Controlling Diastereomers

For some oxiranyllithium species, which we have discussed in Chap. 2, two stereoisomers can exist. In such cases, an oxiranyllithium species has another side reaction that needs attention Isomerization caused by inversion of configuration at its carbon atom bound to lithium in addition to its decomposition involving ring opening. To synthesize a substituted epoxide stereoselectively, we need to suppress such isomerization and allow only one isomer to selectively react with an electrophile. For this purpose, high-resolution reaction time control using a flow microreactor system is effective. 

Such isomerization can be observed when a trisubstimted oxirane shown in Fig. 6.2 is used. In more detail, oxiranyllithium species c-E, which results from deprotonation of (2R, 3S )-2-methyl-2,3-diphenyloxirane (c-D), reacts with methyl iodide to give tetrasubstituted oxirane c-F. When, however, the species c-E isomerizes to t-E, t-E reacts with methyl iodide to give the corresponding 

6 Controlling Isomerization by High-Resolution Reaction Time. .. 

3 Generation and reaction of oxiranyllithium in a flow microreactor system 

When the temperature is lowered further to -48 °C, the isomerization can be suppressed almost completely. As a result of this, the starting material c-D can be deprotonated to generate oxiranyllithium species c-E, which can then react with methyl iodide to selectively give product c-F, without any decomposition and isomerization of c-E. 

---

When both a-positions of the oxime possess active hydrogen, the regiochemistry of the Hoch-Campbell reaction prefers the side with more available hydrogens— indicating the process is kinetically controlled. In case of oxime 36, azirine 37 was not formed. Instead, azirine 38 was obtained exclusively. Addition of the third equivalent of the Grignard reagent delivered aziridine 39 as a mixture of two diastereomers. 

For the kinetically controlled formation of 1,3-disubstituted tetrahydro-P-carbolines, placing both substituents in equatorial positions to reduce 1,3-diaxial interactions resulted in the cw-selectivity usually observed in these reactions." Condensation reactions carried out at or below room temperature in the presence of an acid catalyst gave the kinetic product distribution with the cw-diastereomer being the major product observed, as illustrated by the condensation of L-tryptophan methyl ester 41 with benzaldehyde. At higher reaction temperatures, the condensation reaction was reversible and a thermodynamic product distribution was observed. Cis and trans diastereomers were often obtained in nearly equal amounts suggesting that they have similar energies."... 

In the 1,3-dipolar cycloaddition reactions of especially allyl anion type 1,3-dipoles with alkenes the formation of diastereomers has to be considered. In reactions of nitrones with a terminal alkene the nitrone can approach the alkene in an endo or an exo fashion giving rise to two different diastereomers. The nomenclature endo and exo is well known from the Diels-Alder reaction [3]. The endo isomer arises from the reaction in which the nitrogen atom of the dipole points in the same direction as the substituent of the alkene as outlined in Scheme 6.7. However, compared with the Diels-Alder reaction in which the endo transition state is stabilized by secondary 7t-orbital interactions, the actual interaction of the N-nitrone p -orbital with a vicinal p -orbital on the alkene, and thus the stabilization, is small [25]. The endojexo selectivity in the 1,3-dipolar cycloaddition reaction is therefore primarily controlled by the structure of the substrates or by a catalyst. 

Scheeren et al. reported the first enantioselective metal-catalyzed 1,3-dipolar cycloaddition reaction of nitrones with alkenes in 1994 [26]. Their approach involved C,N-diphenylnitrone la and ketene acetals 2, in the presence of the amino acid-derived oxazaborolidinones 3 as the catalyst (Scheme 6.8). This type of boron catalyst has been used successfully for asymmetric Diels-Alder reactions [27, 28]. In this reaction the nitrone is activated, according to the inverse electron-demand, for a 1,3-dipolar cycloaddition with the electron-rich alkene. The reaction is thus controlled by the LUMO inone-HOMOaikene interaction. They found that coordination of the nitrone to the boron Lewis acid strongly accelerated the 1,3-dipolar cycloaddition reaction with ketene acetals. The reactions of la with 2a,b, catalyzed by 20 mol% of oxazaborolidinones such as 3a,b were carried out at -78 °C. In some reactions fair enantioselectivities were induced by the catalysts, thus, 4a was obtained with an optical purity of 74% ee, however, in a low yield. The reaction involving 2b gave the C-3, C-4-cis isomer 4b as the only diastereomer of the product with 62% ee. 

The enantiomers are obtained as a racemic mixture if no asymmetric induction becomes effective. The ratio of diastereomers depends on structural features of the reactants as well as the reaction conditions as outlined in the following. By using properly substituted preformed enolates, the diastereoselectivity of the aldol reaction can be controlled. Such enolates can show E-ot Z-configuration at the carbon-carbon double bond. With Z-enolates 9, the syn products are formed preferentially, while fi-enolates 12 lead mainly to anti products. This stereochemical outcome can be rationalized to arise from the more favored transition state 10 and 13 respectively ... 

To control the first factor, one of the two lone pairs of the sulfide must be blocked such that a single diastereomer is produced upon alkylation. For C2 symmetric sulfides this is not an issue, as a single diastereomer is necessarily fonned upon alkylation. To control the second factor, steric interactions can be used to favor one of the two possible conformations of the ylide (these are generally accepted to be the two conformers in which the electron lone pairs on sulfur and carbon are orthogonal) [14], The third factor can be controlled by sterically hinder-... 

