Direct Coupling of Pyrroles with Carbonyl Compounds: Short Enantioselective Synthesis of (<i>S</i>)‐Ketorolac

Abstract

An odd couple: The union of pyrroles and carbonyl compounds (ketones, amides, esters, lactones, lactams, see scheme) is described, and by the use of an intramolecular variant of this new method, the nonsteroidal, anti-inflammatory drug (S)-ketorolac has been prepared in a short enantioselective synthesis. Mechanistic underpinnings of this reaction, which couples unfunctionalized C(sp2) and C(sp3) atoms, are also discussed. Pyrroles and indoles are the foundation of innumerable medicines, natural products, and synthetic materials.1 Alkylations, acylations, and transition-metal-mediated couplings are the staple reactions for intermolecular CC bond formation between these heterocycles and other organic substrates.1–3 Typically, organometallic couplings require that one3 or both entities are prefunctionalized (i.e. halogenation or any other disposable functionality).2 Simple and practical methods that eliminate this prerequisite in synthetic design are rare.4 Inspired by a family of naturally occurring indole alkaloids, we recently reported a remarkably simple method for the direct coupling of the C-3 carbon of indoles with the α-carbon of carbonyl compounds as shown in Figure 1.5 We imagined that these results could be extended to the analogous coupling of pyrroles (at C-2) with carbonyl compounds. Herein, we demonstrate the coupling of pyrroles with ketones, esters, amides, lactones, and lactams. An intramolecular variant of this methodology has also been developed and applied to a concise, enantioselective synthesis of the nonsteroidal anti-inflammatory drug (S)-ketorolac (1) that can be performed without the need for protecting groups. The underlying mechanisms of these processes are also discussed. Direct coupling of indoles and pyrroles to carbonyl compounds and retrosynthetic analysis of (S)-ketorolac. Whereas the locus of nucleophilicity of the indole anion is at C-3, it resides at C-2 for the pyrrole anion,1 and it was expected that direct coupling would link pyrroles to carbonyl compounds at that site. By slight modification of our previously reported conditions,5 this proved to be the case as shown in Scheme 1. As with most electron-rich pyrroles, the products shown in Scheme 1 are quite sensitive to air and light. Nevertheless, both the pyrrole and indole5 couplings can be performed on a multigram scale (>100 mmol). Further, complex pyrrolidines, pyrrolidinones, and pyridines are accessible from the pyrrole adducts.1 A range of substitution patterns are tolerated on the pyrrole subunit (3–6) and, as with indoles,5 highly congested pyrroles fused at a quaternary center are easily prepared with complete diastereocontrol (2). Sultam 7 was prepared as a 14:1 mixture of diastereomers, and the structure was confirmed by X-ray crystallographic analysis (colorless needles (1:1 cyclohexane/Et2O), mp 131–133 °C). Although lactones (8) and lactams (10) are easily joined with pyrroles, esters (9) generally couple in lower yield (see Supporting Information for details), and pyrroles that bear electron-withdrawing groups either do not couple or do so in low yield. However, efforts to overcome these limitations are underway. Preparation of pyrrole–carbonyl compounds (LHMDS=lithium hexamethyldisilazide). Yields (%, isolated after chromatography unless otherwise stated) and d.r. values of the products are indicated in parentheses. [*] Yield based on recovered starting material. As shown in Figure 1, the pyrroleacetic acid scaffold is found in the widely marketed pharmaceutical agents tolmetin and ketorolac and the previously marketed agent zomepirac. Our retrosynthetic analysis of ketorolac (1, Toradol and Acular) was influenced strategically by the pioneering syntheses patented in the literature6, 7 and by the fact that the S antipode of ketorolac is more potent and causes fewer side effects.8 Our aim was not to improve upon the extremely efficient and practical five-step Syntex route7 (≈45 % yield from pyrrole, racemic); an enantioselective synthesis9 of 1 would merely serve as an ideal proving ground for the versatility of the current method. Our synthesis (Scheme 2) commences with pyrrole acid 11, which results from the near-quantitative union of pyrrole with butyrolactone on a multigram scale.10 The stage was set for an intramolecular pyrrole–carbonyl coupling after installing the chiral auxiliary to afford 13 (Et3N, MeOCOCl, then 12, 100 %).11 In the event, we were unable to achieve the oxidative annulation of 13 with many standard oxidants (CuII, FeIII, AgI, AgII, TiIV, MnIII, CeIV). After considerable exploration, ferrocenium hexafluorophosphate (14, a practical, recyclable, and commercially available oxidant)12 was found to elicit the cyclization of 13 to 15 in 65 % yield (based on recovered starting material and determined by 1H NMR spectroscopy) as a 4.5:1 mixture of diastereomers. The isolated yield of 15 was about 35 %, however, as 15 was quite sensitive to air and moisture, so the crude reaction mixture that contained 13, 15, and