Description Field of Invention: The invention relates to the regio-selective and enantio- selective conversion of olefins to B-hydroxyamides. More particularly, the invention relates to catalytic asymmetric additions or amidohydroxylations of olefins and other unsaturated substrates using an N-halo carboxamide as an oxidizing agent in the presence of an osmium catalyst and a chiral ligand.

Background : The p-aminoalcohol moiety is one of the most abundant structural units in biologically active compounds. Recent developments by us (Li et al. Angew. Chem. 1996,108,449-452; Angew. Chem. Int. Ed. Engl. 1996,35,451-454; Li et al. Acta Chem. Scand. 1996,50,649-651) and others (Larrow et al. J. Am.

Chem. Soc. 1996,118,7420-7421; Shibasaki et al. Pure Appl.

Chem. 1996,68,523-530) have led to viable metal catalyzed routes for its asymmetric synthesis.

Separate and distinct synthetic methodologies have been developed by Sharpless et al. for the vicinal hydroxyamination of olefins to form the p-aminoalcohol moiety. There are three major groups of oxyamination procedures which produce aminoalcohols (Sharpless et al. J. Am. Chem. Soc. 1975,97, 2305; Sharpless et al. J. Org. Chem. 1978, 43,2628; Sharpless et al. J. Org. Chem. 1980, 45,2257), hydroxysulfonamides (Sharpless et al. J. Org. Chem. 1976, 41,177; Sharpless et al.

J. Org. Chem. 1978, 43,2544; Sharpless et al. J. Org. Chem.

1979, 44,1953; Sharpless et al. Org. Syn. 1980,61,85) or hydroxycarbamates (Sharpless et al. J. Am. Chem. Soc. 1978,100, 3596; Sharpless et al. J. Org. Chem. 1980, 45,2710; Sharpless et al. U. S. Patent No.'s 4,871,855; 4,965,364; 5,126,494; EP 0 395 729). Each oxyamination procedure has unique reaction conditions and includes variations in solvents, auxiliary salts,

nucleophiles, temperature, stoichiometric v. catalytic amounts of osmium species and stoichiometric v. catalytic amounts of ligand. Each procedure is highly dependent on the nature of the substrate and possesses unique properties which afford different yields, chemoselectivities, stereoselectivities, regioselectivities and enantioselectivitive outcomes.

1. Aminoalcohols The first reported oxyamination procedure (Sharpless et al.

J. Am. Chem. Soc. 1975,97,2305) generated aminoalcohols from mono and di substituted olefins, using stoichiometric quantities of a tri-oxo (tert-butylimido) osmium species. The procedure required reductive cleavage of the osmate ester which was performed with lithium aluminum hydride and afforded tertiary vicinal aminoalcohols. Yields were good to excellent, but in some cases, the side product vicinal diol was formed as an undesired by-product. The stereochemistry of addition, in methylene chloride or pyridine, was exclusively cis (Sharpless et al. J. Org. Chem. 1978, 43,2628). In addition, the carbon- nitrogen bond formed was, in every case, at the least substituted olefinic carbon atom. Di and tri-substituted olefins reacted much slower with the generated imido reagent than with monosubstituted alkenes; tetrasubstituted alkenes yielded only the corresponding diol. However, by using a coordinating solvent such as pyridine, higher yields and higher ratios of aminoalcohol to diol were reported. Sharpless et al.

J. Org. Chem. 1980, 45,2257; Sharpless et al. J. Org. Chem.

1976, 41,177; Sharpless et al. J. Org. Chem. 1978,43,2544.

2. Hydroxysulfonamides Sharpless et al. first demonstrated that hydroxysulfonamides could be obtained using either stoichiometric or catalytic amounts of 1% osmium tetraoxide in the presence of 1.5-5 equivalents of Chloramine-T trihydrate (TsSO2NClNa'3H20, Ts = tosylate ; commercially obtained) to effect cis addition of a hydroxyl (OH) and an arylsulfonamide moiety (Ar-SO2NH) across a mono or disubstituted olefinic linkages (Sharpless et. al. J. Org. Chemistry 1976, 41, 177).

Two procedures were developed to effect hydroxyamination of olefins using sulfonamides. (Sharpless et al. Org. Syn. 1980, 61,85). The first procedure used phase transfer catalysis conditions at 55-60 °C with 1% Os04, 1 : 1 v/v, 0.20 Molar CHC13/H20, and benzyltriethylammonium chloride as the phase transfer catalyst. The chloramine T-trihydrate (TsSO2NClNa'3H20) was either added directly or formed in situ in water; this solution was then directly used in the phase transfer mixture.

The in situ procedure, for generating the chloramine salts, involved stirring a suspension of the arylsulfonamide with an equivalent of sodium hypochlorite (Clorox) until a homogenous solution was obtained. The yields were comparable with those obtained with isolated chloramine salts and the procedure was found most effective for monosubstituted and 1,2 disubstituted olefins. The phase transfer method, however, gave poor results with trisubstituted and 1,1-disubstituted olefins and the procedure did not succeed with diethyl fumarate and 2- cyclohexen-1-one. Sharpless et al. J. Org. Chem. 1978,43,2544.

A second procedure was carried out in tert-butyl alcohol at 55-60 °C with 1% OSO4, silver nitrate (with or without) and commercially obtained chloramine T-trihydrate (TsSO2NClNa'3H20) which provided the only source of water. The procedure did not succeed with tetramethylethylene and cholesterol, and negative results were found with most hindered tri-and tetrasubstituted olefins. Sharpless et. al. J. Org. Chemistry 1976,41,177; Sharpless et al. Org. Syn. 1980,61,85. The addition of divalent metal salts such as AgN03 and Hg (N03) 2 improved some reactions, however, other reactions suffered deleterious effects from the addition of the metal salts. Sharpless et al. J. Org Chem. 1978,43,2544; Sharpless et. al. J. Org. Chemistry 1976, 41,177.

Further elaboration on either procedure showed that other sulfonamide derivatives (ArSO2NClNa) could be successfully employed in addition to chloramine T, where Ar = phenyl, o- tolyl, p-chlorophenyl, p-nitrophenyl, and o-carboalkoxyphenyl.

Sharpless et al. J. Org. Chem. 1978,43,2546.

