Unusual Flavones from Primula macrocalyx as Inhibitors of OAT1 and OAT3 and as Antifungal Agents against Candida rugosa

A bioactivity guided program exploring the interaction of phytochemicals in the entire plant Primula macrocalyx with the organic anion transporters (OAT1 and OAT3) and microorganisms led to the elucidation of ten known flavones (1–4, 6–10, 12) and two previously undescribed flavones (5, 11). The structures of the compounds were determined by extensive analysis of spectroscopic data, as well as by comparison with data from previous reports. Two known flavones (9, 12) are reported for the first time from the family Primulaceae. All compounds were evaluated for inhibition of OAT1 and OAT3. Six flavones (2, 3, 6–8, 12) showed potent inhibitory activity on OAT1, while seven flavones (2, 3, 6–9, 12) showed marked inhibitory activity on OAT3, with IC50 ≤ 10.0 µM. Antimicrobial activities of crude fractions against sixteen microorganisms were tested to give a target yeast strain Candida rugosa for further evaluation of MICs on the isolates. Three flavones (7, 8, 12) showed marked antifungal activity with MIC < 2.0 µM. To our knowledge, this study is the first to evaluate these flavones as inhibitors of the OAT1 and OAT3, and as antifungal agents.

kidney disease in folk medicine 2 , making the interaction between OAT1/3 and P. macrocalyx an attractive target for further investigation.
Recent years have seen a resurgence of interest in antimicrobial agents from plants due to their ethnomedicinal uses and low toxicity and side effects. Particularly, developing countries rely on plants for the treatment of infectious and non-infectious diseases 16 . P. macrocalyx powder is in ethnomedicinal use for the treatment of tuberculosis 1 . Herein, we screened four fractions (n-hexane-soluble, dichloromethane-soluble, n-butanol-soluble and water-soluble) of the methanol extract of P. macrocalyx on sixteen kinds of microorganisms as part of an ongoing search for new antimicrobial chemotypes.
In our preliminary studies, the dichloromethane soluble fraction of a methanol extract of entire plant of P. macrocalyx elicited marked inhibition of OAT1 and OAT3 in vitro, and potent antifungal activity against yeast strain Candida rugosa.
In the present study, a bioactivity guided fractionation was performed on the methanol extract of P. macrocalyx collected in Armenia, followed by structure determination of the isolated compounds based on LC-MS and NMR, leading to the elucidation of twelve flavones (1)(2)(3)(4)(5)(6)(7)(8)(9)(10)(11)(12), including two previously undescribed compounds (5,11). To our knowledge, this study is the first to evaluate these flavones as inhibitors of the OAT1 and OAT3. These data may allow an initial elucidation of the structure activity relationships within this group, and may also provide a rational basis for the therapeutic applications of P. macrocalyx in traditional medicine. Additionally, the isolated antifungal agents could play a complementary role in the chemotherapy of fungal infections.
Compound 5 was obtained as a yellow amorphous powder. It showed two quasimolecular ions at m/z 269.0805 [M + H] + (calcd. for C 16 O 4 in the HRESIMS. It was ascribed as having a flavone skeleton 17,18 bearing methoxy and hydroxy substituents as shown by the 1 H and 13 C NMR spectroscopic analysis (Table 1). In the COSY spectrum (Fig. 2 . 2) with the carbon at δ C 150.8, and NOESY correlations ( Fig. 2) with δ H 6.93 (H-3) and δ H 7.08 (H-3′), establishing the location of this substituent group at C-2′. In addition, the downfield hydroxyl was located at C-5′ in the B-ring, supported by HMBC correlations between δ H 9.41 and C-4′ (δ C 119.3), C-6′ (δ C 115.0) and C-5′ (δ C 151.1), and NOESY correlations of δ H 9.41 with δ H 7.31 and δ H 6.96. If the hydroxyl group had been located at C-4′ rather than C-5′, H-3′ would be flanked by two oxygen containing substituents and expected farther upfield around δ H 6.7 PPM 19 . Additionally, this proton would be expected to show NOESY correlations with both the methoxy protons and the proton of the hydroxy group. This is clearly not the case.
All compounds (purity > 95%) isolated from the active fractions of P. macrocalyx were tested for their inhibitory activity on OAT1 and OAT3. When evaluated as inhibitors of OAT1 and OAT3 (Table 2), these flavones showed a concentration dependent inhibition of the transporters with IC 50 's ranging from 3 to 50 µM. The observed activities were consistent with the activity originally observed in the crude extract.
