Reactions of 2-Aryl-1,3-Dithianes and [1.1.1]Propellane

Abstract

Bicyclo[1.1.1]pentanes (BCPs) have sparked the interest of medicinal chemists due to their recent discovery as bioisosteres of aromatic rings. To study the biological activity of this relatively new class of bioisosteres, reliable methods to incorporate BCPs into target molecules are in high demand, as reflected by a flurry of methods for BCP synthesis in recent years. In this work, we disclose a general method for the synthesis of BCP-containing dithianes which, upon deprotection, provide access to BCP analogues of medicinally abundant diarylketones. A broad scope of 2-aryl-1,3-dithianes, including several heterocyclic derivatives, react with [1.1.1]propellane to afford 26 new derivatives in good to excellent yields. Further transformation of the dithiane portion into a variety of functional groups demonstrates the robustness of the products. A computational study indicates that the reaction of 2-aryl-1,3-dithianes and [1.1.1]propellane proceeds via a two-electron pathway.

Graphical Abstract

The two electron path: Propellylation of 2-aryl-1,3-dithianes provides bicyclo[1.1.1]pentyl substituted dithianes via a 2-electron mechanism. Further transformations give access to BCP aryl ketones and BCP aryl difluoromethanes.

Bicyclo[1.1.1]pentanes (BCPs) have recently attracted attention from synthetic and medicinal chemists due to their potential as bioisosteres of phenyl groups in bioactive molecules^[1]. In the past decade, medicinal chemists have demonstrated the high promise of BCPs for the optimization of drug performance, such as in BMS’s BCP-BMS-708163,^[2] BCP-resveratrol,^[3] and BCP-darapladib.^[4] These BCP analogues demonstrate improved pharmacological properties without compromising potency. Despite considerable interest from the medicinal chemistry community, the incorporation of BCPs into a broad variety of structural classes found in bioactive molecules remains an unsolved challenge. Recently, strain-release functionalization has become a promising approach to prepare novel organic compounds,^[5] including BCPs. Many protocols for the monofunctionalization of BCPs^[6] as well as unsymmetrically 1,3-difunctionalized BCP derivatives^[7] have been reported, which fill major synthetic gaps for the BCP analogues of anilines, biaryls, fluoroarenes, and α-aryl ketones. Prior to these disclosures, all of which were reported in the past five years, no reliable methods to access BCP analogues of these scaffolds were known, reflecting the nascency of BCP synthesis for medicinal chemistry. Along these lines, we noticed a major gap in the literature for the synthesis of BCP ketones: the addition of 1-lithio-BCPs to nitriles, as reported by Wiberg, produces an extremely limited scope of diarylketones in modest yields (Scheme 1a).^[8] Wiberg’s pioneering study also describes the addition of acyl radicals to [1.1.1]propellane, but this method suffers from poor chemoselectivity and narrow substrate scope. Despite the abundance of diarylketones in FDA-approved drugs, such as Ketoprofen^[9], Fenofibrate^[10] (Tricor), Mebendazole^[11] (Vermox) and Tolcapone^[12] (Tasmar) (Scheme 1b), no synthetic route to BCP analogues of these targets are known. To address this critical gap, we aimed to develop a method for the synthesis of BCP ketones via the reactivity of 2-aryl-1,3-dithiane anions and [1.1.1]propellane.

Scheme 1.

Previous synthesis of BCP ketones, examples of bioactive diaryl ketones, general reactivity of dithianes, and synthesis of BCP dithianes.

Beginning with Corey and Seebach’s 1965 disclosure of metallated dithianes and their addition to electrophiles,^[13] dithianes have been used to achieve a variety of carbon-carbon bond formations, finding utility from simple synthetic transformations to the synthesis of natural products.^[14] Recent reports demonstrate new and unexpected reactivity of 1,3-dithianes, such as for the synthesis of pyridines and photochemical addition to electrophiles.^[15] A notable case of new reactivity of dithianes is the Smith group’s development of anion-relay chemistry using 2-silyl-1,3-dithianes.^[16] 2-Aryl-1,3-dithianes could be used as pronucleophiles for palladium-mediated arylation^[17] and allylic substitution^[18] in a deprotonative cross-coupling process. To enable synthetic access to BCP ketones and to explore new reactivity of 1,3-dithianes, we describe herein the addition of 2-aryl-1,3-dithianes to [1.1.1]propellane, an unprecedented transformation which furnishes BCP analogues of a valuable pharmacophore. BCP dithianes are readily deprotected to the corresponding ketones. We also demonstrate that these BCP dithianes are readily transformed to geminal difluoromethanes and esters. Computational investigations indicate a two-electron pathway for the reaction, demonstrating 2-aryl-1,3-dithiyl anions to be competent nucleophiles for addition to [1.1.1]propellane. To the best of our knowledge, these computational investigations are the first of their kind to support a purely anionic pathway for C—C bond formation to [1.1.1]propellane.

