Nitrosobenzene-Mediated C-C Bond Cleavage Reactions and Spectral Observation of an oxazetidin-4-one Ring System

Abstract

While bond formation processes have traditionally garnered the attention of the chemical community, methods facilitating bond-breaking remain relatively undeveloped. We report a novel, transition metal-free oxidative C-C bond cleavage process for a broad range of ester and dicarbonyl compounds involving carbanion addition to nitrosobenzene. ReactIR experiments on the nitrosobenzene-mediated oxidative decarboxylation of esters indicate the reaction proceeds via fragmentation of a previously unobserved oxazetidin-4-one heterocycle, characterized by an intense IR stretching frequency at 1846 cm^-1. These mechanistic studies have allowed further expansion of this protocol to ketone cleavage reactions of a diverse array of β-ketoester and 1,3-diketone substrates. The conceptual and mechanistic insights offered by this study are likely to provide a platform for further development of bond-breaking methodologies.

Synthetic organic chemistry is concerned with the assembly of complex chemical structures from relatively simple starting materials. Not surprisingly, processes facilitating the formation of new chemical bonds encompass a preponderance of the literature.¹ However, the inverse process of C-C bond cleavage, while synthetically desirable in many cases, poses a great challenge due to the inherent strength of the C-C bond.^2,3 Accordingly, relatively few procedures have been reported for this transformation and those that have usually rely on transition metal catalysis.⁴ In the present work, we describe a powerful, one step oxidative C-C bond cleavage methodology for a broad range of ester and dicarbonyl substrates utilizing nitrosobenzene as oxidant. Moreover, mechanistic studies using ReactIR spectrometry have allowed the first spectral observation of the oxazetidin-4-one ring system. This methodology marks a breakthrough over related multi-step and metal-mediated procedures, providing a highly robust oxidative C-C bond cleavage protocol for a diverse array of carbonyl compounds.^5,6

Given our long standing work in Diels-Alder chemistry, we became interested in C-C bond cleavage methodology in the context of developing a highly versatile asymmetric ketene equivalent.⁷ In preliminary studies toward an asymmetric route to bicyclo[2.2.1]hept-5-ene-2-one derivatives it was found that treatment of N-hydroxy methyl ester 2 with LiOH provided bicyclic ketone 4 in 75% yield over two steps (after acidic hydrolysis of the imines).^8,9 Interestingly, under identical conditions N-hydroxy ethyl ester 5 gave the corresponding product 6 in only 16% yield. By increasing the lability of the ester it was found that direct addition of nitrosobenzene to the lithium enolate of phenyl ester 7 at -78 °C provided the corresponding N-phenyl imines 6, accompanied by evolution of CO₂, within 5 minutes in 91% yield in one step (Scheme 1).

Scheme 1.

Preliminary Data on Oxidative Decarboxylation

The markedly contrasting results obtained between the related ethyl, methyl, and phenyl bicyclo[2.2.1]hept-5-ene-2-carboxylates prompted us to undertake detailed mechanistic studies. As a starting point, the conversion of N-hydroxy methyl ester 2 to N-phenyl imines 3 was examined. Two conceivable reaction pathways may exist for this transformation. Initial hydrolysis of the methyl ester of 2 would afford carboxylate intermediate 9. Subsequent loss of CO₂ and elimination of hydroxide ion would furnish imines 3.¹⁰ More interestingly, the N-hydroxyl group could first be deprotonated to give 10. Intramolecular attack of the methyl ester by the oxy-anion would give the highly energetic, spirooxazetidin-4-one intermediate 11.¹¹ This strained species should spontaneously fragment, expelling CO₂, to give imines 3 (Scheme 2).^12,13

Scheme 2.

Plausible Pathways for Oxidative Decarboxylation

To ascertain if the reaction proceeded through Pathway 1 (viz. intermediate 9), carboxylic acid 12¹⁴ (for preparation see Supporting Information) was subjected to the identical conditions for which imines 3 had been obtained from N-hydroxy methyl ester 2 (Figure 1A). However, under these same reaction parameters 12 gave only a complex mixture of decomposition products. To verify if Pathway 2 was indeed correct, spectroscopic observation of the oxazetidin-4-one functionality of 11 was required. However, due to the slow rate of conversion of 2 to 3, the low concentration as well as the expected short lifetime of 11 would make detection difficult. The observation that phenyl esters are directly cleaved to yield the corresponding imine products by addition of nitrosobenzene at -78 °C suggested that under these conditions observation of the oxazetidin-4-one intermediate was feasible.

Figure 1.

Control and ReactIR experiments were conducted to determine the mechanistic course of the nitrosobenzene-mediated oxidative decarboxylation reaction. (A) Control experiment showing that conversion of 2 to 3 does not follow Pathway 1. (B) 3D and 2D plot of ReactIR experiment showing Pathway 2 is correct with the first spectral observation of an oxazetidin-4-one heterocycle 14. Immediately after addition of nitrosobenzene to the lithium enolate of 7 at -78 °C, oxazetidin-4-one 14, having an IR stretch at 1846 cm-1, is observed. As the cooling bath is gradually warmed, the disappearance of 14 is marked by the appearance of N-phenyl imines 6. Note: The intensity of the line plot of 6 is reflective of only one isomer of 6.

