Synthesis, biological evaluation, and NMR studies of 3-fluorinated derivatives of 3’,4’,5’-trihydroxyflavone and 3’,4’,5’-trimethoxyflavone (2025)

Published in final edited form as: Bioorg Med Chem Lett. 2020 Nov 28;32:127720. doi: 10.1016/j.bmcl.2020.127720

Introducing a fluorine atom into a biologically active molecule is a valuable approach in the process of drug discovery. This atom has a high electronegativity, a small size, and forms a strong bond with carbon, and these effects can improve the pharmacological and pharmacokinetic qualities of a lead structure.^1,2 Flavones are secondary metabolites found in many fruits and vegetables, and these molecules display many biological activities, including action as antioxidants, which makes them a valuable scaffold in medicinal chemistry.³ Synthetic methods for the construction of fluorinated flavones have appeared from Fuchigami,⁴ Dolbier,⁵ Rozen,⁶ and Zhang,⁷ and all of these reports have focused on the incorporation of a fluorine atom at the 3-position of flavone.⁸ This position is the most difficult to access on the flavone skeleton, because all of the other sites can be modified by standard halogenation methods for substituted benzenes.

Flavones display a three-ring structure of 2-phenyl-4H-chromen-2-one and are labelled with A-, C-, and B-rings. Limited structure–activity data for fluorination of flavones have appeared in the literature and are summarized in Figure 1.^9–11 The placement of fluorine at the 3-position of 3’,4’,7,8-tetrahydroxyflavone provides similar inhibition of telomerase activity compared to the non-fluorinated counterpart.⁹ Sale and coworkers created fluorinated derivatives at the 5-, 6-, and 7-positions of the flavonols, robinetin, myricetin, quercetin, kaempferol, and fisetin, and these modifications typically enhanced cytotoxicity against human prostate carcinoma (22Rv1) cells compared to the parent natural products.¹⁰ Flavonols are subclass of flavones that have a 3-hydroxy group. A synthetic derivative of the flavone, chrysin, with a 4’-fluorine displays a reduced rate of scavenging of reactive oxygen species of 55% compared to the natural product of 99%.¹¹ These data provided the basis for this study to investigate the effects of fluorination on the anti-oxidant and neuroprotective activities of flavones.

Our initial non-fluorinated targets were 3’,4’,5’-trimethoxyflavone (1) and 3’,4’,5’-trihydroxyflavone (2), because each acts as an antioxidant and serves as free radical scavenger in a DPPH assay.^12,13 Flavone 1 was synthesized according to the route in the literature as shown in Figure 2.¹² Briefly, 2-hydroxyacetophenone 3 and trimethoxybenzoyl chloride 4 were coupled in pyridine to give the ester 5 in 73% yield. Treatment of 5 with KOH in pyridine at 65 °C produced the 1,3-diketone 6 through a Baker-Venkataraman rearrangement. Cyclization of 6 by refluxing acetic acid and hydrochloric acid gave the 3’,4’,5’-trimethoxyflavone 1 in 71% yield.¹² In order to improve this transformation, we found that iron(III) chloride in CH₂Cl₂ at rt promotes the cyclization of 6 to 1 in a higher yield of 89%.¹⁴ Finally, complete demethylation of 1 was achieved using hydrobromic acid at 100 °C and gave the 3’,4’,5’-trihydroxyflavone 2 in quantitative conversion.¹⁵

The 3-fluoro-3’,4’,5’-trimethoxyflavone (7) is the 3-fluorinated counterpart of 1, and 7 has been synthesized¹⁰ but has not been tested for antioxidant activity. The 3-fluoro-3’,4’,5’-trihydroxyflavone (8) is not known in the literature, to our knowledge, and is the 3-fluorinated analogue of 2. Initially, fluoroflavone 7 was synthesized according to the route in the literature as shown in Figure 3.¹⁰ The fluorination of the 1,3-diketone 6 with NFSI in a solvent mixture of CH₂Cl₂ and CH₃CN is known to produce the intermediate 9 in a low yield (i.e., 37%) after seven days. After additional investigations, we discovered that using pyridine as solvent and heating the mixture to 50 °C not only increased the yield to 72% but also decreased the reaction time from seven days to 12 h. The compound 9 is then cyclized in refluxing acetic acid with sulfuric acid or iron(III) chloride at rt to give the 3-fluoro-3’,4’,5’-trimethoxyflavone 7.¹⁶ During the course of our synthetic studies, we discovered that treating 6 with two equivalents of NFSI at rt for 12 h followed by heating the mixture to 80 °C for 12 h resulted in formation of 7 in 61% isolated yield. This one-pot transformation is more efficient than the two-step conversion of 6 to 9 and then 9 to 7. The fluoroflavone 7 is demethylated using hydrobromic acid at 100 °C to produce in quantitative yield the 3-fluoro-3’,4’,5’-trihydroxyflavone 8.¹⁷ The synthetic routes allowed the production of sufficient amounts of compounds for biological evaluation as antioxidants and neuroprotective agents.

