Baeyer-Villiger Rearrangement of a Substituted Pyrrole by Oxone

Abstract

Pyrroloxyls have been reported to exhibit very narrow EPR spectral lines, essential for in vivo imaging. En route to pyrroloxyls, we observed an unexpected Baeyer-Villiger rearrangement, leading to loss of aromaticity and formation of a 4,5-dihydro-1H-ketopyrrole.

Introduction

The advent of low-frequency electron paramagnetic resonance (EPR) spectrometers ¹ that can detect and image paramagnetic species in animals in real time creates an urgent need to synthesize stable free radicals (so-called “spin probes”) that can report in vivo physiology (e.g., O₂ concentration in various tissues). ² Owing to their relative ease of synthetic manipulation, nitroxides are attractive as spin probes for physiological EPR spectroscopy and imaging. ³

The utility of nitroxides as EPR imaging probes has been limited by their relatively large spectral linewidth (~1 G), which impacts both imaging sensitivity and spatial resolution. In 1970, Ramasseul and Rassat⁴ described the synthesis of pyrroloxyls that exhibited narrow EPR spectral linewidths in deoxygenated organic solvents (linewidth of one deuterium-substituted pyrroloxyl was reported to be ~0.1 G⁵). In view of their remarkable narrow linewidths, we wished to investigate the potential of pyrroloxyls as in vivo EPR imaging agents. For our initial series of studies,⁶ we decided to prepare 2,5-di-tert-butyl-3-ethoxycarbonyl-1-pyrroloxyl 5a, using methods described by Ramasseul and Rassat.⁴ To our surprise, nickel peroxide oxidation of 4 afforded two products, neither of which was described by Ramasseul and Rassat: (a) 2,5-di-tert-butyl- 3-ethoxycarbonyl-4-hydroxy-1-pyrroloxyl 5b, which is readily oxidized to nonparamagnetic 7 and (b) the biradical 2,2′,5,5′-tetra(tert-butyl)-4,4′-bis(ethoxycarbonyl)-3,3′-bipyrrolyl-1,1′-dioxyl 6 (Scheme 1).⁶ Despite the enhanced stability of biradical 6 as compared to pyrroloxyl 5b, its dimeric nature makes it less than optimal as a contrast agent for in vivo EPR imaging. We therefore sought alternative nitroxides that might exhibit narrow EPR spectral lines. We decided to investigate the synthesis of more water-soluble analogs of the reported pyrroloxyls⁴ with the expectation that they would also exhibit narrow linewidths.

Scheme 1.

a) NaOEt, b) NH₂OH, c) NiO₂, d) O₂

Results and Discussion

Based on our earlier publication,⁶ it is clear that all positions on the pyrrole ring must be substituted prior to the one-electron oxidation of N–OH to N–O•. We considered that preparation of di-ester 8, followed by cyclization with hydroxylamine would afford the desired pyrrole 9 (Scheme 2). A careful review of the literature⁵ revealed that 9 was indeed accordingly obtained, but only in about 5% yield. In our hands, however, attempts to prepare 9 in an identical fashion were unsuccessful.

Scheme 2.

a) Na/Et₂O, b) NH₂OH

An alternative approach was sought. We reasoned that since 4 can be readily prepared in reasonable yield,^4,6 initial protection of the N–OH followed by the Villsmeier-Haack reaction might result in the corresponding formylpyrrole 11. Oxidation of 11, followed by removal of the protective group, should lead to 9. However, after O-benzylation of 4 to yield 10, classic Villsmeier-Haack reaction conditions⁷ did not result in formylation (only 10 was recovered). By optimizing reaction conditions—1 equivalent of 10 in 5 equivalents of N-methyl-N-phenylformamide/ POCl₃ (solvent-free), 50°C, 3 h, followed by hydrolysis with aqueous sodium acetate—we obtained 11 in acceptable yield (Scheme 3).

