The Mimic of Type II Aldolases Chemistry: Asymmetric Synthesis of β-Hydroxy Ketones by Direct Aldol Reaction

Authors


Corresponding authors: Jianlin Han, molab@nju.edu.cn; Yi Pan, yipan@nju.edu.cn

Abstract

An efficient direct aldol reaction has been developed for the synthesis of chiral β-hydroxy ketone using a combination of C1-symmetric chiral prolinamides based on o-phenylenediamine and zinc triflate as catalyst. The reaction was convenient to carry out in aqueous media with up to 98% chemical yields and up to 94% ee values. The current strategy can be regarded as the analogue of aldolase type II, which suggests a new pathway for the designing of new organocatalysts.

The development of stereoselective and enantioselective aldol reaction has become an interesting and challenging topic in modern organic and medicinal chemistry (1–3), because the resulting chiral β-hydroxy ketones belong to an extremely important class of biological compounds. In fact, β-hydroxy ketones can serve as versatile building block for the asymmetric synthesis of carbohydrates, amino acids and many other biomolecules (4–6). They also provide privileged structural functionalities that exist in many important natural products (7–12). For example, the β-hydroxy ketones functionalities exist in macrolide classes of antibiotics such as telithromycin and cethromycin which are targeted primarily against Gram-positive bacterial strains including Streptococcus pneumoniae and S. pyogenes, fastidious Gram-negative strains including Haemophilus influenzae and Moraxella catarrhalis, atypicals Mycoplasma pneumoniae, Chlamydia pneumoniae and Legionella pneumophilia (13). They were also found in conagenin that can improve the antitumor efficacy of adriamycin and mitomycin C against murine leukemias, which suggest its potential utility for cancer chemotherapy (14).

In virtue of the importance of chiral β-hydroxy ketones, several approaches aiming to prepare them have been reported in the last decade (15–27). Among all the catalytic direct aldol, the most attractive one is the mimic of the actions of the enzymes. Although the direct catalytic aldol systems which were assumed to proceed through an enamine mechanism by mimic of aldolase type I have been well explored (28–32), the direct aldol systems by using the methods analogue to the actions of aldolase type II containing a zinc cofactor have still remained challenging. In fact, there were only a few reports for direct aldol reactions promoted by the zinc complexes with N-donor ligands in the presence of water until now (33–38). Herein, we reported a chiral ligands and zinc triflate catalyzed asymmetric direct aldol reaction with water as solvent. To our knowledge, this is the first example that mimics type II aldolase utilizing C1-symmetric organic molecular based on o-phenylenediamine as chiral ligand.

Results and Discussions

According to the mechanism of type II aldolase (33–38), we designed and prepared a series of chiral ligands (Figure 1). These chiral ligands 1–6 were easily prepared from proline and aniline or o-phenylenediamine in high yields. Most of these ligands have C1 symmetry, expect compound 4, which is a C2-symmetric ligand. Then, all these compounds were used in the direct aldol reactions. The reactions were carried out between cyclohexanone and p-nitrobenzaldehyde in aqueous media (Table 1). As shown by Table 1, all the designed chiral prolinamide derivatives worked well in the direct aldol reactions and gave the desired β-hydroxy ketone in good yields and good enantioselectivities. Addition of zinc trifluoromethanesulfonate could obviously improve the diastereoselectivities and enantioselectivities (entry 10 versus entry 9). Good result was obtained when the combined catalyst of zinc triflate and the C2-symmetric prolinamide ligand 4 containing two amino acid moities was used (Table 1, entry 8). However, the best result with respect to yield and enantioselectivity was observed with C1-symmetric prolinamide ligand 5 and Zn(OTf)2 as additive (96% yield and 93% ee, Table 1, entry 10). The loading amount of Zn(OTf)2 in the reaction was also investigated. The use of 5 mmol% of Zn(OTf)2 seemed best for the reaction (Table 1, entry 10). Although lower loading of Zn(OTf)2 slightly decreased both yields and selectivities (Table 1, entries 15 and 16), no improvement was detected when more than 10 mmol% catalyst was used (Table 1, entries 13 and 14).

Figure 1.

 Structure of the catalysts 16.

