e-NEWSLETTER

Issue 10 September 2003

Highlights from Chiral Europe, 2003, London, 12-14^th May 2003

This meeting was sponsored by Dowpharma. Presentations were of a uniformly high standard, but with a varied subject matter. Keynote lectures were given by Professor Barry Trost and Professor Takao Ikariya. The lectures by – John Carey (GSK, Org. Process Res. Dev., 2002, 6(4), 488-491 and 1999, 3(1) 60-63), Ian Lennon (Dowpharma, Org. Process Res. Dev., 7(3), 407-411, 2003, 7(1), 89-94 and 2001, 5(4), 438-441), and Nick Thomson (Pfizer, Org. Process Res. Dev., 7(3), 362-368) contained some material that has been published in recent editions of the Organic Process journal and the talk by Jos Brands (Merck, J. Amer. Chem. Soc., 2003, 125, 2129-2135) has been published in JACS. One tip that is worth picking out of the talks is to consider how the reagents used in the work up of an intermediate, that is going to be subjected to a catalytic reaction, can effect the next step. Ian Lennon mentioned a case where they were trying to carry out an asymmetric hydrogenation on 2-methylenesuccinamic acid, which had been prepared by reaction of itaconic anhydride with ammonia followed by acidification. If the acidification was carried out with hydrochloric acid, the intermediate contained traces of chloride, which slowed down the hydrogenation dramatically. The maximum substrate to catalyst ratio achievable was 1000:1. However if sulphuric acid was used for the acidification, substrate to catalyst ratios of 100,000:1 were achievable.

Crafting Chiral Space for Asymmetric Induction in a Catalytic Synthetic Reaction, Keynote Lecture 1

Barry Trost’s opening lecture contained a comprehensive review of the many of approaches to asymmetric induction that have been studied by his group. One example in the field of asymmetric allylation chemistry that may be of interest to process chemists concerns the chemistry shown below.

The first time this reaction was run the product was obtained in 90% ee. A repeat run gave a product with much lower ee. The difference in the two reactions turned out to be the age of the butyl lithium (BuLi) used. The original reaction contained BuLi from an old bottle, while the second reaction was carried out using fresh BuLi. The result from the original reaction could be reproduced with fresh BuLi if butanol was added to the reaction mixture. In another twist to this tale the reaction was tried with a lower catalyst loading (0.5 mol %) and the ee of the product improved.

This result was later explained in Guy Lloyd-Jones’s excellent lecture on “Some Mechanistic aspects of Selectivity in Pd-Catalysis”. This was a physical organic examination of asymmetric allylic alkylation reaction and provided some fascinating mechanistic insights. Deuterium labelling experiments showed how enantiomeric substrates of a racemic mixture react at different rates and what might otherwise have been thought to be an indifferent catalyst (in asymmetric induction terms) may in fact just have been used on the wrong substrate. So for example an ee of 50% can arise from one enantiomer (of a racemic mixture of allylic acetates for example) giving 100% ee and the other enantiomer not showing any asymmetric induction.

In a study of the example described by Barry Trost above where the ee of the product increases as the catalyst loading decreases, detailed ³¹P nmr experimentation led to an understanding of what was happening. When a catalyst with an asymmetric ligand such as the Trost ligand (see below) is used the monomeric form of the catalyst shown below is the

desired form, which creates the asymmetric pocket that allows product with a high ee to be obtained. At low catalyst loadings this form predominates but at higher catalyst loadings an oligomeric form predominates. Solutions of these two forms exhibit different rotations with the monomer having [a]_D = + 640 and the oligomer having [a]_D = - 150.

