Ruthenium and Iron‐Catalysed Decarboxylative N‐alkylation of Cyclic α‐Amino Acids with Alcohols: Sustainable Routes to Pyrrolidine and Piperidine Derivatives

Abstract A modular and waste‐free strategy for constructing N‐substituted cyclic amines via decarboxylative N‐alkylation of α‐amino acids employing ruthenium‐ and iron‐based catalysts is presented. The reported method allows the synthesis of a wide range of five‐ and six‐membered N‐alkylated heterocycles in moderate‐to‐excellent yields starting from predominantly proline and a broad range of benzyl alcohols, and primary and secondary aliphatic alcohols. Examples using pipecolic acid for the construction of piperidine derivatives, as well as the one‐pot synthesis of α‐amino nitriles, are also shown.

All reactions were carried out under an Argon atmosphere using oven (120 °C) dried glassware and using standard Schlenk techniques. 1-Hydroxytetraphenylcyclopentadienyl-(tetraphenyl-2,4cyclopentadien-1-one)-µ-hydrotetracarbonyldiruthenium(II) (Shvo's catalyst, C1) was purchased from Strem Chemicals, Inc. Complex C2 was synthesized according to the literature procedure. [1] All other reagents were purchased from Sigma-Aldrich, Acros and TCI in reagent or higher grade and were used as received without further purification.

Representative procedures
General procedure for the decarboxylative N-alkylation of amino acids An oven-dried 20 mL Schlenk tube, equipped with a stirring bar, was charged with amino acid (0.5 mmol, 1 equiv.), corresponding alcohol (1 or 2 mmol, 2 or 4 equiv.), Shvo's catalyst (C1, 0.005 mmol, 1 mol%) or Knölker's complex (C2, 0.02 mmol, 4 mol%), and toluene (as a solvent, 2 mL). Solid materials were weighed into the Schlenk tube under air. Then the Schlenk tube was subsequently connected to an argon line and vacuum-argon exchanges were performed three times. Liquid starting materials and solvent were charged under an argon stream. The Schlenk tube was capped and the mixture was rapidly stirred at room temperature for 1 min, then was placed into a preheated oil bath at 120 °C and stirred for a given time (typically, 24 h). Then, the reaction mixture was cooled down to room temperature, the crude mixture was filtered through silica gel, eluted with ethyl acetate, and concentrated in vacuo. The residue was purified by flash column chromatography to provide the pure amine product.

Note 1: Additional description belonging to Supplementary Figures 5
We rationalise the formation of aminonitrile regioisomer 5' via the hydrogen borrowing/ decarboxylative degradation sequence when mandelonitrile substrates are used as carbonyl precursors ( Table 4, main text), in accordance with our general mechanistic proposal (Figure 4). Isomerisation of α-aminonitrile 5' to 5 have been previously investigated and described by Siedel. [2,3] Due to the excess of proline in our system, we assume analogous isomerisation pathways to take place via an azomethine ylide intermediate (Supplementary Figure 5). The formation of the two regioisomers would depend on the charge distribution in the azomethine ylide. Li and co-workers have investigated the energy of resonance structures of the corresponding ylides leading to 5 and 5' in favour of regioisomer 5, [2] which perfectly matches with the obtained experimental data in our work ( Table 4). Alternative pathways, for example the loss of HCN upon reaction of proline with the carbonyl compound formed by dehydrogenation of the mandelonitrile substrate, followed by subsequent direct decarboxylation to the corresponding azomethine ylide would also be possible.

Note 2:
Deuterium incorporation experiment was carried out in the following way: An oven-dried 20 mL Schlenk tube, equipped with a stirring bar, was charged with DL-proline (1a, 0.1 mmol, 1 equiv.), benzyl alcohol-α,α-d2 (2f-d2, 0.2 mmol, 2 equiv.), Shvo's catalyst (C1, 0.001 mmol, 1 mol%), and toluene-d8 (as a solvent, 0.4 mL). Solid materials were weighed into the Schlenk tube under air. Then the Schlenk tube was subsequently connected to an argon line and vacuum-argon exchanges were performed three times. Liquid starting materials and solvent were charged under an argon stream. The Schlenk tube was then capped and the mixture was rapidly stirred at room temperature for 1 min, then was placed into a pre-heated oil bath at 120 °C and stirred for 2 h.
The deuterium is transferred from the benzylic position of the alcohol substrate (2f-d2) to the product confirms that the reaction indeed proceeds through a hydrogen borrowing pathway. The deuterium was incorporated at the α-position to the nitrogen at the endocyclic position and/or the benzylic position of the product, confirming that the azomethine ylide tautomerises during the reaction. Multiple deuterium atoms can be found in the products shown in Supplementary Figure 6. An additional reason for deuterium incorporation could be the existence of a hydrogenation/dehydrogenation equilibrium between IVa and IVb as well as between Vl and Va/Vb

Note 3:
Kinetic experiments were carried out in the following way: An oven-dried 20 mL Schlenk tube, equipped with a stirring bar, was charged with DL-proline (0.3 mmol, 1 equiv.), p-methoxybenzyl alcohol (0.6 mmol, 2 equiv.), Shvo's catalyst (C1, 0.003 mmol, 1 mol%), and toluene-d8 (as a solvent, 1 mL). Solid materials were weighed into the Schlenk tube under air. Then the Schlenk tube was subsequently connected to an argon line and vacuum-argon exchanges were performed three times. Liquid starting materials and solvent were charged under an argon stream. The Schlenk tube was capped and the mixture was rapidly stirred at room temperature for 1 min, then was placed into a pre-heated oil bath at 120 °C and stirred for a specified time (5 min, 15 min, 30 min, 1h, 1h 30 min, 2 h). All reactions were set up in parallel. Then, the reaction mixture was cooled down to room temperature and 1,3,5-trimethoxybenzene (0.1 mmol, 0.3 equiv.) as an internal standard was added. Preparing the sample, 0.6 mL of the reaction mixture was placed to a J-Young NMR tube under argon. All spectra were recorded using Bruker Avance NEO 600 machine.
The obtained data is summarized in the