Iodine-catalyzed electrophilic substitution of indoles: Synthesis of (un)symmetrical diindolylmethanes with a quaternary carbon center

Introduction

Diindolylmethanes (DIMs) represent an important class of indole alkaloids, that are constituents of pharmaceuticals [1-7] and agrochemicals [8,9]. DIM derivatives possess a variety of biological activities (Figure 1) [10]. Unsubstituted DIM (I), for example, exhibits antimicrobial [5], anticancer [11-13], and anti-inflammatory effects (Figure 1) [14]. There is preclinical evidence for activity against several types of cancer [15], and DIM has been clinically evaluated for the treatment of prostate cancer [16] and showed promise for the treatment of cervical dysplasia [17]. The related trisindoline (II) was reported to possess antibiotic activity [18], while DIM derivatives III and IV also showed anticancer activities (Figure 1). Owing to their exciting biological activities, DIM derivatives have recently received increasing attention from synthetic organic chemists, biologists, and pharmacologists.

In general, DIMs can be synthesized via electrophilic substitution of indoles by aldehydes or ketones in the presence of conventional Lewis or Brønsted acids as catalysts [19]. This strategy is straightforward, but it always only provides symmetrical DIMs. The synthesis of unsymmetrical DIM derivatives, however, remains challenging, and merely sporadic examples are reported in literature [20,21].

Methods for the introduction of fluorinated groups into organic molecules are of high interest due to fluorine’s unique physical and chemical properties, such as its small size, high electronegativity, and high C–F bond dissociation energy [22-24]. Organofluoro compounds developed as drug molecules often display increased metabolic stability and bioavailability compared to non-fluorinated analogs [25]. Considering the ever-growing demand for organofluorine compounds, the development of new methodologies that allow the incorporation of fluorine atoms into bioactive molecules is highly desired and will also be addressed herein.

Recently, the use of indol-3-ylmethanols as electrophiles has emerged as a powerful strategy for constructing synthetically valuable indol-3-yl-containing molecules. In particular, the reaction of indol-3-ylmethanols with indoles has become a useful route for the preparation of tertiary unsymmetrical 3,3'-DIMs [26-35]. However, synthetic methods for efficient synthesis of unsymmetrical 3,3'-DIMs with a quaternary carbon center, including trifluoromethyl-substituted 3,3'-DIMs, are still rare.

Sasaki et al. reported the reaction of trifluoromethyl(indolyl)phenylmethanols with indoles in the presence of trifluoroacetic acid (TFA) and CHCl₃ (Figure 2) [36]. Very recently, Ling et al. reported the same reaction in the presence of Ga(OTf)₃ in acetonitrile (Figure 2) [37]. Although these methods are certainly useful, they have several undeniable drawbacks, including the use of heavy-metal catalysts and the necessity of employing indoles bearing bulky substituents at their 2-position (Ling et al.), or the need for chlorinated solvents (Sasaki et al.), as well as difficulty to scale up the reactions to a multigram scale, as well as a generally rather limited substrate scope. Therefore, finding a robust method with a broad substrate scope and functional group tolerance is highly desirable.

As part of our continuous efforts to prepare biologically active DIM derivatives [38], we herein report an innovative approach to synthesize unsymmetrical 3,3'-diindolylmethanes (DIMs) with a fluoromethyl-containing quaternary carbon center via an iodine-catalyzed coupling reaction of trifluoromethyl(indolyl)phenylmethanol with indole derivatives. This method has also been extended to the synthesis of pentafluoro-ethylated and heptafluoro-propylated DIMs in excellent yields. Selected compounds were evaluated in radioligand binding studies for their affinities towards cannabinoid CB₁ and CB₂ receptors.

Results and Discussion

Optimization of the reaction

The reaction conditions were optimized using 2,2,2-trifluoro-1-(5-methoxy-1H-indol-3-yl)-1-phenylethan-1-ol (1a, 5 mmol) and 1H-indole (2a, 5 mmol) as model substrates (Table 1). At first, the reaction was attempted in trifluoroethanol (TFE), and water, respectively, as these solvents had been utilized for the preparation of unsymmetrical DIMs from (1H-indol-3-yl)(phenyl)methanol by Xiao and co-workers [39,40]. However, no product was formed in either solvent even at high temperatures (Table 1, entries 1–3). This is likely due to the steric hindrance of the CF₃-substituted quaternary carbon atom in substrate 1a. Therefore, the solvent was changed to H₂SO₄ (5%) in water (Table 1, entry 4) or glacial acetic acid (entry 5), and the reactions were performed at room temperature. While no reaction occurred in 5% H₂SO₄, traces of product were observed in acetic acid (entry 5). Therefore, the reaction mixture was gradually heated to 50 °C (Table 1, entry 6), 80 °C (entry 7), and 100 °C (entry 8). To our delight, the formation of the expected product was steadily increased to 32, 47, and 56%, respectively. Nevertheless, it was not possible to further increase the yield of the product using this solvent.