Neither the phase transfer catalyst or tert-butyl alcohol procedures succeeded with tetramethyl ethylene, 2,3-dimethyl-2-

octene, diethyl fumarate, or 2-cyclohexen-l-one. Negative results were also obtained with most hindered tri-and tetrasubstituted olefins. Herranz E., MIT Ph. D. Thesis, 1979, 33.

Solvent conditions for the synthesis of the hydroxysulfonamides included organic solvents such as acetonitrile, tert-butyl alcohol, isopropyl alcohol and chloroform which was in contact with the aqueous phase in the phase transfer catalyst procedure.

The tert-butyl alcohol procedure (including other solvents used) was not run with added water; the phase transfer catalyst (PTC) procedure required a biphasic mixture of 1: 1 v/v chloroform/water. Recently, however, an improvement was reported which used a 1: 1 ratio of organic solvent to water in a homogeneous, rather than a biphasic solution or organic solvent with small amounts of water. These conditions were found to provide optimum enantioselectivity, regioselectivity and improved yields from either the previously described t-butyl alcohol or PTC conditions. Sharpless et al. Angew. Chemie Intl Ed. 1996, 35,451.

The use of chiral ligands with sulfonamides provides enantioselectivity and has been observed to both accelerate and decelerate the rate of catalysis. The hydroxysulfonamide process is a stereoselective cis process. The presence of ligands also has a dramatic effect on the regioselectivity. In a study with no ligand present with methyl cinnamate, the two regioisomers were present in a 2: 1 ratio. With the addition of ligand, the ratio was improved to 5: 1 or greater. Another positive effect of the ligand was its ability to suppress formation of diol by-product. Angew. Chemie Intl Ed. 1996, 35, 451.

Preferred ligands for use with sulfonamides have included the use of monovalent cinchona alkaloids or the bivalent phthalazine based, commercially available (DHQ) 2PHAL and (DHQD) 2PHAL alkaloids. Sharpless et al. Angew. Chemie Intl Ed. 1996, 35,451.

Temperature conditions for the hydroxysulfonamide asymmetric aminohydroxylations have varied from 60 °C to 25 °C

for reactions including sulfonamides, auxiliary salts, ligands, phase transfer catalysts and stoichiometric or catalytic osmium species, primarily in organic solvents with small amounts of water. Recently, it has been shown that temperature can been lowered to 0 °C while running the reaction, to obtain product by filtration; many hydroxysulfonamides tend to be highly crystalline Sharpless et al. Acta Chemica Scandinavica 1996 in press.

Cleavage of the sulfonamides, to free aminoalcohols, have been accomplished via standard deprotection conditions including dissolving metals (Na, NH3 ; Sharpless et al J. Org. Chem 1976, 41, 177) and HBr, acetic acid and phenol (Fukuyama et al.

Tetrahedron Lett. in press).

3.Hydroxycarbamates A drawback with the hydroxysulfonamide procedure was that cleavage conditions were too strong for some substrates. The use of carbamates to protect the nitrogen, however, provided a methodology which avoided the use of harsh acids or reducing deprotection problems found with hydroxysulfonamides (Sharpless et al. J. Am. Chem. Soc. 1978,100,3596; Sharpless et al. J.

Org. Chem. 1980, 45,2710; Sharpless et al. Org. Syn. 1981, 61, 93; Sharpless et al. U. S. Patent No.'s 4,871,855; 4,965,364; 5,126,494; EP 0 395 729).

Sharpless first demonstrated the synthesis of hydroxycarbamates with the use of N-chloro-N-argentocarbamates (Sharpless et al J. Am. Chem. Soc. 1978 100,3596). The N- chloro-N-argentocarbamates were generated in situ via the addition of N-chlorosodiocarbamates and silver nitrate to a solution of the olefin in acetonitrile or tert-butanol with trace amounts of water (4.5 molar equivalents based on olefin) and 1% of osmium tetroxide catalyst to generate vicinal hydroxycarbamates in generally good yields. The methodology was reported to be more effective with electron deficient olefins such as dimethyl fumarate and trisubstituted olefins were reported to be less readily oxyaminated with N-chloro-N- argentocarbamates than with the chloramine-T procedures (Sharpless et. al. J. Org. Chem. 1976, 41, 177).

Sodio-N-chlorocarbamates were always first converted to either argento or mercurio salt analogs. The addition of the AgNO3 or Hg (NO3) 2 salts, to make N-chloro-N-argentocarbamates or mercurio salt analogs, was crucial for the reaction to retain its desired properties. (Sharpless et al J. Org. Chem., 1980, 45,2711). This was in contrast to the sulfonamide conditions, where the sodio-N-chloro-sulfonamide salts could be used directly with either the t-butanol or chloroform/water-phase transfer catalyst procedures (Sharpless et al. J. Org. Chem.

1978, 43,2544).

The addition of nucleophiles such as tetraethylammonium acetate were also proven to be beneficial to the reaction in the procedures using the silver and mercury salts of the chloramines from carbonates. Alternatively, the reactivity and yields were enhanced by addition of excess AgN03 and Hg (N03) 2 (over that needed to react with the NaClNCOOR salt) Sharpless et al. J.

Org Chem. 1980,45,2710.

Preferred conditions included employment of ROCONClNa + Hg (N03) 2 + Et4NOAc with N-chloro-N-sodiocarbamates; these conditions were recommended as the best procedure for mono, di and tri substituted olefins even including some olefins unreactive in all of the various chloramine T based processes.

(Sharpless et al. Org. Syn. 1981, 61,93).

Among the carbamates tried, it was found that both benzyl N-chloro-N-argentocarbamate and tert-butyl N-chloro-N- argentocarbamates (or mercurio analogs) were among the most effective oxidants, especially with addition of nucleophiles such as tetraethylammonium acetate. Other carbamates such as isopropyl, ethyl, menthyl and bornyl derivatives were also used, however, chemo, regio and stereoselectivities were lower.

Virtually no asymmetric induction was observed when chiral menthyl or bornyl derived carbamates were employed for hydroxyaminations. (Sharpless et al J. Am. Chem. Soc. 1978 100, 3596).

