Nickel-Catalyzed Decarboxylative Cross-Coupling of Bicyclo[1.1.1]pentyl Radicals Enabled by Electron Donor–Acceptor Complex Photoactivation

Abstract

The use of bicyclo[1.1.l]pentanes (BCPs) as para-disubstituted aryl bioisosteres has gained considerable momentum in drug development programs. Carbon–carbon bond formation via transition-metal-mediated cross-coupling represents an attractive strategy to generate BCP–aryl compounds for late-stage functionalization, but these typically require reactive organometallics to prepare BCP nucleophiles on demand from [l.1.l]propellane. In this study, the synthesis and Ni-catalyzed functionalization of BCP redox-active esters with (hetero)aryl bromides via the action of a photoactive electron donor–acceptor complex are reported.

Graphical Abstract

The operational ease of Pd-catalyzed C(sp²)–C(sp²) coupling reactions along with the availability of coupling partners (in particular, arylboronates and aryl halides) has resulted in a bias toward biaryl-containing molecules in drug development programs.¹ These cross-coupling platforms have enabled the synthesis of diverse biaryl-containing drugs targeting a wide swath of therapeutic areas. However, considerable interest has developed in recent years to investigate sp³-rich aryl isosteres in drug development programs.²

Of particular interest is the bicyclo[l.1.l]pentane (BCP) motif, which has been reported most commonly as a p-disubstituted aryl motif but also as a tert-butyl or alkyne bioisostere, often imparting favorable pharmacokinetic properties, including improved aqueous solubility and membrane permeability.³

Although considerable advancements have been made in the synthesis of BCP-containing targets,^4–6 the methods available to forge BCP–aryl products via direct cross-coupling are somewhat limited. Szeimies and co-workers,⁷ de Meijere and co-workers,⁸ and Knochel and co-workers⁹ have reported Kumada- and Negishi-type couplings between aryl halides and the corresponding BCP organometallic reagents (Figure 1A). Additionally, Kanazawa, Uchiyama, and co-workers¹⁰ reported a Suzuki-type coupling using l,3-difunctional silyl BCP boronates. Shortly after, the Walsh group, in collaboration with scientists at Merck,¹¹ disclosed the use of l,3-difunctional benzylamine BCP boronates in Suzuki couplings (Figure 1B).

Figure 1.

Comparison of two- and one-electron strategies to forgi BCP–aryl cross-coupling products.

Leveraging a radical-based mechanism, Baran and co-workers¹² reported the cross-coupling of a single BCP redox-active ester (RAE) with Ph₂Zn. Anderson and co-workers, in collaboration with Pfizer,¹³ then reported the iron-catalyzed Kumada coupling of various BCP iodides with aryl Grignard reagents (Figure 1C). In 2020, VanHeyst, Qi, and co-workers from Merck and WuXi engaged BCP trifluoroborate salts in Ni/photoredox dual cross-coupling and achieved modest to acceptable yields (Figure 1D).¹⁴ Of note and particular pertinence to the results reported herein, the BCP trifluoroborate salts required preparation through a continuous flow photoborylation method from the corresponding N-(acyloxy)-phthalimide redox active esters (RAEs).^14,15

Although BCP organometallic reagents perform well as cross-coupling partners, their applicability in late-stage functionalization is limited because of their functional group incompatibility and short-term stability. To address these shortcomings, we envisioned employing bench stable, commercially available BCP carboxylic acid feedstocks under mild, Ni-catalyzed photochemical conditions.¹⁶

The initial goal was to use carboxylic acids directly in Ni/photoredox dual cross-coupling. Encouragingly, we established suitable conditions for the decarboxylation of a BCP carboxylate and verified that the resultant BCP radical engages in defluorinative alkylation with a trifluoromethyl-substituted alkene (Scheme S1A of the Supporting Information). However, adapting these conditions to Ni-catalyzed C–C bond formation with 4-bromobenzonitrile failed to generate the desired arylated BCP product.