For subsequent attempts, we investigated the reaction using different Lewis acid catalysts, including AlCl₃ (Table 1, entry 9, 10% yield), FeCl₃ (entry 10, 17%), I₂ (entry 11, 51%), p-TsOH (entry 12, 5%), and InCl₃ (entry 13, traces), in acetonitrile at room temperature. Among these, the presence of I₂ led to the highest yield of 51% (Table 1, entry 11). This trend was consistent: Upon heating to 40 °C, reactions with FeCl₃ (Table 1, entry 14) or p-TsOH (entry 15) yielded 67% and 15% of product, respectively, while 98% of the product was obtained in the presence of I₂ (Table 1, entry 16). However, further increase of the reaction temperature to 80 °C did not significantly affect the generation of the product (Table 1, entry 17). Lowering the amount of catalyst from 10 mol % to 5 mol % reduced the product formation (Table 1, entry 18).

Having optimized the reaction conditions (I_2, 10 mol %, 40 °C for 5 h in MeCN; entry 16, Table 1), we explored the scope of the reaction. At first, we employed differently substituted indole derivatives (Table 2). A large variety of substituted indoles was well tolerated, and their reactions with 1a provided the desired products in good to excellent yields (61–99%). Reaction of 1a with indoles bearing electron-donating substituents, such as methoxy (2b) or hydroxy (2f), afforded the products in good yields (3b: 67%; 3f: 61%). Coupling of 1a with indoles substituted with electron-withdrawing groups, including cyano (2c), fluoro (2d, 2e, 2h), and bromo (2g), likewise resulted in good to excellent yields of the desired products (3c: 65%, 3d: 99%, 3e: 96%, 3g: 95%, 3h: 91%).

Shifting the position of the methoxy group of 1a from position 5 to 4 (1b) or 6 (1c) also led to the formation of the products in very good yields (3i: 77%, 3j: 89%, 3k: 70%, 3l: 72%). The intermediates 1d with 6-fluoro or 1e without substituent on the indole ring reacted with 5,6-difluoroindole (2h) and formed the desired products in excellent yields of 89% (3m) and 99% (3n). Besides, 2-methylindole (2i) smoothly reacted with 1e, affording product 3o in 87% yield.

Next, we studied the substrate scope of the trifluoromethyl(indolyl)phenylmethanols 1 with respect to the substitution of the phenyl ring (Table 3). The derivatives bearing p-tolyl (1f), p-fluorophenyl (1g), or p-bromophenyl (1g) rings reacted with a series of halogenated indoles (2d, 2e, 2f, 2g, 2h, and 2j) providing the unsymmetrical DIMs (3p: 99%, 3q: 98%, 3r: 87%, 3s: 90%, 3t: 82%, 3u: 92%) in excellent yields ranging from 82–99%. It was interesting to see that a compound, in which the phenyl ring of 1 was replaced by the heteroaryl moiety thiophene (1i), reacted efficiently with a series of indole derivatives (2a, 2d, 2h) providing yields of 82–90% (3v–x).

We further extended this protocol to the preparation of unsymmetrical pentafluoroethylated and heptafluoropropylated DIM derivatives (Table 4). The (indol-3-yl)phenylmethanol derivative bearing a pentafluoroethyl residue (1j) was efficiently reacted with a series of indole derivatives (2d, 2e, 2h, 2j, and 2k) substituted either both electron-donating or electron-withdrawing groups, and provided the desired products (3y: 96%, 3z: 89%, 3aa: 90%, 3ab: 86%, 3ac: 92%, 3ad: 94%) in excellent yields (86–96%).

The (indol-3-yl)phenylmethanol derivative bearing a heptafluoropropyl residue (1k) underwent coupling reactions with indole derivatives (2a, 2d, and 2e) and yielded the expected products (3ae: 65%, 3af: 72%, 3ag: 70%) in very good yields (65–72%). It was observed that the yield of heptafluoropropylated DIMs was slightly lower compared to their pentafluoroethylated congeners (compare 3y: 96% vs 3af: 72%, and 3aa: 90% vs 3ag 70%).

Next, the necessity of the fluoroalkyl substituent was investigated to study the scope of the reaction. The non-fluorinated (indol-3-yl)phenylmethanol derivative bearing a methyl residue (1l) and 2c or 2d were reacted applying the optimized conditions. However, the desired products 3ah and 3ai were not produced (Figure 3). These experiments indicated that the presence of a fluoroalkyl substituent was indeed essential for the alkylation of indoles.

We further investigated the new method’s feasibility for large-scale synthesis (Figure 4). Thus, 2,2,2-trifluoro-1-(5-methoxy-1H-indol-3-yl)-1-phenylethan-1-ol (1a, 1.0 g, 3.10 mmol) and 2,2,3,3,3-pentafluoro-1-(5-methoxy-1H-indol-3-yl)-1-phenylpropan-1-ol (1j, 1.0 g, 2.7 mmol) were reacted with indole (2a, 0.40 g, 3.4 mmol; and 0.347 g, 2.9 mmol, respectively). The reactions proceeded without significant loss in efficiency, affording 1.2 g of 3a (92% yield) and 1.10 g of 3ad (87% yield).

