February 2004
This latest meeting in the long running series was attended by 68 people, with 11 nationalities represented. There were 15 presentations including two late changes after two speakers were unable to attend due to illness. Dr Jim Zeller of Pfizer, Holland, MI replaced Professor Peter Beak (University of Illinois at Urbana-Champaign), and Dr Marc Thommen of Solvias, Switzerland replaced Professor James Leighton (Columbia University).
Copies of the full conference proceedings can be purchased from Scientific Update upon request.
The first presentation was a process development case study containing some fascinating chemistry. Key points being a zinc mediated Michael addition to cyclopentenone, leading to improved diastereoselectivity compared to other alternatives.
Conversion of the cyclopentanone carbonyl group in to a gem difluoride was achieved with Deoxofluor™, and this was followed by removal of the acetal protecting group and purification of the acid as the dicyclohexylamine salt. The other half of the molecule was synthesized from 2,6-dibromopyridine by formylation of the Grignard “ate” complex (formed by halogen metal exchange with n-Bu3Mg- Li+). Reductive amination introduced the aminopiperidinyl group and was followed by a copper catalysed amination with ammonia in ethylene glycol under pressure. The origin of this reaction is a lesson to all synthetic chemists - it worked first time as a direct uncatalysed amination in a pressure vessel, but could not be repeated until an analysis of the original reaction liquors showed trace amounts of copper, which must have been present in the pressure vessel. The final amide coupling was carried out with EDC/HOBT in aqueous acetonitrile.
Professor Morken’s lecture was prefaced by a short discussion of suitable analytical methodology for rapid catalyst screening. Standard gc analysis (with a 40 minute run time) of a 96 well plate requires 4 days, so an alternative method was devised – single pulse 13C nmr analysis with each analysis taking ~15 sec, that also allows in situ analysis. This methodology was used to screen catalysts and ligands for a variety of transition metal catalysed reductive coupling reactions such as the silane mediated reductive aldol reaction shown.
The later part pf the lecture concerned the addition of diboron reagents, such as di(catechol-borane) (B2(cat)2), to simple alkenes and the subsequent reactions that can be carried out on the intermediates, e.g. oxidation and/or coupling reactions (see Scheme).
Dr Guzzo described work carried out on a medicinal chemistry collaboration between Albany Molecular Research and AstraZeneca R&D, Mölndal, Sweden aimed at synthesizing three types of chiral molecule – peptide mimetic PAI-1 inhibitors, optically active mandelic acids and GABAB receptor agonists. Syntheses of di- and tetrapeptide mimetics containing either a trans alkene in place of an amide linkage or where the amide linkage was part of a pyrrolidinone ring were presented. The former synthesis utilised Evans chiral auxiliary methodology, while the latter used starting materials from the chiral pool forming the pyrrolidinone ring by cyclising an amide anion onto a sulphonium ion derived from S-methylhomocysteine. Optically active mandelic acids were prepared by resolution of racemates produced by standard chemical routes. Four γ - aminopropylphosphinic acids (1-4), two of them fluorinated analogues were chosen as potential GABAB receptor agonists.
Compound (4) showed particularly good activity and was resynthesized in chiral form from L-serine by the synthesis shown below via a rearrangement involving an aziridinium ion during the fluorination with DAST.
Dr Tao described Pfizer's approaches to the use of biocatalysis to find better routes to various target compounds. An example of a hydrolytic kinetic resolution of an acetylenic substrate, where the solvent had a significant effect on the reactivity of the system and the enantioselectivity, was presented.
The next example was an enzymatic reduction of a 3-phenylpyruvic acid derivative, which significantly reduced the cost of a building block and avoided some inefficient diazonium chemistry that used an expensive starting material.
This system was shown to have a broad substrate spectrum. The use of cyclases to carry out macrocyclisations of substrates such as oligopeptides was discussed. A dramatic improvement to a steroid synthesis was also presented in which a 31 step chemical synthesis, costing $200/g, was reduced to an 11-step synthesis costing $6/g, by the use of an enzymatic ?-hydroxylation at C 11 in a steroid. The final part of the talk concerned some work, which is ongoing, to find a more efficient alternative method to replace the mandelic acid resolution which is the final stage in the manufacturing route for Pregabalin.
An enzymatic hydrolysis has been developed that operates at high concentration and could have significant cost savings over the current manufacturing route. The cost of the enzyme could be further improved, if necessary, by protein engineering.
Atropisomerism was the underlying theme of Professor Clayden’s presentation and how the use of N,N-diisopropyl amides of ortho-substituted aromatic acids allows chiral centres to be generated, in some cases at remote centres, particularly during lithiation reactions where the amide group can assist.
The use of cheap resolving agents such as ephedrine and tolyl menthyl sulphoxide as temporary directing agents to allow dynamic thermodynamic resolution of atropisomers was also discussed. The principle is that a resolving agent is attached to a racemic compound, the diastereomers are allowed to equilibrate leading to the more thermodynamically stable diastereomer and then the resolving agent is removed as shown in the conversion of racemic naphthaldehyde-amide (1) into a single enantiomer of the alcohol (2).