Recent Advances in Asymmetric Transfer Hydrogenation, Keynote Lecture 2

Professor Takao Ikariya discussed his work on including a dynamic kinetic resolution of a–substituted ketones. If a racemic a–substituted ketones is subjected to transfer hydrogenation conditions using a chiral ruthenium catalyst [RuCl[(S,S)-Tsdpen](p-cymene)] and triethylamine / formic acid, four possible products ([R,R], [R,S], [S,R], [S,S]) can be formed as shown below. The relative amounts of each product depends on the relative rates of hydrogenation of the two enantiomeric ketones, the rate of inversion of the ketones and the effectiveness of the catalyst in asymmetric induction.

In many cases (X = OH, OCH₃, OⁱPr) over 93% of the product is the [R,R] diastereomer, for X = OAc 85% of the product is the [R,R] product, whilst for X = OBz the reaction is far less effective. In this case the product contains all 4 possible diastereomers ([R,R] 32%, [R,S] 30%, [S,R] 29%, and [S,S] 9%).

Evolutionary and Combinatorial Methods in Asymmetric Catalysis

Professor Manfred Reetz described work on two areas – mixed chiral monodentate phosphorus ligands and the directed evolution of enantioselective enzymes. Chiral monodentate ligands such as monophos are thought to involve a reactive intermediate that contains the metal and two ligand molecules. But what happens if two different monodentate ligands are used? The answer in many cases is improved ee’s. So for the enamide reduction shown below better results are obtained with a mixed ligand system than either of the pure ligands (see Table 1)

Ligand system	ee % Ar = Ph	ee % Ar = p-Cl-C₆H₄	ee % Ar = 2-naphthyl
Binap-P-CH₃	75.6	73.0	78.2
Binap–P-tBu	13.2	16.2	3.0
Binap-P-CH₃/ Binap–P-tBu	96.1	95.0	97.0

Molecular models of the possible catalyst in the mixed ligand system suggest that the methyl group of the Binap-P-ch3 ligand lies close to one of the t-butyl methyl groups suggesting that bidentate ligand shown below should be a good catalyst. However it is still not as good a catalyst as the mixed ligand system.

Directed evolution enantioselective enzymes involves finding a micro-organism that shows some activity for catalysing a particular reaction and then optimising the enantioselectivity by mutagenesis suing one of a variety of techniques (site directed mutagenesis, error-prone polymerase chain reaction, saturation mutagenesis, cassette mutagenesis, DNA shuffling or staggered extension process). Often several iterations are required to produce a suitable enzyme. In one example of a mutant lipase, 5 mutations were required, but perhaps surprisingly all the new amino acids in the enzyme were remote from the active site.

Design, Performance, Manufacture and Industrial Use of Chiral Ligands

Dr Marc Thommen of Solvias described the synthesis of a number of chiral ligands (such as josiphos, walphos taniaphos, mandiphane, butiphane, rophos and a new ligand – Solvias binap) and their use as hydrogenation catalysts and some new research using these ligands in cross-coupling reactions, asymmetric conjugate acyclic enones, and enantioselective ring opening of heterobicyclic alkenes.

Benefits and Limitations of chiral resolution via preferred crystallisation

Professor Gérard Coquerel from the University of Rouen discussed crystallisation approaches to preparing single enantiomers and in particular the use of conglomerates to allow selective crystallisation of one enantiomer from a racemic solution. Only one in ten to fifteen compounds crystallises as a conglomerate, so in most cases this will not seem to be a viable option. However, if the compound itself does not crystallise as a conglomerate one of it’s salts or derivatives may do, so a screen of salts and/or derivatives may reveal a compound which can be purified by crystallisation.

PLUS

Cyclopropanation via a Simple Barbier Reaction in DMF

Jean Paul Paugam and co-workers from France have published a simple cyclopropanation reaction between α,α-trihalomethyl or α,α-dihalomethyl aromatics and activated olefins. In a typical reaction magnesium is suspended in DMF and dichlorodiphenylmethane (DCDPM) and dimethyl itaconate are added via a dropping funnel. Yields are typically in the region of 50-70%, with substrates such as DCDPM, α,α-dibromotoluene and methyl trichloroacetate reacting with olefins such as itaconate esters, cyclohexenone, ethyl acrylate, vinyl ketones and acrylonitrile. The reaction is thought to proceed by initial conjugate addition of the Grignard reagent to the olefin followed by SN_i cyclisation although an alternative pathway via formation of a carbene from the Grignard reagent cannot be ruled out.(Synlett, 2003, (4), 485-488).

