Chiral oxazoline ligands with two different six-membered azaheteroaromatic rings – synthesis and application in the Cu-catalyzed nitroaldol reaction

Introduction

According to our interest focusing on the design, synthesis, and applications of chiral oxazoline ligands we have already described several ligands 1 which combine in their structure two heterocyclic moieties: chiral oxazoline and 1,2,4-triazine linked by an N-phenylamine unit (Figure 1). The ligands differ in the type of a substituent in the oxazoline ring and in the substitution mode of the 1,2,4-triazine moiety.

Figure 1

Previously investigated 1,2,4-triazine-oxazoline ligands.

These ligands were applied to the copper catalyzed enantioselective nitroaldol reaction of several aromatic and aliphatic aldehydes providing products with enantiopurity up to 92% and yield up to 95% [1–3]. Thus, the ligands with the phenyl substituent in the oxazoline ring and ligands that possess the indane moiety exhibit high stereocontrolling abilities. Substitution in the 1,2,4-triazine ring has a crucial impact on chemical yield of the investigated reaction. From this point of view, the ligands having a phenyl group at the C5 position of 1,2,4-triazine are preferred. Structures of 1-Cu complexes so far are not known, but it can be suggested on the basis of the ligand crystal structures [4] that the nitrogen atoms of oxazoline and amino groups may participate in the complexation of copper ion. Similar complexation mode of zinc was observed by Guiry [5]. In the Henry reaction, a base or counter ion is necessary to abstract the proton from nitromethane. Our catalytic system works without an external base, so the nitronate anion must be generated with the involvement a counter ion or the 1-Cu complex plays the role of a bifunctional catalyst. As an extension of the previous work, now we report the synthesis of chiral oxazoline ligands 2 and 3 as analogs of ligands 1 where the N-phenylamine linkage is replaced with pyridine-2-amine or 2-aminopyridine 1-oxide moiety (Figure 2). The presence of the nitrogen atom or N-oxide moiety in the linkage can change electronic properties of the donating nitrogen atoms, thus having impact on the complexation and stereo-controlling abilities of ligands. They can also act as an additional internal base. On the other hand, the competition between the new electron pair donors and the nitrogen atoms of oxazoline and amino groups in complexation with copper ion can take place.

Figure 2

Chiral oxazoline ligands with six-membered azaheteroaromatic rings.

Results and discussion

The 1,2,4-triazine ring is π-electron deficient and, therefore, susceptible to nucleophilic attack [6]. Highly active 3,3′-dichloro-5,5′-bi-1,2,4-triazine undergoes displacement of chlorine atoms after few minutes of stirring with aliphatic amines or aniline in the absence of an additional base [7]. Reactions of less active 1,2,4-triazines require the presence of metal catalyst to replace halogen or other nucleofugal leaving group [1, 8, 9]. Thus, a two-step synthetic route to ligands 2 employs the Pd-catalyzed Buchwald-Hartwig reaction [10–14] of enantiopure 3-(4,5-dihydro-1,3-oxazol-2-yl)pyridin-2-amines 6 with 3-bromo-5,6-diphenyl-1,2,4-triazine (7a) as the key step (Scheme 1). 3-(4,5-Dihydro-1,3-oxazol-2-yl)pyridin-2-amines 6a–d were obtained by condensation of enantiomerically pure aminoalcohols 5a–d with the cyano group of 2-aminopyridine-3-carbonitrile (4). This reaction was conducted in the presence of a catalytic amount (20 mol%) of ZnCl₂ as Lewis acid [15]. After 4 days of heating in boiling chlorobenzene the appropriate (oxazol-2-yl)pyridin-2-amines 6a–d were isolated with 31–99% yield. Our previously described protocol of condensation of aminoalcohols with 2-aminobenzonitrile requires the use of an excess of ZnCl₂, and after removal of solvent, stirring the mixture with NaOH solution to isolate the product [1]. The use of three-fold excess of ZnCl₂ in the reaction between 4 and 5a–d allowed to shorten the reaction time to 24 h, but isolation of the product from the reaction mixture appeared difficult.

Scheme 1

Synthesis of ligands 2a–d.

After obtaining amines 6, an attempt to react them with 3-bromo-5,6-diphenyl-1,2,4-triazine (7a) was undertaken. To synthesize the ligands 2a–d the application of Buchwald-Hartwig Pd-catalysis was necessary. Thus, the aryl aminations between 7a and 6a–d were carried out using Pd₂dba₃, Xantphos, and K₂CO₃ in boiling dioxane (Scheme 1). This procedure allowed to obtain the ligands 2a–d in moderate yields ranging from 32% to 37%.

Synthesis of ligands of type 3 could not be accomplished by simple oxidation of 2a–d, since 1,2,4-triazine [16] as well as the oxazoline [17] rings are also susceptible to oxidizing conditions. Initial attempt was made to obtain ligands 3 utilizing amination of 3-halo-1,2,4-triazine with N-protected aminopyridine 1-oxides 9 based on aza-Smiles rearrangement (Scheme 2) [18].

Scheme 2

Proposed synthetic approach to ligands 3.