We explored the propellylation of aryl dithianes using 2-phenyl-1,3-dithiane (2a) as a pronucleophile and [1.1.1]propellane (1). Propellylation using 2 equiv LiN(SiMe₃)₂ and THF at room temperature, conditions which previous studies had demonstrated as suitable to generate dithiyl anions, produced BCP product 3a in 5% AY (assay yield) after 16 h (Table 1, entry 1). Increasing the temperature to 60 °C slightly improved the AY to 20% (entry 2). Switching to 2 equiv NaN(SiMe₃)₂ offered a modest improvement to 58% AY of product 3a (entry 3). Further increasing the reaction temperature to 80 °C and switching to DME solvent gave 89% AY (entry 4). Product 3a was isolated in 93% yield by raising the loading of NaN(SiMe₃)₂ to 2.5 equiv (entry 5). A substoichiometric loading of base (20 mol%) promoted the reaction comparatively slowly to generate product 3a in 43% yield with 52% conversion of starting material after 48 h (entry 6). Thus, we employed the conditions in entry 5 with 2.5 equiv NaN(SiMe₃)₂ in DME at 80 °C for 16 h. With the optimized conditions in hand, we explored the scope of propellylation with 2-aryl-1,3-dithiane pronucleophiles (Table 2). Pronucleophiles with electron neutral groups gave the corresponding products in excellent yields (3a–3b, 93–96%). Pronucleophiles with electron donating groups in the 4-position (Me, tBu, OMe, SMe, C≡CH, NMe₂) underwent propellylation to provide good to excellent yields of the corresponding products (3c–3h, 54–96%). Notably, the alkyne-containing pronucleophile 2g reacted exclusively at the dithiane methine position despite the acidity of the C(sp)–H bond, highlighting the excellent chemoselectivity of this method. In the case of 2h which contains the strongly electron donating 4-N,N-dimethylamino group, increasing the base loading to 4 equiv NaN(SiMe₃)₂ gave 54% yield. Pronucleophiles with electron withdrawing groups or halogens (3-OMe, 4-F, 4-Br, 3-F, 3-CF₃, 4-BPin) reacted to form the desired products in excellent yields (3i–3n, 75–97%). Surprisingly, pronucleophile 2n, which bears a boronate ester, was well-tolerated under the reaction conditions. Sterically hindered pronucleophiles bearing polycyclic aromatics (1-naphthyl, 9-phenanthryl, 9-anthracenyl) formed the desired products in good yields using 3 equiv base (3o–3q, 64–71%). The more hindered substrate 2-(2-tolyl)-1,3-dithiane was recovered after subjection to the standard reaction conditions, and 2-(2-bromophenyl)-1,3-dithiane decomposed. To demonstrate the potential value of this method to medicinal chemistry, we next examined the scope of heterocycles in the propellylation reaction. Gratifyingly, a number of heterocycles were tolerated under the optimal conditions. The 3-furyl dithiane 2r reacted to afford 3r in 60% yield. Thiophenes also underwent propellylation to provide good to excellent yields of the corresponding products (3s–3u, 76–95%). Most nitrogen-containing pronucleophiles decomposed under the standard reaction conditions (Table 1, entry 5). By switching to 3 equiv LiN(SiMe₃)₂ base instead of NaN(SiMe₃)₂, good yields of the corresponding products were obtained from 3-pyridyl-, 4-pyridyl-, 5-isoquinolyl-, and N-methyl 5-indolyl-derived dithianes (3v–3y, 65–85%). Notably, the formation of staffanes was not observed under the optimal conditions in any case.

Table 1.

Optimization of propellylation of 1a^{[a], [b]}

Entry	M (equiv)	Solvent	T (°C)	Yield[b] (%)	SM[c](%)
1	Li (2)	THF	rt	5	>90
2	Li (2)	THF	60	20	81
3	Na (2)	THF	60	58	43
4	Na (2)	DME	80	89	8
5	Na (2.5)	DME	80	96 (93[d])	trace
6	Na (0.2)	DME	80[e]	43	52

^[a]Reactions conducted on 0.1 mmol scale and 0.1 M.