Table 1.   Screening of catalysts in the reaction between cyclohexanone and p-nitrobenzaldehydeaThumbnail image of

From Table 1, it was also found that the prolinamides and zinc triflate combined catalyst can work very well in aqueous media. Then, the solvents used for this reaction were examined (Table 2). The chemical yields were increased by the use of water as co-solvent (Table 2, entries 3–5). The best result was observed when the ratio of cyclohexanone to water was 5:1, and 96% chemical yield and 93% ee value were obtained (Table 2, entry 5). Almost no desired product was detected when DMSO was used as solvent (Table 2, entry 7), and only medium yield and enantioselectivity were found with neat cyclohexanone (Table 2, entry 6) or tetrahydrofuran (Table 2, entry 2) as solvent. It is obvious that water is helpful to form product and control the stereoselectivity of the reaction. Reaction temperature was also found to have some effect on the current catalytic system. When the temperature was increased to 50 °C, both the diastereoselectivity and the enantioselectivity decreased (Table 2, entry 8). Decreasing the temperature to 3 °C still did not show any improvement on diastereoselectivity (Table 2, entry 10). From the results above, we found that these new chiral ligands together with zinc triflate could act as catalysts for direct aldol in aqueous media. Both zinc triflate and chiral prolineamides were essential for the catalytic system. In this respect, the combined catalysts could be regarded as mimic of type II aldolase, and an enamine was presumed to be formed in the active site (33–38).

Table 2.   The optimization of the current aldol reaction conditionsaThumbnail image of

After the optimized reaction conditions was obtained, several aldol reactions under the above-mentioned conditions were carried out to check their catalytic activities, and the results are summarized in Table 3. Several aromatic aldehydes were suitable substrates for the reaction and gave corresponding β-hydroxy ketone with good to excellent yields. Especially for aldehydes with strong electron-withdrawing group (NO2) on the aromatic ring, the reaction gave excellent yields, high enantioselectivities (Table 3, entries 1–3). However, for the aldehyde substituted with a CF3 group, only moderate diastereoselectivity was found (anti:syn = 59:41, Table 3, entry 7). In the cases of the aldehydes with electron-donating group (OMe) on the aromatic ring, such as 4-methoxybenzaldehyde and 4-methylbenzaldehyde, only a trace amout of desired products were observed even after the reaction time had been extended to 72 h.

Table 3.   Catalytic aldol reaction between cyclohexanone and aryl aldehydesaThumbnail image of

Acetone was also used as substrate in the direct aldol reaction with p-nitrobenzaldehyde under the same reaction condition (Scheme 1). The corresponding product 8 was formed with excellent chemical yield (90% yield). However, It was obvious that acetone gave a little bit lower enantioselectivity compared to the cyclohexanone substrates, and only 60% ee value was obtained.

Figure Scheme 1:.

 Aldol reaction with acetone.

The mechanism of this catalytic system is believed to be similar to that of previous reported zinc-proline complex catalyzed aldol reactions (39,40), which can be assumed as the mimic of Type II aldolases. The Zn2+ ion may co-ordinate to ketone and can promote the formation of ketone-enolate. Although the asymmetric ligand provides a chiral environment, resulting in the enantioselectivity of the current system.

In summary, a series of C1-symmetric chiral prolinamides based on o-phenylenediamine were designed and were successfully used as chiral ligands in metal-assisted asymmetric direct aldol reactions in aqueous media for the first time. This catalytic direct aldol provides an easy access to the chiral β-hydroxy ketones with good diastereoselectivities (dr up to 88:12) and high enantioselectivities (up to 96% ee). This catalytic results were really good in the area of mimicking the type II aldolase. These new C1-symmetric chiral prolinamides based on o-phenylenediamine explore an interesting area of direct aldol reaction catalyzed by zinc salt and organocatalysts in aqueous media.

Experimental Section

General remarks

Cyclohexanone was distilled and p-nitrobenzaldehyde was sublimated before use. All reagents not listed were used as received from common commercial sources. NMR spectra were obtained on Bruker 300 or 500 MHz spectrometer in DMSO-d6 or CDCl3, and chemical shifts are reported in ppm using tetramethylsilane as internal standard. IR and ESI-MS spectra were measured on Bruker Vector 22 as KBr pellets and Finnigan Mat TSQ 7000 instruments (Finnigan, San Jose, CA, USA), respectively. Microanalyses were obtained on Perkin-Elmer 240 instruments (PerkinElmer, CT, USA), and melting points (mp) were determined with a digital electrothermal apparatus without further correction. Optical rotation measurements were determined at RT using HPLC-grade solvents. Analytical HPLC measurements were carried out on the Perkin-Elmer machine. Compound 1 was prepared according to the reported method (41).