A Simple Protocol for Direct Reductive Amination of Ketones and Aldehydes (including a.b-unsaturated carbonyl compounds)

The method described by Basudeb Basu et al uses potassium formate and catalytic palladium acetate in a transfer hydrogenation. Yields are typically in the range 60-80%. A variety of aromatic aldehydes and ketones were examined as substrates reacting with amines such as cyclohexylamine, morpholine, benzylamine and various anilines. (Synlett, 2003, (4), 555-557).

Pyrrolidine-2-carboxylic acid (L-Proline) as an enantioselective catalyst

A recent Synlett “Spotlight” focuses on the use of L-proline as an enantioselective catalyst in a number of carbon-carbon bond forming reactions. Examples of the aldol reaction, Michael reaction, Mannich reaction, a-amination of ketones with azodicarboxylates, and the Robinson annulation are included. (Synlett, 2003, (4), 582-583).

Superacid Catalysed Hydroxyalkylation of Aromatics with Ethyl Trifluoropyruvate (A New Route to Mosher’s Acid Analogues)

A recent paper by G.K. Surya Prakash and George Olah describes the reaction of ethyl trifluoropyruvate with a variety of activated aromatic and heteroaromatic substrates (such as using trifluoromethanesulphonic acid or gallium trifluoromethanesulphonate as catalyst.

Yields are generally excellent (mainly >90% ), but occasionally lower (84% with p-xylene, 40% with N-methylpyrrole for example) and regioselectivity is 100% for most substrates with the exception of toluene and anisole where the regioselectivity is 94% and 92% respectively. ( Synlett , 2003 (4), 527-531).

Polymethylhydrosiloxane (PMHS) as an Additive in the Sonogashira Reaction

A study by Gallagher and Maleczka has shown that the use of PMHS, CsF, CuX and a palladium catalyst facilitates Sonogashira coupling under mild, relatively neutral, amine-free conditions. All 4 reagents were essential for Sonogashira coupling to occur. Catalytic amounts of copper thiophenecarboxylate (CuTC) or CuCl can be used provided the alkyne is coupled with a triflate or nonaflate. Bromides or chlorides react, but only when stoichiometric quantities of CuTC or CuCl are present. In a typical reaction an alkyne is reacted with 1.5 equivalents of nonaflate or triflate, PMHS (2 equiv), CsF (5 equiv), CuCl or CuTC (5 mol%) and bistriphenylphosphine palladium dichloride (5 mol%). Yields are mainly 80-96% with one or two exceptions. When compared with traditional Sonogashira conditions the PMHS version gives superior results with no homocoupled dialkynes being formed. (Synlett 2003, (4), 537-541).

One-pot synthesis of dihydropyrimidinones catalysed by lithium bromide: an improved procedure for the Biginelli reaction

Maiti and co-workers have found that lithium bromide (10 mol%) efficiently catalyses the three component condensation reaction of an aldehyde, a b ketoester and urea in refluxing acetonitrile to afford the corresponding dihydropyrimidinones in high yields (85-90%). A wide variety of aromatic aldehydes (alkoxy- and dialkoxybenzaldehydes, hydroxy-, nitro-, and chlorobenzaldehydes), furfuraldehyde and n-hexanal all react well under these conditions. Thiourea has also been used successfully in place of urea to generate the corresponding thio derivatives. (Tet. Lett., 2003, 44, 2757-2758).

