Resolution Route

The P-cyanodiester 4 was prepared by condensation of isovaleraldehyde with diethyl malonate followed by the addihon of potassium cyanide. The cyanodiester 4 was hydrolyzed and decarboxylated to give the P-cyano acid 5. Reduction with Raney nickel gave racemic pregabalin (6), which was resolved with (S)-mandelic acid. The diastereomeric salt was split with wet TH F under neutral conditions to give pregabalin, which was recrystallized from isopropanol (IPA) to give the final Active Pharmaceutical Ingredient (API). 

1) For a discussion on why the t-butylamine salt was selected as the starting material for the asymmetric step see reference [9]. 

Catalyst [Substrate] 10 catalyst ratio Pressure Enantiomeric excess 

McQuade et al. [11] have pubhshed a nice synthesis using a microencapsulated nickel-based catalyst for promoting a Henry reaction based upon the work of Evans [12]. Torrens et al. have pubhshed a somewhat longer synthesis from D-mannitol bisacetonide [13]. 

---

Scheme 16.19. The Pfizer enzymatic resolution route to pregabalin (2).

Mercedes Amat and Joan Bosch of the University of Barcelona have been exploring (Cltem. Commun. 2005, 1327) a kinetic resolution route to piperidines. Condensation of a ketone or aldehyde ester such as 7 with an enantiomerically-pure amino alcohol such as 8 with proceeds with high (15 1) diastereoselectivity, to give 9. Reduction of 9 then delivers the piperidine 10 in high enantiomeric excess. 

A comparative cost analysis showed that the classical resolution route (Scheme 8.2) was 12 times cheaper than the discovery route (Scheme 8.1). The classical resolution route was successfully scaled up and used to launch the product and provide the first year s market supply. However, using a final-stage resolution meant that by definition half of the synthetic materials were thrown away. When an E factor analysis [8] was performed on the pregabalin synthesis it was found that 86 kg of waste was being produced for every kilogram of the desired product, and this inspired a search for more efficient chemistries. 

Pathway B utilized conditions from the classical resolution route [7] which used 4 as the starting material instead of 13. Less exploratory work was done on this pathway, mainly because of the higher risk of epimerization of the C3 center under highly alkaline conditions as well as the difficult isolation of 18, which imphed carrying enzyme (as well as enzyme by-products) to the API isolation step. The poisoning of the Raney nickel by the enzyme found in pathway A also made this an undesirable option. 

Classical resolution route (Scheme 8.2) Enzymatic route (Scheme 8.6) ... 

Stage Classical resolution route Energy (MJ/kg 1) Enzymatic route (no recycle of 4) Energy (MJ/kg 1)... 

Scheme 2. Racemic synthesis and enzymatic resolution route.

Figure 16.1 Synthetic routes to enantiomerically pure 2-hydroxycarboxylic acids, via oxynitrilase (hydroxynitrile lyase) catalysed enantioselective hydrocyanation (route A) and (R)-nitrilase (nitrilase) mediated dynamic kinetic resolution (route B).

During the same period, eontinuous process improvements on the resolution route to S -naproxen have achieved dramatic cost reductions. The breakthrough comes from in-process racemization and recycle of the R-naproxen byproduct and the recovery of the resolving agent (Figure 2). In this novel and efficient resolution process, the racemic naproxen is reacted with half an equiva-... 

The bis(phosphane oxide) resolution route to (S)-MeOBIPHEP (S-29), the ligand for the geraniol hydrogenation, was successfully developed to a larger scale process (Fig. 11). In this process, phosphinoylation of 3-bromoanisole was realized... 

A kinetic resolution of an ester requiring separate steps to recover and recycle the less reactive enantiomer has been transformed into a dynamic resolution of the corresponding thioester. While the thioester route requires an additional synthetic step, the efficiency of this step renders the dynamic resolution route to (R)-l shorter. The development of an alternative route provides our company with additional options for commercial-scale roxifiban preparation. 

However, a source of the non-natural 9S isomer (2) was first required. The ready availability of natural crinitol made a racemization/resolution route, as illustrated in Scheme 1, attractive. Racemization was accomplished by Collins oxidation (16,25) to the dicarbonyl compound (14), followed by lithium aluminum hydride (LAH) reduction to give the racemic mixture (1 + 2). Resolution via diastereomeric derivatives seemed plausible. Esterification with enantiomerically pure a-methoxy-a-(trifluoromethyl) phenylacetic acid (MTPA) (17), followed by separation of diastereomers by recycle-HPLC (R-HPLC), had earlier been used to purify enantiomers of ipsenol and ipsdienol (26). A model system, the resolution of -3-nonen-2-ol, a secondary allylic alcohol naturally occurring in Rooibos tea (16,27), also worked satisfactorily. Therefore, the route using the bis-(MTPA) esters was selected for crinitol. 

Here we present the results of head-to-head comparisons where chemoselective routes are found to be better than classical resolution routes and vice versa. This exercise shows that one cannot make statements about plan performance based simply on generalizations. Each plan needs to be examined on its own by a thorough metrics analysis and both its strengths and weaknesses are identified in relation to other competing plans. Tables 9.4 and 9.5 summarize results for the syntheses of (+)-cocaine and (-)-dapoxetine, respectively. Full details of plans are given in the aeeompanying CD-ROM. 

Damon, D. B., Dugger, R. W., Magnus-Aryitey, G., Ruggeri, R. B., Wester, R. T., Tu, M., Abramov, Y. (2006). Synthesis of the CETP inhibitor torcetrapib the resolution route and origin of stereoselectivity in the imininm ion cychzation. Organic Process Research and Development, 10, 464-471. 

Henegar et al. [78] at Pfizer developed an efficient and greener synthesis of the (S, 5)-reboxetine 157, which is being evaluated for the treatment of neuropathic pain and a variety of other indications (Scheme 9.42). The reported chiral synthesis of 157 starts by SAE of cinnamyl alcohol 154 to give R, R)-epoxide 155 in 89% yield (> 98% ee). Reaction of 155 with 2-ethoxyphenol gave crystallized product 156. The overall yield of (S, 5)-reboxetine succinate increased by 9%, compared to resolution method. The catalytic asymmetric process offers use of less solvent and reduces waste generation by approximately 50%, compared to the resolution route. 

In this chapter we describe the DSM aminoamidase processes in more detail. Three different enzymatic resolution routes have been developed for the preparation of natural and synthetic amino acids using biocatalysts from different origin, i.e.. Pseudomonas pu-tida, Mycobacterium neoaurum, and Ochrobactrum anthropi. Scope and limitations and enzyme characterization of these amidases will be presented together with some specific examples. In addition, the use of some of these amino acids in peptide sjmthesis, catalytic asymmetric synthesis, and further synthetic transformations will be given. 

---