Tropane and related alkaloid skeletons via a radical [3+3]-annulation process

The study addresses the long-standing challenge of efficiently accessing N-arylated tropane and homotropane alkaloid frameworks, which are core structures in biologically active natural products and medicinal agents. Classical ionic methods (e.g., Robinson’s synthesis) struggle with N-arylated derivatives and have limited scope. The authors hypothesize that a radical-based formal [3+3] annulation, mimicking biosynthesis/Robinson strategies but using 1,3-bis α-amino radical donors and 1,3-bis radical acceptors under photoredox conditions, can provide direct and flexible access to these bicyclic scaffolds under mild conditions with high selectivity.

The paper surveys α-functionalization of tertiary amines via one- and two-electron pathways, noting that overoxidation to iminium ions complicates radical processes when using stoichiometric oxidants. Prior work established photoinduced generation of α-aminoalkyl radicals and their additions to electron-poor alkenes (Mariano, Pandey, Reiser, Nishibayashi), including mono- and bis-alkylations and α-vinylation/allylation of N-aryl amines (MacMillan, Li). Ready reported regioselective α-functionalization with a notable [3+3] adduct under specific conditions. Despite these advances, a general radical [3+3] annulation to form tropane/homotropane skeletons had not been reported. The authors build on photoredox-mediated α-amino radical chemistry, acid co-catalysis benefits, and known redox/acid–base properties of amine radical cations to design a tandem annulation process.

Overall strategy: Visible-light photoredox activation of cyclic N-arylated amines to form α-aminoalkyl radicals that add to ethyl 2-(acetoxymethyl)acrylate, followed by β-fragmentation and a rapid 6-endo-trig cyclization to forge bicyclic tropane/homotropane skeletons. Key is selective deprotonation of aminium radical cations and control of side reactions. Optimization: Model reaction with N-phenylpyrrolidine and various allylic radical traps (Y = SO2Ph, OAc, OPiv, OTFA, S-alkyl, Br) and photocatalysts (Ir-A = [Ir{dF(CF3)ppy}2(dtbpy)]PF6; Ir-B = [Ir(dtbbpy)(ppy)2]PF6; Ir-C = [Ir(dFppy)2(dtbbpy)]PF6; Ru(bpy)3(PF6)2; 4CzIPN; Eosin Y), bases (CsOAc, CsOPiv, NaOAc, Cs2CO3), solvents (1,2-DCE preferred), and 390 nm LEDs. Optimal conditions minimized mono-allylated (2a), α,α′-bis-allylated (4a), and bis-addition (5a) byproducts: Ir-A (1 mol%), CsOAc (1.2 equiv), ethyl 2-(acetoxymethyl)acrylate (1.1 equiv), 1,2-DCE (0.05 M), 390 nm irradiation, ~20 min. Higher dilution and shorter time suppressed over-allylation. Ir-A outperformed others; allyl acetate trap superior to sulfone/sulfide/halide; allyl bromide stalled at 2a. Cyclization studies: Isolated 2a cyclized to 3a under standard photoredox conditions. Acetic acid (1 equiv), generated in situ during allylation, dramatically accelerated cyclization (3 min, 91%); CsOPiv with AcOH gave near-quantitative conversion in ~2 min. Acid additives not needed in the one-pot process due to in situ AcOH formation. Scope: Explored N-arylpyrrolidines (para-substituents), ortho-substituted aryls, and 2-methylpyrrolidine; N-arylpiperidines (homotropanes), N-phenylmorpholine, N-phenylpiperazine, and N-phenylthiomorpholine (no ring-opening observed); azepanes (lower efficiency). A Baylis–Hillman-derived trap (ethyl 2-(1-acetoxyethyl)acrylate) enabled stepwise formation of 2,3-disubstituted [4.4.1] homotropane via E-selective mono-allylation followed by cyclization. Deprotections and downstream modifications: N-p-methoxyphenyl removal via CAN in MeCN/H2O, followed by NaBH4 reduction of quinone and Cbz protection, affording Cbz-protected nortropane/homotropane. N-(p-pinacolboryl)phenyl dearylation via Na perborate then CAN/CbzCl to the same targets. Base-promoted epimerization (NaOEt/EtOH, 40 °C) converted kinetically formed α-ester isomers to thermodynamic β isomers with high dr. Mechanistic and analytical studies: Proposed two photoredox cycles with selective deprotonation to α-amino radicals, intermolecular Michael addition to the allylic trap, β-fragmentation, and fast 6-endo cyclization; final enolate undergoes exo-selective protonation. DFT (ADF; GGA-VWN/BP; TZ basis, no frozen core; vacuo) provided redox potentials and pKa trends for radical cations, rationalizing selective deprotonation and reaction termination at bicyclic stage. Cyclic voltammetry in MeCN (0.1 M [Bu4N][PF6], GC/Pt/Ag–AgCl, evaluated scan-rate dependence; Fc/Fc+ internal standard) assessed oxidation potentials and base effects, supporting deprotonation kinetic limitations for electron-rich anilines. General experimental procedure: In oven-dried vial, combine ethyl 2-(acetoxymethyl)prop-2-enoate (1.1 equiv), N-arylpyrrolidine or N-arylpiperidine (1.0 equiv), Ir-A (1 mol%), CsOAc (1.2 equiv); degas with N2; add dry degassed 1,2-DCE (0.05 M). Irradiate with 390 nm LED (5 cm distance) until completion. Workup with sat. NaHCO3, extract with CH2Cl2, dry, concentrate, and purify by flash chromatography (neutral Al2O3 or SiO2).

