fig.1B), the ion atm/z427.102 7 [M-H]-was from the direct loss of an A-ring, and the ion atm/z409.092 5 [M-H]-was from the loss of an A-ring and one water molecule.In addition, similar to TF, both peaks had the typical fragment ions of A-ring atm/z125.024 3 and 137.024 2 [M-H]-from the fragmentation of the A-ring, and the fragment ion of the benzotropolone moiety atm/z241.050 1 [M-H]-.Therefore, based on the proposed molecular formulas and the MS fragmentation data, we speculated that these two peaks identified in TF-treated mice were the C-ring cleavage metabolites of TF.

Fig.1 The ion chromatograms under SIM mode (A) and ESI-MS2 (negative ion) spectra (B) of the C-ring cleaved metabolites DH-TF from mouse feces,chemical reactions, and anaerobic incubation, and their chemical structures (C).(A) The ion chromatograms of DH-TF from fecal samples collected from TF-treated CD-1 mice (TF-CD1, A1), fecal samples collected from TF-treated GF mice (TF-GF, A2) and conventionalized mice (TF-SPF, A3), synthesized DH-TF standards (chemical reaction, A4), incubation of TF with the bacteria ATCC25559 (TF-ATCC25559, A5) for 72 h, incubation of TF in medium without ATCC25559 (TF-medium, A6), and blank CD-1 mouse feces (blank, A7).(B) ESI-MS2 (negative ion) spectra of the ion m/z 565.1364 at 10.6 min (B1, B3, B5, B7)and 10.9 min (B2, B4, B6, B8) from CD1 (B1, B2), SPF (B3, B4), Reaction (B5, B6), and ATCC25559 (B7, B8).(C) The structures of DH-TF.

Fig.1 (Continued)

To further confirm that these two peaks were products of cleaving one of the C-rings of TF instead of the reduction of the benzotropolone ketone of TF, the hydrogenation reaction of TF and 3,4,6-trihydroxy-5H-benzocyclohepten-5-one, the compound that has just the benzotropolone core structure, by Pd/C was undertaken.The LC-MS/MS results showed that the reaction of TF produced the targeted products with the same molecular ion,m/z565.135 1 [M-H]-,and the same MS2data with the peaks from TF-treated mice(Figs.1A-1B).However, 3,4,6-trihydroxy-5H-benzocyclohepten-5-one was unable to be hydrogenated under the same condition for TF(data not showed), suggesting the two products of TF are from the C-ring cleavage and not the reduction of the benzotropolone ketone(Fig.1C).All the above evidence confirmed that TF could be degraded through C-ring cleavage, in which a similar process occurred with green tea catechins by gut microbiota [10,12,13].

3.2 The role of gut microbiota on the formation of DH-TF in mice

To confirm that the mouse intestinal microbiota was responsible for the degradation of theaflavin by cleaving the C-ring, TF was administrated to GF and conventionalized mice at 200 mg/kg dose.The feces were collected for 24 h and subsequently analyzed by LC-MS.As shown in Fig.1A, two DH-TF peaks were detected in conventionalized mice but not in GF mice, suggesting the intestinal microbiota cleaved one of the C-rings on thea flavin to generate the two DH-TF metabolites.

3.3 The role of specific bacteria for the formation of DH-TF

The bacteriaE.lentaandAdlercreutzia equolifaciensfrom human and rat intestine, respectively, have been shown to cleave the C-ring of tea catechins to produce the C-ring cleavage metabolites,which undergo further degradation [10,12-14].Therefore, we hypothesized that the same bacteria could cleave the C-ring from TF when incubated together.To test this hypothesis, a commercially available strain ofE.lentawas used to incubate with TF.A targeted search of the two DH-TF metabolites from the incubation samples was employed using LC-MS method.The results showed that the bacterium could degrade a portion of TF into DH-TF after 6 h; while the maximum biotransformation of TF occurred after 72 h (data not shown).The LC-MS/MS data showed there were two major products with the targeted molecular weights atm/z565.135 1 [M-H]-(Fig.1A).The major ions displayed the same retention times and mass data with those fromin vivosample, but not from the controls including incubation of TF without bacterium and only bacterium without TF (Figs.1A-1B).This finding supports our hypothesis thatE.lentacan cleave the C-ring of TF, while under anaerobic conditions.

