Keywords

8.1 Introduction

Flavonoids are ubiquitously distributed in the plant kingdom and typically feature a 2-aryl 4H-chromen-4-one that could be expressed as a flavone skeleton [1]. They are frequently found in the form of a C-glycosylated derivative, where the anomeric position is directly connected to the C6 and/or C8 position(s) of a flavone skeleton [2]. Their unique structures and promising biological activities of these compounds make them attractive targets for synthetic chemists aiming to supply valuable, homogeneous samples for bioassays.

Vicenin-2 (1), a bis-C-glycoside of apigenin, was isolated from an annual plant native to Argentina, Urtica circularis [3], and other plant species [4,5,6,7,8] (Fig. 8.1). It has been reported to exhibit a broad spectrum of bioactivities, such as, anti-cancer, anti-inflammatory, antioxidant, and anti-diabetic effects [9,10,11,12,13].

Fig. 8.1
A chart presents the chemical structures of apigenin, vitexin, and vicenin-2. Each is a trihydroxyflavone.

Structures of vicenin-2 and related compounds

Full size image

This compound shares the bis-C-glycosylated flavone skeleton, of which the benzene core (A-ring) is fully substituted by carbon or oxygen atoms.

8.2 Strategic Analysis

8.2.1 Previous Studies

Among many studies focusing on the synthesis of C-glycosides, most have concentrated on the assembly of mono-C-glycosides. As a facile approach to a C-glycosyl flavonoid, the Friedel–Crafts reaction of a glycosyl donor with a phenol derivatives was investigated well [14]. Nevertheless, the synthesis of vicenin-2 and other bis-C-glycosyl flavonoids remains have been underexplored to date. Sato and co-workers explored scandium triflate-promoted bis-C-glycosylation of 2,4,6-trihydroxyacetophenone [15, 16] or naringenin [17] with unprotected monosaccharides, directly yielding the corresponding bis-C-glycosides, albeit in low yields and requiring tedious purification procedures. Furthermore, a late-stage modification by Shie [18], involving tandem C-glycosylation of flavan derivatives, necessitates multiple manipulation steps to complete the synthesis.

8.2.2 Fluoroarene Strategy

In order to achieve a high-yield C-glycoside formation, we focused on a two-step conversion protocol originally developed by Kraus and Molina, converting from a glucono lactone with an aryl anion (Fig. 8.2) [19].

Fig. 8.2
A chemical reaction for aryl C glycoside formation. oxo tetrahydropyran reacts with alkoxy benzene anion. A Lewis acid facilitates the transformation of this anion into the final product.

Step-wise aryl C-glycoside formation strategy by Kraus and Molina

Full size image

The method performs the nucleophilic addition of an aryl anion to the lactone ring followed by the reductive deoxygenation of the resulting lactol, readily forming an aryl C-glycoside structure. In 1990, Tatsuta applied this method to the synthesis of medermycin and successfully achieved its total synthesis [20].

The question here was whether this method could be applied repeatedly, in particular, for constructing a densely functionalized and sterically hindered multi-substituted benzene unit.

To realize this approach, we came up with an idea to use 1,3,5-trifluorobenzene (2) as a synthetic platform. An alternate latent polarity pattern on the benzene ring (Fig. 8.3) could be considered based on two key reactivities induced by fluorine atoms on a benzene ring (Fig. 8.4): (1) nucleophilic aromatic substitution (SNAr) by oxygen atom nucleophiles, i.e., alkoxides [21] and (2) electrophilic substitution via lithiation followed by alkylation, where a fluorine atom substituted at the ortho position of the reaction site acts as a strong directing group for ortho-metallation [22,23,24,25]. In particular, the pKa value of 2 is ca. 31.5 [26], which is very low compared to common benzene derivatives, therefore allowing for deprotonation easily. We envisioned that combining these methods would enable a facile and regioselective access to multi-functionalized benzene skeleton [27].

Fig. 8.3
Chemical structure for 1, 3, 5 trifluorobenzene. The molecule consists of a benzene ring with three fluorine atoms attached at positions 1, 3, and 5. Two resonance structures are displayed, each with positive and negative charges on different carbons in the benzene ring.

Alternate latent polarity pattern of 1,3,5-trifluorobenzene

Full size image
Fig. 8.4
A chemical reaction illustrating the dual reactive nature of fluoroarenes. E S, represented on the left, involves an electron-deficient molecule attacking the fluoroarene ring to form a new compound. S N A r is displayed on the right and features a nucleophile displacing a leaving group X to create a new molecule.

