Towards the stereoselective synthesis of α-methylated (2S,3aS,7aS)-octahydroindole-2-carboxylic acid

Abstract

A high yielding and remarkably stereoselective α-methylation reaction of the (2S,3aS,7aS) stereoisomer of octahydroindole-2-carboxylic acid, (S,S,S)-Oic, suitably protected is described. The severe steric hindrance imposed by the fused cyclohexane ring, which prevents the application of Seebach’s self-reproduction of chirality methodology, accounts for the formation of (S,S,S)-(αMe)Oic with high selectivity and retention of configuration.

The incorporation of conformationally constrained α-amino acids into bioactive peptides is an effective strategy to circumvent the toughest issues in the development of peptide drugs: proteolytic degradation and poor bioavailability.¹ Quaternary α-amino acids have proved particularly useful to produce modified peptides with drug-like properties.² The introduction of an alkyl chain at the α-carbon of an amino acid elicits, when incorporated into a peptide, conformational constraints that cannot only enhance the biological activity, but also improve the resistance to proteolysis and ultimately increase the bioavailability. Additionally, quaternary α-amino acids are invaluable tools to elucidate the receptor-bound conformation of bioactive peptides and to gain insight on the three dimensional structure of protein receptors that play critical roles in physiological processes.

Quaternary proline analogues³ have received considerable attention as these residues limit the conformational freedom of peptides. The modulation of the cis-trans proline amide geometry can allow stable arrangements of turns, which are fundamental recognition elements for peptide function.

Octahydroindole-2-carboxylic acid (Oic, Fig. 1) is a proline surrogate with a fused bicyclic structure shown to be a very useful scaffold. Among the eight different stereoisomers, (2S,3aS,7aS)-Oic (Fig. 1) has application in the elaboration of medicinally relevant drugs already in the market.⁴ This particular isomer has also been used as a surrogate of proline in large peptide systems such as bradykinin, thereby making feasible the preparation of orally available antagonists to the B₂ receptor of this hormone, with attractive anti-cancer properties and enhanced resistance to degrading enzymes.⁵ It has also served as a β-turn inducer in the design of alternative non-peptidic bradykinin B₂ antagonists.⁶ Additionally, some applications of (S,S,S)-Oic⁷ derivatives have been reported in the field of organocatalysis.⁸ Given the interest of this Oic stereoisomer, considerable efforts have been directed towards its preparation. The strategies reported in the literature include stereoselective processes,⁹ chemical resolutions of diastereomeric salts,¹⁰ and chromatographic¹¹ and enzymatic¹² resolutions.

Figure 1.

Structure of (2S,3aS,7aS)-octahydroindole-2-carboxylic acid.

As quaternary proline surrogates, (αMe)Oic derivatives are highly valuable systems for peptide engineering purposes due to their conformationally well-defined structures. Moreover, a number of quaternary Oic derivatives have been reported to be the core structure of investigational compounds with potential applications for treating osteoarthritis.¹³ However, the preparation of C^α-tetrasubstituted Oic analogues remains almost unexplored.^13,14 The lack of effective methodologies for ready accessibility to their enantiomerically pure forms constitutes the main hurdle for their exploitation.

Recently, we have described a very convenient and concise route leading to enantiopure (R,S,S)-(αMe)Oic¹⁴ (Scheme 1) through a completely diastereoselective α-methylation reaction of the trichloromethyloxazolidinone¹⁵ derived from enantiopure (R,S,S)-Oic [(R,S,S)-1, Scheme 1]. The preparation of the (S,S,S)-(αMe)Oic stereoisomer has only been disclosed in a patent,¹³ where no experimental data to sustain the configurational assignment made for the α carbon are given.

Scheme 1.

Reported synthesis for (R,S,S)-(αMe)Oic (ref. 14).

On approaching the preparation of enantiopure (S,S,S)-(αMe)Oic, Seebach’s self reproduction of chirality methodology¹⁵ could not be applied because the precursor amino acid (S,S,S)-1 fails to undergo condensation with trichloroacetaldehyde¹⁴ to produce the desired oxazolidinone (Scheme 1). The steric hindrance of the cyclohexane group in (S,S,S)-1, on a cis relative orientation with respect to the carboxylic function precludes the formation of oxazolidinone (R,S,S,S)-2. Such a tricyclic system would suffer from severe steric clash¹⁴ between nearby hydrogens at the inner concave face. In fact, we proved¹⁴ that the aforementioned condensation reaction constitutes a very effective tool for the separation of epimeric mixtures of (R,S,S)-1 and (S,S,S)-1.

Although the steric shielding imposed by the cyclohexane ring in (S,S,S)-1 was truly advantageous for the selective preparation of (R,S,S)-(αMe)Oic,¹⁴ it prevents the application of the self-reproduction of chirality strategy¹⁵ when the synthesis of the epimeric (S,S,S)-(αMe)Oic is pursued. In that sense, it is rather unfortunate because this powerful methodology would formally allow α-methylation with complete retention of configuration.