General procedure for the synthesis of 2–6 (Data S2)

2: (2S,4R)-4- hydroxypyrrolidine-2-carboxylic acid (1.15 g, 5 mmol) and TEA (0.61 g, 6 mmol) were dissolved in THF (20 mL). To the solution was added dropwise ethylchloroformate (0.65 g, 6 mmol) at 0 °C. After the solution was stirred for 15 min, aniline (0.46 g, 5 mmol) was added. The resulting solution was stirred at 0 °C for 1 h, at room temperature for 16 h, detected by TLC. The reaction mixture was filtered and washed with THF, and the filtrate was evaporated to dryness. The residue was purified through chromatography on a silica gel column eluted with methanol and dichloromethane to give colorless oil. The colorless product were dissolved in CH2Cl2 (12 mL) and F3CCOOH (3 mL) and stirred for 2 h; the reaction was treated by ammonia solution (10 mL) for 0.5 h; the aqueous layer was extracted with CH2Cl2 (20 mL × 3). The combined organic phase was washed with brine, dried over anhydrous Na2SO4 and removal of the solvent gave a white solid 2 with 90% yield inline image = −29.3 (c = 1.0, CH3OH). 1H NMR (300 MHz, CDCl3): δ 9.79 (s, 1H), 7.61-7.58 (m, 2H), 7.36-7.31 (m, 2H), 7.14-7.09 (m, 1H), 4.49-4.47 (t, = 3.2 Hz 1H), 4.18-4.13 (t, = 8.6 Hz 1H), 3.14-3.09 (dd, = 1.5, 12.2 Hz 1H), 2.92-2.87 (dd, = 3.0, 12.6 Hz, 1H), 2.42-2.34 (m, 1H), 2.09-2.00 (m, 1H).13C NMR (75 MHz, CDCl3): δ 173.2, 137.6, 129.0, 124.2, 119.5, 73.2, 60.2, 55.3, 39.8. IR (KBr disc): 3450 cm−1.ESI-MS: m/z = 207.24[M + H+]. Anal. Calcd (%) for C11H14N2O2: C, 64.06; H, 6.84; N, 13.58; Found: C, 64.03; H, 6.85; N, 13.55.

3: (2S,4R)-4-hydroxypyrrolidine-2-carboxylic acid (1.15 g, 5 mmol) and TEA (0.61 g, 6 mmol) were dissolved in THF (20 mL). To the solution was added dropwise ethylchloroformate (0.65 g, 6 mmol) at 0 °C. After the solution was stirred for 15 min, o-phenylenediamine (0.54 g, 5 mmol) was added. The resulting solution was stirred at 0 °C for 1 h, at room temperature for 16 h, detected by TLC. The reaction mixture was filtered and washed with THF, and the filtrate was evaporated to dryness. The residue was purified through chromatography on a silica gel column eluted with methanol and dichloromethane to give colorless oil ((2S,4R)-N-(2-aminophenyl)-4-hydroxypyrrolidine-2-carboxamide) with 72% yield.

Boc-protected proline (0.65 g, 3 mmol) and TEA (0.31 g, 3 mmol) were dissolved in THF (20 mL). To the solution was added dropwise ethylchloroformate (0.38 g, 3.5 mmol) at 0 °C. After the solution was stirred for 15 min, the intermedia obtained above (0. 96 g, 3 mmol) was added. The resulting solution was stirred at 0 °C for 1 h, at room temperature for 16 h, detected by TLC. The reaction mixture was filtered and washed with THF, and the filtrate was evaporated to dryness. The residue was purified through chromatography on a silica gel column eluted with methanol and dichloromethane to give colorless oil ((R)-tert-butyl 2-((2-((2S,4R)-4-hydroxypyrrolidine-2-carboxamido)phenyl)carbamoyl)pyrrolidine-1-carboxylate) with 65% yield. Then, the Boc-protect 3 was dissolved in CH2Cl2 (12 mL) and CF3COOH (3 mL) and stirred for 2 h; the reaction was treated by ammonia solution (10 mL) for 0.5 h; the aqueous layer was extracted with CH2Cl2 (20 mL × 3). The combined organic phase was washed with brine, dried over anhydrous Na2SO4 and removal of the solvent gave brown oil 3 (0.61 g, 38% yield overall steps). inline image = −30.1 (c = 1.0, CH3OH). 1H NMR (300 MHz, DMSO-d6): 1H δ 9.87 (s, 1H), 7.67-7.61 (m, 2H), 7.15-7.12 (m, 2H), 4.22-4.21 (m, 2H), 3.92-3.87 (m, 2H), 3.74-3.69 (m, 2H), 2.95-2.78 (m, 4H), 2.10-1.98 (m, 2H), 1.88-1.78 (m, 2H), 1.71-1.59 (m, 2H).13C NMR (75 MHz, DMSO-d6): δ 174.4, 174.4, 131.0, 130.7, 125.4, 125.3, 124.4,.124.1, 71.9, 61.1, 60.3, 55.5, 47.2, 40.25, 30.9, 26.4. IR (KBr disc): 3422, 3332, 3234 cm−1. ESI-MS: m/z = 319.33 [M + H+]. Anal. Calcd (%) for C16H22N4O3 C, 60.36; H, 6.97; N, 17.60; Found: C, 60.29; H, 7.01; N, 17.55.