Direct Synthesis of Propylene Oxide with CO₂ as Solvent

The direct oxidation of propylene to give propylene oxide is still one of the holy grails of the bulk chemicals industry. The predominant process for making propylene oxide is via the oxidation of ethylbenzene to give ethylbenzene hydroperoxide, which oxidises propylene and generates styrene as a co-product. More recently direct oxidation with hydrogen peroxide over a titanium silicalite (TS-1) has been investigated but the reaction only works well in the presence of methanol as co-solvent, which unfortunately is responsible for many of the by-products formed (methyl formate, acetone, acrolein, acrylic acid and propylene oxide ring-opening products). Now a report from Beckman et al reveals that much improved selectivity can be realised if carbon dioxide is used as solvent, selectivities up to 94% are possible at moderate conversion levels (6-10%), with the only detectable by-product being propane. The conditions used have the added advantage of generating hydrogen peroxide in situ by passing hydrogen and oxygen over a Pd/TS-1 catalyst. (Angew. Chem. Int. Ed. , 2003, 42(10), 1140-1142).

Direct Sulphonation of Methane to Methanesulphonic Acid with SO₂ Using Ca Salts as Promoters

Mukhopadhyay and Bell have reported the first liquid phase oxidation of methane with SO₂ in triflic acid to methanesulphonic acid using K₂S₂O₈ as oxidant and a catalytic amount of a calcium salt as a promoter. A maximum conversion of SO₂ to methanesulphonic acid of 22% could be achieved. This method is considerably more environmentally than the commercial process for manufacturing methanesulphonic acid in which methylmercaptan is oxidised with chlorine and water generating 6 moles of HCl per mole of methanesulphonic acid. (J. Am. Chem. Soc., 2003, 125, 4406).

Palladium-Catalysed Asymmetric Addition of Pronucleophiles to Allenes

Barry Trost has reported an atom economic method for adding pronucleophiles to benzyloxyallene with high regio- and enantioselectivity. The exact conditions depend on the pronucleophile used. For example with Meldrum's acid as the pronucleophile (the anion of Meldrum’s acid would be the nucleophile) the reaction is carried out in dichloromethane, using 1 mol % ρ allylpalladium chloride dimmer, 1.25 mol % ligand and 1 mol % trifluoroacetic acid, giving the addition product in 81% yield, 100% regioselectivity and 94% ee. The regioselectivity drops in alternative solvents such as THF (3:1 in favour of the branched product).< >However, when the pronucleophile is the azlactone shown below, the best conditions require 2 mol % potassium tert-butoxide in place of 1 mol % trifluoroacetic acid. Under these conditions the product is formed in 80% yield, 7:1 dr and 73% ee. However if the reaction is buffered by addition of 20 mol % hippuric acid, the results are 85% yield, 20:1 dr, and 93% ee. A number of examples with other azlactones are reported with similar results. (J. Am. Chem. Soc. , 2003, 125, 4438-4439).

Highly Enantioselective Catalytic Conjugate Addition of Cyanide to a,b-unsaturated Imides

Jacobsen and Sammis have published the first examples of asymmetric 1,4-conjugate addition to a,b-unsaturated carbonyl compounds with cyanide ion. Chiral aluminium salen catalysts were found to give good enantioselectivity under optimum conditions. The cyanide source is important, with best results being obtained when HCN is generated in situ from trimethylsilyl cyanide and isopropanol. If HCN alone is used as the cyanide source no reaction is observed. Yields over 90% and ee’s over 95% are obtained. Examples are included where R = Me, Et, ⁿPr, ⁱPr, ⁱBu, (CH₂)₃CHCH₂, ^tBu, and CH₂OBn. (J. Am. Chem. Soc., 2003, 125, 4442-4443).