Sixteen microorganisms, including eight bacteria, Staphylococcus aureus, Staphylococcus epidermis, Streptococcus mutans, Pseudomonas fluorescens, Enterococcus hirae, Moraxella (Branhamella) catarrhalis, Pseudomonas aeruginosa, Bacillus subtilis subsp. Spizizenii, and eight fungi, Candida albicans, Aspergillus niger, Saccharomyces kudriavzevii, Penicillium chrysogenum, Candida parapsilosis, Candida rugosa, Candida tropicalis and Rhizopus stolonifer, were used to evaluate antimicrobial activities of the four crude fractions from P. macrocalyx, at a concentration of 1 mg/mL in DMSO, using an agar well diffusion assay. Results showed the dichloromethane fraction had marked antifungal activity against Candida rugosa with a 20 mm diameter zone of inhibition www.nature.com/scientificreports www.nature.com/scientificreports/ (positive control, amphotericin B: 11 mm). Bioactivity guided isolation was performed to isolate the active antifungal compounds. Minimal inhibitory concentration (MIC) determinations are summarized in Table 2.

Discussion
Twelve flavones (1-12) have been isolated and identified from the active dichloromethane soluble fraction of the entire plant of P. macrocalyx. Among them, 5 and 11 are newly described. Compounds 1-3, 6, 7, 10 are reported  www.nature.com/scientificreports www.nature.com/scientificreports/ for the first time from this species while 4 and 8 have been reported previously from this plant 2 , and 6 was previously reported from Primula veris 20 . Compounds 9 and 12 are herein reported for the first time from family Primulaceae. Additionally, 9 is described for the first time as a natural product, having been previously described as a product of chemical synthesis or biotransformation 26,27 . Previous reports and our present research on the phytochemistry of this species indicate that flavones are characteristic chemical constituents in P. macrocalyx. The unusual substitution patterns of the newly described constituents may prove useful in further chemotaxonomic studies of P. macrocalyx or the Primulaceae in general.
Evaluation of the isolated compounds as OAT inhibitors showed that six flavones (2, 3, 6-8, 12) showed good concentration-dependent inhibition on OAT1 mediated 6-CF uptake and seven flavones (2, 3, 6-9, 12) showed good concentration-dependent inhibition on 6-CF uptake mediated by OAT3 with IC 50 ≤ 10 µM. Among these compounds, 2′-methoxyflavone (3) was the most potent inhibitor of OAT1, with IC 50 value of 2.9 µM, while 3′-hydroxy-4′,5′-dimethoxyflavone (6) was of comparable potency with an IC 50 of 3.2 µM. Comparing 2 with 3 and 10, the inhibitory activity on OAT1 increased with the presence of methoxy group and decreased with the presence of hydroxyl group at the 2′ position in the B-ring. Additional B ring substitution appeared to reduce potency. Compound 12, 5,6,2′-trimethoxyflavone, was the most potent inhibitor of OAT3, with an IC 50 of 2.9 µM. The parent molecule, unsubstituted flavone (2) was also a potent inhibitor on OAT3, with an IC 50 of 4.6 µM. Comparing 2 with 3 and 10, the addition of a methoxy or hydroxy group at the 2′ position in the B-ring reduced the inhibitory activity on OAT3. Again, additional substitution on the B ring appeared to reduce potency.
Inhibitory activity on OAT1/3 of these flavones is highly dependent on the number and nature of the substituents attached to the flavone ring system, as well as the substitution pattern. In our previous study, 2′-hydroxy-6, 7,8-trimethoxyisoflavone and 7-methoxyflavone showed inhibitory activity against OAT1 with IC 50 values of 9.1 and 3.6 µM, respectively, and 7,2′-dihydroxy-6,8-dimethoxyisoflavone, 2′-hydroxy-6,7,8-trimethoxyisoflavone, 6,2′-dihydroxy-7,8-dimethoxyisoflavone, and 7-methoxyflavone showed moderate inhibitory activity against OAT3 with IC 50 values of 5.6, 8.7, 17.9, and 5.8 µM, respectively 28 . Morin and luteolin were reported to be potent OAT1 inhibitors (IC 50 values of <0.3 and 0.47 µM, respectively) in a para-aminohippuric acid (PAH) uptake assay in LLC-PK1 cells, while the tested flavonoid glycosides had weak inhibitory effects on OAT1 29 . Apigenin has been reported to inhibit OAT1 (IC 50 value of 0.737 µM) on the uptake of acyclovir in MDCK cells 30 . Wogonin has been reported to inhibit OAT1 (IC 50 value of 5.4 µM), and baicalein and wogonin have been reported to inhibit OAT3 (IC 50 values of 2.4 and 1.3 µM, respectively) in an assay similar to that used in our study 31 . In addition, flavone was identified by UPLC-ESI-MS/MS as the one main active component of Gnaphalium pennsylvanicum extract which showed activity in reducing serum uric acid through OAT1 32 . We have also recently reported the OAT3 inhibitory activity of the more complex dimeric biflavonoids, amentoflavone, cupressuflavone and podocarpusflavone A from Juniperus oblonga 33 , with IC 50 values of 2.0 μM, > 50 μM and 3.8 μM, respectively. Orally administered amentoflavone markedly altered the pharmacokinetic parameters of furosemide, a substrate of Oat3, in rats. Comparison of potency data among these studies is complicated by the use of the number of different bioassays commonly used in this area.
