摘要:

Dynamic Resolutions in Asymmetric Synthesis S. Caddick and K. Jenkins The Chemistry Laboratory University of Sussex Brighton UK BN I 9QJ 1 Introduction Equilibration The introduction of chirality into organic molecules is an increas- ingly important consideration when planning a route to targets of biological importance. Living systems are inherently chiral because many essential biomolecules exist in homochiral form. These include carbohydrates which are involved in cell signalling and recognition processes and act as substrates in many key biochemi- cal pathways; and also the amino acids which are important bio- chemical substrates and are the primary constituents of structural proteins enzymes and receptors. It is not surprising therefore that biological activity is highly enantiomer dependent. Synthesis of a racemic compound in which one enantiomeric form is poorly active or inactive is inefficient. The administration of a racemate as a pharmaceutical is undesirable as it usually requires a higher dose to elicit the desired response. Furthermore the pres- ence of the other enantiomer may have adverse side effects and indeed the commercialisation of racemic drugs is becoming increas- ingly difficult to justify. Thus the demand for methods for the pro- duction of chiral non-racemic compounds has increased rapidly in response to these commercial considerations. Most methods used for the preparation of enantiomerically enriched chiral organic molecules can be classified into two distinct strategies. The first involves the stereocontrolled formation of the new stereogenic centre it. as the new chiral element is formed it is done so in a non-racemic fashion. This demands that the reactive centres experience some stereo-discriminating environment in the transition state. This can originate from an existing stereogenic centre in the substrate (chiral substrate or auxiliaries) or via a chiral reagent or catalyst (chemical or enzymatic). The second approach involves resolution; this utilises a stereoisomeric mixture and does not demand asymmetric induction in the formation of any new chiral element. Thus preparation of a single stereoisomer by reso- lution of a stereoisomeric mixture may be achieved via a conven- tional separation procedure or by exploiting the difference in reactivity (kinetic resolution). The major drawback with the con- ventional resolution approach is that the yields will always be limited to 50%. However this limitation can be overcome if the stereoisomers can interconvert; it may then be possible to selec- tively manipulate one of the isomers and hence effect a dynamic resolution with greater than SO%# yield of the desired isomer (Scheme 1). Stepheti Caddick received his undergraduate education at Paisley College of Technology and then part time at CCAT (Cambridge) whilst working in the luborato- ries of FBC Ltd (Saffron Walden).In 1989 he received a Ph D degree from Southampton University working with P. J. Parsom and then spent two years at Imperial College us a postdoctoral ass istant with W. B. Motherwell. In 1991 he was appointed to the Jta# at Birkbeck College in London and in 1093 moved to the University of Sussex. 447 (@-A (S)-A B' Chiral :El* Chiral resolving resolvingagent agent i (S)-AB*r(@-AB* + (S)-AB' 1 (RJ-AW >50°/o <So% Separation and Separation and dissociation 1 diss;ig!on Non-dynamic resolution Dynamic resolution Scheme 1 2 Dynamic Kinetic Resolution (DKR) This area has recently been the subject of two excellent review arti- cles by Noyoril and Ward.2 It is our intention in this review to high- light examples of dynamic kinetic resolutions which have not previously been covered by these articles. In cases where we have felt compelled to include work already covered by these reviews we have referred to the appropriate review article(s). It may be possible to effect a dynamic kinetic resolution (DKR) between rapidly equilibrating stereoisomers (A B) by virtue of dif-ferential chemical reactivity. For a DKR to be successful the rate of epimerisation (k,,k,) must be a fast process relative to the rate of substrate transformation (kAk ). Also the transformation itself should be essentially irreversible and the products formed need to be reasonably configurationally stable under the reaction conditions in order to avoid a thermodynamically derived product distribution. The difference in reactivity of the substrate stereoisomers leads to the bias in the product for one isomer over the other. As the rate dif- ferential (k lk ) becomes more pronounced the stereochemical bias is increasingly reflected in the products (Scheme 2). N-kK A-kA -B k6 * B (Minor product) k0 (Major product) kA kB >> kB > kA ; kA/ kB = KAB Scheme 2 Kerry Jenkins obtained his BSc @om the University of Exeter (1989- 1992). After spending 12 months in the Structural Chemistry department at G laxo Group Research he returned to academia and obtained an MSc in Medicinal Chemistry jrom the University of SusJex (1993-1994). His research involved novel radical ips0 sub- stitution reactions of heterouryl sulfones. He is currently a post- graduate student with Dr Stephen Caddick and is explor- ing dynamic resolution method- ologies. .) + A B [B']/[A'] = exp [(-AGOAB + ActA -AG',)/RTI (first order/pseudo first-order reactions) Scheme 3 This situation can be interpreted using Winstein-Holness and Curtin-Hammett kinetics The product distribution is related to the transformation rates (kA Jc ) and equilibrium constant (KAB)by the equation [B']/[A']= k KAB/kA Expansion with Gibbs free energy and activation free energy parameters provides the equation given in Scheme 3 Where the substrate stereoisomers are related as enan- tiomers chiral discrimination must occur via a chiral catalystlpro- moter or by using a chiral reagent Since AGO = 0 the relative transition state energies (Act) solely determines the kinetic prefer- ence If the substrate stereoisomers