Two characteristic reactivities of a fluoroarene

Full size image

An illustrative demonstration of the viability of this synthetic approach is evident in the subsequent experimental findings. In this study, we investigated the stepwise conversion of 1,3,5-trifluorobenzene (2) into the hexa-substituted benzene derivative 4 (Fig. 8.5) [28]. The synthesis initiated with SNAr reaction, replacing one of three fluorine atom. Subsequently, a regioselective ortho-lithiation/sulfur-atom incorporation protocol was executed, yielding 3. The second SNAr reaction with a nitrogen nucleophile, followed by the step-wise halogenation resulted in the formation of 4, Notably, this represents the first instance of synthesizing hexa-substituted benzene derivative with six distinct heteroatom substituents.

Fig. 8.5
A chemical reaction. It displays a step-by-step chemical reaction transformation of a trifluorobenzene molecule with three fluorine substituents into a benzene molecule decorated with six distinct functional groups containing various heteroatoms.

Step-wise conversion of 1,3,5-trifluorobenzne to the hexa-substituted benzene derivative, bearing six different hetero-atom substituents

Full size image

8.3 Retrosynthesis

Figure 8.6 delineates our synthetic strategy en route to vicenin-2. The C-ring pyrone unit in 1 could be constructed at a late stage by conducting an intramolecular oxidative oxa-Michael addition of the phenol in I to the enone moiety. The three oxygen substituents on the benzene ring (A-ring) would be incorporated by substituting fluorine atoms with oxygen-atom nucleophiles via SNAr reaction. Subsequently, two C-glucosyl units and the coumaroyl unit would be introduced electrophilically via the iterative reaction with aryl anions, generated via ortho-metalation, with the corresponding carbonyl derivatives II and III.

Fig. 8.6
A reaction. It begins with oxa-Michael addition and oxidation of the enone moiety to form 1 and further processing to form trifluoro benzene. Further addition of O H groups results in carbonyl derivatives 2 and 3.

Our synthetic plan

Full size image

8.4 First C-glycosylation Stage

According to the strategy, our synthesis started with the synthesis of bis-C-glycosylated benzene core. Scheme 8.1 shows the synthesis of the mono-C-glycoside 7 from 1,3,5-trifluorobenzene (2). Upon treatment of 2 with n-BuLi in Et2O (– 78 °C, 1 h), the ortho-lithiation proceeded smoothly, and the nucleophilic attack of resulting aryl lithium species to lactone 5 [29] gave the corresponding lactol 6 quantitatively. Subsequent treatment of 6 with Et3SiH in the presence of BF3⋅OEt2 [30, 31] afforded mono-C-glycoside 7 in high yield with co-production of the corresponding α anomer (10% yield), which could be separated by silica gel column chromatography. It is worth mentioning that when 1,3,5-trimethoxy benzene was employed as an aryl anion precursor instead of 1,3,5-trifluoro benzene, no formation of the desired C-glycosidic product was observed, and a sizable amount of an unexpected α,β-unsaturated lactone 9 was obtained by β-elimination of benzyl alcohol from 5. This result implies a steric repulsion between the nucleophilic species and the lactone ring in 5.

Scheme 8.1
A chemical reaction for synthesizing a mono C glycoside molecule. The step-by-step process starts with labeled materials and progresses through additions of various reagents like F, n B u L i, B F 3 O E t 2, and E t 3 S i H to reach the final product. Yields are presented for each stage of the reaction.

Assembly of mono-C-glycoside

Full size image

8.5 Second C-glycosylation Stage

Next, we addressed the introduction of the second sugar unit. However, the second deprotonation reaction proved more challenging than the first, presumably due to steric hindrance around the reaction site caused by the buttressing effect of the sugar moiety. The ortho-lithiation of C-glycosyl trifluorobenzene 7 using n-BuLi, followed by the reaction with gluconolactone 5, resulted in only an 18% yield of lactol 10 with a sizable recovery of 6 (81%). Note that the reaction proceeded with partial epimerization at the C2 position.

Scheme 8.2
A chemical reaction of bis C glycosylation. Starting material compound 5 reacts with n B u L i and a glycosyl donor to form intermediate 7, which then reacts with another glycosyl donor to yield 18% of the desired product compound 10 and 81% of byproduct compound 8.