4: (2S,4R)-4-hydroxypyrrolidine-2-carboxylic acid (2.32 g, 10 mmol) and TEA (1.22 g, 6 mmol) were dissolved in THF (30 mL). To the solution was added dropwise ethylchloroformate (1.31 g, 12 mmol) at 0 °C. After the solution was stirred for 15 min, o-phenylenediamine (0.54 g, 5 mmol) was added. The resulting solution was stirred at 0 °C for 1 h, at room temperature for 16 h, detected by TLC. The reaction mixture was filtered and washed with THF, and the filtrate was evaporated to dryness. The residue was purified through chromatography on a silica gel column eluted with methanol and dichloromethane to give colorless oil (Bco-protect 4) with 62% yield. Then, colorless Bco-protected 4 were dissolved in CH2Cl2 (12 mL) and CF3COOH (3 mL) and stirred for 2 h. The reaction was treated by ammonia solution (10 mL) for 0.5 h; the aqueous layer was extracted with CH2Cl2 (20 mL × 3). The combined organic phase was washed with brine, dried over anhydrous Na2SO4 and removal of the solvent gave brown oil 4 (0.71 g, 42% yield overall steps).inline image = −45.7 (c = 1.0, CH3OH). 1H NMR (300 MHz, DMSO-d6): δ 9.88 (s, 1H), 7.65-7.62 (m, 2H), 7.15-7.11 (m, 2H), 4.22-4.20 (m, 2H), 3.92-3.87 (m, 2H), 2.83-2.82 (d, 2H), 2.07-2.00 (m, 2H), 1.88-1.80 (m, 2H). 13C NMR (75 MHz, DMSO-d6): δ 174.4, 130.9, 125.4, 124.3, 71.9, 60.3, 55.5, 40.2. IR (KBr disc): 3434, 3340, 3201 cm−1. ESI-MS: m/z = 335.33 [M + H+]. Anal. Calcd (%) for C16H22N4O4 C, 57.47; H, 6.63; N, 16.76; Found: C, 57.55; H, 6.59; N, 16.72.

5: The o-phenylenediamine (1.08 g, 10 mmoL) and pyridine (0.96 g,12 mmol) were dissolved in the THF (20 mL). To the solution was added dropwise 1-naphthoyl chloride in THF (20 mL) in 30 min; the resulting solution was stirred at room temperature for 4 h, detected by TLC. Then, the mixture was filtered to get the solution of N-(2-aminophenyl)-1-naphthamide.

Boc-protected proline (1.61 g, 8 mmol) and TEA (0.82 g, 8 mmol) were dissolved in THF (20 mL). To the solution was added dropwise ethylchloroformate (1.08 g, 10 mmol) at 0 °C. After the solution was stirred for 15 min, the prepared solvent mentioned above was added. The resulting solution was stirred at 0 °C for 1 h at room temperature for 16 h and detected by TLC. The reaction mixture was filtered and washed with THF, and the filtrate was evaporated to dryness. The residue was purified through chromatography on a silica gel column eluted with methanol and dichloromethane to give colorless oil as Boc-protect 5. Then, the colorless Boc-protect 5 was dissolved in CH2Cl2 (12 mL) and CF3COOH (3 mL), and stirred for 2 h; the reaction was treated by ammonia solution (10 mL) for 0.5 h; the aqueous layer was extracted with CH2Cl2 (20 mL × 3). The combined organic phase was washed with brine, dried over anhydrous Na2SO4 and removal of the solvent gave brown oil 5 (2.34 g, 65% yield overall steps). inline image = 3.03 (c = 1.0, CH3OH). 1H NMR (300 MHz, DMSO-d6): δ 10.38 (s, 1H), 10.33 (s, 1H), 8.46-8.44 (d, 1H), 8.19-8.16 (d, 1H), 8.14-8.11 (d, 1H), 8.06-8.03 (m, 1H), 7.98-7.96 (d, 1H), 7.68-7.61 (m, 3H), 7.46-7.44 (m,1H), 7.36-7.30 (m, 1H), 7.22-7.17 (m, 1H), 3.80-3.76 (m, 1H), 2.92-2.84 (m, 1H), 2.75-2.68 (m, 1H), 2.12-2.00 (m, 1H), 1.89-1.78 (m, 1H), 1.68-1.56 (m, 2H). 13C NMR (75 MHz, DMSO-d6): δ 174.1, 168.5, 134.3, 134.2, 133.7, 131.0, 128.8, 128.4, 127.7, 126.4, 126.0, 125.4, 124.3, 122.0, 61.4, 47.2, 30.9, 26.4. IR (KBr disc): 3464, 3249 cm−1. ESI-MS: m/z = 360.25 [M + H+].Anal. Calcd (%) for C22H21N3O2 C, 73.52; H, 5.89; N, 11.69; Found: C, 73.46; H, 5.61; N, 11.64.