are related as diastereoisomers the stereoselection arises as a function of the transition state energies (As B*)relative to the ground state energies (A,B)of the substrate isomers 2.1 Photochemical Promoted DKR Circularly polarised light has been shown to effect chiral enrich- ment in the photochemical cyclisation of ruc-1 to the hexahelicene 2 (Scheme 4)I Both the substrate and product possess planes of chirality as a consequence of their restricted conformations Circularly polarised light (hv') -/ t I2102 '\ rac -1 2 Scheme 4 Precursor ruc-1 rapidly equilibrates at ambient temperature The chiral influence of the circularly polarised light favours the forma- tion of one enantiomer over the other as evidenced by a small but significant optical rotation (85% ca 0 2% ee) The ability of circularly polarised light to influence the stereo- chemical outcome of a reaction is an important observation in its own right but general application to chemical synthesis has yet to be demonstrated 2.2 Reduction of P-Keto EsterdAmides The propensity for a-substituted P-keto esterdamides to undergo enantiomerisationvia keto-enol tautomerism has long been recog- nised This potentially undesirable property has been exploited to great effect in DKR by catalytic hydrogenation Tail demonstrated a modified nickel catalysed heterogeneous hydrogenation of 2-alkyl-3-oxobutyrates using (R,R)-tartaric acid as a chiral modifying agent (Scheme 5) Addition of hydrogen from the catalyst surface (Sr face) in the preferentially bound 2s-3 complex furnishes the 2S,3R-4 syn product 00 OH 0 3 2.533-4,70 7% 79 9% ee Scheme 5 CHEMICAL SOCIETY REVIEWS 1996 The work by Noyori' and coworkers using the homogeneous Rut'-BINAP catalytic system has greatly extended the synthetic utility of this methodology Very high levels of enantio and diastereo-control can be achieved and some examples have found industrial-scale application The DKR of acyclic P-keto esters gen- erally affords anti selectivity when the C-2 substituent is alkyl Diastereoselectivity is often compromised but enantioselectivity can be excellent If the C 2 substituent can hydrogen bond to the ester moiety syn stereoselectivity is usually observed along with high enantioselection The sense and efficiency of diastereoselection is very substrate and solvent dependent whilst the enantio-selection seems to be less variable and IS related to the absolute configuration of the BINAP ligand The C-2 geometry is fixed when hydride is delivered to the C-3 carbonyl Coordination of the ester carbonyl to the ruthenium acts as a trigger for the delivery of hydride The (R) BINAP ligand typically affords the R configuration at the C 3 hydroxy Very good levels of absolute and relative stereocontrol can be achieved when the C-2 alkyl substituent forms part of a ring due to constraints in the transition state As before the chirality of the BINAP ligand controls the ruthenium-carbonyl facial selectivity Much attention has been focused on enzyme-mediated DKR of P-keto esters As early as 1976 Deol et a/ I had demonstrated that the yeast-mediated reduction of ethyl 2-oxocyclohexanecarboxy-late (5) proceeded to give enantiomerically pure lR,2S-ethyl 2 hydroxycyclohexanecarboxylate (6) as the sole product in 69% yield (88% based on conversion of 5) Buisson and Azerad' * have studied the reductions of P-keto esters using various fungal strains Excellent enantioselection (>95% ee) has been achieved along with moderate to excellent diastereoselection Perhaps most significantly they have shown that some organisms provide predominantly untiltrans diastereomersin contrast to yeasts (Scheme 6) 0 OH U U 58% de >99% ee Geotrichrum candldum * 96% de 91% ee Scheme 6 Given the value of such DKR processes the capability of having complementary approaches to different stereoisomers is extremely desirable To be able to generate a molecule with two chiral non racemic centres from a racemic compound in one step by DKR is particularly elegant The accessibility of racemic a-substituted-P- keto carboxylic acid derivatives coupled with the synthetic utility of chiral a-substituted P-hydroxy carboxyl compounds renders this technology an extremely powerful tool With the array of enzyme systems and chiral transition metal catalysts available the synthetic chemist has a variety of alternatives at his/her disposal many of which are complementary in terms of their relative and/or absolute stereocontrol The complex nature of biocatalytic systems makes it difficult to propose accurate predictive models It is apparent that whole cell systems may utilise a range of oxidoreductase enzymes some of which lead to opposing enantioselection Isolated enzymes can often produce higher enantio and diastereo-selectivities than the organ ism as a whole however isolated enzymes can suffer from reduced substrate compatibility Not surprisingly the substrate structure has much influence on the diastereoselectivity and enzyme activity The nature of the C 2 substituent and the alkoxy group for example have DYNAMIC RESOLUTIONS IN ASYMMETRIC SYNTHESIS -S CADDICK AND K JENKINS been shown to have significant effects in microbial reductions Seebach3 has noted increased stereoselectivity in yeast-mediated reductions of P-ketoesters by using non-fermenting conditions 23 Other Enzyme-mediated DKR Transformations Biotransformations have been recognised since the pioneering work on fermentation by Pasteur in the 1850s Historically there has been a reluctance on the part of the synthetic organic chemist to employ biological processes routinely in organic syntheses As the impetus for chemists to synthesise enantiomerically pure molecules has increased over recent decades biotransformations have begun to emerge as viable alternatives to traditional methods of asymmetric synthesis Much of the interest in biotransformations has been driven by the requirements of the chemical industry in particular the pharmaceutical