In Situ ReactIR technology was employed to monitor the conversion of phenyl ester 7 to the corresponding N-phenyl imines 6. As can been seen in Figure 1B (3D & 2D plots), less than one minute after addition of nitrosobenzene to the lithium enolate of 7 a new, sharp peak appears at 1846 cm^-1. The intensity of this peak gradually decreases as the cooling bath is warmed to -20 °C. Importantly, the disappearance of the 1846 cm^-1 band is marked by the appearance of two bands at 1695 and 1679 cm^-1 corresponding to N-phenyl imines 6. We believe the new IR absorption at 1846 cm^-1 is attributable to the existence of spiro-oxazetidin-4-one intermediate 14. IR data of similar cyclic compounds reveals that a stretch of 1846 cm^-1 is highly characteristic of this type of 4-membered, spiro ring system.¹⁵ Thus, we can now conclude that the conversion of 7 to 6 follows a mechanistic course analogous to Pathway 2. While oxazetidin-4-ones have been postulated as reaction intermediates in only a handful of synthetic and theoretical papers^12,13,16, direct evidence for the existence of this functional group has hitherto been unreported.¹⁷

The reaction scope of this novel process was subsequently examined. Table 1 summarizes these results. The present method tolerates a range of bicyclic, cyclic, sp²-sp³, and sp²-sp² α,α’-di-substituted phenyl esters to provide either ketone or imine products in generally excellent yields. In general, ketimines containing at least one sp² α-subsitutent were stable to silica gel while those having only sp³ α-subsitutents (entries 1-3) were immediately hydrolyzed. Further, malonic ester derivatives could also be applied to provide the corresponding α-imino phenyl esters in uniformly high yields as well.

Table 1.

Nitrosobenzene-Mediated Oxidative Decarboxylation^a

Entry	Substrate	Product	Yield (%)c
1			91d
2			96d
3			92
4			98
5			96
6			98
7			97
8			96
9			83
10			92

^aProducts in entries 4-10 were hydrolytically stable to SiO₂. See Supporting Information for additional details.

^bBase = LDA (lithium diisopropylamide) or NaHMDS (sodium bis(trimethysilyl)amide).

^cIsolated Yield.

^dDetermined by ¹H NMR with MeNO₂ as internal standard.

Unexpectedly, when β-keto phenyl ester 15 was subjected to the standard reaction conditions, a ~1:1 mixture of α-imino-ester 16 and α-keto-imine 17 was obtained, presumably from attack of the aminooxy-anion at either the ketone or ester functionalities (Scheme 3). Guided by our previous findings, use of the corresponding less labile β-keto methyl ester 18 resulted in exclusive cleavage of the ketone moiety furnishing α-imino-methyl ester 19 in 93% yield. Gratifyingly, this process could be applied to ring-opening cleavage reactions of a diverse array of mono and bicyclic β-ketoesters to afford highly functionalized dioxo (acidic workup) or oxo-imino acids (workup under buffered conditions) in quantitative yields (Table 2, entries 1-9). Further, a simple extension of this ketone cleavage methodology to symmetric 1,3-diketone substrates allows for a regioselective route to α-ketoimines (Table 2, entries 11-13). It should be noted that in both oxidative decarboxylation and ketone cleavage reactions spectroscopically pure compounds can be obtained after simple aqueous workup procedures obviating the need for chromatographic purification.

Scheme 3.

Initial Results for Cleavage of β-Ketoesters

Table 2.

Nitrosobenzene-Mediated Ketone Cleavage^a

Entry	Substrate	Product	Yield (%)b
1			>95%
2			>95%
		c	85%
3			>95%
4			>95%
5			>95%
6			>95%
7			>95%
8			>95%
9			>95%
10			75% (R=Me) 83% (R=Ph)
11			83%
12			83%
13			84%

^aSee Supporting Information for details.

^bProducts in entries 1-9 were isolated after acidic workup as their crude acids which were >95% pure by ¹H NMR. Products in entries 10-13 were isolated after column chromatography.

^cIsolated after workup with pH 6.5 citric acid/phosphate buffer. Contains ca. 7% of the corresponding dioxo acid.

In summary, we have developed an efficient single step oxidative decarboxylation reaction of esters involving nitrosobenzene addition to the corresponding enolate. A series of control and spectroscopic experiments have elucidated the mechanism of this novel cleavage process, providing the first spectral evidence for the existence of an oxazetidin-4-one ring system. Based on these findings, this methodology could be generalized to cleavage reactions involving a wide scope of dicarbonyl compounds affording highly functionalized products in excellent yields and in short reaction times. Current work in our laboratory is aimed at expanding the full scope and applications of this new C-C bond cleavage methodology. We believe the findings from the present study will provide impetus for further research into alternative, metal-free bond fission methodologies.

Supplementary Material

1_si_001. Supporting Information Available.

Experimental procedures, X-ray crystallographic data, and spectral data. This material is available free of charge via the internet at http://pubs.acs.org.