Analogous protocol described by us previously, proved to be effective for amination of pyridine and pyrimidine halides [18, 19]. This approach assumes formation of intermediary salt 10 in reaction of formamidine protected 2-aminopyridine-3-carbonitrile 1-oxide 9 with 3-chloro-5,6-diphenyl-1,2,4-triazine (7b). Base-induced rearrangement of 10 should give derivative 11, the condensation of which with aminoalcohols 5 would produce ligands 3. A reaction of N-oxide 8 with 3 equivalents of N,N-dimethylformamide dimethyl acetal (DMF-DMA) was carried out at room temperature without any solvent. After 80 h of stirring, the amidine protected compound 9 was isolated in 92% yield. To our disappointment, the next step, the preparation of salt 10 appeared difficult. A typical procedure for amination of pyridine and pyrimidine halides with formamidine protected aminopyridine 1-oxides involving stirring of the substrates in dry DMF at room temperature in the case of 1,2,4-triazine halide proved to be ineffective. Application of different solvents and changing the reaction temperature did not result in formation of product 10. Difficulties in formation of salt 10 make the selected synthetic strategy unsuccessful.

Searching for another method of synthesis of compounds 3 we envisaged a more reliable procedure involving amination of 3-chloro-5,6-diphenyl-1,2,4-triazine (7b) with (oxazol-2-yl)pyridin-2-amine 1-oxides 12. Thus, the synthesis of ligands of type 3 began with preparation of N-oxide 8 by oxidizing 2-aminopyridine-3-carbonitrile (4) with m-chloroperbenzoic acid in dichloromethane (Scheme 3).

Scheme 3

Synthesis of chiral N-oxides 12a–e.

Condensation of N-oxide 8 with aminoalcohols 5a–e yielded the desired 3-(oxazol-2-yl)pyridin-2-amine 1-oxides 12a–e in good yields. The reaction time was controlled to avoid transformation of the products to oxazolines 6 (see Experimental part). The preparation of ligands 3 was first attempted by heating oxides 12a and 12c with 3-chloro-5,6-diphenyl-1,2,4-triazine (7b) in dioxane in the presence of K₂CO₃ at 95°C. Under these conditions the unexpected respective products 13a and 13c were generated (Scheme 4).

Scheme 4

Reaction of N-oxides 12a and 12c with 3-chloro-5,6-diphenyl-1,2,4-triazine (7b).

The N-oxide function strongly activates the amino group in compounds 12, so the substitution of the chlorine atom in 7b can proceed under metal-free conditions. Moreover, the secondary NH group in 3a and 3c is also active enough to react with a second molecule of 3-chlorotriazine 7b to give compounds 13a and 13c. To inhibit the reaction of the initially formed compounds 3a and 3c with 3-chlorotriazine, K₂CO₃ was eliminated from the reaction mixture. Under base-free conditions at 95°C product 3a was obtained in 22% yield accompanied by undesired compound 14a isolated in the yield of 54% (Table 1, entry 1). When the temperature was lowered to 20°C, no reaction between 12a and 7b occurred (Table 1, entry 2). In DMF at 50°C the expected product 3a and unexpected compound 14a were formed in 15% and 48% yield, respectively (Table 1, entry 3). Under these conditions, N-oxide 12b reacted with chlorotriazine 7b giving compounds 3b and 14b in 10% and 48% respective yields after chromatography (Table 1, entry 4). Chemically pure compounds 3a,b and 14a,b were characterized by several analytical techniques including ¹H NMR, ¹³C NMR, IR, HR MS, which confirmed their structures.

Undesirable compounds 14a,b were formed by transformation of initially formed products 3a,b. Since the reaction of (oxazol-2-yl)pyridin-2-amine 1-oxides 12 with 3-chloro-5,6-diphenyl-1,2,4-triazine (7b) is conducted in the absence of base, evolving hydrochloride protonates the nitrogen atom of the oxazoline ring in 3a,b activating these compounds towards nucleophilic attack of chloride ion. This leads to opening of the oxazoline ring giving rise to products 14a,b (Scheme 5).

Scheme 5

Formation of compounds 14a,b.

The observed reactivity was confirmed in the reactions of three other (oxazol-2-yl)pyridin-2-amine 1-oxides 12c–e with 3-chloro-5,6-diphenyl-1,2,4-triazine (7b). In each case the formation of two products 3c–e and 14c–e was observed. Attempts to separate the products 3c–e and 14c–e by chromatography failed. They were not isolated in chemically pure form. Formation of 3c–e and 14c–e was shown on the basis of analysis of mass spectra of their mixtures. The ESI HR mass spectrum of a mixture of 3c and 14c shows signals at m/z 453.2039 and m/z 489.1804 corresponding to the values expected for [3c+H]⁺ and [14c+H]⁺. The mass spectrum of the mixture of 3d and 14d indicates the presence of ions of m/z 501.2039 and m/z 537.1780 which match the calculated masses of [3d+H]⁺ and [14d+H]⁺. Analogously, ions with m/z 501.2036 and m/z 537.1798 are observed for the mixture of 3e and 14e. Their masses are virtually identical with the masses calculated for [3e+H]⁺ and [14e+H]⁺.

