Multicomponent Macrocyclization Reactions (MCMRs) Employing Highly Reactive Acyl Ketene and Nitrile Oxide Intermediates

Abstract

An efficient synthesis of spiro-fused macrolactams by a multicomponent macrocyclization reaction (MCMR) is reported. The use of highly reactive, transient intermediates in this MCMR permits short reaction times, even at high dilution. The methods employed for this MCMR were first developed as a four component strategy for the synthesis of β-ketoamide isoxazolines.

Macrocycles occur widely in nature and have important applications in medicine, material science, and supramolecular chemistry.¹ Their intrinsic three-dimensional geometry is analogous to tertiary protein structure, lending to site-specific recognition, and macrocycles constitute a unique class of privileged scaffold.^1a,2 Despite the promise of macrocycles, they are under-exploited in libraries for screening;³ one possible cause is the limited availability of macrocycle diversification chemistry. An area of study that begins to address this issue is macrocycle forming reactions involving the union of multiple reactive components, pioneered in the laboratories of Zhu,⁴ Wessjohann,⁵ Severin,⁶ and Luis.⁷ Herein, we demonstrate a unique multicomponent-macrocyclization reaction (MCMR), which derives from a new multicomponent reaction (MCR).

While acyl ketenes provide a powerful ring closing strategy in the synthesis of macrocycles (the Boeckman approach),⁸ we are unaware of any multicomponent reactions forming macrocycles with this methodology. As depicted in Scheme 1, we envisioned macrocyclization occuring by intramolecular condensation of a primary amine with an acyl ketenes. The resulting macrolactam, is poised to condense with a formaldehyde equivalent, yielding a methylene derivative that reacts as a dipolarophile in a KHCO₃-initiated nitrile oxide (from a chlorooxime) 1,3-dipolar cycloaddition and would yield a spiro-fused macrolactam product. This MCMR strategy could potentially access a range of macrocycles that complement those made with previously reported strategies.^4–7

Scheme 1.

Multicomponent macrocyclization plan

Spiro-heterocycles and isoxazolines are important biorelavent scaffolds that have been extensively used in the synthesis of natural products.⁹ However, there are few examples of MCRs leading to spirocyclic products¹⁰ and none, to our knowledge, including a macrocyclization step. Our long-standing interest in 1,3-dipolar cycloadditions prompted development, and report herein, of the novel 4CR summarized in Scheme 2.

Scheme 2.

β-Ketoamide-based 4CRs

MCRs are attractive because they enable formation of complex structures while minimizing synthetic effort – reducing synthesis length and optimizing atom economy.^3b,11 This 4CR strategy (Scheme 2) allows, in principle, for the incorporation of four independently variable components. Both aryl- (e-donating, e-withdrawing, and heterocyclic) and carboalkoxy functionalized nitrile oxides are tolerated. Aromatic and alky (1° or 2°) amines can be used in the amide bond-forming step. The keto R-group (−R = −CH₃ or – C₆H₄OCH₃) in the final product is easily modified by exploiting dioxinone 1 (Scheme 2a) or Meldrum’s acid analog 11 (Scheme 2b; prepared by EDC coupling of a carboxylic acid with Meldrum’s acid) as the acyl ketene precursor. Importantly, we have found that the acyl ketene’s high reactivity toward the amine reactant effectively out-competes potential side reactions between the amine and chlorooxime reactants.

We also investigated a 3CR variant of this 4CR employing commercial β-ketoamide 18 (Scheme 3a). The yield of this 3CR was significantly improved over the 4CRs in Scheme 2 (85% compared to yields of 36–50%). Attempts at using benzaldehyde in place of Eschenmoser’s salt in this 3CR gave condensation product 22 (19:1 undefined alkene selectivity) but no cycloaddition to 23 (Scheme 3b). Condensation with Eschenmoser’s salt delivers effective MCR dipolarophiles (Scheme 2), whereas the dipolarophile derived from benzaldehyde (e.g., 22) fails to yield the multicomponent product.

Scheme 3.