and agrochemical industries The implication of azlactone formation in the racemisation mech- anism of N-acylated peptides and amino acids has generated much interest in this class of heterocycle Condensation of activated N-acyl amino acids gives rise to the configurationally labile oxazolin- 5(4H)-ones The acidity of the 4-proton is enhanced by the adjacent iminyl and carbonyl groups This chiral lability has recently been exploited in DKR processes Bevinakatti* has exploited this racemi- sation mechanism in the transformation of racemic azlactones into enantiomerically enriched amino acid derivatives trans-Acyl ring opening by nucleophiles in the presence of suitable lipase enzymes affords the enantiomerically enriched products with in situ racemi-sation of the azlactone BuOH (2 0 ec I Ph 7 Y H~N*CO~H L-( S)-tert-leucine 9 Scheme 7 An excellent example of this type of DKR has been developed by PFL FOAcI 7d OAc (S)-12 83% 90% ee Scheme 8 Williams and Allen7 have developed an intriguing DKR which uniquely combines a transition metal catalysed racemisation process with an enzymatic resolution The racemic phenyl substi- tuted cyclohexenyl acetate 13 was enantiomerised in situ using 5mol% PdCI,(MeCN) to catalyse a [ 1,3]-sigmatropic acetate shift Enantioselective hydrolysis to the corresponding allylic alcohol 14 was achieved by P juorescens lipase (PFL) in 8 1% yield and 96% ee (Scheme 9) In this strategy it is essential to identify a suitable enzyme to effect the kinetic resolution and which is not adversely affected by (or causes adverse effects on) the palladium catalyst PFL 5 mol% PdCI2(MeCN)2 -$:-OHboAcpH 7 0,O1 M phosphate13 buffer 37-40 "C 19 d 14 81Yo,96% ee Absolute configuration not correlated Scheme 9Turner et a14 (Scheme 7) Treatment of rac-2-phenyl-4-tert-butyl-oxazolin-5(4H)-one (7) with butanol in the presence of the immo- bilised Lipozyme@ (Mucor rniehei) enzyme gave (9-N-benzoyl tert-leucine butyl ester (8) in excellent yield and enantiomeric excess The process is amenable to scale-up and is a useful method for the preparation of the unnatural a-amino acid L-(S)-tert-leucine (9) Some interesting (non-enzymatic) DKRs of azlactones have been studied by Weygands and Miyazawa' using chiral amino acid ester nucleophiles Enantiomerically enriched (S)-cyanohydrin acetates have been made via the DKR of racemic cyanohydrins using a lipase mediated acylation Oda and coworkers1 have shown that cyanohydrins can be formed reversibly from aryl aldehydes under basic conditions and this provides a rapid racemisation process Dynamic kinetic res- olution is effected by enantioselective acylation of the cyanohydrin by a lipase from Pseudomonas sp Rayner et al have employed a similar strategy for the enantiose- lective preparation of hemithioacetals They developed a DKR based on the observation that silica gel column chromatography of hemithioacetals promoted dissociation to the starting thiol and alde- hyde Methyl glyoxalate was treated with thiol10 in tert-butyl methyl ether (TBME)and the intermediate hemithioacetal 11 was kinetically resolved using P Juorescens lipase (PFL) and vinyl acetate in the presence of SiO Hemithioacetal acetate (S)-12was obtained in 83% isolated yield with 90% ee (Scheme 8) A range of thiol/aldehyde combinations were similarly coupled with moderate to excellent chemical and enantiomeric yields (63-90% 55-95% ee) 2.4 Configurationally Labile Alkyl halides AIkyl halides which have a halogen at the asymmetric centre are generally configurationally stable However in certain cases epimerisation can be induced for example many a-halo carboxyl compounds and anomeric glycosyl halides exhibit configurational lability usually induced by additives such as polar solvents base or halide sources In 1993 Durst et a1 made the observation that when a diaster- eomeric mixture of 15 (1 1 S,R R,R) was treated with benzylamine in THF the resulting proline derivative 16 was formed as a 7 1 diastereomeric mixture in which the S,R diastereomer predomi- nated (Scheme 10) They concluded that the R,R-15 diastereomer must react with benzylamine significantly faster than the S,R-15 0 0 Ph 15 16 61Yo isolated yeld Scheme 10 diastereomer and that rapid epimerisation of the a-iodide presum- ably catalysed by liberated halide enabled conversion of the slower reacting diastereomer to the faster reacting diastereomer Similarly impressive DKRs were performed on a variety of sub- strates with a range of amine nucleophiles It was found that a-bromo esters gave best results if a catalytic (0 2 equiv ) amount of tetrahexylammonium iodide was added to facilitate epimerisation The generality of this DKR using the pantolactone auxiliary has since been extended to the preparation of a-hydroxy esters9 and C symmetric 25-disubstituted pyrrolidines lo Other chiral auxiliaries have been studied in related DKRs Nunami et a1 1 introduced tert-butyl (4s)- 1 -methyl-2-oxoimidaz-olidin-4-carboxylate,17 as an effective chiral auxiliary for this class of DKR High levels of stereocontrol were observed in nucleophilic displacements carried out on an epimeric mixture of 18 in polar sol-vents under conditions of base-catalysed epimerisation (Scheme 11) DKR of 18 with benzylamine gave 2(R)-19 as the major C02Bu' C02Bu' 17 18 Polar solvent Nucleophilei k02Bu' C02Bu' Nu = PhCH2NH-2(R)-19 (major) 2(S)-19 Nu = (Me02C)&H-2( S)-20 2(R)-20 (major) 0 2(S)- 21 (major) 0 Scheme 11 product (96% 88% de) whilst DKR with sodium dimethyl- malonatei2 and potassium phthal~midel~ gave 2(R)-20(92% 76% de) and 2(S)-21 (90% 94% de) respectively Durst et a1 l4 have reported aminations of a-bromoliodo-acyl derivatives of the imi- dazolidinone 17 using catalytic tetrabutylammonium iodide to effect epimerisation The de values were generally greater than 98% with the R configurationI4 predominating at the site of substitution Yields for alkylation of benzylamine with a variety of substrates varied from 