Second cycle asymmetric dihydroxylation reaction

Abstract

A process for catalyzing asymmetric dihydroxylations of olefins employs an Os(VI) complex as a catalytic intermediate in the formation of chiral vicinal diol products. The process requires a chiral bidentate ligand that favors diol formation in the “second cycle” of asymmetric dihydroxylation.

Description

FIELD OF THE INVENTION

[0002] The invention relates to osmium catalyzed dihydroxylations of olefins. More particularly, the invention relates to osmium catalyzed second cycle asymmetric dihyroxylations of olefins.

BACKGROUND

[0003] Osmium-catalyzed asymmetric dihydroxylation (AD) of olefins using cinchona alkaloid-derived ligands has proven to be a highly effective and reliable process across nearly the entire range of olefin types and substitution patterns on both laboratory and industrial scales (Kolb, H. C.; et al. Chem. Rev., 1994, 94, 2483-2547; H. C. Kolb, K. B. Sharpless, in Transition Met. Org. Synth., Vol 2, (Eds. M. Beller, C. Bolm), Wiley-VCH, Weinheim, 1998, 219-242).

[0004] The most helpful mechanistic insight in osmium catalysts, from the viewpoint of these process improvement endeavors developing the AD, was the realization that there are two catalytic cycles producing diol (FIG. 1A) (Wai, J. S. M.; et al. J. Am. Chem. Soc. 1989, 111, 1123). In homogeneous conditions, when N-methylmorpholine N-oxide (NMO) is employed as a reoxidant, the second cycle, leading to the reduced enantioselectivities, dominates because two possible hydrolysis steps, h1 and h2 are much slower than the three red-ox steps, r1, r2, and r3. As a consequence, the Os (VI) bisglycolate (ii) becomes the most stable, resting form of the catalyst.

[0005] It is known that certain classes of olefins exhibit unique reactivity in the osmium-catalyzed aminohydroxylation and dihydroxylation (Rubin A. E.; Sharpless, K. B. Angew. Chem. Int. Ed. Engl. 1997, 36, 2637-2640; Pringle, W.; Sharpless, K. B. Tetrahedron Lett. 1999, 40, 5150-5154; Fokin, V. V.; Sharpless, K. B. Angew. Chem. Intl. Ed. Eng., 2001, in press). Unlike most substrates, these special olefins undergo rapid and nearly quantitative conversion to aminoalcohols and diols, correspondingly, with very low catalyst loading in the absence of added ligands and with only one equivalent of the oxidant. This is in sharp contrast to other olefins, whose turnover is crucially dependent on the Ligand Acceleration Effect (for a review of ligand accelerated catalysts, see Berrisford, D. J.; et al. Angew. Chem. Int. Ed. Engl. 1996, 35, 451-454). The more recently, the remarkable reactivity of unsaturated carboxylic acids has been characterized (Fokin, V. V.; Sharpless, K. B. Angew. Chem. Intl. Ed. Eng., 2001, in press).

[0006] In all of the above special cases, only racemic products are formed, even when a chiral ligand is added in large excess. This and other available evidence (H. -T. Chang, Ph.D. dissertation, The Scripps Research Institute, 1997.) suggest that these olefins turn over almost exclusively in the 2nd catalytic cycle, where osmium (VI) bis(glycolate) (ii) (or the bis (azaglycolate) in case of aminohydroxylation) is the most stable intermediate—so stable that it is the only detectable osmium complex present under steady-state conditions. According to the current mechanistic hypothesis, proximal carboxylate groups facilitate the hydrolysis of this complex, which is the rate-determining step. This explains the dramatically increased reactivity of these substrates.

[0007] Although it is desirable to avoid the 2nd cycle at all costs, deleterious as it is to enantioselectivity, the enticing possibilities it offers for a new way to control osmium(VIII) catalysis have been clear since the time of its discovery in 1982. Although early attempts to obtain enantioselectivity with 2nd cycle ligands failed, the recent enormous jump in effectiveness of the 2nd cycle systems (Rubin A. E.; Sharpless, K. B. Angew. Chem. Int. Ed. Engl. 1997, 36, 2637-2640; Pringle, W.; Sharpless, K. B. Tetrahedron Lett. 1999, 40, 5150-5154) has shown that its inherent advantages may be exploited to develop new catalytic processes. To gain control over the 2nd cycle, one needs to design a ligand that (a) is chiral and capable of controlling stereochemistry in the olefin oxidation step r3; (b) aids in the hydrolytic release (h2) of the product from the Os(VI) complex (ii) formed by oxidation of olefin; and (c) is not itself hydrolytically removed from the osmium coordination sphere. This restriction is not really so severe—it only has to dominate the catalysis. This can be achieved with a mobile ligand too (e.g., ligand-accelerated catalysis or simply equilibrium favoring the desired osmium complex in the olefin oxidation step).