This type of approach can also be used in a relay fashion where the stereochemistry of the ephedrine aminal controls the addition at a remote centre 23 atoms away as shown in the scheme below (diastereoselectivity: >90:10).
Oxidative aryl coupling is one of the more common C-C bond forming reactions carried out in nature and with the advent of chiral binaphthalene ligands it has become an important synthetic process for organic chemists. Dr Kozlowski presented work on the oxidative enantioselective homo-coupling of a wide variety of 3-substituted 2-naphthols of the type shown below, where R = CO2Me, COMe, COPh, SO2Ar, PO(OMe)2, COH(Ph)2, etc.
Enantiomeric excesses of 75-95% are common, but some substrates, particularly where R is an electron-donating group, only give ee’s around 50%.Where there are two oxidisable positions polymers can be prepared.
Similar polymers can be generated from 2-hydroxy-6-ethynylnaphthalene-3-carboxylates, with coupling occurring at the 1-position and at the alkyne terminus. Applications of the oxidative biaryl coupling towards the total synthesis of some bi-naphthoquinone natural products such as calphostin and cercosporin and naphthopyrones, such as nigerone were presented. The final section of the presentation concerned the oxidative coupling of a variety of groups to 2-hydroxy-4-tert-butylbenzaldehyde such as quinolines and alkylamines to produce precursors for novel salen ligands with a basic functionality attached.
Dr Zeller described second generation process development work aimed at producing alternative, cheaper route(s) to atorvastatin (Lipitor®), run in parallel to the investigation of potential improvements to the existing manufacturing route.
The existing route, shown above, has two cryogenic steps (the cost of liquid N2 for these 2 steps contributes $3.50/kg to the cost of TBIN ) and a tedious removal of borate esters during the work up of the sodium borohydride reduction step. Initial approaches centred on condensation of 3-aminopropanal diethyl acetal with the diketone of atorvastatin, but all attempts to develop catalytic asymmetric additions to the unmasked aldehydes failed. The most successful route used an aminopropionamide condensation followed by elaboration to a β,ó-diketoester and subsequent selective reduction to atorvastatin (see schemes below). The reduction is run with Bayer’s Cl-MeO-BIPHEP ligand, which gives 98% ee at C-5, with a mixture at C-3. This is overcome by dehydration to an α,β-unsaturated lactone which is subjected to a stereoselective oxy-Michael addition to give the benzyl ether of atorvastatin, which is hydrogenolysed to the desired product, atorvastatin lactone. Unfortunately just as the process was nearing the cost target, the price of the hydroxynitrile starting material for the existing manufacturing route dropped significantly and the cost target moved. This project is currently on the “Back Burner”.
Production
of Sugar Polyols by Fermentation
Dr D. Demirjian, zuChem, USA
This presentation centred on the development of a biocatalytic process for the production of mannitol to improve upon the existing chemical hydrogenation of fructose, which produces a 50:50 mixture of mannitol and sorbitol. A heterofermentative process using a lactobacillus was selected after initial screening work. The fermentation can be run at fructose concentrations up to 300g/L under pH-controlled conditions with good selectivity and productivity for mannitol. The fermentation can be speeded up by carrying it out in fed batch mode with continuous addition of fructose and this process is currently being scaled up to pilot plant and is estimated to yield up to 25% cost improvement over the chemical route. A second-generation process, based upon a continuous bioreactor process, is also being developed and this could yield a cost improvement in excess of 50%.
Professor Shi started his lecture by giving an overview of the asymmetric epoxidation of olefins with oxone catalysed by sugar derived ketones and followed this with a discussion of modifications to the catalyst to generate more stable catalysts, that do not degrade by Bayer-Villiger oxidation.
The scope of the process has been extended to include α,β-unsaturated esters and enol acetates. Epoxides of enol acetates can be opened up in the presence of acids or Lewis acids leading to either enantiomer of an α–hydroxyketone.
The reaction can now be run using hydrogen peroxide and an organic nitrile as oxidant rather than oxone, thus reducing the potential waste load and making scale up easier.
This presentation concerned the development of manufacturing routes to non-nucleoside reverse transciptase inhibitors as potential treatments for AIDS. The first part of the talk centred on two unsymmetrical cyclic ureas (DMP 850 and DMP 851). Tartaric acid provided a cheap, chiral starting material and could be converted to the bis-hydrazone, by protection of the diol functionality as an acetonide, Dibal reduction of the carboxylate groups to the aldehydes oxidation state and reaction with N,N-dimethylhydrazine. Addition of tolyl lithium was stereospecific! Raney nickel reduction generated the diamine.
