Metallated Aziridines for Cross-Coupling with Aryl and Alkenyl Halides via Palladium Catalysis

Abstract

The palladium catalyzed coupling of an aziridinylzinc chloride intermediate with alkenyl and aryl halides has been demonstrated. The method provides products with retention of aziridine stereochemistry. The utility of the coupling procedure is illustrated in the synthesis of structures related to L-furanomycin.

Lithiated aziridines 1 (Scheme 1, Prot = SO₂R, CO₂R, alkyl) are useful intermediates for the preparation of functionalized aziridines 4.¹ However, transition metal catalyzed coupling reactions involving 1–3 have not been reported to our knowledge. Prior work in our laboratory showed that stannylated aziridine 5 lacks suficient reactivity while lithiated aziridine 6 produces complex reaction mixtures under palladium catalyzed conditions. Herein, we show that Negishi coupling can be achieved in good yields using the protected aziridinylzinc chloride 7 and an improved palladium catalyst.

Scheme 1.

Initial attempts to couple 7 with methyl 4-bromobenzoate using Pd(PPh₃)₄ (THF, 25 °C, 18 h) produced only traces of the desired product. The source of difficulties became clear when a control experiment was conducted under identical conditions, but without any palladium catalyst. Quenching the control with CH₃OD gave a complex mixture containing the deuterated aziridine 9 (37%) as the major component along with variable yields and ratios of imine 10 and the known acrolein imine 11² (ca. 10–30 % of 10 + 11 combined), and more consistently, the enol ether 12 (33%). Yet another product was sometimes isolated from this reaction, apparently resulting from a dimerization pathway, but this product was not formed reproducibly and the yield never exceeded 10%.^3,4 Imines 10 and 11 are thought to arise from intramolecular ring opening of 7 to produce an intermediate 8, with subsequent protonation or elimination of a benzyloxy group to form 10 or 11, respectively.⁵ Because 10 was obtained without any significant deuterium incorporation after chromatographic purification, we suspect that the initial product of CH₃OD quenching is 8 with M = D, and that deuterium/proton exchange as well as conversion to 10 and 11 take place during workup. The origin of the enol ether 12 was also somewhat perplexing, but the reason for the apparent incorporation of methanol became more clear when freshly prepared 7 was quenched with CH₃OD at −78 °C. This afforded 9 and 12 without any of the other products (ca. 1:1 9:12, 75% combined yield after chromatography). Therefore, the other products (10, 11, dimer) result from the thermal decomposition of 7 to 8, while 9 and 12 reflect deuteration or methanolysis directly from 7 without prior decomposition.

Although details of the methanolysis step were not investigated, the conversion to 12 can be understood as an example of zinc-assisted leaving group displacement by CH₃OD to generate 13, followed by 1,2-elimination. Prior examples of analogous displacement have been noted, but they are limited to simple structures such as XZnCH₂X (X = halide) and have not been investigated in depth.⁶ In the present example, we suggest that CH₃OD adds to 7 to generate a transient zincate intermediate, followed by deuterium transfer to nitrogen and internal methoxide transfer to carbon as the aziridine C-N bond is cleaved.

Fortunately, the complex background reactions resulting from the thermal decomposition of 7 at 25 °C could be suppressed by using the more reactive catalyst Pd₂(dba)₃/[(tBu)₃PH]BF₄ in place of Pd(PPh₃)₄ in the Negishi coupling. This simple variation was so effective that none of the undesired products 9–12 were detected if sufficient time was allowed for complete conversion of 7 (ca. 3 h, rt). The scope of coupling was evaluated in several representative examples using a stoichiometry of 1 equiv 5, 1.1 equiv nBuLi, 1.2 equiv ZnCl₂, 1.5 equiv halide, 2.5 mol % Pd₂(dba)₃ and 10 mol % [(tBu)₃PH]BF₄ (Table 2).

The aromatic halides iodobenzene (entry 1) and methyl 4-bromobenzoate (entry 2) gave satisfactory yields of 15 and 17, respectively. The alkenyl halides vinyl bromide (entry 3) or iodoacrylate 20 (entry 4) also afforded coupled products under the standard conditions. Similarly, iodoindole 22 reacted with 7 to produce the desired aziridine 23, although in moderate yield.

To demonstrate the synthetic potential of aziridinylzinc chloride 7, we have studied the conversion of aziridine 21 (Table 1, entry 4) into structures related to the antibacterial agent L-furanomycin 24 and the analogue 25 (Scheme 2).⁷

Table 1.

Scope of the Negishi Coupling Procedure^a

entry	R-X	product	% yield
1	14	15	65
2	16	17	75
3	18	19	64
4	20	21	80
5	22	23	48

^aReaction conditions: 1 equiv 5, 1.1 equiv nBuLi, THF, −60 to −20 °C; 1.2 equiv ZnCl₂, −78 °C; 1.5 equiv halide, 2.5 mol % Pd₂(dba)₃ and 10 mol % [(tBu)₃PH]BF₄, 25 °C.

Scheme 2.

We planned to assemble the dihydrofuranyl moiety of 25 by manipulation of the appended acrylate of 21 via Grignard addition and intramolecular ring opening of the aziridine. Thus, 21 was treated with CH₃MgBr to afford the tertiary alcohol 26 and subsequent acid catalyzed aziridine cleavage gave 27 (52% from 21). Reductive cleavage of the trityl protecting group of 27 with TFA/Et₃SiH⁸ produced the corresponding amine, which was protected as the Boc derivative 28 (61% over the two steps). To avoid the risk of hydrogenating the double bond under H₂/Pd/C conditions, benzyl deprotection of 28 was accomplished using Na/NH₃ reduction to produce the alcohol 29 in 97% yield.⁹ Oxidation of the primary alcohol to the corresponding aldehyde using the Dess-Martin periodinane, followed by Pinnick oxidation gave the protected amino acid 30 (86%). The relative stereochemistry and substitution pattern of 30 was established by comparing NMR data with the enantiomeric structure used earlier as an intermediate in the synthesis of 25,⁷ while the absolute configuration follows from the configuration of 5 and is confirmed by optical rotation comparison at the stage of 30.^1,7

In summary, we have demonstrated that aziridinylzinc chloride 7 is a useful intermediate for palladium couplings with aryl and alkenyl halides. The thermal decomposition of 7, resulting in complex background reactions, can be suppressed by using the more reactive catalyst Pd₂dba₃/[(tBu)₃PH]BF₄. The synthetic utility of the coupling procedure was demonstrated by stereospecific synthesis of the protected dihydrofuranyl amino acid 30.