66-87% An analogous DKR utilising an imidazolidinone chiral auxiliary has been developed by our group (see also section 3.4) l5 (4R5S)-I 5-Dimethyl-4-phenylimidazolidin-2-one(22) is readily prepared in one step via the thermal fusion of (-)-ephedrinium chloride and urea Treatment of a diastereomeric mixture (4555 2S 2R) of 23 with benzylamine and catalytic tetrabutylammonium iodide resulted in the predominant formation of 2(R)-24 in quantitative yield (74% de) Conversely treatment of 23 with sodium dimethyl- malonatei6 and tetrabutylammonium bromide gave 2(R)-25in 78% yield with 55-60% de (Scheme 12) A striking feature of the DKRs so far reported with imidazolidi-none derived chid auxiliaries is the anomalous diastereoselectiv- ity observed with amine nucleophiles It would be reasonable to assume that the substrates would preferentially adopt a ground state conformation whereby the imidazolidinone and acyl carbonyl groups are opposing one another and the alkyl substituent R3 would be expected to adopt the geometry shown The transition state energy for halide displacement is lowest when the a-halogen IS perpendicular to the acyl carbonyl group One would expect the faster reacting substrate to be 2(R) whereby the nucleophile CHEMICAL SOCIETY REVIEWS 19% 0 CH3NK,,U cH3 CH3Q Q 22 23 BU4NX [NucleophileTHF 0 Nu. 0 NuA cH3 Q cH3 Q Nu = PhCH2NH-2(R)-24 (major) 2(S)-24 Nu = (Me02C)2CH-2(S)-25 2(R)-25 (major) Scheme 12 approaches from the least sterically hindered face as dictated by the auxiliary directing group(s) (Fig 1) However the major products in all of the aminations have the 2(R)configuration which implies that the 2(S) substrates are the faster reacting diastereomers in S,2 displacements with amine nucleophiIes Nunamii3l7 proposed a model to account for the unusual stereo- selection observed with amine nucleophiles The model invokes a transition state in which the amine nucleophile hydrogen bonds to the ester carbonyl of the auxiliary (4s)tert-butyl carboxylate group thus guiding the amine to displace the halide from the sterically hindered face [Fig 2(a)] In an attempt to support this model the authors prepared the ether analogue 26 which was shown to be sig- '\4' R2 Figure 1 Predicted ground state geometry Figure 2 Predicted conformation for 2(S) diastereomers in amination DKR nificantly less stereoselective in the DKR (Scheme 13) Whilst the selectivity was undoubtedly reduced relative to the ester analogue 18,the sense of stereoselection still results from preferential attack from the more hindered side Furthermore in light of our results with the 1,5-dimethyl-4-phenylirnidazolidin-2-oneadduct 23 which is unable to participate in this 'amine guiding' process this model does not adequately explain the anomalous stereoselection An alternative mechanism invokes a bifurcated hydrogen bond bridging the acyl carbonyl and the auxiliary carbonyl via a six membered chelate The amine nucleophile displaces the halide DYNAMIC RESOLUTIONS IN ASYMMETRIC SYNTHESIS -S CADDICK AND K JENKINS 45 1 CH3 PhCH2NH2 D Et3N HMPA OBu' \OBu' 26 93% 34% de Scheme 13 which occurs faster in the substrate in which the dr orbital of the C-Hal bond is least hindered [Fig 2(b)] At present this model is only supported by circumstantial evidence but may be useful in assisting the prediction of stereochemical outcome in these systems Ward's groupI8 have reported a highly efficient DKR of an a-bro- mopropanoic acid derivative using Oppolzer's chiral camphorsul- tam When a diastereomeric mixture of 27 was treated with dibenzylamine(10 equiv ) in acetonitrile (reflux) or Me,SO (60 "C) 2R-28was formed as the exclusive diastereomer in quantitative yield (Scheme 14) xs Bn2NH MeCN Reflux DMSO 60 "C Br Ph 27 28 >98% yield >98% de Scheme 14 Devine et a1 I9 have used a DKR for the enantioselective prepa- ration of 2-aryloxy carboxylic acids towards a synthesis of a potent endothelin antagonist The substrate 29 was treated with preformed sodium or lithium aryloxide in THF to furnish 30 in good to excel- lent yield (78%-86%) with excellent selectivity (88%-92% de Scheme 15) The analogous substrate using ethyl lactate as a chid auxiliary was found to react sluggishly and with diminished selec- tivity (60%-75% de) 29 CH30eOM \(M = Na LI) 30 78-86% 88-92% de Scheme 15 Matteson and Man2 have achieved a DKR of racemic (l-bro- moalky1)boronic esters by diastereocontrolled reaction with chiral N-acyloxazolidinone enolates under iodide catalysed racemisation (Scheme 16) Alkylation of the lithium enolate of (S)-4-(1-methylethyl)-3-propanoyloxazolidin-2-one~31 with rac-32 in the presence of catalytic NaI and 18-crown-6 afforded 33 (100% con- version >97% de >94%ee) The importance of carbohydrates in biomolecular recognition processes has driven the development of new methods for glycosi- dation Glycosyl halides represent one of the most widely used gly- 31 rac-3218-Crown-6 I ABr I 33 >97%de >94% ee Scheme 16 cosy1 acceptors and many variations of coupling glycosyl halides with donors exist One of the most appealing strategies was intro- duced by Lernieux*O nearly three decades ago The formation of suitably protected p-glycopyranosides from a-haloglycopyranosyl precursors has been a long established procedure However the preparation of a-linked oligosaccharides from the p-haloglycopy- ranosides is marred by the thermodynamic instability of the p-halides over the a-halides by virtue of the strong a-anomeric effect Lemieux and coworkers20 made the observation that glycosyl halides could be induced to anomerise in the presence of tetraalkyl- ammonium halide Although the equilibrium lies heavily in favour of the thermodynamically more stable @-(axial) anomer a signifi- cant concentration of the kinetically more reactive p-(equatorial) anomer is maintained Reaction can proceed preferentially via the more reactive P-halide to give the a-glycoside linkage This process has been utilised in the stereoselective