Sharpless disclosed the use of stoiciometric amounts of a first generation monovalent alkaloid ligand with a tert-butyl derived N-chloro-N-argentocarbamate for hydroxyamination in a series of patent applications directed to ligand accelerated

catalytic asymmetric dihydroxylation. These disclosures illustrated an hydroxyamination on trans-stilbene with the use of 1.0 equivalent (stoichiometric to olefin) of monovalent DHQD- p-chlorobenzoate (DHQD= hydroquinidine) ligand, 1 mol % osmium tetroxide, silver nitrate (figure) or mercuric chloride (. 80 equivalents; in protocal), 0.09 Molar acetonitrile (93.11 volume % acetonitrile)/water mix (6.89 volume % water) and tertbutyl derived N-chloro-N-argentocarbamate (1.45 equivalents) at 20 °C (figure) or 60 °C (protocal) for 1 hour. The disclosure reported a 51% ee with a 93% yield of aminoalcohol. (Sharpless et al.

U. S. Patent No.'s 4,871,855; 4,965,364; 5,126,494; EP 0 395 729).

In a review on ligand accelerated catalysis, Sharpless et al. noted that a 92% ee had been achieved in a stoichiometric reaction of trioxo- (tert-butylimido) osmium with stilbene in the presence of DHQD-CLB at ambient temperatures (Sharpless et al.

Angew. Chem. Int. Ed. Engl. 1995,34,1059, ref. 80"unpublished results"); this mention did not disclose reaction conditions.

Recently, an oxyamination reaction for the hemisynthesis of taxol and analogs was reported using a tertbutyl derived N- chloro-N-argentocarbamate, excess silver nitrate or other metallic salts, with the use of either catalytic or stoichiometric amounts of osmium and the addition of stoichiometric amounts of monovalent DHQD (hydroquinidine), DHQ (hydroquinine) ligands in an unsuccessful attempt to influence the diastereoselectivity and the regioselectivity of the aminohydroxylation process. Solvent conditions varied from acetonitrile, toluene or pyridine, and the reactions were carried out at 4 °C to room temperature, in the dark. The study reported that quinuclidine ligands had no effect on the amino alcohol yields but found that the addition of chiral tertiary amines had some beneficial effect on the yields of the various amino alcohol isomers formed. (Mangatal et al. Tetrahedron 1989 45,4177). However, the two pseudoenantiomeric alkaloid ligands (i. e. DHQ-OAc and DHQD-OAc ; OAc = acetate) gave a mixture of stereo and regioisomeric products. The result indicates that this particular hydroxyamination process (be it stoichiometric or catalytic was unclear) had exhibited no"asymmetric"effects.

The procedure can therefore not be regarded as an asymmetric aminohydroxylation.

As a whole, the prior art uses hydroxycarbamates which always run at room temperature with either argento or mercurio salt analogs, monovalent ligands, stoichiometric or catalytic osmium species and organic solvents with trace amounts of water.

(Sharpless et al. J. Am. Chem. Soc. 1978,100,3596; Sharpless et al. J. Org. Chem. 1980, 45,2710; Sharpless et al. U. S.

Patent No.'s 4,871,855; 4,965,364; 5,126,494; EP 0 395 729).

Cleavages of the hydroxycarbamates, to free aminoalcohols, are well known in the art and include mild acid or base hydrolysis and catalytic hydrogenolysis, depending on the attached functionality to the carbamate. (Greene, Protective Groups in Organic Synthesis, 1981, Wiley, 1st edn. pp. 223-249).

Currently, there is nothing in the art which directly converts achiral olefinic substrates to asymmetric B- hydroxyamides. What is needed, then, is a method for catalyzing the asymmetric amidohydroxylation of olefin substrates using a cost efficient oxidant which supplies both a carboxamide radical and a hydroxyl radical directly achieving asymmetric R- hydroxyamides in enhanced yields, enantiomeric efficiency, and regio-selectivity while reducing material and labor costs.

Summary of Invention The invention is directed to a method for converting olefinic substrates to asymmetric B-hydroxyamide products. The method of the invention employs an asymmetric addition reaction involving the addition of a carboxamide radical and a hydroxyl radical to the olefinic substrate. Enhanced yields, regioselectivity, and enantioselectivity may be achieved according to the method of the invention.

The asymmetric addition reaction is carried out in a reaction solution which includes the olefinic substrate, an osmium catalyst, a chiral ligand for enantiomerically and regioselectively directing the asymmetric addition, an N-halo carboxamide, a base, and a solvent in homogeneous or heterogeneous conditions. The N-halo carboxamide serves as a source for the carboxamide radical. The olefinic substrate and

N-halo carboxamide are present and soluble within the solvent or cosolvent in approximately stoichiometric amounts as defined below. The osmium is present within the solvent or co-solvent in catalytic amounts.

One aspect of the invention is directed to a method for converting an olefinic substrate to an asymmetric amidoalcohol product by means of a one step osmium-catalyzed asymmetric addition. During the conversion, a carboxamide radical and a hydroxyl radical are added to the olefinic substrate. The reaction mixture employs a solvent having both an organic component and an aqueous component. The aqueous component of the solvent serves as the source of the hydroxyl radical. An N- halo carboxamide serves as the source of the carboxamide radical. The reaction mixture also includes the olefinic substrate, the N-halo carboxamide, an osmium-containing catalyst, a chiral ligand for enantiomerically directing said asymmetric addition, and a base.