- Developed a visible-light photoredox formal [3+3] annulation between cyclic N-arylamines and ethyl 2-(acetoxymethyl)acrylate, forming N-arylated tropane (8-azabicyclo[3.2.1]octane) and homotropane (9-azabicyclo[3.3.1]nonane) skeletons. - Optimal conditions: Ir-A (1 mol%), CsOAc (1.2 equiv), ethyl 2-(acetoxymethyl)acrylate (1.1 equiv), 1,2-DCE (0.05 M), 390 nm LEDs, ~20–90 min. Allyl acetate outperformed sulfone/sulfide/halide traps; other solvents/bases and catalysts were inferior. - Tropane scope: N-arylpyrrolidines gave 3a–3d in 42–57% yield with α/β ≥ 5:1; ortho-substituted aryl (3e) low yield (21%, long time). 2-Methyl substrate furnished 3f in 40% yield, dr 8:1; isolated cyclization from 2f gave 3f in 66% yield. - Homotropane scope: N-phenylpiperidine to 8a in 67% (56% on 3 mmol; dr >20:1). Tolyl variants 8b–8d in 46–61% (dr 20:1). Electron-rich p-anisyl 8e: 32% (long time). Electron-poor aryls 8f–8j in 30–66%, often shorter times; p-ester 8g 65% (1 mmol: 63%, catalyst 0.5 mol%); p-Bpin 8j 66% (3 mmol: 58%, 0.5 mol%). Heterocycles: morpholine/piperazine/thiomorpholine to 8k–8m in ~64% with good dr, no ring-opening. Azepane gave 8n in 25% after 14 h; mono-allylated 7n formed rapidly (65% in 10 min), cyclization limiting. - Baylis–Hillman trap: Stepwise sequence via E-9 (73%, E/Z 7:1; >20:1 after recryst.) to bicyclic 10 (66%, dr >20:1); one-pot yielded mixtures convertible to 10 (71% crude with minor byproduct). - Acid accelerates cyclization: AcOH (1 equiv) or CsOPiv/AcOH gave 2a→3a in 2–3 min vs 3 h without acid (72% → 91–>95%). - Deprotections: N-p-MeOPh removal from 3c/8e via CAN in MeCN/H2O, NaBH4, Cbz protection afforded 12 (90%) and 13 (80%) retaining dr. N-p-BpinPh (8j) dearylation with Na perborate then CAN/CbzCl gave 13 in 71%. - Epimerization: NaOEt/EtOH at 40 °C converted α-ester 3a/8a to thermodynamically favored β isomers with high diastereoselectivity. - Mechanistic insights: Selective deprotonation of aminium radical cations governs formation of α-amino radicals at less substituted sites; bicyclic amine radical cations are significantly less acidic (higher pKa), preventing further functionalization and stabilizing products. CV data show electron-rich anilines have quasi-reversible oxidation and base affects deprotonation propensity. - Double annulation proof-of-concept: From N,N-dimethylaniline (or N-methyl-N-trimethylsilylmethylaniline), a bis-annulated homotropane 14 isolated in 13–22% yield with good stereocontrol.

The results validate a radical formal [3+3] annulation strategy that directly assembles N-arylated tropane and homotropane scaffolds under mild photoredox conditions. The method overcomes limitations of the Robinson synthesis for N-arylated targets and enables high diastereocontrol with an embedded C-3 ester handle for diversification. Mechanistically, the approach hinges on (i) regioselective deprotonation of aminium radical cations to form α-amino radicals at less substituted positions, (ii) rapid intramolecular 6-endo-trig cyclization that outcompetes intermolecular processes, and (iii) termination at bicyclic products due to the comparatively low acidity (high pKa) of bicyclic radical cations, preventing further α-functionalization. DFT-calculated redox/pKa values and cyclic voltammetry support the proposed selectivity and the observation that electron-rich anilines are kinetically limited by deprotonation rather than oxidation. The capacity to scale reactions with reduced catalyst loading and to deprotect N-aryl groups to secondary amines, combined with base-promoted epimerization, underscores utility for medicinal chemistry and alkaloid synthesis.

A general, mild, and diastereoselective photoredox-enabled radical [3+3] annulation provides N-arylated tropane and homotropane cores from simple cyclic anilines and ethyl 2-(acetoxymethyl)acrylate. The method complements classical ionic approaches by delivering N-arylated bicycles directly and placing an ester at C-3 for downstream derivatization. Mechanistic analysis clarifies why the process terminates at the bicyclic stage, contributing to its robustness. Preliminary data indicate the allylic trap can bear alternative electron-withdrawing groups (e.g., nitrile, sulfone, boronic ester). Future work includes expanding trap and amine scope, enantioselective variants to access optically pure targets, and applications to complex, biologically relevant alkaloids.

- Electron-rich N-aryl amines (e.g., p-anisyl) react slowly and require longer times and/or stronger base (CsOPiv), indicating deprotonation of radical cations can be rate-limiting. - Ortho-substituted N-aryl substrates gave low yields and long reaction times (e.g., 3e at 21%). - Azepane annulations to [4.3.1] bicycles are inefficient; 6-endo cyclization is slow, leading to modest yields (8n 25%). - Allyl bromide traps fail to proceed beyond mono-allylation; sulfide traps complicate isolation and produce malodorous thiols. - One-pot sequences with 2-substituted allyl acetates can give mixtures; stepwise approaches may be required. - Double annulation to 14 is low-yielding (13–22%), and attempts to isolate intermediates (piperidine stage) afforded complex mixtures. - Reaction performance is sensitive to solvent/base/catalyst; non-optimized conditions reduce yields or favor side products (2a, 4a, 5a). - Some substrates require prolonged irradiation, and isolation can be complicated by closely related side products.