Intramolecular [4+2] Cycloadditions of iminoacetonitriles: A New Class of Azadienophiles for Hetero Diels-Alder Reactions

A recent paper from Danheiser et al describes the preparation and intramolecular Diels-Alder reactions of iminoacetonitriles, a class of electron deficient imines whose cycloadditions have not previously been described. As a result iminoacetonitriles can be considered as valuable building blocks for the synthesis of nitrogen heterocycles. Iminoacetonitriles are readily prepared by a Mitsunobu reaction between an alcohol and HN(Tf)CH₂CN, followed by base catalysed elimination of trifluoromethane sulphinate. Cycloadditions are carried out at 85-120^oC in toluene in a sealed tube using BHT as a radical scavenger. E- and Z-imines seem to react at similar rates, in most cases giving a product with an exo-oriented cyano group. It is believed that the initially formed epimeric cycloadducts equilibrate to give the axial cyano isomer, which is favoured as a result of the “a–amino nitrile anomeric effect”. (J. Am. Chem. Soc., 2003, 125, 4970-4971).

Novel Small Organic Molecules for a Highly Enantioselective Direct Aldol Reaction

Gong and co-workers have examined the use of pyrrolidine carboxamides with a terminal hydroxyl group as catalysts (20 mol %) for the direct aldol reaction between nitrobenzaldehyde and acetone. The compound shown below was by far the best catalyst with chemical yields of 60-80%, and ee’s of 93% being obtained. This catalyst was used with a wide variety of other aldehydes and all products were obtained in consistently high ee’s (80-98%), although chemical yields were variable. In general aromatic aldehydes gave best yields (60-95%), whilst aliphatic aldehydes reacted less well (10-50%). The exception is cylohexanecarboxaldehyde which gives aldol products in high yields (77-85%) and ee’s (98% ee) even at lower catalyst loadings. Other related catalysts resulted in good yields, but poorer enantioselectivity (30-80% ee). (J. Am. Chem. Soc., 2003, 125, 5262-5263).

Selective C-Arylation of Free (NH)-Hetroarenes via Catalytic C-H bond Functionalisation

Sezen and Sames have reported a novel palladium catalysed arylation of unsubstituted heteroaromatics without the need for prior halogenation of the heteroaromatic. Under optimised conditions the reaction of the Grignard reagent formed by the reaction of indole with ethyl magnesium bromide or magnesium oxide is reacted with for example 4-iodotoluene (1.2 equiv), palladium acetate (5 mol %), triphenylphosphine (20 mol %), in dioxane to give exclusively 2-(4-tolyl)indole with no trace of other isomers or bis-arylated products. The reaction seems to tolerate electron withdrawing and electron donating groups in the para position of the aryl iodide (such as Me, OMe, F, CF₃, COMe), with yields in the region 75-90%. Interestingly when the reaction is carried out with 2-iodotoluene 2 equivalents of 2-iodotoluene are required and a mixture of 2-(o-tolyl)indole (53%) and 1-(o-tolyl)indole (17%) is obtained. The reaction has been applied to other NH-heteroaromatics such as pyrrole, pyrazole, and imidazole with a single product being obtained in each case.

Specifically pyrrole gave 2-phenylpyrrole in 86% yield, pyrazole gave 2-phenylpyrazole in 81% yield, and imidazole gave 4-phenylimidazole in 72% yield The regioselectivity of the arylation of imidazole can be altered by adding a co-catalyst such as copper iodide. Under these conditions the reaction produces 2-phenylimidazole in 83% yield. Similarly 2-phenylbenzimidazole was formed in 90% yield and 2-phenylpyrimidine was formed in 78% yield. N-Arylation is also possible if an alkali metal base is used in place of a Grignard reagent provided a suitable ligand (such as DPPF, tBu₃P etc) is added. (J. Am. Chem. Soc., 2003, 125, 5274-5275).

The Dienyl Pauson-Khand Reaction

A study aimed at determining whether intermediates in the [4+2] cycloaddition of dienynes could be trapped with CO has led to the development of a dienyl Pauson-Khand reaction by Wender and co-workers. In principle multifunctional compounds such as (1) can react by a number of pathways, but by optimising the reaction conditions and the solvent a high yield (89%) of (2) can be realised.