The unusual substitution patterns seen in compounds 1-12 may help to define the flavone chemotype as a pharmacophore for OAT inhibitors, providing the basis for future structure activity relationship studies. Identification of these active compounds may also help establish the chemical basis for use of this herb in the treatment of kidney disease.
Evaluation of these flavones as antifungal agents against C. rugosa showed that compounds 7, 8, and 12 possessed marked antifungal activity against this organism with MIC values of 0.4 µM, 1.2 µM, and 2.0 µM, respectively. These three flavones share the common structural features of methoxylation at positions 5 and 6 of the A ring and position 2′ of the B ring. While 11 which is inactive also shares these features, it is also methoxylated at C-8, suggesting that substitution at C-8 greatly reduces or eliminates activity. Although flavone (2) shows minimal activity against C. rugosa (MIC = 500 µM), none of the other compounds lacking the methoxylation at 5, 6, and 2′ showed any antifungal activity. Most known flavones are oxygenated at positions 5 and 7 on the A ring but are not noted for having antifungal activity, suggesting that the unusual 5,6-disubstitution pattern found in our isolates may be required for this activity. Further study is needed to define optimal substitution on the B ring and to explore the effect of substituents other than methoxy groups on the activity of these compounds. The pressing need for new antifungal chemotypes, especially for non-albicans Candida infections including the emerging "superbug", C. auris, highlights the importance of the discovery and potential development of a new pharmacophore such as this.
All solvents used were HPLC grade (Tianjin Concord Technologies; Tianjin Guangfu Technology Development Co., Ltd., Tianjin, China). 6-Carboxyfluorescein Extraction and isolation. Plant samples comprising the whole plant of P. macrocalyx were freed of extraneous matter, air dried in the shade and then ground to a coarse powder. A 1 kg portion of this sample was extracted three times with methanol (6 L, 1 day each) at room temperature to give the methanol extract (108 g) on removal of the solvent in vacuo.
The sample was dissolved in methanol and water (300 mL, 9:1 v/v), and extracted with n-hexane (300 mL × 3) to get an n-hexane fraction. The residual methanolic phase was freed of methanol in vacuo, suspended in water (300 mL), extracted with dichloromethane (300 mL × 3) followed by water-saturated n-butanol (300 mL × 3), to obtain a dichloromethane fraction, an n-butanol fraction and an aqueous fraction. The four fractions were each freed of solvents in vacuo and subjected to preliminary evaluation of inhibitory activity on OAT1 and OAT3 and antimicrobial activities against sixteen microorganisms. The dichloromethane fraction elicited strong inhibition of OAT1 and OAT3 and strong antifungal activity against C. rugosa, in vitro.
The active dichloromethane soluble fraction (3.5 g) was then fractionated by chromatography on silica gel Active fractions 4-5 from the silica gel column were pooled to form Fraction A, which consisted of one major peak (1) with a retention time of 7.45 min. Fraction A (216 mg) was purified by recrystallization using dichloromethane-methanol mixtures (7:3, 10 mL) three times to afford compound 1 (10 mg).
Active fraction 25 from the silica gel column afforded additional quantities of 8.