Preferred N-halo carboxamides include N-fluoro-(C1- C15 (alkyl))-amide, N-chloro-(C1-Cl5 (alkyl))-amide, N-bromo- (Cl- C15 (alkyl))-amide, N-iodo- (Cl-Cls (alkyl))-amide, N-fluoro- (aryl)- amide, N-chloro- (aryl)-amide, N-bromo- (aryl)-amide, N-iodo- (aryl)-amide, N-fluoro-2-chloro-(C1-Cls (alkyl)) amide, N-fluoro-2- bromo- (Cl-Cl5 (alkyl)) amide, N-fluoro-2-iodo- (Cl-Cl5 (alkyl)) amide, N-bromo-2-chloro- (Cl-Cl5 (alkyl)) amide, N-bromo-2-bromo-(C1- Ci5 (alkyl)) amide, N-bromo-2-iodo- (Cl-Cl5 (alkyl)) amide, N-iodo-2- chloro- (Ci-Cis (alkyl)) amide, N-iodo-2-bromo- (Cl-Cl5 (alkyl)) amide, N-iodo-2-iodo-(C1-Cl5 (alkyl)) amide, N-fluoro-alkoxybenzamide, N- chloro-alkoxybenzamide, N-bromo-alkoxybenzamide, N-iodo- alkoxybenzamide, N-fluoro-2-alkoxyacetamide, N-chloro-2- alkoxyacetamide, N-bromo-2-alkoxyacetamide, and N-iodo-2- alkoxyacetamide. More particularly, preferred N-halo carboxamides are N-bromoacetamide, N-chloroacetamide, N- bromobenzamide, N-chlorobenzamide, N-fluoro-2-chloro-acetamide, N-fluoro-2-bromo-acetamide, N-fluoro-2-iodo-acetamide, N-bromo- 2-chloro-acetamide, N-bromo-2-bromo-acetamide, N-bromo-2-iodo- acetamide, N-iodo-2-chloro-acetamide, N-iodo-2-bromo-acetamide, N-iodo-2-iodo-acetamide, N-chloro-p-methoxy benzamide, N-chloro- 2-methoxy acetamide, N-bromo-p-methoxy benzamide, N-bromo-2-

methoxy acetamide, N-iodo-p-methoxy benzamide, and N-iodo-2- methoxy acetamide. The preferred concentration of the N-halo carboxamides is within a range of 0.50 to 10 equivalents. In a preferred mode, the N-halo carboxamide is present in near stoichiometric amounts.

Preferred osmium containing catalysts include potassium osmate dihydrate, osmium tetroxide, osmium (IV) oxide, osmium (IV) oxide dihydrate, osmium (III) chloride, and osmium hexachlorooxmate (IV). In a preferred mode, the catalytic concentration of the osmium is within a range of 0.50-20 mole%.

Preferred chiral ligands include p-phenylbenzoyl dihydroquinidine; acetyl dihydroquinine; dimethylcarbamoyl dihydroquinine; benzoyl dihydroquinine; 4-methoxybenzoyl dihydroquinine; 4-chlorobenzoyl dihydroquinine; 2-chlorobenzoyl dihydroquinine; 4-nitrobenzoyl dihydroquinine; 3-chlorobenzoyl dihydroquinine; 2-methoxybenzoyl dihydroquinine; 3- methoxybenzoyl dihydroquinine; 2-naphthoyl dihydroquinine; cyclohexanoyl dihydraquinine; p-phenylbenzoyl dihydroquinine; methoxydihydroquinidine; acetyl dihydroquinidine; dimethylcarbamoyl dihydroquinidine; benzoyl dihydroquinidine; 4- methoxybenzoyl dihydroquinidine; 4-chlorobenzoyl dihydroquinidine: 2-chlorobenzoyl dihydroquinidine; 4- nitrobenzoyl dihydroquinidine; 3-chlorobenzoyl dihydroquinidine; 2-methoxybenzoyl dihydroquinidine; 3-methoxybenzoyl dihydroquinidine; 2-naphthoyl dihydroquinidine; cyclohexanoyl dihydroquinidine; and ligands represented by the following structures: wherein R1 is radical selected from a group consisting of a group represented by one of the following structures:

In a preferred mode, the chiral ligand is present and soluble within the reaction solution at a catalytic concentration within a range of substantially 0.50 mole % to 10 mole %. In a preferred example, the catalytic concentration of the osmium is within a range of 0.50-20 mole% and the chiral ligand has a catalytic concentration of approximately 5 mole%.

Preferred bases include LiOH, NaOH, KOH, NH40H, Na2CO3, K2CO3, CaCO3, and BaC03.

In the preferred aqueous/organic solvent system, preferred organic compounds include methanol, ethanol, n-butanol, n- pentanol, n-propanol, 2-propanol, 2-butanol, tert-butanol, ethylene glycol; acetonitrile, propionitrile; tetrahydrofuran, diethyl ether, tert-butyl methyl ether, dimethoxyethane, 1,4- dioxane; dimethyl formamide, acetone, benzene, toluene, chloroform, and methylene chloride. The aqueous/organic solvent system may be either a homogenous or heterogeneous mixture. The aqueous component is water wherein the aqueous component of the solvent has a range between 10% and 90% on a volume basis. In a preferred mode, the aqueous and organic components of the solvent are each approximately 50% on a volume basis.

Exemplary olefins include cis stilbene, trans stilbene, ethyl acrylate, styrene and C1-C6 (alkyl)-cinnamate ester. A preferred reaction temperature is within a range between-5.0 and 5.0 °C.

Brief Description of Figures Figure 1 illustrates preffered cinchona alkaloid ligand derivatives.

Figure 2 shows a generic scheme indicating that a variety of olefins, as defined below, can be reacted smoothly with 1.1 equivalents of the oxidant/nitrogen donor to give the vicinal aminoalcohols in good yield and with high enantiomeric excess.

Figure 3 illustrates Table 1: Acetamide based asymmetric aminohydroxylation of various olefins wherein the indicated notations are defined as follows [a] Conditions : see experimental procedure. [b] From reaction catalyzed by (DHQ) 2- PHAL. [c] Determined by 1H-NMR. [d] Determined by chiral HPLC (entry 1-4) or GC (entry 5). [e] Negative values indicate the formation of the opposite enantiomer (product from reaction catalyzed by (DHQD) 2-PHAL). [f] Isolated yields of the pure products 1-5 after chromatography on silica gel. [g] tButanol/water 1: 1 was employed. [h] KOH as base and 1- Propanol/water 1: 1 as the solvent was used. [i] CH3CN/water 1 : 1 was used as a solvent.

Figure 4 illustrates a large scale synthesis based on acetamide as the oxidant for the amidohydroxylation.

Figure 5 illustrates regioselectivity preferences using AQN ligands and N-bromobenzamide.

Figure 6 illustrates the reversal of the regioselectivity with styrene derivatives wherein the indicated notations are defined as follows: [a] Conditions: see experimental procedure.

[b] The DHQD-derivative was used. [c] Determined by 1H-NMR. [d] Determined by chiral HPLC. [el Yields refer to the mixtures of isomers after chromatography on silica gel. [f] Not determined.