^[b]Assay yields determined by ¹H NMR spectroscopy of the crude reaction mixture using 0.1 mmol CH₂Br₂ as an internal standard.

^[c]Amount of dithiane starting material remaining as determined by ¹H NMR spectroscopy.

^[d]Yield of isolated product.

^[e]48 h reaction time.

Table 2.

Scope of propellylation ^[a]

^[a]Reactions conducted on 0.1 mmol scale and 0.1 M.

^[b]4 equiv NaN(SiMe₃)₂

^[c]3 equiv LiN(SiMe₃)₂.

^[d]3 equiv NaN(SiMe₃)₂.

To evaluate the scalability of our method, the propellylation of 2-(3-thiophene)-1,3-dithiane 2s was conducted on a 4.0 mmol scale, providing the corresponding product 3s (94% yield, 1.01 g; 0.1 mmol scale = 95% yield), demonstrating excellent reaction efficiency upon scale-up.

To demonstrate the chemical robustness of BCP dithianes, we studied their conversion to novel BCP-containing products. Deprotection of dithianes worked well for a variety of BCP dithianes, affording ketone products in good yields without the need for substantial reoptimization of Barik’s dithiane deprotection procedure (Scheme 2a).^[19] BCP aryl ketones derived from dithianes 3b, 3k and the heterocycle 3s were obtained in good yields from deprotection (75–81%). Additionally, using a modified version of Katzenellenbogen’s procedure,^[20] aryl dithianes, including thiophene heterocycles, were successfully converted to BCP aryl difluoromethanes in good yield. BCP dithiane 3e underwent difluorination smoothly to provide 5e in 83% yield within 15 min whereas heterocycles 5t and 5u required longer reaction times (1 h) to furnish the difluorinated products in good yield (72% and 86% yield, respectively, Scheme 2b). Access to the BCP difluoromethyl core is unprecedented in the current literature. Notable diaryl difluoromethanes, for which BCP aryl difluoromethanes are potential bioisosteres, include Ledipasvir (GS-5885), a potent inhibitor for the treatment of hepatitis C.^[21]

Scheme 2.

Synthesis of BCP ketone derivatives, synthesis of gem-difluoro BCP compounds, transformations of BCP dithioacetal [a] reaction time 15 min. i) 8 equiv mCPBA, CHCl₃, rt, 16 h. ii) 4 equiv mCPBA, CHCl₃, 50 °C, 24 h. note: SDS = sodium dodecylsulfate.

1,3-Dithianes derived from other thiols were also successful in the propellylation reaction. A BCP product derived from thiophenol was prepared as 3z (see SI for details). Hydrolysis of dithioacetal 3z to ketone 4z using 8 equiv m-CPBA in CHCl₃ at room temperature proceeded in excellent yield (91%). Ketone 4z could be further transformed to ester 6z via Baeyer-Villiger oxidation^[22] using 4 equiv mCPBA at 50 °C in excellent yield (96%, Scheme 2c).

To evaluate the scalability of our method, the propellylation of 2-(3-thiophene)-1,3-dithiane 2s was conducted on a 4.0 mmol scale, providing the corresponding product 3s (94% yield, 1.01 g; 0.1 mmol scale = 95% yield), demonstrating excellent reaction efficiency upon scale-up.

To demonstrate the chemical robustness of BCP dithianes, we studied their conversion to novel BCP-containing products. Deprotection of dithianes worked well for a variety of BCP dithianes, affording ketone products in good yields without the need for substantial reoptimization of Barik’s dithiane deprotection procedure (Scheme 2a).^[19] BCP aryl ketones derived from dithianes 3b, 3k and the heterocycle 3s were obtained in good yields from deprotection (75–81%). Additionally, using a modified version of Katzenellenbogen’s procedure,^[20] aryl dithianes, including thiophene heterocycles, were successfully converted to BCP aryl difluoromethanes in good yield. BCP dithiane 3e underwent difluorination smoothly to provide 5e in 83% yield within 15 min whereas heterocycles 5t and 5u required longer reaction times (1 h) to furnish the difluorinated products in good yield (72% and 86% yield, respectively, Scheme 2b). Access to the BCP difluoromethyl core is unprecedented in the current literature. Notable diaryl difluoromethanes, for which BCP aryl difluoromethanes are potential bioisosteres, include Ledipasvir (GS-5885), a potent inhibitor for the treatment of hepatitis C.^[21]

1,3-Dithianes derived from other thiols were also successful in the propellylation reaction. A BCP product derived from thiophenol was prepared as 3z (see SI for details). Hydrolysis of dithioacetal 3z to ketone 4z using 8 equiv m-CPBA in CHCl₃ at room temperature proceeded in excellent yield (91%). Ketone 4z could be further transformed to ester 6z via Baeyer-Villiger oxidation^[22] using 4 equiv mCPBA at 50 °C in excellent yield (96%, Scheme 2c).