Scheme 3.

a) Bn-Br, K₂CO₃, DMSO (70% yield); b) HCON(CH₃)C₆H₅/POCl₃ (47% yield). Bn = benzyl

While there are various methods for oxidizing aldehydes to acids and esters, the recent procedure of Travis, et al,. ⁸ wherein Oxone^® was used to convert aryl aldehydes to the corresponding ethyl esters, seemed an attractive approach to pyrrole 12. One potential problem was that electron-rich molecules, such as 4-hydroxybenzaldehyde, can also undergo the Baeyer-Villiger reaction, resulting in a formate ester, which upon hydrolysis leads to the corresponding phenol.⁸ Because pyrrole 11 is not electron-rich, we were optimistic that the mild experimental conditions described in Travis, et al.,⁸ might favor oxidation to the desired ester and not a Baeyer-Villiger rearrangement.

When 11 and Oxone^® (2:1 molar ratio of KHSO₅ to substrate) were stirred in absolute ethanol at room temperature for 16 h, no reaction occurred. When the ratio of KHSO₅ to 11 was increased to 6, and the reaction was vigorously stirred at room temperature for 3 days, TLC analysis indicated the formation of a new compound. The ¹H-NMR spectrum of the isolated product showed multiple resonances inconsistent with the highly symmetrical structure of 12. The X-ray crystallographic structure⁹ of the isolated product revealed a rearrangement of 11 by Oxone^®, resulting in 13 rather than the predicted pyrrole 12 (Scheme 4). The presence of a chiral center in 13 implies the generation of stereoisomers. Indeed, 13 crystallizes as a racemate, with an asymmetric unit comprising a pair of enantiomers (Fig. 1).

Scheme 4.

a) Oxone/EtOH (48% yield). Bn = benzyl

Figure 1.

Enantiomers of 13 constituting the asymmetric unit in the X-ray crystallographic structure.⁹

We speculated that 11 could have undergone a Baeyer-Villiger-type oxidation mediated by Oxone^®, with subsequent rearrangement to 13. To gain further insight into the mechanism underlying the formation of 13, we changed experimental conditions to avoid using a protic solvent that could act as a nucleophile: reaction with 3 molar equivalents of Oxone^® in DMF for 16 h at room temperature transformed 11 into a new product, 16. The ¹H-NMR spectrum of 16 suggested the presence of a formate ester. The structure of 16 was determined by X-ray crystallography (Fig. 2),¹⁰ and is seen to be the formate ester expected from Baeyer-Villiger oxidation of 11 (Scheme 5).

Figure 2.

Formate ester 16 formed from Baeyer-Villiger oxidation of 11.

Scheme 5.

a) Oxone/DMF (65% yield). Bn = benzyl.

Oxone is an acidic triple salt comprising potassium peroxymonosulfate, potassium hydrogen sulfate, and potassium sulfate (2KHSO₅·KHSO₄·K₂SO₄). Therefore, if formate ester 16 did form in the ethanolic Oxone reaction, acid-catalyzed transesterification with ethanol could have removed the formyl group to yield a hydroxypyrrole which, in turn, could have undergone rearrangement to yield the dearomatized compound 13. To investigate this possibility, formate ester 16 was vigorously stirred with a mixture of anhydrous KHSO₄ and Na₂SO₄ in absolute ethanol at room temperature for 40 h. The sole isolated product was hydroxypyrrole 17 (Scheme 6). Thus cleavage of the formyl group did not trigger subsequent rearrangements under acidic, non-oxidizing conditions in anhydrous ethanol. In view of the last finding, we surmise that in the presence of Oxone in ethanol, formate ester 16 and/or hydroxypyrrole 17 must undergo further oxidative dearomatization and addition to generate pyrrolinone 13.

Scheme 6.

Conditions: KHSO₄, Na₂SO₄, EtOH (62% yield). Bn = benzyl

To our knowledge, this is the first time that a hindered and fully substituted pyrrole has undergone a Baeyer-Villiger reaction to yield a product which, in the presence of an appropriate solvent, rearranges with loss of aromaticity. Attempts to oxidize 13 to the corresponding nitroxide using typical oxidants such as m-chloroperbenzoic acid, hydrogen peroxide/sodium tungstate and dimethyldioxirane^11,12 all failed, presumably owing to 13 being highly hindered.¹¹