Strain-Release [2π + 2σ] Cycloadditions for the Synthesis of Bicyclo[2.1.1]hexanes Initiated by Energy Transfer

Abstract

Saturated bicycles are becoming ever more important in the design and development of new pharmaceuticals. Here a new strategy for the synthesis of bicyclo[2.1.1]hexanes is described. These bicycles are significant because they have defined exit vectors, yet many substitution patterns are underexplored as building blocks. The process involves sensitization of a bicyclo[1.1.0]butane followed by cycloaddition with an alkene. The scope and mechanistic details of the method are discussed.

Graphical Abstract

In recent years, a greater emphasis has been placed on the incorporation of saturated building blocks into chemical libraries for drug development.¹ For example, bicyclo[1.1.1]pentane has garnered significant attention in the literature because it has attributes similar to those of a benzene ring.² Many of the synthetic advances in this area are driven by the strain-release ring-opening reaction of [1.1.1]propellane. The bicyclo[2.1.1]hexane scaffold is significant because of its rigid nature and defined exit vectors, yet it has received significantly less attention (Scheme 1A).³

Scheme 1.

Cycloaddition Approaches toward Bicyclo[2.1.1]hexanes

The most common method for the synthesis of bicyclo[2.1.1]hexanes is via crossed [2 + 2] cycloaddition of 1,5-dienes.⁴ An alternative approach involves [2π + 2σ] cycloaddition of bicyclo[1.1.0]butanes (BCBs).^5–7 Only three examples of the reaction of BCBs with alkenes are known.^7,8 Among those, two are intermolecular and were reported in 1966 by Blanchard and 1986 by De Meijere.^7a,b In each study, it was demonstrated in one example that bicyclo[2.1.1]hexanes could be generated by heating a mixture of either butadiene or an α-thioacrylonitrile. More recently, in 2006 Wipf reported a thermally induced intramolecular cycloaddition to arrive at complex tricyclic compounds (Scheme 1B).^7c On the basis of these reports, it is clear that cycloadditions of BCBs to prepare bicyclo[2.1.1]hexanes are quite limited. Herein we provide a solution to this challenge by introducing a new mode of activation of bicyclo[1.1.1]butanes. The approach allows for intermolecular cycloaddition and thus access to a diverse range of bicyclo[2.1.1]hexanes with new substitution patterns.⁸

The reaction design is illustrated in Scheme 2A. We envisioned that the strained central C–C bond of BCB could be cleaved by single electron transfer (SET), thermolysis, direct excitation, or energy transfer to generate a diradical or radical anion intermediate. Capture of this intermediate with an alkene would lead to the formation of the desired product via stepwise cycloaddition. It should be noted that other strain-release methods to enable cycloaddition are known. The vast majority of known examples involve monocyclic systems (e.g., cyclopropane). For example, the chemistry of donor–acceptor cyclopropanes, which typically proceeds via two-electron pathways, is well-established.⁹ Reactions involving radical intermediates are also known. For example, single-electron reduction of cyclopropyl ketones followed by reaction with an alkene has been demonstrated in a variety of contexts.¹⁰ Finally, radical-initiated ring-opening reactions of substrates derived from cyclopropylamines,¹¹ alcohols,¹² and alkenes¹³ are known. Some of these methods have also been extended to cyclobutanes.¹⁴ In one very recent study, the Stephenson group utilized bicyclo[1.1.1]pentanes in a strain-release cycloaddition with an alkene.¹⁵

Scheme 2.

Initial Investigations

After considerable experimentation with thermally induced reactions as well as photoredox processes, no product from reaction between a BCB and an alkene was observed. In most cases, polymerization of the BCB was observed.¹⁶ At this stage, we decided to explore an alternative approach involving triplet energy transfer (Scheme 2B).¹⁷ In this scenario, the substituent bound to the BCB should be capable of being sensitized. We reasoned that an aryl ketone should undergo energy transfer with an appropriate sensitizer, as demonstrated previously.¹⁸ Initial studies with phenyl ketone 1 did not result in productive product formation (Scheme 2B). We reasoned that the similarity of the triplet energies of phenyl ketone 1 and styrene (E_T = 61 kcal/mol) was likely the issue. Therefore, naphthyl ketone 2 was examined, as the calculated triplet energy (E_T = 54 kcal/mol) indicated that this substitution should lead to more facile sensitization relative to styrene. An initial hit was observed when 2-isopropylthioxanthone (ITX, 5) was used, delivering the product in 46% yield.

