Some novel studies in stereochemistry

Abstract

The thesis entitled “Some Novel Studies in Stereochemistry” describes, in five chapters, studies aimed at developing novel methodologies in organic asymmetric synthesis, clarifying the mechanism of formation of the interesting cavitand cyclotriveratrylene, and, lastly, understanding the origin of optical activity in nature.

Chapter 1. This briefly surveys current activity in organic stereochemistry. Important and interesting developments in terms of new definitions and nomenclature are described. Recent trends in asymmetric synthesis are reviewed. A few examples pertaining to recent developments in stereochemical methodology based on new conceptual insights are described.

Chapter 2. This chapter, in two sections, describes the design and development of a new chiral auxiliary, viz., an oxazoline based on camphor. The studies were motivated by the well?known usefulness of both chiral oxazolines and the camphor framework in asymmetric synthesis. Section 1 describes the synthesis and use of the oxazoline in asymmetric alkylation. The present representative is a tricyclic oxazoline (3) prepared in two steps from the known apocamphylamine (1), by NaBH? reduction of the ketone followed by treatment of the resulting aminoalcohol (2) with ethyl iminopropionate hydrochloride in the presence of base (Scheme 1). The chiral apocamphylamine was itself derived from 1S-(+)-10?camphorsulfonic acid, by a vastly improved version of the reported procedure. The potential of 3 in diastereoselective alkylations was then explored, based on the well?known studies reported by Meyers and co?workers, the final objectives being chiral ??alkylpropanoic acids (5) (Scheme 2). Lithium diisopropylamide (LDA) deprotonation of 3 at ?78?°C in THF solution, followed by quenching of the resulting anion with various alkyl halides, resulted in totally diastereoselective alkylations as observed by NMR spectroscopy, which detected only one (4) of the two possible diastereomeric alkylation products. Acidic hydrolysis of the oxazoline moiety in 4 to obtain 5 by the available methods was accompanied by considerable racemisation; partial acidic hydrolysis to the intermediate ester followed by attempted saponification resulted in a facile O?to?N acyl transfer. This was circumvented by developing a new strategy: initial treatment of 4 with acetic anhydride followed by saponification of the resulting amide–ester. The overall chemical yields of (2R)?5 obtained were >50%, and the enantiomeric excesses >95%, as determined polarimetrically. These results are explicable on the basis of diastereoselective deprotonation of 3 to yield a Z?olefinic vinylaza anion (not shown), which is alkylated on the less hindered Re face (i.e., the endo side of the camphor moiety). Section 2 deals with diastereoselective aldol condensations of 3 with various aldehydes (Scheme 3). The aldols (6a and 6b) formed in a syn:anti ratio of 1:2; although attempts to improve selectivity by changing base, solvent, or ionic strength were unsuccessful, addition of tetramethylethylenediamine (TMEDA) altered the ratio to 1:3. Diastereoselectivity at the newly formed stereocentres was apparently total, as far as could be observed by NMR. Hydrolysis of 6 with 6?N H?SO? furnished the corresponding ??hydroxy acids in fair yields. However, it was observed that ??hydroxy acids derived from aromatic aldehydes were completely racemised.

Chapter 3. This chapter describes a general strategy for catalytic deracemisation, particularly as applied to two ketones. In general, deracemisation with a chiral catalyst is impossible: because enantiomers have equal free energies in solution, they must be present in equal amounts at equilibrium, regardless of the catalyst (microscopic reversibility). Interestingly, catalytic deracemisation is feasible in select cases—two of which form part of this study—the following arguments being novel as far as we know.

Meso–dl manifold under kinetic control: For a compound with two chiral centres in its meso form, it is, in principle, possible to selectively invert one centre under kinetic control with a chiral catalyst. If the selectivity is total, this process establishes an equilibrium between the meso form and only one of the enantiomeric forms. The meso form enjoys an entropic advantage of R?ln?2 over any one enantiomer and would thus predominate by a factor of 2 at normal temperatures in this “kinetically controlled equilibrium.” Partial selectivity yields an equilibrium favouring the meso form by a factor greater than 2, and total equilibration over longer durations results in complete racemisation with meso and dl forms present in equal amounts. (Enthalpy effects are ignored here for simplicity.) Interestingly, the action of certain deracemase enzymes appears to confirm the above analysis.

Heterogeneous interphase mechanism: Catalytic deracemisation can also occur in a heterogeneous process where the racemate is in one phase and an enantiomer in another. Here the free energies differ (racemate vs. single enantiomer), so a selective catalytic process at the inter?phase could equilibrate the racemate with only one enantiomer. A solid–liquid equilibrium is preferable experimentally; an important requirement is that the racemate be a racemic compound, not a conglomerate.