Initial attempt to bis-C-glycosylation

Full size image

In pursuit of more suitable deprotonation conditions, we carried out the deuterium incorporation experiment using mono-C-glycosyl trifluoro benzene 7 (Scheme 8.3). Upon treatment of 7 with n-BuLi (1 equiv) at – 78 °C in Et2O (1 h), methanol-d1 was added to incorporate a deuterium into the benzene ring. The rate of deuterium incorporation (D%) was assessed by 1H-NMR to be 34% with 95% yield of 7-d1. When N, N, N′, N′-tetramethylethylene diamine (TMEDA) was used as an additive (1 equiv), the value of D% was increased to 50%. Using THF instead of Et2O as a solvent, it showed no any noticeable effect. In the end, we found that the use of a more strong base, i.e., t-BuLi, led to the optimal incorporation of a deuterium (84%) into 7 with a high yield.

Scheme 8.3
A chemical reaction for the ortho-lithiation conditions. A process introducing a lithium atom to a specific location on an aromatic ring on the top. On the bottom, it has a table with five columns and four rows. The column headers are as follows. Run, R L i, conditions, D percent, combined yield of 7 slash 7 minus d 1 slash percentage.

Optimization for the ortho-lithiation conditions

Full size image

With successfully finding the optimal conditions, we re-examined the reaction of 7 and lactone 5 (Scheme 8.4). The reaction via the ortho-lithiation (Et2O, – 78 °C, 1 h) followed by the nucleophilic attacking to lactone 5 smoothly proceeded to give 10 in 83% yield. Subsequently, the Lewis-acid promoted reduction of lactol 10 yielded C2-symmetrical bis-β-C-glycoside 11 in high yield (82%).

Scheme 8.4
A chemical reaction. It displays the step-by-step process for creating a bis C glycoside molecule. It features the initial sugar molecules, necessary reagents like t B u L i and B F 3 O E t 2, and the final chemical structure of the bis-C-glycoside product, including the 83% reaction yield.

Assembly of bis-C-glycoside

Full size image

8.6 Introduction of the Side Chain Unit

Next, we investigated the introduction of the side chain, constructing the fully substituted (hexa-substituted) benzene ring. To evaluate the reactivity, we examined model reactions of bis-C-glycoside 11 with various electrophiles (Fig. 8.7). First, the deuterium incorporation experiment was conducted. Treatment of 11 with t-BuLi (– 78 °C, 1 h) followed by adding methanol-d1 resulted in a complete incorporation of deuterium. As a comparison experiment, we also attempted the reaction with ortho-dimethoxybenzene derivative 13, resulting in a poor incorporation of deuterium (15%). Comparing these two results proves that fluorobenzene is particularly effective in generating the corresponding aryl lithium species. We next attempted the reaction of the bis-C-glycosyl trifluorobenzene derivative 11 with various carbon electrophilic units. The use of N,N-dimethylformamide (DMF) as an electrophile gave the corresponding aldehyde 12 (R = CHO) in 80% yield, while the reaction with N,N-dimethylacetamide (DMA) led to no reaction. Using acetic anhydride (Ac2O) also showed poor reactivity, recovering 11. In contrast, ethyl acetate as an electrophile gave the acetylated product 12 (R = Ac) in 31% albeit with 50% recovery of the starting material 11. These results suggested that the competitive deprotonation inevitably occurred at a α-position of the carbonyl group in the electrophile or the C-acylated product.

Fig. 8.7
A chemical reaction between a bis C glycoside and an electrophile. The experiment explores product formation at negative 78 degrees Celsius, depending on the electrophile used. It has a table with seven columns and five rows. The column headers are as follows. Run, electrophile, T, time, R, yield %, and recovery %.

Reaction of bis-C-glycoside 11 via ortho-lithiation

Full size image

Based on these results, we then attempted to react with the α,β-unsaturated amide with the carbon units necessary for the total synthesis of 1 (Scheme 8.5). Pleasingly, generation of the lithiated species from 11 followed by in situ trapping with Weinreb amide 14 [32, 33], which has no acidic α-proton of the carbonyl group, led to the acylated product 15 in 80% yield along with a partial recovery of 11 (15%). This reaction serves as an effective approach for building chalcone skeletons that were once accessed primarily by Claisen–Schmidt condensation between benzaldehyde derivatives and acetophenones [34].

Scheme 8.5
A chemical reaction of the chalcone unit via ortholithiation displays the formation of a chalcone unit from a benzene ring with bromine substituents, a methoxy group, and a fluorine atom. The reaction occurs at negative 50 degrees Celsius and negative 78 degrees Celsius.