In addition to the promising results, we were also drawn to the straightforward synthesis and favorable properties of the naphthyl ketone-substituted bicyclo[1.1.0]butane 2 (Scheme 3). While known routes to BCBs typically rely on the use of t-BuLi,¹⁹ we explored a route that avoids use of this pyrophoric reagent. The synthesis commences with Weinreb amide formation followed by reduction and mesylation to generate 7.²⁰ These first three steps were conducted without any precautions to exclude air or moisture. The transannular ring formation could be easily carried out by treatment with potassium tert-butoxide to provide Weinreb amide 8. Completion of the synthesis was accomplished by the addition of 2-naphthyllithium to 8. The final product is a solid (mp ≈ 60 °C (decomp); see the Supporting Information (SI) for DSC data) that is stable in air for <1 month at room temperature and indefinitely stable at −30 °C.

Scheme 3.

Synthesis of Bench-Stable Bicyclo[1.1.0]butane

With the identification of ITX as an effective sensitizer, optimization was conducted with this as a starting point. It was reasoned that the use of a thioxanthone-based sensitizer with a lower triplet energy might be beneficial to allow for more selective energy transfer to ketone 2 as opposed to styrene. It should be noted that with the use of ITX, extensive dimerization/polymerization of styrene was observed. Previous studies by Booker-Millburn showed that addition of electron-donating groups at the 2- and 2′-positions of thioxanthone lowers the triplet energy.²¹ Thus, we examined 2-OMeTX (9) and 2,2′-OMeTX (10) and found that the latter resulted in the formation of the product in 76% yield (Table 1, entries 1–3). For each sensitizer evaluated, a light source was used on the basis of prior efforts.²² While we proceeded with 10 to evaluate the substrate scope, further examination of Ir-based sensitizers²² (Table 1, entries 4–7) revealed that [Ir(ppy)₂dtbbpy]PF₆ (14) could also provide product 4, albeit in lower yield (Table 1, entry 7). While the absolute triplet energies of the catalysts evaluated in the Ir series and the TX series are not correlated, the efficiency of energy transfer is due to factors beyond simple comparison of triplet energies.²³

Table 1.

Optimization of the Reaction Conditions^a

entry	sensitizer	ET (kcal/mol)	X	hυmax (LED)	yield (%)
1	2-iPrTX (5)	63.5	10	395	46
2	2-OMeTX (9)	57.8	10	395	28
3	2,2′-OMeTX (10)	55.2	10	450	76
4	Ir[(dFCF3ppy)2dtbbpy]PF6 (11)	60.1	1	450	58
5	Ir[(dFppy)2dtbbpy]PF6 (12)	57.1	1	450	60
6b	Ir[(Fppy)2dtbbpy]PF6 (13)	53.3	1	450	18
7	Ir[(ppy)2dtbbpy]PF6 (14)	51.0	1	450	70

^aReactions were run on a 0.1 mmol scale. Yields were determined by ¹H NMR analysis of the unpurified reaction mixtures with CH₂Br₂ as an internal standard.

^bThe limited solubility of the sensitizer resulted in a poor yield.

Under the optimized set of conditions, a variety of alkenes were examined with BCB 2 (Scheme 4). With simple vinyl arenes, substitution with electron-donating (products 16, 31) or electron-withdrawing substituents (products 17–19, 26, 30) were tolerated. In addition, sterically demanding groups did not greatly diminish the yield (products 22, 26). The reaction of heterocycles was met with mixed success (products 23–25, 27–29): while the desired product was observed in every case evaluated, the yields were generally lower. Overall, the yields tended to be better with pyridine-derived heterocycles (products 28 and 29). While reactions of 1,2-disubstituted alkenes also led to product formation (products 33–35), albeit in low to moderate yield, 1,1-disubstituted alkenes generally worked better (products 39–41). For reaction of cis- and trans-β-methylstyrene, stereoconvergence was observed (product 33). For substrates beyond vinyl arenes, success was found with a vinylBpin (product 38). Other terminal alkenes such as vinyl acetate (product 36) and vinylsilane (product 37) did lead to product formation, but in low yields. Electron-deficient alkenes and dienes do not allow for product formation (see the SI for details). In cases where low yields were observed, polymerization of the BCB was occurred.¹⁶ Finally, substitution on the BCB was also examined. 2-Methyl (product 43) and 2-phenyl (product 44)-derived BCBs were successfully synthesized, with more favorable results for the former. The lower yield for the 2-phenyl substrate may be due to competitive cleavage of the cyclobutane. In the case of 3-methyl (product 45) and 3-phenyl (product 46) derivatives, the yield was higher for the latter. It should be noted that the 3-substituted variants were prepared by a route analogous to that shown in Scheme 3,²⁴ whereas 2-substituted analogues were synthesized by intramolecular cyclopropanation (see the SI for details).²⁵