The unsubstituted diindolylmethane (I, Figure 1) was previously reported to bind to the cannabinoid receptors, CB₁ (K_i 4.3 µM) and CB₂ (K_i 1.1 µM) [41]. Both receptors are considered important therapeutic targets, e.g. for neurodegenerative and inflammatory diseases. Selected final products (3a, 3b, 3e, 3g, 3h, 3n, 3ad) were tested for their binding affinities towards human CB₁ and CB₂ receptors (Table 5).

At the CB₁ receptor, compound 3a with a methoxy substituent on one of the two indole rings showed equipotent affinity to lead compound I, while introducing an additional 4-methoxy moiety into the second indole ring reduced binding affinity (3b). Compounds bearing 5-OMe,6'-F (3e), 5-OMe,7'-Br (3g), and 5-OMe,5',6'-diF substitution (3h) exhibited similar binding affinities to lead compound I. 5,6-DiF-DIM derivative 3n (CB₁: K_i 2.04 µM) showed a slightly improved binding affinity compared to lead compound I. These results suggest that compounds with small substituents like fluoro on only one indole ring are favorable for CB receptor binding. The pentafluoroethylated DIM derivative 3ad was the best CB₁ ligand of the present series with a K_i value of 1.82 µM. The binding curves of 3e and 3ad are depicted in Supporting Information File 1, Figure S1.

Compound 3e showed similar binding affinities at both CB₁ and CB₂ receptor. Therefore, it was selected to determine and compare its functional activity at both receptor subtypes. Compound 3ad was selected due to its high CB₂ receptor affinity and selectivity. It is well known that CB₁ receptors exhibit high constitutive activity [41]. Compound 3e reduced the basal activity of CB₁ receptors (Supporting Information File 1, Figure S2A) but not that of CB₂ receptors indicating that this compound acts as an inverse agonist (EC₅₀ 0.786 ± 0.233 µM) at CB₁ receptors (Supporting Information File 1, Figure S2B). This effect was less pronounced for 3ad. Non-transfected cells used as controls also did not show any effect after treatment with 3ad (Supporting Information File 1, Figure S2C). DIM was previously shown to be weak inverse agonist at CB₁ receptors which is consistent with our current findings for DIM derivatives 3ad and especially 3e [42]. Next, we investigated the antagonistic effect of 3e at CB₁ receptors (Supporting Information File 1, Figure S3A). Compound 3e blocked CB₁ receptor activation with an IC₅₀ value of 5.68 ± 0.54 µM, while it was weaker in inhibiting CB₂ receptor activation. Similarly, 3ad was also able to fully block CB₁ receptor activation (IC₅₀ value of 5.22 ± 0.68 µM). Our results indicate that the new DIM derivatives act as potent CB₁ receptor antagonists with inverse agonistic activity, i.e., they stabilize the inactive receptor conformation. Further optimization is warranted. This class of compounds also possesses potential for the development of CB₂-selective or dual CB₁/CB₂-receptor antagonists.

Based on previous reports [25-35,43-53], a plausible reaction mechanism is proposed for the synthesis of 3a as an example, as depicted in Figure 5. We suggest that the reaction is initiated by iodine-mediated activation of the secondary alcohol in compound 1a (A), followed by elimination of HOI to generate the vinyliminium ion species B (see mesomeric structure C) [52]. This electrophilic intermediate undergoes a C3-selective Friedel−Crafts reaction with 2a to deliver intermediate D, and the catalyst I₂ is regenerated by the reaction of HOI and I⁻ (see C to D in box highlighted by dashed line). The intermediate D is stabilized by aromatization yielding product 3a and H₂O.

Conclusion

The described novel and efficient synthetic protocol provides a convenient access to a wide range of unsymmetrical trifluoromethylated 3,3'-diindolylmethanes via I₂-catalyzed Friedel–Crafts alkylation reaction of trifluoromethylated (indol-3-yl)-1-phenylethan-1-ols with substituted indoles. The method was also extended to the synthesis of pentafluoroethylated and heptafluoropropylated-DIMs. It constitutes an important addition to the active field of DIM syntheses facilitating the preparation of unsymmetrical quaternary DIMs without the need for chlorinated solvents, high temperatures, or heavy-metal catalysts. A broad range of substrates is tolerated and the reaction is suitable for large-scale preparation of the target compounds. The outlined methodology allows for the rapid generation of structurally diverse DIM derivatives to study structure–activity relationships, to optimize biological activity and other properties in order to prepare tool compounds and future drugs. Several compounds displayed micromolar binding affinities toward CB₁ and CB₂ receptors acting primarily as CB₁ receptor antagonists/inverse agonists. We are confident that our straightforward new approach will enable us and others to extensively investigate these bioactive molecules and their targets in future studies.