Under dilute reaction conditions (0.025 M instead of 0.1 M), the corresponding BCP aryl ester was observed exclusively (Scheme S1B of the Supporting Information). This result can be rationalized by the high s character of the BCP–carboxylate bond (~sp²¹),⁶ resulting in a slower rate of decarboxylation, thus favoring an energy-transfer-dependent C–O coupling.^17,18

These results motivated us to investigate the activation of BCP–N-(acyloxy)phthalimide RAEs, bench-stable solids that are readily prepared from the corresponding carboxylic acids via a quantitative Steglich esterification. These derivatives undergo decarboxylative radical fragmentation upon single-electron reduction¹⁹ through photochemical strategies that are unavailable to the parent carboxylic acids.²⁰ In this vein, our group recently reported the nickel-catalyzed cross-coupling of primary and secondary alkyl RAEs with (hetero)aryl halides using Hantzsch ester (HE) as a potent organic photoreductant.²⁰ The intermediacy of a photoactive electron donor–acceptor (EDA) complex^21–25 was envisioned to facilitate the generation and subsequent functionalization of BCP radicals in Ni-catalyzed cross-electrophile paradigms, bypassing the need for preformed carbon nucleophiles as well as electron transfer events from exogenous photoredox catalysts. In addition, the developed protocol would provide a low barrier for practical implementation in medicinal chemistry settings.

Subsequently, we successfully adapted our previously reported conditions to the cross-coupling of a BCP RAE (1,3-dioxoisoindolin-2-yl bicyclo[1.1.1]pentane-1-carboxylate) with 4-((4-bromophenyl)sulfonyl)morpholine in modest yield (Table S1 of the Supporting Information). The addition of a mild base, such as NaHCO₃ or K₂HPO₄, improved the cross-coupling yield by diminishing the loss of the BCP radical to Minisci-type addition (SI-21) to Hantzsch pyridine, which is generated upon photoaromatization.

With suitable conditions established (see the Supporting Information for full optimization details), we proceeded to investigate the scope of aryl bromides amenable to the cross-coupling protocol (1–16, Scheme 1). In general, the best yields were obtained using aryl bromides bearing electron- withdrawing groups. Modest product formation was observed with electron-rich and electron-neutral aryl bromides because of competitive proto-debromination. Electrophilic and protic functional groups were well-tolerated (in contrast with cross-coupling methods based on organometallic reagents), including several synthetic handles found in substrates, such as boronate 4, chloride 5, ester 7, and amide 15. Selected heteroaromatic bromides were accommodated and did not engage in Minisci-type side reactivity, giving modest to synthetically acceptable cross-coupling yields (17–23). Of particular note, pyridine 20 bears a 2-Cl handle for diversification via S_NAr, and the successful preparation of furan 23 was made possible by this net-reductive cross-electrophile platform. Importantly, these reductively and oxidatively sensitive systems are traditionally challenging structures in cross-couplings mediated by external photoredox catalysts, further underscoring the selectivity using the EDA paradigm. Finally, comparable reactivity was observed for product 16 using 3.3 mmol (1.0 g) of aryl bromide instead of 0.5 mmol.

Scheme 1. Cross-Coupling Scope: Evaluation of BCP RAEs and (Hetero)aryl Bromides^a.

^aValues refer to isolated yields. Unless otherwise noted, reactions were performed using ArBr (1 equiv, 0.5 mmol), RAE (1.5 equiv, 0.75 mmol), HE (2 equiv, 1.0 mmol), Ni(dtbbpy)Br2 (20 mol %, 0.10 mmol), NaHCO₃ (4 equiv, 2.0 mmol), and dry, degassed DMA (5.0 mL). Irradiation was performed at room temperature using a 390 nm Kessil lamp with fan cooling. ^bThe reaction was performed using 3.3 mmol of ArBr (1.0 g).