The DPPH assay is routinely used to characterize potential antioxidant activity, and the flavones 1, 2, 7, and 8 were tested in this assay to determine effects of radical scavenging (Table 1).¹⁸ Vitamin C (L-ascorbic acid) was used as the positive control, and a 1:1 solution of ethanol-acetone provided the optimal solubility for the flavones in assay. Serial dilutions of each of the test substrate were treated with DPPH (a stable free radical) across 20 minutes and then absorbance was measured. Six replicates were performed at each concentration, and EC₅₀ values were calculated with the standard error. Flavones 1 and 2 displayed EC₅₀ values (μg/mL) of 71 and 0.33, respectively, which are similar to reported values.^12,13 The fluoroflavones 7 and 8 displayed improved activity as radical scavengers with EC₅₀ values of 37 and 0.24, respectively. The results support that fluorination at the 3-position of the flavone scaffold may enhance antioxidant activity. These structure–activity data at the 3-position correlate well with the similar increases in potency in the telomerase assay observed by Menichincheri and coworkers.⁹ The trihydroxyflavone 2 and the 3-fluorinated trihydroxyflavone 8 display the most promising activity as antioxidants.

Flavones are currently investigated as source of potential neuroprotective agents, because new lead structures are needed for many neurodegenerative disorders.¹⁹ The flavones 1, 2, 7, and 8 were tested in oxidative glutamate toxicity assays with rat cortical neurons (Figure 4).^20,21 The neurons were treated with glutamate in the presence of a test substrate and compared to the controls, DMSO and MK801 (GluN receptor antagonist). Glutamate provides an excitotoxic insult to the neurons, and the cells are incubated for 24 h before the addition of the MTS reagent used to evaluate cell viability. The production of formazan dye by viable neurons is measured, and the data were compared the controls to determine statistical significance using a one-way ANOVA with post hoc Dunnett’s test (Figure 4A). The trihydroxyflavone 2 and the 3-fluorinated trihydroxyflavone 8 display neuroprotective effects similar to MK801, and these data also correspond to the antioxidant activity in the DPPH assay. As glutamate excitation induces calcium imbalances and reactive oxygen stress that leads to neuronal death, the ability of these compounds to reduce reactive-oxygen species (ROS) generation in the presence of glutamate stimulation was assessed (Figure 4B). Test substrate (i.e., 1, 2, 7, or 8) was compared to the vehicle-control treated (DMSO) and glutamate-stimulated neurons. Only the 3-fluorinated trihydroxyflavone 8 reduces levels of ROS below baseline in the rat cortical neurons. These data suggest a potential beneficial effect for 3-fluorination when compared to the non-fluorinated counterpart 1 which only displays activity in the MTS assay.

The scavenging of radical species by flavones is commonly attributed to the presence of hydroxyl groups on the ring B, which converts to an ortho-quinone during the process.²² Sawai and coworkers demonstrated that ¹³C NMR can observe the formation of the two carbonyl groups of this ortho-quinone intermediate upon the addition of DPPH to individual polyphenols present in tea leaves (Camellia sinesis).^{23, 24} Another report more recently recorded the formation of the ortho-quinone from phenolic acids treated with DPPH using ¹H NMR.²⁵ We investigated if an ortho-quinone intermediate could be observed from the fluoroflavone 8 following its quenching of free radicals. Accordingly, two equivalents of DPPH were added to a solution of fluoroflavone 8 in DMSO–d₆ and ¹³C NMR data were acquired after 20 min (Figure 5). The appearance of three new carbonyl signals at 197, 194, and 193 ppm suggest that ring B of the flavone has changed in the process. The data from a DEPTQ135 analysis confirmed the quaternary nature of these carbon atoms.²⁶ The carbonyl signal assigned to the 4-position of the flavone 8 (i.e., 169 ppm) also decreased in the ¹³C NMR, and it is important to note that none of the other peaks assigned to the parent flavone 8 disappeared following the addition of two equivalents of DPPH. In order to assign the new carbonyl peaks to the product, data was acquired by 1D selective HSQMBC NMR and verified that the carbonyl signals at 194 and 193 were on the same structure, which we assigned to the ortho-quinone.²⁶ Additionaly, the experiment was analyzed by mass spectrometry (i.e., ESI–TOF), and three peaks were observed in the negative ion mode at 285.020, 287.028, and 287.036.²⁶ The major peak at 287.036 is assigned to the starting material. The minor peak at 285.020 corresponds to the ortho-quinone, whereas the minor peak at 287.028 likely arises from a partially oxidized by-product. These data suggest that the ortho-quinone can form from 8, but there is not a complete conversion of the catechol into the quinone.

These synthetic and mechanistic studies demonstrate the utility of fluorination of the 3-position on the flavone backbone. Antioxidant activity may be enhanced from this structural change whereas as neuroprotective activity is conserved. An improved synthetic route for the installation of the 3-fluorine was developed to ensure production of sufficient quantities for biological evaluation. NMR analysis of the 3-fluorinated flavone in the presence of a free radical source suggests that quenching occurs by conversion to the ortho-quinone. Future work is focused on enhancing the sensitivity of these probes for conducting mechanistic investigations on cultured neurons following the application of excitotoxins and will be reported in due course.