Formation of the chalcone unit via ortho-lithiation

Full size image

8.7 Substitution of Fluorine Atoms to Oxygen Atoms on the Benzene Ring

Having prepared a key synthetic intermediate 15 possessing all the carbon chains necessary for the synthesis of the target natural product, we next investigated SNAr reactions to replace three fluorine atoms into three oxy-functions (Scheme 8.6). One of the possible ways to achieve this goal is to utilize a conventional SNAr reaction, in particular, with fluoroarenes that should have an electron-withdrawing substituent at the ortho-position (Eq. 1 in Scheme 8.6) [35]. However, we could demonstrate a similar reaction by employing 1,3–5-trifluorobenzene derivative with no any EWG substituents at ortho-positions.

Scheme 8.6
A chemical reaction for the conventional S n A r reaction of fluoroarenes with electron-withdrawing groups. It displays the transformation of fluoroarene compounds through nucleophilic attack by either alkoxide or hydroxide ions, resulting in the substitution of fluorine atoms and the formation of complex molecules.

Conventional SNAr reaction of fluoroarenes with a EWG group at the ortho-position

Full size image

As a feasibility study, we evaluated the reactivity of 1,3,5-trifluorobenzene (2) in the SNAr reaction (Scheme 8.7), and, thus treatment of 2 with sodium benzyloxide smoothly proceeded at 0 °C to give mono-alkoxide 16. Other substituted products proceeded over reactions were not detected. Furthermore, the repeated SNAr reaction with BnONa also worked well but needed a relatively high temperature (25 °C), giving dialkoxide 17 in high yield. The third replacement of the one remaining fluorine atom was possible only if the reaction temperature was raised above 100 °C.

Scheme 8.7
A step-wise S n A r reaction. At 0 degrees Celsius, the addition of B n O N a transforms the initial molecule 2 into 16, replacing one fluorine with an O B n group. Subsequent addition of B n O N a at 25 degrees Celsius yields molecule 17, with two O B n groups replacing 2 fluorine. After heating with B n O N a it results in molecule 18.

Step-wise SNAr reaction of 1,3,5-trifluorobenzene

Full size image

With these promising results in mind, we extended our investigation to bis-C-glycosyl trifluorobenzene 11, which has no electron-withdrawing group at an ortho-position to fluorine atoms (Scheme 8.8). Despite the presence of excess sodium methoxide, no reaction occurred initially. However, upon elevating the reaction temperature from 0 °C to 100 °C, the reaction proceeded step-wise manner. The reaction was terminated before complete exchange of all three fluorine atoms, giving only a disubstituted product 19. Subsequent attempts to further advance the reaction by extending the reaction time at 100 °C did not yield any additional products.

Scheme 8.8
A chemical reaction of S N A r. In compound 11, it features multiple B n O and O B n groups alongside a benzene ring with three fluorine atoms and reacts with N a O M e in N M P solvent at 0 to 100 degrees Celsius for 1.5 hours. This yields compound 19, with two O M e groups added and one fluorine atom retained, achieving a 54% yield.

Initial attempt of SNAr reaction from bis-C-glycosyl trifluorobenzene 11

Full size image

8.8 Pyran-Ring Formation and the Following Replacement of Fluorine Atoms by SNAr Reaction

Next, hoping to facilitate the reaction, we tried the reaction using the acylated substrate 15 with the aim of replacing all fluorine atoms with alkoxy or hydroxy groups (Scheme 8.9). Unfortunately, this approach proved unfruitful, and the desired trialkoxide A could not be obtained. Unexpectedly, we identified an unexpected intermolecular oxa-Michael addition followed by the retro-aldol condensation and/or the carbon–carbon bond cleavage between the C4 and C10 positions, leading to undesired compounds B and C derived from mono- and di-alkoxylated intermediates, respectively. We reasoned that these reactions occurred due to the presence of a carbonyl group unable to be conjugated to the benzene ring because of steric repulsion between the two ortho substituents as well as the strong electron-withdrawing properties of the aryl fluoride unit. Consequently, the hard nucleophiles, such as –OH or –OR, employed in the above reactions, reacted preferentially at hard electrophilic reaction sites.

Scheme 8.9
A chemical reaction of a failed attempt at a nucleophilic aromatic substitution reaction. The nucleophile attacks an aromatic ring, but a bulky group hinders the reaction.