synthesis of the blood-group substance H (type 1):' 37,from the L-fucosyl halide 34 (Scheme 17) Other blood-group determinants have similarly been prepared 34a 34P I Ph Ph BzO OH OBzl35Ph '0 a ACNHI OBzl Deblocking-a-L-Fuc(l-2)-P-~-Gal BzlO (1-+~)-D-GIcNAcOBzl 3736 Scheme 17 2.5 DKR of Configurationally Labile Anions The configurational stability of carbanions in organometallic inter- mediates has received a great deal of attention in recent decades. The main emphasis has been concerned with methods of generating stable chiral anions from chiral ,non-racemic precursors which can then be used in stereoselective reactions. There has been a gradual increase in the realisation that configurationally labile anions can be utilised in asymmetric synthesis by using principles of DKR. Hoffman has developed a useful test which enables the configura- tional stability of a carbanion to be evaluated in relation to reaction with an electrophile.22 The test involves the generation of the anion as a racemate. Treatment of the racemic anion with a chiral racemic electrophile will give a diastereomer ratio that reflects the kinetic difference of formation of the (racemic) product diaster- eomers (Scheme 18). R' R' R' rac -anion rac -diastereomerA rac -diastereomer B M = metal; E = electrophile [RZC-A]/[~UC-B]x k,/k k = rate constant of formation of diastereomer A k = rate constant of formation of diastereomer B Scheme 18 Providing the diastereomeric ratio is significantly different from 1:1 (ideally 1.5-3.0) the electrophile is suitable for the test (if not a different electrophile probe must be used). Treatment of the racemic anion with an equimolar amount of the chiral ,non-racemic electrophile must result in the formation of the (enantiomerically pure) produce diastereomers in a 1 1 ratio if the anion is stable on the reaction timescale at 100% conversion (Scheme 19). If the Provided that [ruc-A]l[rac-B] # 0 R' R' R' Chiral,non-racemic E' R3 * R3**Ap + R2a.A E*R3 R2 R2 rac -anion Chiral non-racemic Chiral non-racemic diastereomer A diastereomer B [A]/[BJ = 1.O if anion is conjigurationally stable on reaction timescale [A]/[B] # 1.O if anion is conjigurationally labile on reaction timescale Scheme 19 product diastereomer ratio differs from 1:1 then this is an indication that the anion is conjgurationally labile on the reaction timescale and that product formation is susceptible to a dynamic kinetic res- olution by virtue of the difference in rates of formation of the product (as determined in the first experiment). The reactions must approach 100%conversion for the results to be meaningful other- wise in the latter experiment kinetic resolution may complicate the interpretation of the data. The use of organometallic bases to effect asymmetric deprotona- tions in the presence of chiral complexing agents such as (-)-sparteine (38)has been the focus of much attention in asymmetric synthesis. There are some notable examples in which it is apparent that enantioselection is achieved not by enantioselective deproto- nation to form a stabilised chiral anion but by deprotonation and subsequent complexation of the interconverting anions by the chiral ligand. In principle there are two possibilities by which an asym- metric transformation can occur hereon. The ligand may selectively complex one of the anion enantiomers such that a single diaster- eomeric complex forms by virtue of the in situ enantiomerisation of the other uncomplexed anion. This complex is then trapped by rapid reaction with an electrophile. Alternatively the rapidly intercon- verting anion enantiomers may both form complexes with the chiral CHEMICAL SOCIETY REVIEWS 1996 ligand such that both diastereomeric complexes exist in rapid equi- librium in solution. Diastereoselective reaction of one of the com- plexes with an electrophile accompanied by in situ equilibration of the unreacted complex could result in enantiomerically enriched products. In practice it may be difficult to establish which process is operating and it may not be possible to classify the examples included in this section. One of the first examples of this type of asymmetric transformation was illustrated by Nozaki and cowork- er~~~in 1971. Lithiation of ethylbenzene in the presence of (-)-sparteine (38) followed by quenching with excess CO gave a mixture of hydrotropic acid 39 and ethylbenzoic acids 40 which were isolated with 87% conversion. The mixture was optically active and based on the proportion of hydrotropic acid (1 8%) and the optical rotation it was estimated that (-)-(R)-hydrotropic acid was formed with 30% ee (Scheme 20). Bu"Li ' Li (-)-sparteine 38 - Hexane (-)-sparteine = Li' xs cop1 902H 40 39 10%. 30% ee Scheme 20 Beak24 and coworkers have demonstrated that the intermediate sparteine complexes 42-( -)-38 derived from the lateral lithiation of 41 with BusLi/( -)-sparteine (38) are configurationally labile as evi- denced by the Hoffman The nature of the electrophile had a profound effect on the sense of stereoselection; chlorides such as allychloride gave the R* products whereas tosylates such as allyl- tosylate gave the S* products (Scheme 21). It is suggested that the 41 42 Scheme 21 enantioselection arises through the difference in transition state energy for electrophilic substitution between the two diaster- eomeric anion-sparteine complexes [R-42.