We then found that the steric shielding introduced by the cyclohexane moiety could inflict unambiguous facial stereodifferentiation on the enolate derived from (S,S,S)-1. On this basis, it seems reasonable to consider that this spatial effect could effectively govern the stereochemical outcome of direct alkylations performed on a suitably protected derivative of (S,S,S)-1 such as (S,S,S)-3 (Fig. 2).

Figure 2.

Proposed precursor for the synthesis of (S,S,S)-(αMe)Oic.

Compound (S,S,S)-3 was prepared from enantiomerically pure (S,S,S)-1, which, in turn, was obtained in a single step and high yield from commercially available (S)-indoline-2-carboxylic acid as already described.¹¹ Protection of the carboxylic acid and amino functionalities in (S,S,S)-1 was carried out by treatment with methanol in the presence of thionyl chloride and subsequent reaction with di-tert-butyl dicarbonate (Scheme 2).

Scheme 2.

Synthesis of enantiomerically pure (S,S,S)-3. Reagents and conditions: (a) See ref. 11; (b) SOCl₂, MeOH, r.t.; (c) Boc₂O, iPr₂EtN, 4-dimethylaminopyridine, THF, r.t.

Next, we addressed the study of the α-methylation reaction of (S,S,S)-3 through the generation of an intermediate enolate. It appears that the treatment of (S,S,S)-3 with LDA in THF at −78 °C generated a lithium enolate that successfully reacted with methyl iodide in high yield. Thus, a mixture of the two possible diastereoisomers, (S,S,S)-4 and (R,S,S)-4, was isolated in 92% yield, after 12 h of reaction at −60 °C (Scheme 3). In comparison, the use of either LHMDS or KHMDS as a base, led to poor conversions (∼50%) under similar reaction conditions.

Scheme 3.

α-Methylation reaction of (S,S,S)-3. Reagents and conditions: (a) LDA, MeI, THF, −78 to −60 °C; (b) 3N HCl / EtOAc, r.t.

A 89:11 diastereomeric ratio was established by analysis of the relative intensities of appropriate signals in the ¹H NMR spectra of the amino ester hydrochlorides (S,S,S)-5 and (R,S,S)-5,¹⁶ that were quantitatively obtained after treatment of the crude mixture of (S,S,S)-4/(R,S,S)-4 with a saturated solution of hydrogen chloride in ethyl acetate.

The stereoselectivity attained is rather significant given that the direct α-methylation reaction of equally protected L-proline furnishes racemic (αMe)Pro.¹⁷ In the case of N-Boc prolines with additional substituents at the ring, the reported diastereoselectivities are low to moderate, with few exceptions.¹⁸ In our case, the high steric hindrance imposed by the cis-fused cyclohexane ring allows an efficient facial stereodifferentiation that translates into a good diastereoselectivity.

The major diastereoisomer of the (S,S,S)-4/(R,S,S)-4 mixture was isolated pure¹⁹ in 75% yield after column chromatography (Sheme 4). Subsequent acidic treatment led to the corresponding amino ester hydrochloride, which furnished colourless single crystals suitable for X-ray diffraction analysis.²⁰ The crystalline structure (Fig. 3) showed an (S,S,S) stereochemistry, with a cis relative disposition of the hydrogen atoms at the ring junction and the α-methyl group, thus confirming that the alkylation reaction proceeded with retention of configuration at the quaternary stereocenter.

Figure 3.

X-ray crystal structure of (S,S,S)-5·HCl. Most hydrogens have been omitted for clarity.

Additionally, the basic hydrolysis of (S,S,S)-4 furnished the N-Boc amino acid (S,S,S)-6 in excellent yield (Scheme 4). This Oic derivative is suitably protected for incorporation into peptides.

Scheme 4.

Synthesis of enantiopure (S,S,S)-(αMe)Oic derivatives. Reagents and conditions: (a) LDA, MeI, THF, −78 to −60 °C; column chromatography of the mixture (SiO₂, EtOAc/Hexanes 1:10) furnishes pure (S,S,S)-4; (b) 3N HCl / EtOAc, r.t.; (c) 1N KOH, MeOH / H₂O, reflux.

In conclusion, this initial study shows that enantiopure (S,S,S)-(αMe)Oic can be accessed by means of a diastereoselective α-methylation reaction on (S,S,S)-3, which is easily derived from (S,S,S)-Oic. The facial stereodifferentiation imposed by the cyclohexane ring on the nucleophilic enolate carbon derived from (S,S,S)-3 accounts for the selective outcome of the α-methylation reaction. Such a spatial effect can thus be regarded as an effective tool for stereocontrol induction during the functionalization at the α carbon for those sterically hindered proline-type substrates where the formation of a trichloromethyloxazolidinone is not possible.