Figure 7 illustrates the effects on regio and stereochemistry using acetamide base amidohydroxylation using (DHQD) 2PHAL (entries 1-3) or the (DHQD) 2AQN (entries 4-6) ligands on the respective olefins (substrate olefins not shown).

Figure 8 illustrates heterogenous conditions with a phase transfer catalyst wherein the stereochemistry is reversed using the indicated AQN ligand.

Detailed Description: The a-aminoalcohol moiety is a widespread structural motif in natural products and synthetic drugs. Its generation in an enantioselective manner via metal catalysis represents a major challenge for synthetic organic chemists. The catalytic asymmetric aminohydroxylation (AA) represents an even more elegant approach. Alkenes are therein converted into protected (3-amino alcohols in a single step via a syn-cycloaddition catalyzed by osmium salts and chiral quinuclidine-type ligands derived from cinchona alkaloids. Three major aspects are to be addressed in this reaction: namely, chemo-, regio-, and enantioselection. The most important variable is the ultimate oxidant/nitrogen source for the generation of the active osmium (VIII) imido species responsible for aminohydroxylation.

As described in the background, several different reactants have been introduced in the literature thus far: the akali metal salts of (1) N-halosulfonamides (for the synthesis of P-hydroxysulfonamides), and the 2) N-halocarbamates (for the synthesis of a-hydroxycarbamates). What is needed, however, is a direct method to synthesize P-hydroxycarboxamides, the amide form of a-aminoalcohols, which are important chemical moieties found in such important targets as the anticancer taxol.

The invention is therefore directed to a novel method for the regio-selective and enantio-selective conversion of olefins to R-hydroxyamides using an N-halo carboxamide as the oxidizing

agent in the presence of an osmium catalyst and a chiral ligand.

Alkali metal salts of N-chloro carboxamides have been well- known for their proclivity to undergo Hofmann rearrangement (Hofmann et al. Ber. 1881,14,2725; Wallis et al. Org. React.

1967,3,267-306). For the invention, however, were pleased to find that this undesirable competing reaction could be completely supressed by operating at 4 °C while also using the more stable N-bromo derivative (the acetamide derivative, N- bromoacetamide is commercially available from Lancaster, but it should be recrystallized in a mixture of chloroform/hexane 1: 1 before use. We recommend preparation via the published procedure by Oliveto et al. Org. Synth., Coll. Vol. IV 104-105. The purity of this oxidant was checked via acid-base titration using the procedure of Bachand et al. J. Org. Chem. 1974,39,3136-3138 and Virgil et al. (N-Bromoacetamide) in Encyclopedia of Reagents for Organic Synthesis, Vol. 1 (Ed.: L. A. Paquette), John Wiley & Sons, 1995, p. 691). By suppressing the competing Hofman rearrangement, therefore, we were able to obtain the direct conversion of olefins to R-hydroxyamides using the N-halo carboxamide as oxidizing agent in the presence of an osmium catalyst.

In particular, in the presence of 4 mol-% of potassium osmate dihydrate and 5 mol-% of the second generation alkaloid ligands (Kolb et al. Chem. Rev. 1994,94,2483-2547) olefins reacted smoothly with 1.1 equivalents of the oxidant/nitrogen donor to give the vicinal aminoalcohols in good yield and with high enantiomeric excess as generally shown in Figure 2.

For each olefin class given in Figure 3, the reaction parameters were optimized in terms of ligand, solvent, base and base/oxidant ratio; since the base/oxidant ratio should not exceed 1: 1, the amount of hydroxide resulting from the initial K2OsO2 (OH) 4-oxidation to OsVIII-, releasing two equivalents of base, was taken into account. Both product enantiomers can be obtained by using either'pseudoenantiomers'of the alkaloid

ligands (Kolb et al. Chem. Rev. 1994,94,2483-2547). A major advantage of this system is that only a stoichiometric amount of oxidant is needed, rather than an excess of oxidant, which greatly simplifies isolation and purification of the product. In fact, use of excess of oxidant does not improve the reaction in any way.

As in the sulfonamide and carbamate based transformations, cinnamates are among the best substrates (Figure 3, entries 1-3; the use of isopropyl cinnamates instead of the methyl esters is preferable in terms of greater stability towards hydrolysis and enhanced regioselectivity under the reaction conditions. As in the case of compounds 1 and 2, which are quite soluble in the reaction medium, the use of more than 50 % (v/v) water may result in slightly increased regioselectivities). While cis- stilbene gave mostly diol, trans-stilbene afforded the threo- aminoalcohol 4 in 50 % yield (together with 10 % diol) and high enantiomeric excess (Figure 3, entry 4). Ethyl acrylate gave rise to the isoserine derivative as the major regioisomer detected by lH-NMR (isoserine/serine >20 : 1, Figure 3, entry 5).

It is important to note that, with styrenes as substrates (Figure 5), the regioselectivity is highly dependent on the choice of solvent and ligand. For example, alcoholic solvents resulted in a slight preference (8/7 = 1.1-2.5: 1) for introduction of the nitrogen substituent at the benzylic position (regioisomer 8, Figure 6), whereas acetonitrile significantly favored the other regioisomer (7) (7/8 = 2-13: 1).

In most cases, the benzylic amides were formed with higher asymmetric induction (8,85-96 % ee) than the benzylic alcohols (7,62-94 % ee), but poor regioselectivity was observed when phthalazine ligands were used. However, our recently introduced anthraquinone (AQN) ligands (Becker et al. Angew. Chem. 1996, 108,447-449; Angew. Chem. Int. Ed. Engl. 1996,35,448-451).

Both (DHQ) 2-AQN and (DHQD) 2-AQN were more effective, leading to good regioselectivities with this olefin class (Figure 6, entry 3,6,9). The AA protocols now enable selective synthesis of

either of the regioisomeric a-aminoalcohols derived from styryl olefins. The phenyl glycinols (8) arise using the carbamate AA/phthalazine ligand combination and the adrenaline-type regioisomers (7) are preferentially derived using the acetamide AA/anthraquinone ligand combination. Experiments with other substrates indicate that a reversal of regiochemistry occurs when anthraquinone derived ligands instead of phthalazines are used.