Attempts at the SNAr reaction with the acylated bis-C-glycosyl trifluorobenzene 15

Full size image

To circumvent these unfavorable reactions, we systematically screened potential oxygen nucleophiles and identified the anion of oxime as a promising candidate (Scheme 8.10). Consequently, achieving regioselective substitution of one of three fluorine atoms, we utilized the alkoxide generated from benzaldoxime and t-BuOK in THF at room temperature, yielding 20 in high yield [36, 37]. In contrast, the reaction led to the facile retro-aldol condensation of 15 when a powdered potassium hydroxide without benzaldoxime was employed in THF at room temperature. Notably, no formation of di- and tri-hydroxylated products under these reaction conditions were observed. The reaction proceeded stepwise [38], initially undergoing a nucleophilic attack of the alkoxide of the oxime (Ph–CH=N–O–) to form the corresponding oxime ether D, which subsequently allowed the deprotonation and the elimination of the phenolate from D, giving mono-phenol 20, poised for the flavone-ring formation.

Scheme 8.10
A chemical compound synthesis scheme displaying how to convert a fluorocarbon to a phenol. The scheme involves reacting a fluorocarbon with t-B u O K in T H F, followed by a reaction with benzene acetaldehyde oxime. The product is a phenol with a much higher p K a than the starting material, indicating its less acidic.

Direct conversion of fluoroarene15 to the phenol 20

Full size image

8.9 Endgame

Having obtained a pivotal cyclization precursor, we progressed to accomplish the total synthesis (Scheme 8.11). The flavone skeleton was oxidatively constructed by employing I2 as a catalyst in DMSO at the heating conditions (140 °C), giving flavone 20 in 90% yield [39]. The reaction proceeded via the electrophilic activation of the enone moiety by iodine, thereby forming the iodinated pyran-ring E. Subsequent elimination of HI leads to the flavone 20. Iodine is regenerated by the dehydrative oxidation of a hydrogen iodide by DMSO [40].

Scheme 8.11
A chemical reaction for the formation of a six-membered ring containing an oxygen atom. A lithium-ion acts as a catalyst in this process. Reactants on the left combine to form a product with the new ring structure on the right. On the bottom, an additional chemical reaction involving the substitution of a fluoro substituent for an iodo substituent.

I2-catalyzed pyran-ring formation

Full size image

At the final stage, we investigated the substitution of the remaining two fluorine atoms with hydroxy groups (Scheme 8.12). Pleasingly, flavone 21 successfully reacted with two moles of benzyl alkoxide [41]. The solvent choice was crucial at this juncture. When treating 21 with KOH and benzyl alcohol at the heating conditions in 1,4-dioxane (88 °C, 2 h), the reaction proceeded to give bis-benzoxylated product 22 in high yield. However, the use of dipolar aprotic solvent, such as DMSO, DMF, or NMP unexpectedly suffered from ether cleavage of the incorporated alkoxy group(s), i.e., –OBn, at C4’, C5, and C7 positions, forming mono- and di-hydroxy byproducts and dibenzyl ether. At this stage, the coplanarity of the carbonyl group with the benzene ring facilitated the nucleophilic substitution of the fluorine atoms by alkoxide at C5 and C7.

Scheme 8.12
A total synthesis scheme. It features multiple rings including a fluorine atom and various side chains, and undergoes an S N A r reaction with B n O H and K O H in 1, 4-dioxane at 88 degrees Celsius for 2 hours, yielding an intermediate compound with 89% yield.

Endgame of the total synthesis through SNAr reactions

Full size image

At this stage, a rigorous 1H-NMR assignment of dialkoxide 22 was very difficult owing to the slow or restricted rotation around the C-glycosidic bonds. The signals became broad even at elevated temperatures on the NMR time scale (500 MHz) (295–373 K) [42]. Despite these challenges, we were delighted to complete the total synthesis by conducting hydrogenolysis of 22 employing ASCA-2® catalyst [≃ Pd(OH)2/C] in a mixed solvent of EtOH and EtOAc (room temperature, 12 h), affording vicenin-2 (1) in high yield (92%). All spectroscopic data ([α]D, 1H- and 13C-NMR, IR, and HRMS) were confirmed to be identical to those of the reported data [43, 44].

8.10 Conclusion

In summary, we have successfully accomplished the total synthesis of vicenin-2 (1), a bioactive bis-C-glucosyl flavonoid. This achievement was made through the bis-β-C-glucoside formation of 1,3,5-trifluorobenzene followed by the replacement of three fluorine atoms with oxygen-atom substituents. The current approach establishes a versatile synthetic pathway for bis-C-glycosyl natural products. Furthermore, our research not only propels the field forward but also unveils new possibilities for innovative synthetic organic applications involving fluoroarenes, especially in the synthesis of complex natural products.