Cascade Synthesis of (E)-2-Alkylidenecyclobutanols

Abstract

A facile, one-pot reaction cascade condenses 1,1,1-trichloroalkanes with α,β-unsaturated ketones to unexpectedly furnish moderate to good yields of (E)-2-alkylidenecyclobutanols

In recent years, our laboratories¹ and others² have introduced an assortment of organochromium reagents and exploited their unique physical/chemical properties for access to a wide range of natural products and high value targets.³ In continuation of these studies, we sought to extend the utility of select chromium reagents via in situ transmetalation and subsequent reaction with electrophiles. In one such example, chromium carbenoid 2 was generated from 1,1,1-trichloroalkane 1 using excess anhydrous CrCl₂, except both copper cyanide and an α,β-unsaturated ketone 3 were present. We anticipated the (E)-vinylchromium(III)-intermediate 2 would undergo transmetalation and subsequent 1,4-conjugate addition with 3. Unexpectedly, however, (E)-2-alkylidenecyclobutanol 4 was isolated as the major product in moderate to good yields (Scheme 1).

Scheme 1.

Synthesis of (E)-2-Alkylidenecyclobutanols

Alkylidenecyclobutanols, and the cyclobutanols which are readily derived from them, appear as substructures⁴ in many architecturally interesting and/or bioactive natural products.⁵ They also display unique reaction manifolds that make them useful as synthetic intermediates.⁶ Access to these strained ring systems is generally restricted to [2+2]-cycloadditions,⁷ ring expansions or contractions of the corresponding homologs,⁸ Wittig ⁹ and, to a lesser extent, via intramolecular alkylations.¹⁰

To better understand the implications of this unusual cascade reaction, we investigated its scope and possible mechanism and report our findings herein. The reaction parameters were systematically optimized using 1,1,1-trichloroalkane 5, α,β-unsaturated ketone 6, CrCl₂ (6 equiv) and CuCN (1.2 equiv) as the benchmark system. Yields of 7 were best in THF (Table 1, entry 1), somewhat lower in DME, CH₃CN and dioxane, and poor in DMF, HMPA, DMSO and EtOAc. The reaction was also highly dependant upon the copper salt. CuCN was superior to all others for producing alkylidenecyclobutanols; little, if any, 7 or conjugate addition was observed with CuI, CuBr, CuCl, PhSCu, or CuTc whereas CuOTf gave a 35% yield of the 1,4-adduct 28, but no alkylidenecyclobutanol (Scheme 2). Adjuvants, e.g., NiCl₂, BF₃·Et₂O, and KCN were likewise unhelpful as were higher (70 °C) or lower (4 °C) reaction temperatures. The amount of CrCl₂ could be reduced from 6 equiv to 1 equiv using Mn(0) powder as a regeneration agent,¹¹ although the yield of 7 declined to 24%. Substoichiometric amounts of CuCN also led to significantly lower yields.

Table 1.

Synthesis of (E)-2-Alkylidenecyclobutanols^a

^asee footnote ¹² for general procedure

Scheme 2.

1,4-Conjugate Adduct

Both allylic 8 (entries 2 and 3) and benzylic 12 (entry 4) trichloroalkanes behaved analogously to 5 and afforded adducts 9, 11, and 13, respectively, from ketones 6 and 10.¹² Importantly, the cascade was compatible with silyl ether 14 (entries 5–8), electron-rich napthalene 16 (entry 6), and even the aryl bromide 18 (entry 7). X-ray analysis (see Supporting Information) of adduct 17, following desilylation, confirmed its identity and the E-olefinic geometry. The latter was a key insight that must be accommodated by any proposed annulation process (vide infra).

It should be noted that benzylic trichloromethylcarbinols, e.g., 20 (entry 8), which are readily prepared from aldehydes, were also suitable precursors for the casacde, albeit with slightly diminished yields of adduct.¹³ Addition to α-substituted α,β-unsaturated ketone 22 proceeded smoothly to furnish 23 as a 1:1.8 diastereomeric mixture (entry 9) and, notably, the polymerization-prone exocyclic ketone 24 was transformed into fused bicyclic 25 (entry 10). In contrast, analogous efforts using the β-substituted analogue 26 (R = Me, Ph) failed to give any 27 (entry 11).

While the mechanistic details remain undefined at present, we speculate that one-electron reduction of enone 3¹⁴ to enol radical 29 occurs concurrently with the production of α-halovinylidene chromium carbenoid 2 (Scheme 3).^1j Subsequent copper mediated Kharasch-type addition¹⁵ and loss of copper chloride from the resultant adduct 30 deliver (E)-vinylchromium 31. Enol quench, perhaps by the previously identified internal proton return process^1j,16 or adventitious water, gives 32 from which 4 is obtained by intramolecular ketone vinylation.¹⁷

Scheme 3.

Proposed Mechanism

In summary, we have demonstrated a convergent, (E)-selective synthesis of 2-alkylidenecyclobutanols based upon mechanistically unique, synergistic chemistry not achievable using either CrCl₂ or CuCN alone.