A bio-inspired approach to proline-derived 2,4-disubstituted oxazoles

Introduction

Since their isolation from marine products in the late 1980s, oxazole-containing peptidomimetics have been studied extensively by synthetic organic and medicinal chemists [1], [2], [3], [4], [5]. It is believed that many of these natural products are derived from enzymatic post-translational modifications of peptide-based precursors containing serine or threonine moieties [2]. It is not surprising, therefore, that 2,4-disubstituted oxazoles of general formula 1 derived from α-amino acids have attracted much attention as promising building blocks for drug discovery [6], [7], [8], [9], [10], [11], [12], [13] or total synthesis of natural products including plantazolicins [14], [15], telomestatin [16], ulapualides [17], [18] and diazonamide A [19], [20] and as ligands for enantioselective catalysis [21], [22]. In most cases, a ‘biomimetic’ approach to the synthesis of 1 has been used, including the formation of oxazoline 2 and its subsequent aromatization (Scheme 1) [7], [8], [9], [10], [11], [12], [13], [14], [15], [16], [17], [18], [19], [20], [21], [22], [23]. However, this method is effective only when the substituent R¹ in the oxazole ring formed is an electron-withdrawing group (e.g. R¹=CO₂Me). Even for R¹=Ph, the aromatization step requires harsh conditions (NiO₂, 150°C, μW) [13]. Recently, an alternative approach has been reported for the preparation of the compounds 1 with primary amino group (R³=H), which is based on oxidation of phthaloyl derivatives of the type 3 (R¹=H, Ph, Bn; R²=i-Pr; R³/PG=phthaloyl) with the Dess-Martin reagent, followed by cyclodehydration and deprotection [6]. Apart from the limited scope, the method offers a moderate yield of the key step (34–50%) and can be used only at a milligram scale. Moreover, partial racemization occurs upon removal of the phthalimide protective group.

Scheme 1

Biomimetic approach to oxazole building blocks.

In this work, we have taken inspiration from the ‘biocore’ concept described by Kombarov and co-workers [24]. This concept refers to a combination of (hetero)aliphatic and hydrophilic aromatic rings providing privileged structures for drug discovery (Figure 1). Analysis of the literature data published on ‘biocores’ obtained by a combination of oxazole and common heteroaliphatic amines (pyrrolidine/piperidine) shows that 2-(2-pyrrolidinyl)oxazoles 4 are most often encountered in papers and patents [25]. As in the case of the general structure 1, however, nearly all known compounds 4 are substituted with R¹=aryl or CO₂R, or are derived from them [7], [8], [9], [10], [11], [12], [21], [22], [23].

Figure 1

‘Biocore’ concept and its implementation into the target compounds 4.

Here, we describe the multigram synthesis of oxazole-substituted pyrrolidines 4 with aliphatic substituents at C-4 position (R¹=H, alkyl; R²=H) starting from enantiopure (S)-N-Boc-proline (5 in Figure 1) [26]. Instant JChem was used for the prediction of the physico-chemical properties of the compounds. Conformational restriction provided by the saturated heteroaliphatic ring of 4 and an improved hydrophilicity are valuable features in lead-oriented synthesis in medicinal chemistry [27], [28].

Results and discussion

Our strategy for the preparation of the secondary amines 4a–e was related to the ‘biomimetic’ approaches shown in Scheme 1. Initially, the synthesis followed the literature analogs and involved dipeptides 6, which were prepared from 5 in an 85–95% yield using the mixed anhydride activation method (Scheme 2) [29]. However, reduction of 6a–e with NaBH₄ was accompanied by a problematic isolation of the target products 7a–e, which was related to their high aqueous solubility and formation of stable boron complexes. Although compounds 7 were isolated in a 50–85% yield, the attempted scaling up of this procedure was not successful. Therefore, we switched to a one-step synthesis of 7 directly from 5 using coupling with β-amino alcohols 8 (either commercially available or prepared from the corresponding α-amino acids). Products 7a–f were obtained with overall yields of 60–86%. The procedure did not require chromatographic purification and was amenable to a multigram scale-up.

Scheme 2

Synthesis of the compounds 4a–e·2HCl.

The next step, oxidation of 7 to aldehydes 9, was performed first with the Dess-Martin reagent, by analogy with the literature examples [6], [29] and our previous in-house experience. It was found that in the case of 7a, the yield of the product 9a varied significantly in different experiments [35–70% by ¹H nuclear magnetic resonance (NMR)]. Moreover, the product 9a showed limited storage stability. A detailed investigation revealed that the Swern reaction [dimethyl sulfoxide (DMSO), (COCl)₂, Et₃N, −60°C, 1 h] gives more reproducible results, and the crude products 9a–d were obtained in a 55–85% yield (by ¹H NMR). In the case of 7e (R¹=H), the use of the Dess-Martin reagent appeared to be more fruitful than the Swern oxidation (35% vs. 5% yield of 9e by ¹H NMR). Unfortunately, the product 9f (R¹=Ph) could not be obtained at all using either method.