3.4 The formation of C-ring cleavage metabolite tetrahydrothea flavin (TH-TF) in mice

Since there were two C-rings present in the TF structure, a targeted search for double C-ring cleaved metabolite of TF (TH-TF) was undertaken for all samples using the LC-MS.In CD-1 mouse feces,one major peak with them/zat 567.150 8 was observed at 10.7 min,which showed the same retention time and MS2data with the peak from chemical reaction, as seen in Figs.2A and 2B.The MS2fragments atm/z429.120 4, 411.108 7, 137.024 4, and 125.024 3 were from the loss of the entire A-ring or the fragmentation of the A-ring,which was similar to DH-TF.Its structure was tentatively identified as shown in Fig.2C.However, no TF metabolites with both C-rings being cleaved were detected in the fecal samples from TF-treated conventionalized mice or from the incubation of TF withE.lenta.

Fig.2 The ion chromatograms under SIM mode (A) and ESI-MS2 (negative ion) spectra (B) of the C-ring cleaved metabolite TH-TF from mouse feces(TF-CD1, A1, B1) and chemical reaction, (A2, B2) and its chemical structure (C).

3.5 The formation of further degradation phenolic metabolites

After C-ring cleavage, catechins could be further degraded into small phenolic metabolites by gut microbiota [10,12-14].To investigate for this kind of TF degradation, a targeted search of TF-related phenolic metabolites was performed.Three of these metabolites in TF treated mouse feces were detected, including 5-(3’,4’-dihydroxyphenyl)-γ-valerolactone or 5-(3’,5’-dihydroxyphenyl)-γ-valerolactone (DHVL) withm/zat 207.066 3 [M-H]-, which had MS2fragments atm/z163.076 4 and 122.037 2 (Fig.3A); 5-(3’,4’-dihydroxyphenyl)valeric acid or 5-(3’,5’-dihydroxyphenyl)valeric acid (DHVA) withm/zat 209.081 9[M-H]-, which had MS2fragments atm/z191.071 0, 165.091 9, and 123.081 4 (Fig.3B); and 5-(3’,4’,5’-trihydroxyphenyl)valeric acid(THVA) withm/zat 225.076 8 [M-H]-, which had MS2fragments atm/z163.078 4, 123.045 1, and 101.024 2 (Fig.3C).All these data, as seen in Fig.3, were consistent with those reported in literature [8].

Fig.3 Ion chromatograms under SIM mode (1), ESI-MS2 (negative ion) spectra (2), and chemical structures (3) of small phenolic metabolites of thea flavin in mice.(A) DHVL, (B) DHVA, and (C) THVA.

Fig.3 (Continued)

3.6 Dose-dependent formation of DH-TF and TH-TF in mice

Thea flavin was adminstered at various doses to mice to determine the rate of which the gut microbiota was able to metabolize TF through the cleaving of its C-ring.The resulting metabolites, found in the feces, were then compared.

The LC-MS results, as seen in Fig.4A, revealed that the single C-ring cleaved metabolites, DH-TF, were clearly detected in all treated mouse fecal sample extracts, but not in the control samples.The total peak areas of both peaks showed a dose-dependent response from 100 mg/kg to 400 mg/kg.Similar to the single C-ring cleavage,the double C-ring cleaved metabolite, TH-TF, was also clearly detected in all treated mouse fecal extracts, but not in the control samples (Fig.4B).The peak areas of the double C-ring cleaved metabolite showed a dose-dependent response from 100 mg/kg to 400 mg/kg, albeit the peak areas of TH-TF were smaller than those of DH-TF.For the small phenolic metabolites of TF, the peak areas of DHVL and THVA showed a clear dose-dependent formation (Figs.4C and 4D).But no dose-dependent response was observed for the metabolite, DHVA.