We next turned our attention to the reaction mechanism. While most reactions involving dithiane pronucleophiles are proposed to proceed via two electron processes, examples involving one electron processes are well known.^[22] To better understand the mechanism of the propellylation reaction, computational studies were undertaken for the reaction of 2a with [1.1.1]propellane (1). DFT calculations were performed using Gaussian 09. All geometry optimizations and vibrational frequency calculations of stationary points and transition states (TSs) were carried out at the UB3LYP level of theory with 6–31+G(d,p) basis set for all atoms with an SMD solvation model (1,4-dioxane).

Calculations suggest that the reaction of dithiane 2a with [1.1.1]propellane proceeds via a two-electron pathway (Figure 1). The first step of the process involves the deprotonation of substrate 2a by NaN(SiMe₃)₂ to obtain INT1. This anion reacts with [1.1.1]propellane via TS_Equat or TS_Axial to afford INT2. TS_Axial is energetically preferred due to the higher 1,3-diaxial interactions with propellane in the axial position (TS_Equat). A subsequent proton transfer from 2a leads to formation of the desired product and regeneration of INT1. Calculations suggest that the reaction can be catalytic in base, as supported by the experimental results reported in Table 1 – entry 6. However, the catalytic version is kinetically slower and stoichiometric base was found to give higher yields.

Figure 1.

Energy profile for the reaction of 2a with [1.1.1]-propellane. Relative energy values calculated with SMD-1,4-dioxane-UM06–2X/6–311+G(d,p)//UB3LYP/6–31+G(d,p).

A second mechanistic pathway, involving a single-electron transfer from INT1 to [1.1.1]propellane^{[15b, 23]} was also considered. This pathway would proceed through an electron-transfer from INT1 to [1.1.1]propellane (see SI for details). Calculations showed that this process is highly unlikely due to the high free energy of these radical intermediates (57.8 kcal/mol).

Inspired by the abundance of one-electron pathways that are proposed for reactions of [1.1.1]propellane,^{[7e, 24]} a pathway involving 2-aryl-1,3-dithiyl radicals with [1.1.1]propellane was also considered. However, formation of the 2-aryl-1,3-dithiyl radical is not evident under the reaction conditions described above. Thus, the two-electron pathway (Figure 1) is most likely to be operative. To our knowledge, this computational study is the first to invoke an exclusively two-electron pathway for σ-bond cleavage of [1.1.1]propellane by an anionic nucleophile. To support these computational results, we conducted the propellylation of 2a with nBuLi in DME and obtained 50% of 3a. Since DME might be sufficiently acidic to serve as a proton source in the propellylation of 2a, due to the high basicity of a BCP carbanion (INT2), the solvent was changed to cyclohexane. Under these conditions, near quantitative deprotonation was observed, but only 4% of 3a was produced (see SI for details). Despite a modest barrier for addition of anionic INT1 to [1.1.1]propellane, we propose that the equilibrium between INT1 and INT2 heavily favors INT1, according to our calculations. Thus, we propose that protonation of INT1 by dithiane or hexamethylsilylamine to the product consumes INT2 irreversibly and provides a driving force for the formation of INT2 from INT1 despite the energetic favorability of INT1. This result is supported by the low yield obtained when no proton sources are present (e.g. n-BuLi as the base rather than NaN(SiMe₃)₂ (see SI for details).

In conclusion, we have disclosed an efficient procedure to synthesize 26 new dithiane BCPs, in good to excellent yields. The reaction tolerates a broad substrate scope and is scalable. These products can be transformed into their corresponding ketones, which are potential bioisosteres of diaryl ketones. BCP dithianes can also be transformed to gem-difluoro BCPs and BCP esters. All products presented are novel and fill a substantial gap in both the synthetic and medicinal chemistry literatures by providing access to BCP analogues of pharmacophores with demonstrated biological activity. Computational results suggest that a two-electron pathway is most likely for this propellylation reaction.