Two strategies to generate unsymmetrical ureas were investigated. Dibenzylation (by reductive amination with benzaldehyde) followed by cyclic urea formation with phosgene generated a symmetrical intermediate, which could be selectively mono-debenzylated with sodium in liquid ammonia / THF. Alternatively, mono-trifluoroacetylation generated an unsymmetrically substituted intermediate more efficiently.
The rest of the presentation related to the development of efficient, chiral routes to some second-generation analogues (1, 2, 4 and 5) of Sustiva™ (Efavirenz, 3), which was discovered and developed by DuPont Merck (now part of BMS).
The overall synthetic approach is similar to that used for Efavirenz in that the chirality is introduced by adding the lithium salt of cyclopropylacetylene to a trifluoromethyl substituted sp2-hybridised carbon.
However, the new targets (1, 2, 4 and 5) require addition to an imine rather than a ketone, to provide the second nitrogen in the heterocyclic ring. Initial attempts utilised a chiral auxiliary susbtituent (a-methylbenzyl) on the nitrogen to direct the addition. This worked for the chloro-derivatives, but failed for the difluorinated derivatives. A new asymmetric process was developed for the difluoro compounds, using a terpene derived chiral moderator, and this method also worked for the chloro compound. The main issue concerned the hydrolytic instability of the imine precursor, which was formed by dehydration with benzenesulphonic acid in mixed xylenes. Thermal dehydration gave a higher yield and a product that was hydrolytically stable.
Repeating the asymmetric addition, which gave 80% yield and >99.9% ee previously, gave a product with substantially lower ee (~74%). Addition of catalytic amounts of benzenesulphonic acid restored the enantioselectivity and improved the reaction time! Reduction with lithium aluminium hydride provided the two compounds with the trans-olefin moiety instead of an alkyne.
Professor Augustine’s presentation concerned the tethering of homogeneous catalysts on to solid supports such as alumina, silica, montmorillonite etc using heteropolyacids to “link” the support and the metal atom of the catalyst. Rhodium complexes with ligands such as DiPAMP, DuPhos, Josiphos, Skewphos and BoPhoz were anchored in this way and gave solid catalysts that showed almost identical activity to the homogeneous equivalents and could be used repeatedly with little loss of activity. The rhodium content of the product solutions was ²1.5 ppm. In some cases, such as the hydrogenation of dimethyl itaconate using Rhodium DuPhos tethered on carbon, the turn over frequency (TOF) increases with each use of the catalyst; this is apparently a fairly common phenomenon with carbon supported catalysts.
Dr Fotheringham described work carried out at Ingenza and the University of Edinburgh on the development of combined biocatalytic and chemical approaches to deracemisation of amino acids – taking a racemic mixture and converting one enantiomer in to the other. This can be achieved by using an amino acid oxidase that converts one enantiomer into an imine that is then reduced back to a racemic mixture with a non-stereoselective chemical reducing agent. D-amino acid oxidases and L-amino acid oxidases have both been developed so that a racemic mixture of an amino acid can be converted in to either enantiomer. Examples included conversion of piperazinecarboxylic acid to L-piperazinecarboxylic acid in 86% yield, >99% ee. Improvements to the original system, where 500 equivalents of NaBH4 were required, include using a transfer reduction system with Pd/C and formic acid. This technique can also be applied to amines, but this requires novel activity as natural amine oxidases show little or no enantioselectivity. Those enzymes which show low enantioselectivity were improved by directed evolution to give for example 77% yield, 95% ee in the deracemisation of α–methylbenzylamine (more examples can be found in Angew. Chem. Internat. Ed., 2003, 42, 4807).
The presentation by Dr Chorghade concerned the use of the various approaches to synthesizing chiral compounds as applied to chiral switches in the pharmaceutical industry (developing and marketing single enantiomer versions of compounds previously marketed as racemates). Examples of successful chiral switches, such as omeprazole to esomeprazole were described along with some unsuccessful attempts (e.g. dexfenfluramine) and aborted attempts (e.g. fluoxetine hydrochloride). The second half of the presentation included some examples of chiral switches carried out in India, such as asomex (S-Amlodipine) and finally some new approaches using radical mediated diallylation followed by ring closing metathesis for spiro-cyclopentannulation.
Dr Thommen described the development of a number of asymmetric catalysts and their applications to industrial problems as well as comparing some of the different catalysts that are now available. Examples of some of the ligands are shown below.
The use of a Rhodium Walphos complex in a chiral hydrogenation reaction, a key step in the synthesis of SPP 100, a drug candidate being developed by Speedel Pharma, was described (see below).
Co-operation with some academic groups is finding new applications for some of the Solvias ligands. For example, the asymmetric conjugate reduction of acyclic enones can be carried out using polymethylhydrosiloxane in the presence of 1mol % copper (I) chloride, sodium tert-butoxide (1 mol % each) and Josiphos. The enantioselective ring opening of heterobicyclic alkenes can be catalysed by another Josiphos ligand (see below).
The core of this presentation was the development of a chiral enolate addition to the nitrostyrene shown below. The magnesium salt was found to exhibit some selectivity,