The cationic activated oxazoline ring-opening process under an attack of nucleophile gained synthetic importance, for example in the synthesis of polymers [20, 21]. To circumvent the problem with conversion of 3 to 14, the use of organic bases in the reaction of 12 with 7b was considered. It was predicted that aliphatic amines such as triethylamine or DABCO can form salts with 3-chlorotriazine 7b, as observed previously [22] and their use was excluded. Therefore, the reaction of 12b and 7b was repeated with 2 equivalents of N,N-dimethylaniline, m-nitroaniline or p-nitroaniline. Unfortunately, no significant improvement in yield of the desired product 3b was achieved. Compound 3b isolated in yield 17–24% was still accompanied by its transformation product 14b (45–52%, Table 1, entry 6, 7, and 8). As can be seen from the results summarized in Table 1, in the reaction of 12b and 7b approximately 70% of the substrates are initially converted to the desired product 3b but about 50% of the product is transformed to the open-chain compound 14b.

Initial study was focused on the screening of ligands 2 and 3 in the copper catalyzed enatioselective nitroaldol reaction. Several aldehydes were treated with nitromethane in the presence of in situ generated complexes of ligands 2 and 3 with Cu(OAc)₂·H₂O. The reactions were carried out in 2-propanol at room temperature and lead to β-nitro alcohols with moderate to good chemical yield but no enantioselectivity was observed (Table 2). Enantiopure compounds 13c and 14b were also screened in the attempted enantioselective addition of nitromethane to m-nitrobenzaldehyde. The β-nitro alcohols formed in the respective yields of 42% and 56% and were isolated as mixtures of enantiomers, which indicates that the compounds 13c and 14b are also not good enantiomeric catalysts in the nitroaldol reaction (Table 2).

Conclusion

A new class of chiral compounds 2 with three different heterocyclic rings in their structure have been prepared by employing Pd-catalyzed Buchwald-Hartwig cross-coupling reaction between the enantiopure 3-(4,5-dihydro-1,3-oxazol-2-yl)pyridin-2-amines 6 and 3-bromo-5,6-diphenyl-1,2,4-triazine (7a) as the key step. Two different synthetic routes towards compounds 3 containing N-oxide functionality were investigated: amination of 3-chloro-5,6-diphenyl-1,2,4-triazine with N-protected aminopyridine 1-oxide 9 based on aza-Smiles rearrangementand and amination of 3-chloro-1,2,4-triazine with enantiopure 3-[4-R/Ar-4,5-dihydro-1,3-oxazol-2-yl]pyridine-2-amine 1-oxides 12. The use of the former compound failed due to difficulties in formation of salt 10. The second method encountered problems with unexpected subsequent reactions of initially formed product 3. Under basic conditions, product 3 acts as good nucleophile and undergoes reaction with a second molecule of 7b giving compounds 13. In the experiments carried out in the absence of a base or in the presence of a weak organic base, initially formed product 3 undergoes the oxazoline ring opening leading to compound 14. These new classes of oxazoline ligands 2 and 3 do not exhibit enantio-controlling abilities in the copper catalyzed enantioselective Henry reaction. This result may indicate a different mode of copper complexation with ligands 2 and 3 in comparison to ligands 1. The nitrogen atom of the pyridine ring or the oxygen atom of the N-oxide function can be involved in complexation with the copper ion. The stereogenic center of the complex may be located at some distance of the Lewis acid center and does not have any influence on the stereochemical course of the nitroaldol reaction. This issue needs further explanation and will be discussed in due course.

Experimental

¹H and ¹³C NMR spectra were determined at 400 MHz and 100 MHz, respectively, on a Varian 400 spectrometer. Chemical shifts (δ) are reported in part per million from tetramethylsilane with the solvent resonance as the internal standard. Coupling constants are given as absolute values expressed in Hertz. Mass spectra were obtained by using AMD 604 (AMD Intectra GmbH, Germany) and GC/MS QP 5050 Shimadzu (30 m×0.25 mm ID-BPX 5 0.25 mm) spectrometers. Infrared spectra were obtained by using a Magna FTIR–760 Nicolet apparatus and Shimadzu FT IR Affinity-1 spectrometer. Elemental analyzes were obtained with a Elementar Vario EL III CHNS analyzer. Optical rotation values were measured at room temperature with a Perkin-Elmer polarimeter. The ee values were checked by HPLC (Knauer) analysis by using a chiral stationary phase column (Chiralcel OD-H), and eluting with 2-propanol–hexanes. Thin layer chromatography (TLC) was carried out on aluminium sheets coated with silica gel 60 F₂₅₄ (Merck). Column chromatography separations were performed by using Merck Kieselgel 60 (0.040-0.060 mm). Solvents were dried and distilled according to standard procedures. 3-Chloro-5,6-diphenyl-1,2,4-triazine [23] and 3-bromo-5,6-diphenyl-1,2,4-triazine [1] were synthesized according to literature procedures.