Irrespective of the regiochemical outcome, the facial selectivity of these acetamide AA reactions agrees with the face selectivity rule for the AD (so far, only a few exceptions to our mnemonic device for the AD have been reported: Hale et al.

Tetrahedron Lett. 1994,35,425-428; Krysan et al., Tetrahedron Lett. 1996,37,1375-1376; Boger et al. J. Am. Chem. Soc. 1996, 118, 2301-2302; Salvadori et al. J. Org. Chem. 1996,61,4190- 4191; Vanhessche et al. J. Org. Chem. 1996,61,7987-7979).

Thus, dihydroquinine derived ligands (DHQ class) give rise to oxidation from the a-face of the olefin, whereas the diastereomeric dihydroquinidine (DHQD) ligands effect attack on the 5-face of the double bond.

With this efficient new process in hand, we focused on its application for a large scale synthesis of (2R, 3S)-3- phenylisoserine (the yield might be further increased (5-10 %) by optimizing the isolation procedure ; we believe that this new amide-AA-based protocol is superior to our earlier AA-or AD- based approaches: Wang et al. J. Org. Chem. 1994,59,5104- 5105) a precursor for the side chains of the anti-cancer drugs Taxol and Taxotere@.

A 630-fold scale-up of our standard one mmol recipe was undertaken with a somewhat lower catalyst loading (1.5% osmate salt and 1% (DHQ) 2PHAL), which did not affect the outcome of the catalysis. The fact that less ligand than osmium can be used, demonstrates once again the great advantages of a ligand accelerated catalysis (LAC); for a review see D. J. Berrisford, C. Bolm, K. B. Sharpless, Angew. Chem. 1995, 107,1159-1171;

Angew. Chem. Int. Ed. Engl. 1995,34,1059-1070. The AA product was isolated by crystallization of the crude reaction mixture from ethyl acetate/hexane, and a second crop from diethyl ether.

Subsequent hydrolysis furnished the enantiomerically pure aminoalcohol as the hydrochloride salt 89) in 68 % yield over two steps (Figure 4).

The amidohydroxylation employing the use of N- bromoacetamide as the nitrogen source with styrenes as the olefin shows that the regiochemistry of aminoalcohol formation is reversed from that delivered by the carbamate version of the AA. An efficient large scale synthesis of enantiomerically pure 3-phenylisoserine highlights the prodical potential of this latest advance in the osmium-catalyzed asymmetric aminohydroxylation process.

Generally, alkali metal salts of N-chloro carboxamides are unstable and readily undergo Hofmann rearrangement. Thus the N- bromo derivatives are preferred oxidants for amidohydroxylation reactions run at 4 °C. The standard substrates are smoothly converted into protected amidoalcohols with N-bromo alkali (K+ or Li+) salts of acetamide. When working with styrene derivatives and alkaloid ligands that contain an anthraquinone spacer (AQN ligands), the regiochemistry is reversed compared to the carbamate procedure with phthalazine (PHAL) ligands, see Figure 7.

1.1 equivalents of the oxidant/nitrogen source suffice for complete conversion, and for isopropyl cinnamate as the olefin, the catalyst loading can be significantly lowered without affecting yield or enantioselecitvity.

Lowering the amount of catalyst and/or ligand does not influence yield, regioselectivity or ee of the amidohydroxylation (example: acetamide amidohydroxylation of isopropyl cinnamate on a large scale). This is in contrast to the carbamate procedures.

Only stoichiometric amounts of nitrogen sources are

necessary, in contrast to the carbamate recipes where 2-3 equivalents with respect to the olefinic substrate are needed for complete conversion and high ee/regioselectivity.

Electronic tuning of the nitrogen source by using amides of different pKa's is possible (and thus the reactivity can be varied), in contrast to the carbamate nitrogen sources where the acidity of the carbamate nitrogen protons stays more or less the same no matter which carbamate is used.

When using anthraquinone-type (AQN) ligands such as (DHQ) 2AQN or (DHQD) 2AQN and acetamide, the major product is formed in high ee bearing the nitrogen substituent in the terminal position (adrenaline type products). The carbamate recipe gives similar regioselectivities in those cases but the ee's are low.

Under biphasic conditions with benzamides as nitrogen sources, electron poor cinnamates (e. g. ethyl 3-nitrocinnamate) can also be converted to the products bearing the nitrogen substituent a to the ester group in high regioselectivity. The carbamate recipe gives low regioselectivities (typically 1: 1) in those cases.

Nitrogen Sources (Oxidant) The amide-based nitrogen oxidant source, the N-halo carboxamide, is generally used in near stoichiometric amounts (eg. 0.90-1.2 equivalents ; defined herein as near stoichiometric amounts) but can operate efficiently in the range of 0.50 equivalents (less for difficult purification conditions and inexpensive olefins) to 10 equivalents (for less reactive olefins) wherein the N-halo carboxamide can be commercially purchased or synthesized according to standard procedures well known in the art as disclosed vida supra. The addition of an alkali metal in situ, forms the alkali metal salt of the N-halo carboxamide. One can generate beforehand this salt, if desired by premixing the N-halo carboxamide in a basic solution containing the alkali metal (eg. LiOH, NaOH, etc.).

The genus of amide-based nitrogen sources which are preferred with the invention include the following N-halo carboxyamides: N-fluoro-(C1-Cl5 (alkyl))-amide, N-chloro- (C1- C15 (alkyl))-amide, N-bromo-(C1-Cl5 (alkyl))-amide, N-iodo-(C1- C15 (alkyl))-amide, N-fluoro- (aryl)-amide, N-chloro- (aryl)-amide, <BR> <BR> <BR> N-bromo-(aryl)-amide, N-iodo-(aryl)-amide, N-fluoro-2-chloro-(C1- Cis (alkyl)) amide, N-fluoro-2-bromo- (Cl-C, 5 (alkyl)) amide, N-fluoro- 2-iodo-(C1-Cl5 (alkyl)) amide, N-bromo-2-chloro- (C1-C15 (alkyl)) amide, N-bromo-2-bromo-(C1-Cl5 (alkyl)) amide, N-bromo-2-iodo-(C1- C15 (alkyl)) amide, N-iodo-2-chloro-(C1-Cl5 (alkyl)) amide, N-iodo-2- bromo-(C1-Cl5 (alkyl)) amide, N-iodo-2-iodo-(C1-Cl5 (alkyl)) amide, N- fluoro-alkoxybenzamide, N-chloro-alkoxybenzamide, N-bromo- alkoxybenzamide, N-iodo-alkoxybenzamide, N-fluoro-2- alkoxyacetamide, N-chloro-2-alkoxyacetamide, N-bromo-2- alkoxyacetamide, and N-iodo-2-alkoxyacetamide wherein aryl is defined as any substituted or unsubstituted aromatic containing amides and alkyl is defined as any Cl-C,. 5 aliphatic containing amides.