SUMMARY

[0008] One aspect of the invention is directed to a process for catalyzing an asymmetric dihydroxylation reaction for converting an olefin substrate to a chiral vicinal diol product. The process employs the step of mixing the olefin substrate under reaction conditions with a catalytic amount of osmium, a stoichiometric amount of N-methylmorpholine oxide as a co-oxidant, and a suitable amount of chiral bidentate ligand for ligating, together with the olefin substrate, to the osmium for forming an Os(VI) complex as a catalytic intermediate to the formation of the chiral vicinal diol product.

[0009] Another aspect of the invention is directed to a process for catalyzing a second cycle dihydroxylation reaction for converting an alkene substrate to a chiral vicinal diol product. The process employs the step of mixing the alkene substrate under reaction conditions with a catalytic amount of osmium, a stoichimetric amount of N-methylmorpholine oxide as a co-oxidant, and a suitable amount of chiral ligand for facilitating the asymmetric dihydroxylation reaction. The chiral ligand is represented by the following structures: 1

[0010] In the above structure, R is a radical selected from hydrogen, carboxylate, phenyl, 1-naphthyl, 2-naphthyl, alkyl(C1-C12), cyclo-alkyl (C3-C12), carbamoyl, N-alkyl(C1-C12) carbamoyl, or N, N-alkyl(C1-C12) carbamoyl. The phenyl, 1-naphthyl, and 2-naphthyl radicals may optionally have substituents in any available position. R1 is a radical selected from carboxylate, carbamoyl, N-alkyl(C1-C12) carbamoyl, or N,N-dialkyl(C1-C12) carbamoyl. R2 is a radical selected from H, alkyl(C1-C12), aryl, or heteroaryl. X and Y are radicals independently selected from hydroxyl, amino, N-alkyl(C1-C12)sulfonylamino, N-arylsulfonylamino, or N-heteroarylsulfonylamino. In one of the preferred modes, the chiral ligand is an α-hydroxy-β-N-sulfonyl-amino acid. In another of the preferred modes, the chiral ligand is an β-hydroxy-α-N-sulfonyl-amino acid. Good results can often be achieved if the reaction is carried out with 1 mol % to 10 mol % in chiral ligand. However, 5 mol % of chiral ligand is more usual. The range for the osmium concentration is between 0.1 mol % and 1 mol %. However, an osmium concentration of 0.2 mol % is more usual. The best olefin substrates are often electron deficient. The optimal pH for the reaction mixture is often approximately 5.

BRIEF DESCRIPTION OF DRAWINGS

[0011] FIG. 1A illustrates the first cycle and second cycle of asymmetric dihydroxylation.

[0012] FIG. 1B illustrates the first cycle and second cycle of asymmetric dihydroxylation when using these “second cycle” ligands.

[0013] FIG. 2 illustrates a graph of percent ee of product as a function of the amount of ligand used. The ee's remained constant throughout the course of reactions and, more importantly, as low as 1.5-2 mol % ligand was sufficient to attain the ceiling ee.

[0014] FIG. 3 illustrates a reaction using N-(p-toluenesulfonyl) threonine (2S,3R)-6 which has been found to be particularly effective for dihydroxylation of cinnamate esters, with 70% ee realized in one case.

[0015] FIG. 4 illustrates the structures of preferred chiral ligands.

[0016] FIG. 5 illustrates a table that shows dihydroxylation of styrene with novel ligands.

DETAILED DESCRIPTION

[0017] Described here are the first ligands found to induce asymmetry in the osmium-catalyzed dihydroxylation proceeding in the 2nd catalytic cycle. As a simple model for screening the ligand candidates shown in the table of FIG. 5, the dihydroxylation of styrene under the Upjohn conditions is used. It is noteworthy that overoxidation of the product diol, a fairly common side reaction in the Upjohn dihydroxylation, was not observed.