The C2 symmetry of sulfide 13 means that a single diastereomer is formed upon alkylation (Scheme 1.10). Attack from the Si face of the ylide is preferred as the Re face is shielded by the methyl group cis to the benzylidene group (28). Metzner postulates that this methyl group also controls the conformation of the ylide, as a steric clash in 27b renders 27a more favorable [16]. However, computational studies by Goodman revealed that 27a was not particularly favored over 27b, but it was substantially more reactive, thus providing the high enantioselectivity observed... 

In the case of sulfide 7 the bulky camphoryl moiety blocks one of the lone pairs on the sulfide, resulting in a single diastereomer upon alkylation. One of the conformations (29b) is rendered less favorable by non-bonded interactions such that conformation 29a is favored, resulting in the observed major isomer (Scheme 1.11). The face selectivity is also controlled by the camphoryl group, which blocks the Re face of the ylide. 

A synthetically useful diastereoselectivity (90% dc) was observed with the addition of methyl-magnesium bromide to a-epoxy aldehyde 25 in the presence of titanium(IV) chloride60. After treatment of the crude product with sodium hydride, the yy -epoxy alcohol 26 was obtained in 40% yield. The yyn-product corresponds to a chelation-controlled attack of 25 by the nucleophile. Isolation of compound 28, however, reveals that the addition reaction proceeds via a regioselective ring-opening of the epoxide, which affords the titanium-complexed chloro-hydrin 27. Chelation-controlled attack of 27 by the nucleophile leads to the -syn-diastereomer 28, which is converted to the epoxy alcohol 26 by treatment with sodium hydride. 

The alkynyl reagent 9 was recently introduced for the dia stereoselective synthesis of tertiary propargylic alcohols144. 9 can be prepared as a solid 1 1 complex with tetrahydrofuran by treatment of 9-methoxy-9-borabicyclo[3.3.1]nonane with (trimethylsilylethynyl)lithium, followed by addition of boron trifluoride-diethyl ether complex. The nucleophilic addition of reagent 9 to (R)-2-methoxy-2-methylhexanal (10) afforded a mixture of the diastereomers 11 with a considerable preference to the nonchelation-controlled (3S,4R)-isomer144. 

Thus chelation control " may lead to either product, depending on the relative stabilities of the respective ot- and /(-chelates. In cases with predominant formation of the anri-diastereomer, it is often difficult to establish whether the formation of a /(-chelate or an open-chain Felkin - Anh transition state is responsible for the observed stereochemistry the decision usually rests on plausibility considerations. Thus, with regard to the results obtained for a-alkoxy carbonyl... 

With the amide 8 (R1 = C 11,) derived from 2-oxopropanoic acid and amine E, the (2 R)-diastereomer is predominantly formed, regardless of the solvent, through chelation-controlled Re-side attack of the organometal14. Presumably, the weaker steric interaction between the pyrrolidine moiety and the methyl substituent of the amide (R1 = CH3) compared to the phenyl substituent (R1 = C6H5) facilitates the preferential formation of the chelated conformer S-m-8. 

Control experiments, performed with the ( + )-(R)-diastereomer of 1, which differs only in the configuration of the stereogenic center at the metal, afford the enantiomeric homoallylic alcohol, (S)-3-methyl-1-phenyl-3-butenol, also with high enantiomeric excess, indicating that the chiral cyclopentadienyl ligand has no dominating influence1-2. 

Crystalline, diastereomerieally pure syn-aIdols are also available from chiral A-acylsultams. lhe outcome of the induction can be controlled by appropriate choice of the counterion in the cnolate boron enolates lead, almost exclusively, to one adduct 27 (d.r. >97 3, major adduct/ sum of all other diastereomers) whereas mediation of the addition by lithium or tin leads to the predominant formation of adducts 28. Unfortunately, the latter reaction is plagued by lower induced stereoselectivity (d.r. 66 34 to 88 12, defined as above). In both cases, however, diastereomerieally pure adducts are available by recrystallizing the crude adducts. Esters can be liberated by treatment of the adducts with lithium hydroxide/hydrogen peroxide, whereby the chiral auxiliary reagent can be recovered106. 

Complete chelation control but lower simple diaslereoselectivity is observed when the corresponding ( ,)-enolsilane reacts with (S)-2-benzyloxypropanal the ratio of diastereomers [(2S,3.S,45)/(2/ .3.S,4S)] is 85 153. 

All four possible diastereomers are formed from the addition of the same Z-azaenolate to a series of aldehydes. Both the ratio of topside (major)/bottomside (minor) attack (4 1, controlled by the dihydroisoxazole substituents) and the diastereofacial selectivity (syn/anti ratio) are nearly independent of the structure of the aldehyde used26. 

Upon carefully controlled hydrolysis with hydrochloric acid at room temperature, the corresponding serine methyl esters 4 are obtained in reasonable yields. Higher yields of 4 arc obtained by hydrolyzing with dilute trifluoroacetic acid5. In some cases, the diastereomeric ratio of 4 does not exactly correspond to the d.r. of the adduct 3, which is attributed to different kinetics in the hydrolysis of the diastereomers 4. Subsequent treatment of the methyl ester with excess 5 N hydrochloric acid and methyloxirane as an acid scavenger results in the free amino acid 54,7. 

Michael addition of the enolate of (42 )-4-rm-butyl-3-methyl-2-oxetanone to dimethyl (Z)-butenedioate yields a single diastereomer. This provides a method to control two new vicinal stereogenic centers one quaternary and one tertiary. The topicity of the addition is u with respect to the 3,3 -bond and l with respect to the 3, 4 -bondI09. 

---