6: The o-phenylenediamine (1.08 g, 10 mmol) and pyridine (0.96 g,12 mmol) were dissolved in the THF (20 mL). To the solution was added dropwise 1-naphthoyl chloride in THF (20 mL) in 30 min, the resulting solution was stirred at room temperature for 4 h and detected by TLC. Then, the mixture was filtered to get the solution of N-(2-aminophenyl)-1-naphthamide.

(2S,4R)-4-Hydroxypyrrolidine-2-carboxylic acid (1.95 g, 8 mmol) and TEA (0.82 g, 8 mmol) were dissolved in THF (30 mL). To the solution was added dropwise ethylchloroformate (1.08 g, 10 mmol) at 0 °C. After the solution was stirred for 15 min, the solution obtained above (0.54 g, 5 mmol) was added. The resulting solution was stirred at 0 °C for 1 h, at room temperature for 16 h and detected by TLC. The reaction mixture was filtered and washed with THF, and the filtrate was evaporated to dryness. The residue was purified through chromatography on a silica gel column eluted with methanol and dichloromethane to give colorless oil as Boc-protected 6. Then, the colorless Boc-protected 6 was dissolved in CH2Cl2 (12 mL) and CF3COOH (3 mL) and stirred for 2 h; the reaction was treated by ammonia solution (10 mL) for 0.5 h; the aqueous layer was extracted with CH2Cl2 (20 mL × 3). The combined organic phase was washed with brine, dried over anhydrous Na2SO4 and removal of the solvent gave brown oil 6 (1.73 g, 46% yield overall steps). inline image = −30.4 (c = 1.0, CH3OH). 1H NMR (300 MHz, DMSO-d6): δ 10.34 (s, 1H), 10.29 (s, 1H), 8.42-8.39 (m, 1H), 8.15-8.10 (m, 2H), 8.05-8.02 (m, 1H), 7.96-7.93 (m, 1H), 7.67-7.59 (m, 3H), 7.45-7.42 (m,1H), 7.34-7.29 (m, 1H), 7.21-7.12 (m, 2H), 3.98-3.87 (m, 1H), 2.82-2.68 (m, 3H), 2.10-2.03 (m, 1H), 1.84-1.76 (m, 1H). 13C NMR (75 MHz, DMSO-d6): δ 174.0, 168.5, 134.3,134.1, 133.7, 131.0, 130.3, 129.1, 128.8, 128.4, 127.6, 127.4, 127.2, 126.9, 126.4, 126.2, 125.9, 125.5, 124.4, 122.1. 71.9, 60.6, 55.4, 40.2, 28.6. IR (KBr disc): 3422, 3338 cm−1; ESI-MS: m/z = 376.25 [M + H+] Anal. Calcd (%) for C22H21N3O3 C, 70.38; H, 5.64; N, 11.19; Found: C, 70.30; H, 5.60; N, 11.25.

Typical procedure for the direct aldol reaction

Catalyst 5 (5 mol%, 0.025 mmol, 8.99 mg) and Zn(OTf)2 (5 mol%, 0.025 mmol, 9.08 mg) were stirred in the solvent (1.2 mL, cyclohexanone:H2O = 5:1) for 10 min. The 4-nitrobenzaldehyde (0.5 mmoL, 75.5 mg) was then added and the resulted mixture was stirred for the given time and temperature. The aqueous layer was decanted from the precipitated products and extracted with ether. The desired products 7 and 8 were obtained by flash chromatography and have been characterized to be identical to those of known samples (42) (Data S1).

Acknowledgments

We gratefully acknowledge National Natural Science Foundation of China (Grant No. 20772056) and Jiangsu 333 program (for Pan) for the generous financial support. The research funds for Pan from the Qing-Lan program of Jiangsu Province and the Kua-Shi-Ji program of the Education Ministry of China are also acknowledged.

Ancillary