[0018] Even such simple ligand as tartaric acid, although needed in 25 mol % quantity, showed some asymmetric induction. Finding a ligand with a higher affinity for osmium was an obvious requirement to reduce the amount of ligand. Since N-sulfonyl-1,2-hydroxyamines (vicinal hydroxysulfonamides) have much higher binding constants for osmium than analogous 1,2-diols, a number of N-toluenesulfonyl derivatives of α,β-hydroxyaminoacids were screened resulting in the improvement of the ee to 42%. The ee's remained constant throughout the course of reactions and, more importantly, as low as 1.5-2 mol % ligand was sufficient to attain the ceiling ee (FIG. 2). To ensure that the maximum possible ee afforded by each ligand was observed, 5 mol % of ligand was routinely used. It has been demonstrated from different examples that 1.5-2.0 mol % is sufficient. The finding that dependence of the enantioselectivity on the amount of the ligand shows saturation behavior points to the fact that the process in its present form operates at the maximum ee afforded by the ligand.

[0019] Simple structure-activity studies have revealed that a free carboxylate group appears to be an essential component of a successful ligand. Thus, only racemic diol was obtained when the methyl ester of 1 was used as a ligand. Location of the HO— and TsNH—groups was found to play an important role as well. For example, the phenylisoserine-based ligand 1 afforded higher ee than its regioisomer 2. The absolute configuration of the diol products appears to be determined by the stereochemistry of the α-carbon of the ligand (cf. entries 1, 3 and 6). Additional investigations have shown that modification of the substituents on the sulfonamide group (R—SO2NH—) has only a minor effect on the stereochemical outcome of the reaction. However, the ligand should preferentially contain an N-sulfonyl moiety, as its replacement with an amide (as in 5) or a carbamate resulted in very low to no enantioselectivity.

[0020] Most of the ligands discussed above can be readily prepared in enantiomerically enriched form using previously developed catalytic olefin transformations (Li, G.; et al. Angew. Chem. Int. Ed. Engl. 1996, 35, 451-454; H. C. Kolb, K. B. Sharpless, in Transition Metals for Fine Chemicals and Organic Synthesis, Vol. 2 (Eds.: M. Beller, C. Bolm), Wiley-VCH, Weinheim, 1998, 243-260; G. Schlingloff, K. B. Sharpless, in Asymmetric Oxidation Reactions: A Practical Approach, (Ed.: T. Katsuki), Oxford University Press, Oxford, in press). Furthermore, some hydroxyaminoacids are commercially available compounds, and can be easily converted to their corresponding N-sulfonyl derivatives (Brenner, M.; et al. Helv. Chim. Acta, 1951, 36, 2102-2106). For example, N-(p-toluenesulfonyl) threonine (2S,3R)-6 has been found to be particularly effective for dihydroxylation of cinnamate esters, with 70% ee realized in one case (FIG. 3). Acidification of the reaction mixture with sulfuric or acetic acid to approximately pH 5 considerably accelerates the reaction without jeopardizing the enantioselectivity.

[0021] Accordingly, it is demonstrated herein that confining the osmium-catalyzed asymmetric dihydroxylation to the 2nd catalytic cycle is a viable concept. The exemplary processes disclosed herein provide to modest to good enantioselectivities. However, the concept of the invention is broader than the examples and offers many variables for optimization. The 2nd cycle-based osmium-catalyzed oxidations disclosed herein facilitates the asymmetric oxidation of olefins that were heretofor difficult to oxidize.

[0022] Experimental Procedure:

[0023] Typical dihydroxylation procedure as exemplified on methyl 4-nitro-cinnamate: 4-nitrocinnamic acid methyl ester (207 mg, 1 mmol) and N-(4-toluenesulfonyl)-(L)-threonine (13.6 mg, 5 mol%) (Brenner, M.; et al. Helv. Chim. Acta, 1951, 36, 2102-2106) were dissolved in 6 ml of 1:1 tBuOH/H2O mixture. NMO (50 wt % in water, 228 μl, 1.1 mmol) and OSO4 (0.1M in acetonitrile, 20 μl , 0.002 mmol) were added successively. The pH was adjusted to 5 by addition of 150 p82 l 2N H2SO4, and the reaction mixture was stirred vigorously for 24 hrs, at which time the pH was adjusted to 5 again. After additional 24 hrs (≧98% conversion by LC), methyl (2R,3S)-(+)-2,3-dihydroxy-3-(p-nitrophenyl)-propionate (Denis, J. A.; et al. J. Org. Chem. 1990, 55,1957) was obtained in 70% ee (HPLC: Chiralcel OG, 20% iPrOH/hexane). The reaction time can be reduced to ca. 24 hrs by maintaining constant pH using a pH-stat. A 10 mmol scale reaction, performed under similar conditions, afforded product as white solid in 75% yield (1.8 g) and 70% ee. Recrystallization from ethanol produced needle-shaped crystals in 57% yield and 81% ee.