These ideas were explored experimentally with two ketones: meso?2,4?diphenylpentan?3?one (7) and 2?phenylcyclohexanone (8) (Scheme 4), employing the chiral base thebaine (not shown) for selective inversion of the centres ? to the carbonyl groups via deprotonation. Optical purity was generated by the catalytic process for diphenylpentanone (7), although the ee could not be ascertained as the specific rotation of 7 is unknown. For 2?phenylcyclohexanone (8) (heterogeneous equilibration), enantiomeric enrichment of ~9% was observed in hexane at 0?°C. Further studies are indicated.

Chapter 4. This chapter aims to elucidate the mechanism of formation of cyclotriveratrylene (15) from veratryl alcohol (9) (Scheme 5), first observed in 1915 but remaining partly mysterious. The central puzzle is the exclusive formation of 15 from three molecules of 9, rather than 11 from two molecules of 9. The present studies are based on the supposition that the preferred formation of 15 is due to the intermediacy of the o?xylylene 12, formed via deprotonation of the carbocationic precursor 10—o?xylylenes being known to be overwhelmingly favoured energetically over the corresponding xylyl cations. The essential idea is that the geometry around the more substituted dimethane unit, being trans vis?à?vis the other such unit, precludes cyclisation in 12 (to 11) but not in 13 (leading to 15), because of the longer chain in 13. This mechanism suggests that reaction of 9 in MeCO?D might lead to deuterated 15, via possible equilibration of 10 and 12; surprisingly, no deuterium incorporation was observed. However, when ?,??dideuterioveratryl alcohol was used as the substrate in MeCO?H, partially deuterated 15 was formed, indicating proton incorporation. These apparently conflicting findings are explicable by invoking a transannular proton shift in 14, the immediate precursor to 15 (as shown); also, the observation that o?xylylenes are overwhelmingly favoured over xylyl cations is vindicated by the lack of deuterium incorporation mentioned above. Attempts to trap xylylene 12 via cycloaddition with maleic anhydride were unsuccessful. It is possible that 12 exists largely as the isomeric benzocyclobutene (not shown) under the reaction conditions, which would also explain the reported observations.

Chapter 5. This chapter describes studies aimed at testing an important current theory on the origin of optical activity in nature, based on the amplification of parity?violating energy differences (PVEDs). PVEDs arise from the electroweak force, a universal chiral influence because of which enantiomeric molecules possess slightly different energies. The difference is infinitesimal, however, and is typically manifested only upon amplification. The Yamagata mechanism proposes that such amplification can occur during the formation of chiral crystals from achiral molecules. The present study is based on the spontaneous generation of optical activity found during the formation of certain urea inclusion complexes. Although urea is achiral, the crystals of the inclusion compounds are hexagonal and chiral, and the generation of optical activity suggests that the Yamagata mechanism may be operating. However, the optical activity generated is not chirally consistent, being sometimes dextrorotatory and sometimes levorotatory, apparently at random. It is likely that a random “chance factor” (e.g., accidental nucleation) is superimposed upon the Yamagata process. If so, a relatively large number of trials should reveal the deterministic (Yamagata) effect as a persistent chiral bias. A test is experimentally feasible because, in the inclusion of a racemic guest into urea, one enantiomer is selectively included, enriching the supernatant liquid in the other enantiomer. Thus, the chiral bias in crystal formation is mirrored in the supernatant, facilitating polarimetric measurement. A large amount of data can thus be collected relatively easily for statistical analysis. This contrasts with a reported experiment that concluded, in support of the Yamagata theory, that there is a 1% enantiomeric excess of l?quartz in nature, but only after analysing ~17,000 crystals. In the present study, three guests were chosen: the n?hexyl, n?heptyl, and n?octyl esters of 2?methylhexanoic acid. A solution of each of these (0.1?M) in 5?M urea in methanol was allowed to crystallise at 0–5?°C over 24?h. The supernatant was then checked for optical activity by polarimetry. The contents of the polarimeter cell were returned to the crystallisation vessel and the operations repeated to obtain the next reading. A total of about 150 readings could be obtained fairly easily by this method. The results indicate a small bias, but not unambiguously. Initially, a very large bias (32/45 trials) was observed in the case of the hexyl ester, but this almost disappeared upon filtration of the solution prior to crystallisation. A similar, but marginal, bias was observed in the other cases (typically 20/36 trials). The inclusion of the levorotatory ester is favoured in all cases, indicating a persistent bias in favour of the corresponding chiral urea–ester crystalline complex. Interestingly, the relation between configuration and rotational sign is the same in all the above esters, so a consistent chiral bias in the crystallisations is indicated. However, it is known that chiral impurities, particularly of biological origin (hair, pollen, etc.), induce chiral bias in these crystallisations. Thus, an element of ambiguity remains. All the same, the studies define a new strategy for investigating a fundamental question of enduring fascination, viz., the origin of optical activity in nature.