It was shown that the storage stability of products 9a–e is strongly affected by the steric volume of the substituent R¹. In particular, the product 9a has limited stability in CH₂Cl₂ solution (ca. 1 h) and as a neat product at −24°C (ca. 8 h). On the contrary, compounds 9b–d with relatively bulky alkyl substituents are relatively stable in both cases (at least for 24 h). The product 7e is rather unstable either in CH₂Cl₂ solution or as a neat compound at −24°C (less than 1 h). Other factors influencing the stability of 9a–e include temperature which should be kept below 20°C during all manipulations with 9 including the evaporation of the solvent. It can be suggested that the stability of the products 9a–f is also related to their reactivity toward self-condensation, which is diminished upon an increase in the size of the alkyl substituent R¹. In the case of R¹=Ph, this side reaction becomes especially fast, presumably due to a concomitant increase in the acidity of α-CH adjacent to the aldehyde moiety. It was apparent that the use of acidic or basic reagents should be avoided during the synthesis of 9e, which might improve the yield of this product. For this reason, we have synthesized amide 7g (60% yield) and subjected it to oxidative cleavage with NaIO₄. Using this procedure, the product 9e was obtained in a 45% yield.

Taking into account the low stability of aldehydes 9, the crude products 9 were used immediately in the next step, which is the dehydrative oxazole ring formation. The Appel reaction was used for this reason, by analogy with the literature results published previously [6], [29]. It was found that the use of the CBr₄-PPh₃ system gave a low yield of the target products 11 (less than 20% by ¹H NMR), whereas with C₂Cl₆-PPh₃, the oxazoles 11a–e were obtained in a 50–90% yield.

The structure of compound 11a was confirmed by X-ray single crystal analysis (CCDC-1570445 [30], Figure 2). It should be noted that only one enantiomer of 11a was observed in the crystal structure.

Figure 2

Molecular structure of compound 11a.

In the final step of the synthesis, deprotection of compounds 11a–e was performed by treatment with HCl. It was found that using a solution of HCl in Et₂O resulted in precipitation of 11·HCl, so that a complete conversion could not be achieved. Fortunately, the use of HCl in Et₂O-MeOH furnished the target products 4a–e as dihydrochlorides in a 76–85% yield. It should be noted that the method can be used for the synthesis of up to 45 g of the products 4·2HCl in a single run.

To analyze the optical purity of the products, compound 4a·2HCl was derivatized with the Mosher’s reagent, which gave a single diastereomer according to ¹H NMR and ¹⁹F NMR analyses. In addition, the enantiomer of compound 11c was synthesized from D-N-Boc-proline and used as a control for chiral stationary phase high-performance liquid chromatography (HPLC). It was found that enantiopure product 11c showed no detectable peak of its enantiomer in the chromatogram [≥98% ee, (ChiralPack^® IA column, hexanes – 2-propanol (95:5) as an eluent]. Under the same conditions, single peaks (≥98% ee) were also observed in the chromatograms of other products 11.

Conclusions

Synthesis of (S)-4-alkyl-2-(pyrrolidin-2-yl)oxazoles was performed starting from l-Boc-proline and using an approach inspired by biosynthesis of oxazole-containing peptidomimetics isolated from natural products. Optimization of the procedures was performed for each step, in order to make the method scalable. It was shown that the key intermediates, 1-Boc-N-(1-oxoalkan-2-yl)pyrrolidine-2-carboxamides (9), demonstrate low to moderate stability, which is increased with the increasing size of the alkyl group near the aldehyde moiety. Therefore, it is strongly recommended to submit them to the next step immediately after isolation.

The target products were obtained in only four steps from derivatives of natural α-amino acids in a 17–51% overall yield, with ≥98% ee and at up to a 45-g scale. The overall yield of the product is increased upon an increase in the lipophilicity of the compound and hence follows the ‘LogP drift’ effect proposed by Churcher and co-workers (Figure 3) [27]. In our opinion, this effect might be related to isolation issues observed for more hydrophilic compounds.

Figure 3

Overall yields of products 4a–e·2HCl as a function of their lipophilicity.

In conclusion, (S)-4-alkyl-2-(pyrrolidin-2-yl)oxazoles can be considered as building blocks readily available to the chemical community. We believe that synthetic accessibility and favorable physico-chemical characteristics of these oxazole-derived ‘biocores’ make them valuable starting points for early drug discovery.

Experimental

Solvents were purified according to standard procedures [31]. Compounds 6a–e [29], 8b–d and 8f [32] were prepared using reported methods. All other starting materials were purchased from commercial sources. Analytical thin-layer chromatography (TLC) was performed using Polychrom SI F254 plates. Column chromatography was performed using Kieselgel Merck 60 (230–400 mesh) as the stationary phase. NMR spectra were recorded on a Bruker 170 Avance 500 spectrometer (500 MHz for ¹H and 125 MHz for ¹³C). Signal assignments for compounds 4a–e·2HCl were made using ¹H-¹H correlation spectroscopy (COSY), heteronuclear single quantum coherence (HSQC) and heteronuclear multiple bond correlation (HMBC) experiments. Mass spectra were recorded on an Agilent 1100 LC/MSD SL instrument [atmospheric pressure electrospray ionization (APESI)]. The stability of aldehydes 9 in CH₂Cl₂ solution or as neat products at −24°C was studied by recording ¹H NMR spectra for the aliquots hourly; the compound was assumed to be stable if the relative intensity of the aldehyde signal changed by no more than 10%.