[0024] Detailed Description of Figures:

[0025] FIG. 1A illustrates the first cycle and second cycle of asymmetric dihydroxylation. There are two catalytic cycles producing diol. In homogeneous conditions, when N-methyl-morpholine N-oxide (NMO) is employed as a reoxidant, the second cycle, leading to the reduced enantioselectivities, dominates because two possible hydrolysis steps, h1 and h2 are much slower than the three red-ox steps, r1, r2, and r3. As a consequence, the Os (VI) bisglycolate(ii) becomes the most stable, resting form of the catalyst.

[0026] FIG. 1B illustrates the first cycle and second cycle of asymmetric dihydroxylation. There are two catalytic cycles producing diol. In this diagram, the “first cycle” is not entered unless the ligand is hydrolyzed off of the osmium in step h′. “L” in the diagram can represent any monodentate ligand species in the reaction mixture.

[0027] FIG. 2 shows a graph of percent ee of product as a function of the amount of ligand used. The ee's remained constant throughout the course of reactions and, more importantly, as low as 1.5-2 mol % ligand was sufficient to attain the ceiling ee.

[0028] FIG. 3 shows a reaction using N-(p-toluenesulfonyl) threonine (2S,3R)-6 which has been found to be particularly effective for dihydroxylation of cinnamate esters, with 70% ee realized in one case. N-sulfonyl-1,2-hydroxyamines (vicinal hydroxysulfonamides) have much higher binding constants for osmium than analogous 1,2-diols, a number of N-toluenesulfonyl derivatives of a,b-hydroxyaminoacids were screened resulting in the improvement of the ee to 42% (in the dihydroxylation of styrene).

[0029] FIG. 4 shows the structures of the ligands examined. Simple structure-activity studies have revealed that a free carboxylate group appears to be an essential component of a successful ligand. Thus, only racemic diol was obtained when the methyl ester of 1 was used as a ligand. Location of the HO— and TsNH—groups was found to play an important role as well. For example, the phenylisoserine-based ligand 1 afforded higher ee than its regioisomer 2. The absolute configuration of the diol products appears to be determined by the stereochemistry of the α-carbon of the ligand (cf. entries 1, 3 and 6). Additional investigations have shown that modification of the substituents on the sulfonamide group (R—SO2NH—) has only a minor effect on the stereochemical outcome of the reaction. However, the ligand should preferentially contain an N-sulfonyl moiety, as its replacement with an amide (as in 5) or a carbamate resulted in very low to no enantioselectivity.

[0030] FIG. 5 is a table that shows dihydroxylation of styrene with novel ligands. All reactions were performed on 1 mmol scale at 0.5 M concentration in tBuOH/H2O (1:1) with 1.1 eq. NMO and 0.2 mol % of OsO4. The progress was monitored by GC and ees were determined by HPLC (Chiralcel OB, 10% iPrOH/Hexane); the absolute configuration of styrene diol was assigned by comparison with authentic samples.

Claims

1. A process for catalyzing an asymmetric dihydroxylation reaction for converting an olefin substrate to a chiral vicinal diol product, said process comprising the step of mixing the olefin substrate under reaction conditions with a catalytic amount of osmium, a stoichiometric amount of N-methylmorpholine oxide as a co-oxidant, and a suitable amount of chiral bidentate ligand for ligating, together with the olefin substrate, to the osmium for forming an Os(VI) complex as a catalytic intermediate to the formation of the chiral vicinal diol product.

2. A process for catalyzing a second cycle dihydroxylation reaction for converting an olefin substrate to a chiral vicinal diol product, said process comprising the step of mixing the olefin substrate under reaction conditions with a catalytic amount of osmium, a stoichiometric amount of N-methylmorpholine oxide as a co-oxidant, and a suitable amount of chiral ligand for facilitating the asymmetric dihydroxylation reaction, the chiral ligand being represented by the following structure:

3. A process according to claim 2 where the chiral ligand is an α-hydroxy-β-N-sulfonyl-amino acid.

4. A process according to claim 2 where the chiral ligand is an α-N-sulfonyl-β-hydroxy-amino acid.

5. A process according to claim 2 where the reaction is carried out with 1 mol % to 10 mol % in chiral ligand.

6. A process according to claim 5 where the reaction is carried out with 5 mol % of chiral ligand.

7. A process according to claim 6 where the osmium concentration is between 0.1 mol % and 1 mol %.

8. A process according to claim 7 where the osmium concentration is 0.2 mol %.

9. A process according to claim 8 where the olefin substrate is electron deficient.

10. A process according to claim 9 where the pH of the reaction mixture is approximately 5.