Uptake assay. The interactions of 12 flavones with the uptake of 6-CF in HEK-OAT1 and HEK-OAT3 cells were evaluated. This cell uptake assay was performed as previously described 14,35 . A density of 5 × 10 4 cells were seeded per well in 96-well culture plates precoated with poly-D-lysine. Approximately 85% confluency of cells was obtained after growing 24 h. The cells were washed twice and preincubated for 5 min with preheated (37 °C) uptake buffer (135 mM NaCl, 5 mM KCl, 2.5 mM CaCl 2 , 1.2 mM MgCl 2 , 0.8 mM MgSO 4 , 28 mM glucose, and 13 mM Hepes, pH 7.2) for the following uptake experiments. The uptake buffer containing 4 µM 6-CF in the presence or absence of test compounds and probenecid (a classic inhibitor of OAT1 and OAT3, used as a positive control) was incubated for 5 min to allow uptake. Uptake was terminated by adding 100 µL ice-cold uptake buffer, and quickly washing the cells in each well three times with ice-cold phosphate-buffered saline (PBS). The cells were lysed with 100 µL of 20 mM Tris-HCl containing 0.2% TritonX-100. A 50 µL aliquot of lysate was used to quantify fluorescence using a Tecan Infinite M200 plate reader with excitation and emission wavelengths at 485 and 528 nm, respectively. The protein content of the cell lysate was quantified using a BCA Protein Assay Kit. The intensity of fluorescence was standardized against total protein content, and measured in triplicate. The stock solutions of tested compounds were dissolved in DMSO with a final concentration of 50 mM and dilutions were made using uptake buffer. In our initial screening, the inhibition of the 12 compounds was evaluated at a concentration of 50 µM on OAT1 and OAT3. As shown, the majority of compounds showed marked inhibitory effects on OAT1 and OAT3. An inhibitor is defined as a compound that results in > 70% inhibition of 6-CF uptake. Concentration-dependent inhibition experiments were carried out on all selected inhibitors to determine IC 50 (50% inhibitory concentration) values on OAT1 and OAT3. statistical analysis. IC 50 values were summarized in Table 2 were estimated by non-linear regression analysis and expressed as mean ± standard error of mean. Statistical analysis was performed using GraphPad Prism version 7.0. For the uptake experiments, data were analyzed with a two-tailed unpaired Student's t-test.
Well diffusion antimicrobial assay. Antimicrobial activities of four crude fractions (n-hexane fraction, dichloromethane fraction, n-butanol fraction and aqueous fraction) against sixteen bacteria and fungi were performed based on the method reported previously 36,37 with slight modification using agar plates.
Briefly, nutrient ager (NA), brain heart infusion agar (BHIA) and potato dextrose agar (PDA), YM agar (YMA) were mixed in appropriate proportions with ultrapure water and autoclaved for 20 min at 121 °C. Approximately 15 mL aliquots of each medium were dispensed into petri plates and allowed to solidify, then incubated for 24 h to ensure sterility prior to use. Microbial suspensions (0.5 mL) at a density of 5 × 10 6 cfu/mL were inoculated onto the sterile media and allowed to stand for 10-20 min before wells were cut into the agar using a sterile pipet tip. A 10 µL aliquot of a DMSO solution of plant extracts (1 mg/mL) or fractions (2, 0.2, 0.02 mg/mL) were pipetted into each well. Similarly, 10 µL portions of DMSO and a solution of an appropriate antibiotic standard were applied to each plate as negative and positive controls, respectively. All treated plates were incubated for 24 h in incubators at appropriate temperatures. Zones of inhibition were measured in millimeters. The assays were performed in triplicate. Proper media, temperature and antibiotic standards for each microorganism are shown in Table 3.
Broth microdilution antifungal assay. Antifungal activity against C. rugosa (ATCC 10571) was measured as previously described using a broth microdilution method [36][37][38] with minor modifications. The assay was performed in sterile 96-well microtiter plates. Briefly, YM broth was prepared and autoclaved for 20 min at 121 °C, then allowed to cool to room temperature (25 °C). C. rugosa was dispersed in YM broth to an optical density of 1 UA at 600 nm (approximately 1 × 10 7 cfu/mL) using uninoculated broth as the blank. A 198 µL aliquot of the resulting fungal suspension was dispensed into each well of a 96-well plate. Stock solutions of the twelve www.nature.com/scientificreports www.nature.com/scientificreports/ compounds were prepared in DMSO at a concentration of 50 mM. The stock solutions were diluted to a range of final concentrations (500-0.14 µM) in YM broth such that the final DMSO concentration was no higher than 1%, v/v. A 2 µL aliquot of the resulting test samples was transferred to each well of the plate. Plates were incubated at 25 °C for 24 h prior to measurement of the optical density of each well at 600 nm. Minimal inhibitory concentrations (MICs), defined as the lowest concentration required to visibly inhibit the fungal growth compared to the untreated control, were measured in triplicate. Amphotericin B was used as positive control.

Data Availability
The datasets generated in this study are available from the corresponding author on reasonable request.  Table 3. Microorganisms, media, incubation temperature, and antibiotic controls.