( -)-38/S-42-( - )-381 and that the nature of the electrophile determines the preference for inversion or retention on electrophile trapping. This use of a non- covalently bound chid auxiliary is particularly attractive as it allows the production of either enantiomer of a desired product from the same auxiliary. An interesting warm-cool protocol has been introduced by with similar types of substrate. The lability of DYNAMIC RESOLUTIONS IN ASYMMETRIC SYNTHESIS-S CADDICK AND K JENKINS 453 sparteine-anion complexes can be very temperature dependent Hence the mechanism of stereoselectioncan be affected depending on the specific protocol used to perform the reaction Beak has also suggested that the enantioselective electrophilic substitution of the lithio dianion of N-methyl-3-phenylpropionamidein the presence of (-)-sparteine may occur via DKR 26 Hoppe et a1 27 have demonstrated the lability of the a-thioalkyl-lithium sparteine complexes 44a-c in contrast to their a-alkoxyl-lithium analogues Trappingof 44a-cf -)-38 complexeswith CO or chlorotrimethylsilanegave the enantiomerically enriched prod-ucts in good yield (77-95%) and with enantiomeric excesses between 40-6076 (Scheme 22) The apparent independence of 43--78OC BdLi EtzO 1(-)-spartene 38 I-1 XN',X O% 45a-c R EX E Yield (%) Ee (96) a Pr TMSClco CWH3 TMS 91 91 47 46 b CH CO COP 89 40 TMSCl TMS 95 40 c Pr' CO cop 77 60 Carboxylic acids converted to methyl esters using diazomethane Scheme 22 enantiomeric excess on electrophile led the authors to suggest that the product ratio was a reflection of the thermodynamicratio of the equilibratingdiastereomericcomplexes (1 e the equilibrationrate is slow compared to electrophilic trapping) Schlosser and Limat2*have shown that the configurationally labile (-)-sparteine (38) complex of the lithio anion of N-Boc-N-methylbenzylamine (46) gives rise to enantiomerically enriched products on reaction with electrophiles Intriguingly the sense of enantioselection was dependent on the solvent used For example N-Boc-N-methylphenylalanine47 was prepared by treatment of 46 with BusLi,(-)-sparteine and then CO (Scheme 23) In hexane and C02H I) Bu'Li (-) sparteine 38 "CH3 II) co;!rAo Hexane Et200rTHF * I I 46 + 43-65% 47 + Hexane (R)-47 82% ee Et20 (R)-47 67% ee THF (59-47 85%ee Scheme 23 diethyl ether the R enantiomer predominated (82% ee and 67% ee respectively) whilst in THF the S enantiomer was formed in excess (85% ee) The solvent effect was ascribed to a mechanisticcrossover from electrophile substitution on a contact species with hexane and diethyl ether (retention) and a solvent separated ion pair with THF (inversion) An interesting observation was made that in the case where THF was the solvent the ee increased slowly with increasing conversion until the point at which precipitation was observed At this point the enantiomeric excess rose rapidly Heterogeneity is another potential factor in what is already a mechanistically complex class of reaction (see section 3 2) Kumada' utilised the configurational lability of secondary Grignard reagents in an asymmetric Grignard cross-coupling reac-tion catalysed by chiral transition metal complexes The chiral fer-1-rocenyl ligand (S)-NAN-dimethyl-[(R)-2-(dipheny1phosphino)-ferrocenyllethylamine (48) was used in conjunction with NiCI to catalyse the coupling of racemic I -phenylethylmagnesiumchloride (49) with vinyl bromide This gave (!?)-3-phenylbut-l-enein >95% yield and 66% ee A number of variants on this cross-couplingreac-tion have since been developed Hayashi and Ito et a1 have used the C symmetric ligand (+)-50 in a palladium catalysed coupling of 1-phenylethylzincchloride 51 and vinyl bromide (R)-3-phenyl-but-l-ene was obtained in quantitative yield and 93% ee (Scheme 24) &:Mez CH3 H 4&jMe2 FeI PPhz L' 48 L' 50 49 M=Mg 51 M =Zn Scheme 24 3 Crystallisation Induced Dynamic Resolution (CIDR) A very practical and efficient method of asymmetrictransformation is by dynamic resolution of chirally labile substrates by way of selective crystallisation This approach is of general synthetic appeal from small scale preparations (mg-g) to pilot and process scale production (kg-tonnes) The earliest example of CIDR was observed by D~brunfaut~~in 1846who was studying crystallisation of glucose from solution There is great appeal in being able to transform a racemate (or epimeric mixture) to a pure isomer quan-titatively by crystallisation from solution ?O-72 A compound which is either spontaneouslylabile in solution or in which lability can be induced may be able to undergo CIDR through interaction with some stable chiral influence Some enantiomers can be resolved by crystallisation from solution by the action of a homochiral seed crystal which may be added or which may be the first enantiomer to randomly crystallise from the racemic solution Enantiomers may also be resolved by the formation of diastereomeric salts or com-plexes (non-covalently bound) with stable chiral resolving agents (CRAs) The CIDR of enantiomericanions with chiral complexing agents is dealt with in a separate category in this review Substrates which are related as diastereomers via one labile stereocentre may be resolved by crystallisation The use of chiral auxiliaries cova-lently bound to a racemic substrate can be a viable method of achieving CIDR when the enantiomers are not suitable for direct CIDR by seeding. Ideally the chiral auxiliary should be inexpen- sive to prepare should impart crystallinity to the substrate and should be readily cleaved and recovered. A distinct advantage of using chiral auxiliaries in a CIDR approach as opposed to control- ling asymmetry in reactions is that the chiral framework does not need to be tailored for steric and electronic interactions. It is not even essential for the chiral auxiliary to be proximal to the labile stereocentre in CIDR since the chiral influence is largely directed intermolecularly in the crystal lattice and at the solution surface interface. However if the auxiliary is too distant from the labile stereocentre the difference in physical properties of the diaster- eomers is likely to be reduced. 3.1 CIDR Enantiomers An impressive total resolution of racemic nanvedine (52) has been carried out by Shieh and Carlson.33 Under basic conditions narwe- dine can racemise via a retro-Michael reaction. Essentially pure (-)-52 can be crystallised in 84% yield from racemic 52 in ethanol/Et,N (9 1) using 2.5% (-)-52 seed crystals (Scheme 25). It is generally accepted that such an approach is limited to conglom- erates.3*J1 EtOHEtsN (9~1) I I1 80 "C i) 68 "C 2.5% (-)-52 ii) 40%. 