In particular, the following species of N-halo carboxyamides work optimally well with the procedure: N- bromoacetamide, N-chloroacetamide, N-bromobenzamide, N- chlorobenzamide, N-fluoro-2-chloro-acetamide, N-fluoro-2-bromo- acetamide, N-fluoro-2-iodo-acetamide, N-bromo-2-chloro- acetamide, N-bromo-2-bromo-acetamide, N-bromo-2-iodo-acetamide, N-iodo-2-chloro-acetamide, N-iodo-2-bromo-acetamide, N-iodo-2- iodo-acetamide, N-chloro-p-methoxy benzamide, N-chloro-2-methoxy acetamide, N-bromo-p-methoxy benzamide, N-bromo-2-methoxy acetamide, N-iodo-p-methoxy benzamide, and N-iodo-2-methoxy acetamide.

The following table contains additional data on N-bromo-N- lithio amide nitrogen sources other than from acetamide (isopropyl cinnamate as olefin, standard conditions as described in the paper attached to the claim, solvent tbutanol/water 1: 1, no ee's determined. A= major regioisomer, B= minor regioisomer, C= diol byproduct):

Amide Convrsn/% A/B (A+B)/C n-butyramide 99 25: 1 13: 1 benzamide 94 1.8: 1 1: 1.6 p-methoxy benzamide 75 2: 1 3: 1 2-methoxy acetamide 55 15: 1 3: 1 Olefin Classes The asymmetric amidohydroxylation (AA) works well with three olefin classes: 1) monosubstituted; 2) cis-disubstituted, and 3) trans-disubstituted olefins. The 1, 1 disubstituted and trisubstituted types of olefins give only racemic or low ee's while the tetrasusbstituted class, does not provide any signs of turnover; a/ß unsaturated esters, a/ß unsaturated amides, aromatic olefins and heteroaromatic olefins work particularly well with the invention.

Regioselectivity High regioselectivity is one of the more useful features of the amidohydroxylation. The chemistry exhibits a strong preference for nitrogen attachment to the olefinic carbon bearing an aromatic substituent or, in the case of olefins conjugated with a strong electron withdrawing group (EWG), the nitrogen is strongly directed to the olefinic carbon distal to the EWG.

The alkaloid ligand is responsible for high regioselectivity. When the ligand is omitted, there is little preference for either regioisomer. Beyond probable contributions from"binding pocket"effects, the strong regioselection phenomenon requires the operation of powerful electronic determinants.

Solvent Variations : Preferred solvents include acetonitrile, n-propanol, tert- butanol and suitable solvents include methanol, ethanol, n- butanol, n-pentanol, n-propanol, 2-propanol, 2-butanol, tert- butanol, ethylene glycol; nitriles: acetonitrile, propionitrile;

ethers: tetrahydrofurane, diethyl ether, tert. butyl methyl ether, dimethoxyethane, 1,4-dioxane ; miscellaneous: dimethyl formamide, acetone, benzene, toluene, chloroform, methylene chloride.

Solvent Concentration Varations : In its present form the process starts to give lower selectivities for some substrates when the concentration of olefin (which of course prescribes the standard concentration of all the other species) gets much above 0.1 molar.

Ligand Variations : Preferred ligands are shown in Figure 1 and are commercially available or synthesized by procedures well known in the art.

Additional preferred ligands, which can be equally used with the invention with the specified concentrations as described herein, are as follows: p-phenylbenzoyl dihydroquinidine; acetyl dihydroquinine; dimethylcarbamoyl dihydroquinine ; benzoyl dihydroquinine; 4-methoxybenzoyl dihydroquinine; 4- chlorobenzoyl dihydroquinine; 2-chlorobenzoyl dihydroquinine ; 4- nitrobenzoyl dihydroquinine; 3-chlorobenzoyl dihydroquinine ; 2- methoxybenzoyl dihydroquinine ; 3-methoxybenzoyl dihydroquinine ; 2-naphthoyl dihydroquinine; cyclohexanoyl dihydraquinine; p- phenylbenzoyl dihydroquinine; methoxydihydroquinidine; acetyl dihydroquinidine ; dimethylcarbamoyl dihydroquinidine; benzoyl dihydroquinidine; 4-methoxybenzoyl dihydroquinidine; 4- chlorobenzoyl dihydroquinidine: 2-chlorobenzoyl dihydroquinidine ; 4-nitrobenzoyl dihydroquinidine ; 3- chlorobenzoyl dihydroquinidine ; 2-methoxybenzoyl dihydroquinidine; 3-methoxybenzoyl dihydroquinidine; 2-naphthoyl dihydroquinidine ; and cyclohexanoyl dihydroquinidine.

The ligand can range from ca. 0.5 to 10 mol % (less is appropriate for lower temperatures; eg. 0.5 % might be enough at O °C and 10 % would probably be needed to keep the % ee at

reasonable levels if the temperature reaches 35 or 40 °C. In practice, the molarity of the ligand matters and the amount of ligand needed to realize the"ceiling ee"scales directly with the reaction concentration (ie if twice the volume of solvent is used, then the mol% of ligand added must also double to keep its molarity constant and correspondingly if the reaction is run twice as concentrated as usual (see general recipe below) then half of the usual mol % ligand gives the needed ligand molarity). Because the crucial binding of the ligand is an extremely rapid bimolecular process, the equilibrium constant is highly sensitive to temperature which is why the molarity of ligand needed, increases rapidly with temperature.

