www.acsprf.org

Mannich bases (b-aminoketones) and their derivatives are important synthetic intermediates, particularly in the preparation of biologically active molecules. The traditional method for their preparation is the Mannich reaction, however, this method has many drawbacks such as long reaction times, poor regioselectivity, no enantioselectivity, and competition from unwanted side reactions.

We are developing new methods for the preparation of Mannich bases from 2-acylaziridines through intermediary proximal b-amino silyl enol ethers. These functionalized silyl enol ethers should react with a variety of electrophiles to provide Mannich bases of varied structural complexity.

Over the course of this project, we have focused on developing two separate but related methods for the stereo- and regioselective preparation of the Mannich bases and proximal b-amino silyl enol ethers, respectively. The central reaction common to both methods is the reaction of a 2-acylaziridine with a silyllithium reagent, which triggers a Brook rearrangement with concomitant opening of the adjacent aziridine ring.

For the direct formation of simple Mannich bases, an excess of silyllithium reagent is used and conditions have been adjusted to promote the in situ cleavage of the intermediate silyl enol ether, which provides the Mannich base on work up.

Optimization studies were performed using N-acetyl-2-benzoyl-3-methylaziridine and 2-acetyl-3,3-dimethyl-N-tosylaziridine as representative aryl and alkyl ketones substrates, respectively. As anticipated, it was found that solvents such as THF that strongly bind to lithium ions gave the highest yield of Mannich base. Use of nonpolar solvents such as ether or toluene suppressed the necessary desilylation, resulting in lower yields of the desired Mannich base. The most commonly encountered silyllithium reagent, dimethylphenylsilyllithium, was found to efficiently effect the reaction of the aryl ketone substrate, N-acetyl-2-benzoyl-3-methylaziridine. However, reaction of the alkyl ketone substrate, 2-acetyl-3,3-dimethyl-N-tosylaziridine, required the use of methyldiphenylsilyllithium to achieve an acceptable yield of Mannich base. The lack of an electron-withdrawing group at the reaction site of the alkyl ketone necessitated additional electron withdrawing ability from the second phenyl substituent on the silyl group. Overall, the combination of excess methyldiphenylsilyllithium in THF at -40 °C proved to be a convenient system that gave the desired Mannich base in good yield from either the alkyl or aryl substituted substrate. Several differentially substituted 2-acylaziridine substrates have been prepared and reacted using these optimized conditions to determine the scope of the method. Both aryl and alkyl 2-acylaziridines reacted to give the corresponding Mannich bases in good to excellent yields. The method is compatible with amide, carbamate and sulfonamide protecting groups on the aziridine nitrogen, but ineffective for use with an unprotected nitrogen. The Mannich basses can be isolated and purified using silica gel chromatography.

Substantial progress has also been made on the second method of the project, the preparation of proximal b-amino silyl enol ethers. Optimization studies have been performed using N-acetyl-2-benzoyl-3-methylaziridine as an investigative substrate.

Use of nonpolar solvents such as ether or toluene suppresses the desilylation providing mixtures of Mannich base and the desired b-amino silyl enol ether. Variations on the silyllithium reagent have also been investigated. Reaction of N-acetyl-2-benzoyl-3-methylaziridine with dimethylphenylsilyllithium provided the desired b-amino silyl enol ether in low yield. Use of methyldiphenylsilyllithium resulted in a substantially higher yield of the silyl enol ether, but was accompanied by a substantial amount of undesired Mannich base. Increasing the steric bulk of the silyl group was investigated as a potential means to suppress the undesired desilylation of the silyl enol ether. Use of t-butyldiphenylsilyllithium was found to completely suppress the cleavage of the silyl enol ether. Application of these conditions (t-butyldiphenylsilyllithium in toluene at -78 °C) to other aryl 2-acylaziridines has given the corresponding b-amino silyl enol ethers in good yield. Reaction of the alkyl 2-acylazirizine substrate 2-acetyl-3,3-dimethyl-N-tosylaziridine resulted in a disappointing yield of 49%, with a significant recovery (35%) of starting material. This is likely due to the decrease in nucleophilicity associated with the larger tert-butyl group on the silicon, allowing for the competing enolization of the alkyl ketone. The enolate is thus shielded from further reaction with the silyllithium reagent and is recovered as starting material after workup. The b-amino silyl enol ethers thus prepared have so far been difficult to purify. Substantial decomposition is observed during chromatographic purification using silica gel. This has been mitigated to some extent by pretreatment of the silica gel with triethylamine, however some decomposition is still observed. Thus, the yields reported here for the preparation of the b-amino silyl enol ethers are derived from ¹H-NMR using hexamethylbenzene as an internal standard.

Future work on both methods will include the reaction of additional substrates to further demonstrate the generality of each method. For the preparation of b-amino silyl enol ethers, additional optimization of purification technique will also be performed.