In further investigations using 4-((4-bromophenyl)sulfonyl)-morpholine as a standard aryl bromide, the method was demonstrated to be amenable to a wide range of BCP RAEs with varying bridgehead substitutions (24–31). In particular, very few BCP–aryl compounds with amino-(27),^26–29 Cl-(28),⁸ CF₃-(29),³⁰ CN-(30),²⁸ and F-(33)³¹ bridgehead substitutions have been reported, and to our knowledge, none has been prepared via direct cross-coupling. Furthermore, the current method is proven to be more versatile than that of VanHeyst, Qi, and co-workers,¹⁴ who were unsuccessful in employing NHBoc-, CF₃-, and CN-BCP trifluoroborates in Ni/photoredox cross-coupling. Finally, we underscored the utility of the method by engaging the BCP radical with several bromides bearing functionally dense, medicinally relevant structures (37–44). Under the developed conditions, late- stage functionalization of diverse scaffolds can be accomplished, including aryl chloride 37, imidazole 38, quinoline 42, quinazoline 43, and urea 44. Notably, tertiary amines (40 and 42), often present in biologically active substances to modulate pharmacokinetic properties but typically susceptible to SET oxidation with traditional photoredox catalysts, can be accessed, albeit in low yields.³⁴

To lend evidence for the intermediacy of an EDA complex, we measured ultraviolet/visible (UV/vis) absorption spectra for individual reaction components and mixtures thereof (Figure 2B). Although the RAE (violet line) and HE (golden line) absorb in the visible light region, they undergo a bathochromic shift (blue line) when combined, indicating the presence of charge-transfer aggregates. Indeed, the color change is apparent to the naked eye (Figure 2A). The association constant of the EDA complex and analysis via Job’s method demonstrated a 1:1 molar stoichiometry (panels C and D of Figure 2).

Figure 2.

(A) Visual appearance of individual reaction components and mixtures thereof. (B) UV/vis absorption spectra measured in DMA (0.1 M). Aryl bromide, 4-((4-bromophenyl)sulfonyl)morpholine; RAE, 1,3-dioxoisoindolin-2-yl bicyclo[1.1.1]pentane-1-carboxylate; Ni complex, Ni(dtbbpy)Br₂; and HE, diethyl 2,6-dimethyl-1,4-dihydropyridine-3,5-dicarboxylate. (C) Benesi–Hildebrand plot.³² (D) Job plot³³ for a mixture of RAE and HE (0.2 M).

With these results and with insight from our previous report,²⁰ we propose the following mechanism. Visible light excitation of the EDA complex triggers a single-electron transfer (SET) event from HE to the RAE, which fragments to generate phthalimidate, CO₂, and BCP radical III. This BCP radical is then captured by Ni^II oxidative addition complex II. Alternatively, the low-valent Ni⁰ complex I could intercept BCP radical III and then undergo subsequent oxidative addition with the aryl bromide. Both pathways lead to Ni^III intermediate IV, which undergoes reductive elimination to furnish the cross-coupled product. Finally, photoexcited Hantzsch ester reduces Ni^I to the catalytically active Ni⁰ species to close the cycle (Figure 3).

Figure 3.

Proposed cross-coupling mechanism.

In conclusion, we have established a general route toward the synthesis of functionalized bicyclo[l.1.l]pentanes through Ni-catalyzed C(sp³)–C(sp²) bond formation enabled by photoactive EDA complex activation. The developed cross-electrophile protocol evades the need for expensive transition-metal-based photoredox catalysts, preformed organometallics or boronate partners, and stoichiometric metal reductants. Under this photochemical paradigm, the generation of BCP radicals through direct visible-light excitation followed by subsequent cross-coupling with diverse (hetero)aryl halides is feasible. The commercial availability of carboxylic acids, (hetero)aryl bromides, and HE enables the rapid assembly of diverse scaffolds. In addition, the mild reaction conditions facilitate late-stage modification of drug-like molecules with high functional group tolerance. Key spectroscopic studies highlight the necessity for EDA photoactivation for efficient BCP radical generation and subsequent functionalization.