16 h NCH3 iii) 25 OC Crystallisation 52 (-)-52 04% Scheme 25 32 CIDR Diastereomeric Organometallic Complexes Dynamic kinetic resolutions utilising configurationally labile car- banions were reviewed in section 2.5. An interesting alternative CHEMICAL SOCIETY REVIEWS 1996 Pentand BuLi Cyclohexane (-)-sparteine 38 t -1 T I I NPS2 NPt2 (-)-38 (R)-54 (-)-a8(5')-54 0 56 90%,90% ee Scheme 26 PI; Ph asymmetric transformation has been described by H~ppe.~~ Deprotonation of the ally1 carbamate 53 with BuLi/( -)-sparteine (38)gives rise to the configurationally labile diastereomeric com- plexes (S)-54.( -)-38 and (R)-54.(-)-38. Preferential crystallisation of (S)-54-(-)-38 can be achieved by addition of cyclohexane. The stable solid complex can be converted to the configurationally stable titanate (R)-55by rapid quenching with pre-cooled tetraiso- propyltitanate (TIPT). The titanate (R)-55 affords the homoaldol adduct 56 in 90% yield and 90% ee on treatment with isobu- tyraldehyde (Scheme 26). H~ppe~~has used a similar CIDR approach to selectively crys- tallise the (R)-lithio.( -)-sparteine complex derived from 1-methylindene. Electrophilic quenching of the crystal suspension led to 1 ,I-disubstituted indenes in moderate to good yields (52-79%) and excellent enantiomer excesses (>95% ee). 33 CIDR Diastereomeric Salts Reider et a1.2 used a highly efficient CIDR approach to the non-pep- tidal CCK antagonist L-364,718 developed at the laboratories of Merck (MSD). The amine precursor 3(RS)-57 could be readily racemised in the presence of a catalytic amount of aryl aldehyde. This promoted the racemisation process via the more acidic inter- mediate imine. Hence the spontaneous total resolution of 3(RS)-57 was effected by the addition of catalytic 3,5dichlorosalicaldehyde (3 mol%) and then (1s)-(+)-10-camphorsulfonic acid [(+)-CSA 92 mol%]. It was important to add less than a full equivalent of (+)-CSA in order to maintain a concentration of free amine 57 which catalyses the racemisation of the imine. The diastereomeric salt 3(S)-57.(+)-CSA was obtained in 91% yield and >98% optical purity (Scheme 27). N-Acylation with the indole-2-carboxylic acid moiety provided L-364,178 in 79% yield and 99.8% ee. This prepa- ration could be carried out on a 10 kg scale. 3(R)-57 3(5)-57 Crystallisation H03+&1 1 (+)-CSA PA 3(S)-57*(+)-CSA Scheme 27 Toda and Tanakaj6 observed higher than theoretical yields in the resolution of racemic cyanohydrins with the basic alkaloid brucine. The cyanohydrins can racemise via base induced reversible disso- ciation to hydrogen cyanide and prochiral ketone. Controlled evap- oration of the solvent allowed the formation of diastereomerically pure salts in up to 100% yield. 3.4 CIDR Epimers Ethyl 5-(I '-methyl-5'-methylthiopyrrol-2-ylcarbonyl)-I ,2-dihydro-6-methyl-3H-pyrrolo[1,2-a]pyrrole-1-carboxylate 5% is a potent antiinflammatory analgesic compound of the 5-aroyl- I ,2-dihydro-3H-pyrrolo( 1,2-a]pyrrole- 1-carboxylic acid class (Fig. 3).37Esterification of the racemic parent acid 58a with (R)-(+)-a-methylbenzyl alcohol gave rise to a diastereomeric mixture 58b which could be epimerised in the presence of a catalytic amount of DYNAMIC RESOLUTIONS IN ASYMMETRIC SY NTHESIS-S CADDICK AND K JENKINS 58a R=OH FH3 b R= O*Ph C R = OEt Figure 3 1,5-diazabicyclo[43 Olnon-5-ene (DBN) Crystallisation of the diastereomeric mixture from hot ethyl acetate-hexane ( 1 3) with catalytic DBN (ca 10 mol%) gave essentially one diastereomer (86%yield) This preparation has been carried out on >150 g scale In our group15 we have shown that 2RS-23 can be transformed to 2R-23 in 91 % yield by crystallisation from THF in the presence of catalytic tetrabutylammonium bromide (Scheme 28) I. CH3 0 2(RS)-23 Bu,NBr(20 mol%) t THF f 2(4-23 91% >98% de SN2 Nucleophile I '. CH3 0 Scheme 28 The bromide source induced epimerisation and with slow evap- oration of the solvent the less soluble 2R-23 was selectively deposited The tetrabutylammonium bromide can be removed from the precipitate by aqueous extraction leaving the stable 2R-23 diastereomer which can undergo nucleophilic displacement with amines to give amino acid derivatives of the opposite configuration to those derived from the DKR of the same substrate (section 2.4) The principle of being able to transform a racemate to either enan- tiomeric form of a product via complementary dynamic resolution processes using one chiral form of the auxiliary is quite appealing We are currently exploring the generality of such complementary dynamic resolutions During our study we have attempted DKRs and CIDRs on analogous substrates using Evans' oxazolidinone We have observed1" that the oxazolidinones are more prone to trans acyl ring cleavage and whilst CIDR was shown (85% de) the increased chiral lability compromised stereoselectivity during work-up and subsequent S,2 reactions Induced configurational lability is the key to effective CIDR 4 Concluding remarks Amongst the existing armoury of asymmetric synthetic strategies dynamic resolutions are gaining increasing recognition There is a certain elegance in being able to utilise configurational lability at a chiral centre in the efficient transformation to isomerically enriched or pure products Traditionally the presence of potentially labile stereocentres in intermediates along a synthetic pathway would present cause for concern but given consideration they may provide access to enantiomerically enriched products Although dynamic resolutions are confined to suitably labile substrates this review illustrates the diversity of useful products that can be obtained often with unsurpassed efficiency and from relatively accessible sub- strates The variety of approaches that have been employed in both dynamic kinetic resolutions and in crystallisation induced dynamic resolutions has also been