Fig.4 Dose-dependent formation of C-ring cleaved metabolites of TF in mice.Ion chromatograms of (A) DH-TF, (B) TH-TF, (C) DHVL, (D) THVA under SIM mode and its peak areas in feces collected from TF treated CD-1 mice at (1) 0 mg/kg, (2) 100 mg/kg, (3) 200 mg/kg, and (4) 400 mg/kg.

Fig.4 (Continued)

4.Discussion and conclusions

Microbial metabolism plays important roles to help us understand the beneficial health effects of tea.Numerous studies have shown that tea catechins can be metabolized by gut microbiota through C-ring cleavage and further degradation to small molecular phenolic metabolites.Several bacterial strains have been identified to cleave the C-ring of catechins.Takagaki et al.[10]isolated a strain from rat feces, which was tentatively identified asA.equolifaciens, that could convert (–)-epigalocatechin (EGC) into its C-ring cleavage metabolite 1-(3,4,5-trihydroxyphenyl)-3-(2,4,6-trihydroxyphenyl)propan-2-ol.Kutschera et al.[13]isolatedE.lentarK3 from human fecal suspension, cleaved the heterocyclic C-ring of both (+)-epicatechin(EC) and (+)-catechin [12].Two otherE.lentastrains,Eggerthellasp.CAT-1 andEggerthellasp.SDG-2 [15], were also isolated from human feces and validated to cleave C-ring of catechins.Our study reports for the first time that TF, the dimer of EGC and EC, was able to be metabolized by gut microbiota to the corresponding C-ring cleaved metabolites in mice.The absence of the C-ring cleaved metabolites in TF-treated GF mice clearly demonstrated that they are microbial metabolites of TF.Using a commercially available strain ofE.lentawe demonstrated that the catechin-converting bacteria have the capacity to metabolize TF.

In this study, we observed that the single C-ring cleaved metabolites, DH-TF, are the major microbial metabolites of TF, which are five times higher than the double C-ring cleaved metabolite, TH-TF,and the peak area of THVA, the phenolic metabolite of TF, in 400 mg/kg dose was 5 times higher than DH-TF, and over 30 times higher than TH-TF.This may be due to the conversion of DH-TF to THTF is a slower reaction than the conversion of TH-TF to phenolic metabolites.A potential metabolic pathway of TF by gut microbiota is proposed in Fig.5.

Fig.5 Potential formation pathways of the degradation metabolites of thea flavin in mice by gut microbiota through C-ring cleavage.

The TH-TF was not detected in conventionalized 129/SvEv mice treated with TF, while detected in CD-1 mice.The difference in the cleavage of DH-TF to TH-TF may be due to genetic variation,microbiota composition and husdandry conditions since these experiments were performed in two different locations and mice backgrounds.Nonetheless, DH-TF was able to be detected in CD-1 mice, conventionalized 129/SvEv mice but not in GF 129/SvEv mice,which clearly suggest the role of gut microbiota on the cleavage of the C-ring of TF.

In conclusion, we found that TF is degraded through C-ring cleavage by gut microbiotain vivoand further into small phenolic metabolites (Fig.5).Further studies are needed to determine whether DH-TF and TH-TF are bioavailable and bioactive, how DH-TF and TH-TF especially the benzotrophlone core are degraded into the phenolic metabolites, and what bacteria are responsible for their degradation.

Conflict of interest

The authors declare no competing financial interest.

Acknowledgements

The authors wish to thank Mr.Hunter Snooks who assisted in the proofreading and editing of the manuscript.The authors gratefully acknowledge financial support from NIH R01 grant AT008623.