Scheme 4.

Substrate Scope^a

^aReactions were run on a 0.2 mmol scale. Yields of isolated and purified products that are averages of two separate experiments are shown.

The products generated by this strategy can be further manipulated in various ways (Scheme 5). For example, the naphthyl group can be easily transformed to acid 47 by Baeyer–Villiger oxidation followed by hydrolysis. The acid can also be derivatized by a Curtius rearrangement to allow for synthesis of amine 48. In addition, the formation of redox-active ester 49 (NHPI = N-hydroxyphthalimide) followed by either decarboxylative Minisci reaction²⁶ or borylation²⁷ provides access to 50 or 51, respectively. Finally, oxidation of the Bpin group results in the formation of alcohol 52. Thus, the naphthyl ketone products can be easily elaborated to acid, amine, alcohol, and boronic ester functionality.

Scheme 5.

Further Transformations^a

^aReaction conditions: (a) (i) DPPA, Et₃N, PhMe, 110 °C, 2 h; (ii) 3 M HCl, AcOH, 0 °C to rt, 16 h. (b) NHPI, DCC, 10 mol % DMAP, CH₂Cl₂, rt, 12 h. (c) 4-Methylquinoline, TFA, 20 mol % NaI, 30 mol % PPh₃, acetone, rt, 16 h, blue LED. (d) (Bcat)₂, DMAc, rt, 16 h, blue LED, then pinacol, Et₃N. (e) NaBO₃·4H₂O, THF/H₂O, rt, 3 h.

The mechanism of the reaction was also explored (Scheme 6). On the basis of a comparison of the triplet energies of 2,2′-OMe-TX (E_T = 55 kcal/mol) and 2 (E_T ≈ 54 kcal/mol) with the redox potentials ($E_{\text{red}}^{*}\left[{\mathbf{10}}^{+}/{\mathbf{10}}^{*}\right]=-1.29\text{ V}$ vs E_red = −1.93 V, respectively), energy transfer seems more likely.²⁸ Regardless, the most significant data suggesting that the reaction proceeds via excited-state intermediates are shown in Scheme 6. It was demonstrated that the reaction proceeds in 60% yield upon direct irradiation with 365 nm LEDs (emission spectral window of ~350 to ~385 nm) in the absence of the sensitizer.²⁹ UV/vis measurements showed that the absorption spectrum of BCB 2 overlaps with the emission spectrum of the 365 nm LEDs (e.g., the molar absorptivity (ε) for 2 at 360 nm is 87.1 M⁻¹·cm⁻¹), whereas that of styrene does not (Scheme 6B). In addition, it was confirmed that an electron donor–acceptor complex between 2 and styrene does not occur, as evidenced by the lack of a change in the UV/vis spectrum (Scheme 6B).³⁰

Scheme 6.

Mechanism Studies

On the basis on the studies outlined above, it is proposed that the reaction proceeds via energy transfer to generate the T₁(π−π*) state of 2 (Scheme 7). This intermediate can also be accessed by direct excitation of 2 to the S₁(n−π*) state followed by rapid intersystem crossing (ISC). Strain-release-induced bond cleavage then occurs to generate triplet diradical 53. Capture of the secondary radical with the alkene leads to the formation of 54. Finally, ISC and bond formation lead to the formation of the desired product.

Scheme 7.

Proposed Mechanism

In conclusion, a new strategy for the synthesis of bicyclo[2.1.1]hexanes has been presented. The reaction operates by energy transfer, which is a new approach for strain-release-driven transformations. While this strategy has been demonstrated for the synthesis of bicyclo[2.1.1]hexanes, we expect that the principles delineated here will be more broadly applicable to access a diverse range of saturated bicyclic structures.