shown to be wide-ranging Many of the examples included in this article have been developed with a view to industrial-process scale or for large-scale laboratory preparation of important research chemicals hence cost efficiency practical and environmental considerations are of paramount importance Resolution of stereoisomers has always been a reliable method of obtaining enantiomerically pure compounds but the fundamental limitation of yield has made it thoroughly unfashionable The use of dynamic resolutions in asymmetric synthesis seems to be increasing and it would appear that chemists are considering these resolution strategies as a serious a1 ternative to conventional methods for asymmetric synthesis Added in proof Recently a CIDR of dipeptide-derived oxazolones has been described H T Stock and N J Turner Tetrahedron Lett 1996,37,6575 Acknowledgements We gratefully acknowledge the University of Sussex for a maintenance award (to K J ) We are also grateful to the EPSRUBBSRC Glaxo- Wellcome Rh6ne-Poulenc Rorer Parke- Davis and Zeneca for support of our programme We are also grate- ful to Professors Beak (Illinois) Hoppe (Munster) and Williams (Bath) and Drs Devine (Merck) Scott and Jackson (Sussex) for useful comments References 1 R Noyori M Tokunaga and M Kitamura Buff Chem Soc Jpn ,1995 68,36. and references therein 2 R S Ward Tetrahedron Asymmetry 1995 6 1475 and references therein 3 D Seebach,S Roggo T Maetzke H Braunschweiger J Cercus and M Krieger Hefv Chim Acta 1987,70 1605 4 N J Turner. J R Winterman R McCague J S Parratt and J C Taylor Tetrahedron Lett ,1995.36 1 113 5 F Weygand W Steglich and X Barocio de la Lama Tetrahedron Suppf 8 Part I 1966,9 6 S Brand M F Jones and C M Rayner Tetrahedron Lett 1995,36 8493 7 J V Allen and J M J Williams Tetrahedron Lett ,1996,37. 1859 8 K Koh R N Ben and T Durst Tetrahedron Lett 1993.28,4473 9 K Koh and T Durst J Org Chem ,1994,59,4683 10 K Koh R N Ben and TDurst Tetrahedron Lett 1994,35,375 11 K Nunami H Kubota and A Kubo Tetrahedron Lett 1994,35,8639 12 A Kubo M Takahashi H Kubota and K Nunami Tetrahedron Letf 1995,36,6251 13 A Kubo H Kubota M Takahashi and K Nunami Tetrahedron Lett 1996,37,4957 14 J A O'Meara M Jung and T Durst Tetrahedron Lett 1995,36,2559 corrigendum SO96 15 S Caddick and K Jenkins Tetrahedron Lett 1996,37,1301 16 S Caddick and K Jenkins unpublished observations 17 H Kubota A Kubo M Takahashi R Shimizu T Da te K Okamura and K Nunami J Org Chem ,1995,60,6776 18 R S Ward A Pelter D Goubet and M C Pritchard Tetra hedron Asymmetry 1995,6,469 and references therein 19 P N Devine U H Dolling R M Heid Jr and D M Tschaen Tetruhedron Lett 19%,37,2683 20 R U Lemieux,K B Hendr1ks.R V StickandK James,./ Am Chem Soc 1975,97,4056. and references therein 21 H Paulsen and C Kolar Chem Ber 1979,112,3190 and references therein 22 R Hirsch and R W Hoffmann Chem Ber 1992,125,975 23 H Nozaki ,T Aratani ,T Toraya and R Noyori ,Tetrahedron 1971 ,27 905 24 S Thayumanavan,S Lee C Liu and P Beak J Am Chem Soc ,1994 116,9755 25 A Basu and P Beak,./ Am Chem Soc ,1996,118,1575 26 P Beak and H Du,J Am Chem Soc 1993,115,2516 CHEMICAL SOCIETY REVIEWS 1996 27 B. Kaiser and D. Hoppe Angew. Chem. Int. Ed. Engl. 1995 34 32 M. M. Harris,Progr. Stereochem. 1958,2,157. 323. 33 W. Sheih and J. A. Carlson J. Org. Chem. 1994,59,5463. 28 M. Schlosser and D. Limat J. Am. Chem. SOC. 1995,117,12342. 34 0.Zschage and D. Hoppe Tetrahedron 1992,48,5657. 29 A. P. Dubrunfaut,C. R. Acad. Sci.. Paris 1846,23,38. 35 I. Hoppe,M. Marsch,K. Harms,G. Boche and D. Hoppe,Angew. Chem. 30 E. L. Eliel S. H. Wilen and L. N. Mander Srereochemistry of Organic Int.Ed. Engl. 1995,34,2158; and references therein. Compounds Wiley 1993. 36 F. Toda and K. Tanaka,Chem. Lett. 1983,661. 31 J. Jacques A. Collet and S. H. Wilen Enantiomers Racemates and 37 W. K. Hagmann Synth. Commun. 1986,16,437. Resolutions Wiley 1981.

ISSN:0306-0012    

DOI:10.1039/CS9962500447

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Index of Authors Barton, Derek H. R., 237 Benniston, Andrew C., 427 Bernardi, Fernando, 32 1 Bernath, Peter F., 111 Bishop, Roger, 3 1 1 Blaschke, G., 141 Bowden, Keith, 171 Boyd, Derek R., 289 Buckel, Wolfgang, 329 Burgess, John, 85 Caddick, S., 447 Cane, David E., 297 Caneschi, Andrea, 101 Cape, J. Neil, 361 Cardillo, Guiliana, 117 Cativiela, C., 209 Chankvetadze, B., 141 Clark, James H., 303 Collison, D., 25 Cornia, Andrea, 10 1 Cotton, Simon A., 219 Daniels, Vincent D., 179 Chemical Society INDEX Volume 25,1996 Davidson, R. S.,24 1 Diamond, Dermot, 15 Easton, Christopher J., 163 Endresz, G., 141 Kresge, A. J., 275 Krygowski, Tadeusz Marek, Lammel, Gerhard, 36 1 71 Randzio, Stanistaw L., 383 Rebek, Julius Jr., 255 Robb, Michael A., 321 Roberts, M.W., 437 Exner, Otto, 7 1 Fujio, Mizue, 129 Garcia, J. I., 209 Gamer, C. D., 25 Gatteschi, Dante, 101 Golding, Bernard T., 329 Grubbs, Edward J., 171 Hadjiivanov, Konstatin I., 61 Harriman, Anthony, 41 Hepburn, John W., 28 1 Jenkins, K., 447 Lincoln, Stephen F., 163 Luo, Guanglin, 297 Macquarrie, Duncan J., 303 Madden, Paul A., 329 Majoral,J. A ., 209 Marston, George, 33 McCaffery, Anthony J., 49 McKervey, M. Anthony, 15 Mehrotra, Ram C., 1 Negishi, Ei-ichi, 4 17 Nowick, James S., 401 Salvatella, L., 209 Sandhoff, Konrad, 37 1 Sauvage, Jean-Pierre, 4 1 Scott, Stephen K., 265 Sessoli, Roberta, 10 1 Sharma, Narain D., 289 Singh, Anirudh, 1 Smith, Eric M., 401 Timms, Peter L., 93 Tomasini, Claudia, 1 17 Tsuno, Yuho, 129 Johnson, Barry R., 265 Joule, J. A., 25 O’Brien, Sean C., 393 Olivucci, Massimo, 32 1 Wainwright, Mark, 35 1 Williams, R. J. P., 77 Kao, Camilla, 297 Otero Arean, C .,I 87 Wilson, Mark, 339 Khosla, Chaitan, 297 Klissurski, Dimitar G., 61 Pairish, Mason, 401 Parkin, I. P., 199 Yus, Miguel, 155 Zecchina, A., 187 Kolter, Thomas, 37 1 Perkins, M. John, 229 Kondakov, Denis Y., 417 Pieper, Rembert, 297 457

ISSN:0306-0012    

DOI:10.1039/CS9962500457

出版商:RSC    

年代:1996

数据来源: RSC