ferrocene was carried forward without purification (benzoyl chloride, 70 °C, remaining 13 and ferrocene easily separable).13 Hydrolysis of the resulting benzoylated pyrrole by using tetrabutylammonium hydroperoxide14 furnished S-ketorolac (1, 90 % ee determined by chiral HPLC, 38 % isolated yield over 2 steps) along with recovered auxiliary. From the vantage point of synthetic design, certain details are worth noting: 1) the oxidation state of 11 is conserved (reduction, decarboxylation, and halogenation processes avoided),15 2) protecting groups are absent, 3) decent stereocontrol is observed in the ring closure despite the readily enolizable9 nature of the newly formed stereocenter, and 4) overall brevity of the sequence (≈25 % overall unoptimized yield from pyrrole in four operations). Short, enantioselective synthesis of (S)-ketorolac. Reagents and conditions: a) Et3N (1.1 equiv), MeOCOCl (1.0 equiv), THF (0.1 M), 0 °C, 1 h; then 12, 100 %; b) LHMDS (1.2 equiv), Et3N (2.0 equiv), THF (0.01 M), −78 °C, 30 min; then 12 °C, 14 (0.75 equiv), 5 min, 65 % bsm; (c) BzCl, 70 °C, 4 h; then TBAH (2.0 equiv), H2O2 (2.0 equiv), 2-methylbut-2-ene (3.0 equiv), DME (0.25 M), −10 °C, 3 h, 38 %. Xc=camphor sultam auxiliary, bsm=based on recovered starting materials, Bz=benzoyl, TBAH=tetra-n-butylammonium hydroperoxide, DME=1,2-dimethoxyethane. Although conceptually similar, the direct coupling of carbonyl compounds with pyrroles (Scheme 1) by using a CuII oxidant probably differs mechanistically from the intramolecular cyclization (13→15) by using FeIII-based 14. As shown in Figure 2, we believe that in the former case an intermediate CuIII-chelated species (A) may be involved.16 Reductive elimination and loss of CuI should lead to B, followed by tautomerization to yield the product. Several observations support this tentative mechanistic model: 1) dimerization of the pyrrole is never observed, as expected from geometrical constraints, 2) N-protected pyrroles do not react, 3) only 1 equivalent of oxidant is necessary, although the use of 1.5 equivalents gives a slight improvement in yield, and 4) the characteristic red-brown color of copper(I) salts is often observed at the end of the reaction. The same trends are seen for the analogous coupling with indoles and implies that a similar mechanism may be active.5 To effect the conversion of 13 into 15, oxidant 14 was employed because it has been firmly established by the scholarly studies of Jahn et al. to convert enolates into radical species by an outer-sphere single-electron-transfer pathway.12 Remarkably, FeIII-based oxidant 14 is ineffective for the couplings in Scheme 1 as are CuII-based oxidants for the annulation (13→15). Proposed mechanism for the direct coupling of pyrroles with carbonyl compounds by using CuII. In summary, we have developed a new method for the one-step construction of a variety of pyrroles that would ordinarily require multistep sequences to synthesize. The method is scalable, practical, and reliable. The intermolecular heteroarylation of enolates (Scheme 1) uses a CuII-based oxidant, whereas the intramolecular variant proceeds by a different mechanism that functions using the FeIII-based oxidant 14 (Scheme 2). The waste associated with protecting groups, prior modification of the substrates (such as halogenation or any other disposable functionalities), and expensive metals is eliminated. These studies point to a unique approach for the synthesis of π-electron-rich heterocyclic systems through coupling of unfunctionalized C(sp2) and C(sp3) atoms.17 General Procedure for Direct Pyrrole Coupling: The carbonyl compound (0.25 mmol) was dissolved in benzene (1.0 mL) and the solvent was removed in vacuo. Pyrrole (0.75 mmol) was then added and the starting materials were dissolved in THF (8.0 mL). The solution was cooled to −78 °C and a solution of LHMDS (0.50 M, 2.0 mL) was added. The reaction mixture was allowed to stir for 30 min, after which time the septum was removed and copper(II)-2-ethylhexanoate (131 mg, 0.38 mmol) was rapidly added as a solid and then the septum was replaced. The reaction was allowed to warm to −60 °C and stirred for 3 h. The reaction was subsequently warmed slowly to ambient temperature and quenched by pouring into 5 % aqueous NH4OH (15 mL). The aqueous layer was partitioned with EtOAc (20 mL). The organic layer was separated and washed successively with water (15 mL) and then brine (15 mL), dried (MgSO4), and filtered, and the solvent was removed in vacuo. Flash chromatography (silica gel) of the crude reaction mixture afforded pure coupled product. With the exception of compound 2, all the pyrroles began to darken in color immediately after concentration, but NMR spectroscopic analysis showed no loss in purity. In general it was found that the couplings of ketones, amides, lactones, and lactams worked well with the exception of esters, which proved to be more unpredictable. Supporting information for this article is available on the WWW under http://www.wiley-vch.de/contents/jc_2002/2005/z462048_s.pdf or from the author. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.