Beta-ketoamide 3CRs

Having thus established these 3 and 4CR, we next investigated the corresponding MCMR. Intermediates 28a–c (Scheme 4) were prepared by opening anhydride 25 with carbamate-protected amino-alcohol 24a (P = Cbz) or diamines 24b/c (P = Boc) to give 26a–c. Meldrum’s acid was next coupled to these carboxylic acids using EDC. Attempts at macrocyclization from 27a gave hydrolyzed [e.g., ~C(=O)CH₂CO₂H] then decarboxylated [e.g., ~C(=O)CH₃] product – even with rigorous drying as analogously reported by Hoye.^8d Removal of the nitrogen protecting group in 27a–c delivers zwitterions 28a–c, which we hoped would avoid hydrolysis/decarboxylation problems.

Scheme 4.

Synthesis of macrocycle precursors 24a–c

With 28a in hand, we addressed the challenge of forming a large ring from a 5-acyl Meldrum’s acid/amine-based precursor¹² with minimal conformational bias toward cyclization (Scheme 5). We were pleased to find that macrocyclization 28a → 29a proceeded in excellent yield (94%) portraying the MCMR potential of this ring closing strategy. We explored both Cbz and Boc protecting group strategies and found that Boc deprotection with SnCl₄ proceeds more cleanly than Cbz hydrogenolysis; that said, both strategies accommodate the Meldrum’s acid moiety. As with 28a, zwitterion 28b cyclizes in excellent yield (28b → 29b in 92%). We speculate that zwitterions 28a and b form an internal ion pair taking on geometries that resemble the respective macrocyclic products. This preorganization perhaps aids ring closure, explaining in part these high macrocyclization yields. Additionally, deprotonated Meldrum’s acids are reported to be thermally stable¹³ and decomposition of 28a and 28b occurs at >180 °C and >220 °C, respectively. However, in solution, proton transfer presumably occurs allowing for effective cycloreversion. We are unaware of any examples of 5-acyl Meldrum’s acid-based macrolactamizations.¹²

Scheme 5.

12-Membered macrocycle-forming reactions

Our MCMR results are delineated in Scheme 6. Most macrocyclizations must be carried out under high dilution or with slow addition of a reactant/reagent to avoid bimolecular or oligomerization reactions.¹⁴ In contrast, MCRs are often carried out at high concentration to overcome entropic barriers.^11c,11d,15 This portends a fundamental dichotomy when incorporating a macrocyclization within an MCR. In fact, our MCMR addresses this issue by employing highly reactive intermediates – e.g., transient acyl ketene and nitrile oxide species – at concentrations of 0.0125 M. The reactive acyl ketene intermediate enables the MCMR to proceed effectively at this concentration, while still proceeding within a short reaction time. This methodology further benefits from microwave-assisted conditions.¹⁶ Collectively, the MCMRs depicted in Scheme 6 show the facile formation of spiro-fused macrolactam products with 12- and 14-membered ring macrocycles and considerable structural diversity. We also note that the yields are very good (70–74%) – indeed, much better than the yields of the 4CRs depicted in Scheme 2 (36–50%) and comparable to the 3CR yield reported in Scheme 3 (85%).

Scheme 6.

Four related MCMRs

With the potential for two regioisomers in the 1,3-dipolar cycloaddition, the selectivity of these nitrile oxide cycloadditions were confirmed by X-ray crystallographic analysis (Figure 1). The products depicted in Schemes 2/3/6 are the only regioisomers observed.

Figure 1.

Crystal structures of (a) 3CR 20 and (b) MCMR 30.

The presumed MCMR sequence of events from Meldrum’s acid analog 28a (Scheme 7) begins with a proton transfer step to 28a’ as the literature suggests that deprotonated 5-acyl Meldrum’s acids will not form an acyl ketene.¹³ Thermal cycloreversion of 28a' forms the acyl ketene, which subsequently condenses with the ω-1°- amine. The amide-enol form of the b-ketoamide then condenses with Eschenmoser’s salt to give the unsaturated ketoamide allowing for a sequence-ending 1,3-dipolar cycloaddition.

Scheme 7.

Preposed MCMR sequence to a spiro-fused macrocycle.

These results constitute a new MCMR protocol resulting in a facile route to spiro-fused macrolactams wherein macrolactamization and a new multicomponent reaction have been successfully co-joined in a single transformation. Entropic barriers are overcome at dilutions of 0.0125M and reaction times are reduced by using highly reactive, transient intermediates.