Interestingly the yield was improved by reducing the catalyst loading ([RhCl(CO)(PPh₃)₂] – AgSbF₆) and this also allowed the reaction to be carried out under 1 atmosphere of CO. In addition this must be one of the few examples in the literature where the isolated yield of product (89%) is higher than the GC yield in solution (88%), albeit only just! (Angew. Chem. Internat. Ed.. 2003, 42, 1853-1857).

The Pauson-Khand Reaction: the Catalytic Age Is Here

Whilst on the subject of the Pauson-Khand reaction, a review of catalytic versions of the reaction that have been developed over the last 2-3 years has come out recently. The review also includes up to date references on asymmetric variants on the process such as those shown below which was developed by Buchwald and Hicks, and Shibata and Tagaki respectively. (Angew. Chem. Internat. Ed., 2003, 44, 1800-1810).

Construction of Quaternary Stereocenters: New Perspectives through Enantioselective Michael Reactions

Christoffers and Baro have reviewed recent work on enantioselective Michael reactions, which enable quaternary stereocentres to be generated with >90% ee. Compounds such as the four examples shown below can be produced by reaction between the appropriate 1,3-dicarbonyl compound and an enone catalysed by Pd(II)-diaquadiphosphanes such as (R)-tol-binap-Pd(OH)₂ and (R)-binap-Pd(OH)₂. Other catalysts such as the La- binol system developed by Shibasaki and the zinc versions using linked-binol ligands are also covered.

Also included are examples of chiral auxiliary mediated asymmetric Michael reactions. (Angew. Chem. Internat. Ed., 2003, 42, 1688-1690).

Diastereoselective Temporary Silicon-Tethered Ring-Closing-Metathesis Reactions with Prochiral Alcohols: A New Approach to Long-Range Asymmetric Induction

Evans et al have shown that long range asymmetric induction is possible via the temporary silicon tethered ring closing metathesis reaction. Typically in examples of the type shown

ratios of (1):(2) are 20:1 or better depending on the nature of R (nPr, c-Hex, iBu, PhCH₂, Ph(CH₂)₂, BnOCH2, BnO₂CH₂), with ratios of >99:1 being observed in some cases (R = 2-naphthyl, Ph).With larger rings the d.r. is not only lower but also reversed with homologues of (2) (prepared from RCH(OH)(CH₂)_nCH=CH₂, n = 1-4) being preferred to homologues of (1) . (Angew. Chem. Internat. Ed. , 2003, 42, 1734-1737).

A Water-Soluble and “Self-Assembled” Polyoxometalate as a Recyclable Catalyst for Oxidation of Alcohols in Water with Hydrogen Peroxide

Several catalytic methods for the oxidation of alcohols to carbonyl compounds have been published in recent years and another has just been reported. Na₁₂[WZnZn₂-(H₂O)₂(ZnW₉)₃₄)₂] which can be prepared by drop-wise addition of zinc nitrate to a nitric acid solution of sodium tungstate, is an effective catalyst for which oxidises secondary alcohols to ketones (yields ~ 95%) and activated primary alcohols to carboxylic acids (benzyl alcohol gives 100% benzoic acid). Unactivated primary alcohols such as pentanol do not give such good results (66% conversion to a mixture of pentanal (9%) and pentanoic acid (91%) , and 61% conversion to a 41:59 mixture of pentanal and pentanoic acid in the presence of TEMPO). 1,3-Diols, such as 2-ethyl-1,3-hexanediol, are oxidised to b –ketoacids in good yield. All the examples cited were carried out in a two phase system with the aqueous polyoxometalate oxidant and the neat (liquid) alcohol. Other examples include the successful oxidation of 2-butyl-4-chloro-5-hydroxymethylimidazole to the corresponding carboxylic acid (> 95% yield), but the reaction failed with 1-cyclohexyl-3,3,3-trifluoro-2-propanol. ( J. Am. Chem. Soc. , 2003, 125, 5280-5281).