Osmium Variations: The osmium containing catalyst can be selected from any one of the various commercially available osmium sources including, but not restricted to potassium osmate dihydrate, osmium tetroxide, osmium (IV) oxide, osmium (IV) oxide dihydrate, osmium (III) chloride, and osmium hexachlorooxmate (IV). The amount of Os catalyst can range from 0.5% (probably even less in the very best cases, and in any case the number will drop as the process if further improved) to 10 or even 20%. The general procedure conditions uses 4% to have fast reaction times, but 2% is good for most cases. The high loadings of 20%, for example, is needed to achieve reasonable rates with very poor substrates (this conclusion follows from the extensive experience by us and others with the AD, where in desparate situations 20 or more % Os catalyst is needed.

As a typical example of the ranges that are preferred, the following table contains data on the amount of osmium and ligand one can use for the isopropyl cinnamate AA with N-bromo acetamide.

Mol-% Mol-% Conversion A/B (A+B)/C ee/% K, [OsO, (OH) 4] (DHQ) 2PHAL/% 4 5 99 25: 1 25: 1 99 1.5 1 99 25: 1 25: 1 99 0.9 0.6 99 20: 1 11: 1 99 0.4 0.5 89 15: 1 2.4: 1 n. d.

0.2 0.25 36 2: 1 1.4: 1 n. d.

Temperature Variations: For most cases, the hydroxyamide AA process is run between-5.0 and +5.0 degrees Celcius. There may be cases where up to 35 to 40 degrees may be advantageous depending on substrate.

Heterogeneous Condition Variation: For most cases, the hydroxyamide AA process is run in homogeneous conditions, however, applying heterogeneous phase transfer conditions as illustrated in Figure 8, enhanced chemoselectivity was found for some olefins (see synthesis for: methyl (2S, 3R)-2- (benzamido)-3-hydroxy-3- (4-fluoro-3- nitrophenyl) propanoate 16, vida infra). In particular, we found that the AQN ligands furnished a complete reversal of the regioselectivity with cinnamates as olefins.

Phase Transfer Catalysts: For heterogeneous conditions, a commercially available phase transfer catalyst such as tetra-n-butylammonium hydroxide is preferred, however other standard phase transfer catalysts (eg. tetra-alkyl ammonium or phosphonium salts) work well equally with the invention when heterogeneous conditions are required.

Base Variations : For most cases, the hydroxyamide AA process is run in LiOH or KOH, however other alkali earth bases such as NaOH and CsOH are acceptable using the conditions and concentrations as indicated vida infra. Preferred bases are LiOH, NaOH, KOH, NH40H, alkyl or arylammonium hydroxides, and alkali or earth alkali salts of carbonates and bicarbonates (eg. Na2CO3, K2CO3, CaCO3, BaC03 et.)

General Conditions: While a preferred form of the invention has been shown in the description, drawings and synthetic protocals, variations in the preferred form will be apparent to those skilled in the art using the disclosed variety of substrates and reagents with the range of conditions indicated herein. The invention should not be construed as limited to the specific form shown and described, but instead is as set forth in the following claims.

Synthetic Protocals NMR spectra were recorded on Bruker AMX-500, AM-300, or AM- 250 instruments. The following abbreviations were used to explain the multiplicities : s, singlet; d, doublet; t, triplet ; q, quartet ; m, multiplet; apt, apparent; b, broad; obs, obscured. IR spectra were recorded on Nicolet 205, Perkin Elmer 1600 or Galaxy 2020 series FT-IR spectrophotometers. Optical rotations were recorded using a Perkin Elmer 241 polarimeter.

High-resolution mass spectra (HRMS) were recorded on a VG ZAB- ZSE mass spectrometer under Fast Atom Bombardment (FAB) conditions, at the Scripps Research Institute.

All reactions were monitored by color change, HPLC, GC or thin-layer chromatography carried out on 0.25 mm Whatman silica gel plates (K6F-60 A) using W light, p-anisaldehyde, or 7% ethanol phosphomolybdic acid and heat as developing agent. E.

Merck silica gel (60, particle size 0.040-0.063 mm) was used for flash column chromatography. Tetrahydrofuran (THF) and ethyl ether were distilled from sodium-benzophenone and methylene chloride, benzene and toluene were distilled from calcium hydride. All reagents were obtained from Aldrich Chemical Co.

Inc. unless otherwise noted. Solvents used for workup, chromatography, and recrystallizations were reagent grade from Fisher Scientific and were used as received.

General Procedure using homogenous condtions: Carry out all procedures in a well-ventilated laboratory, an wear disposable vinyl or latex gloves.

Equipment * Stainless steel reaction vessel (40 L) * Mechanical stirrer w. stainless steel stirring shaft (1 m) and four blade rotor (6 cm) * Immersion cooler w. cold finger probe * Low temperature thermometer

* Erlenmeyer flask (2 L) * Teflon-coated magnetic stirbar (4 cm) * Erlenmeyer flask (1 L) * PVC tubing (1 m x 10 mm) * 2 Erlenmeyer flasks (4 L) * Glass filter (15 cm) with folded filter paper * Single-necked round bottomed flask (2 L) * Sintered glass filter funnel (2 L, medium porosity) * Recovery flask (4 L) * High vacuum pump * Oil bath (0.5 L) * Single-necked round bottomed flask (1 L) * Reflux condensor * Sintered glass filter funnel (1 L, medium porosity) Materials Hydroquinine 1,4-phthalazinediyl diether, (DHQ) 2PHAL, 9.738 g, 12.5 mmol * t-Butanol, 7.5 L * Lithium hydroxide monohydrate 56.12 g, 1.337 mol * Water, 11.25 L *'Potassium osmate dihydrate' (K2 [OsO2 (OH) 4]), 6.909 g, 18.75 mmol * Isopropyl cinnamate 237.8 g (96 % pure by GC), c.

1.250 mol * N-Bromoacetamide 195.6 g (97 % pure by titration) 1.375 mol * Sodium sulfite, 150 g * Sodium chloride, 1.5 kg * Ethyl acetate, 22.55 L

* Anhydrous sodium sulfate, 1.3 kg * Chloroform, 0.4 L * Silica gel for filtration * Hexane, 0.3 L * Diethyl ether, 0.6 L * t-Butyl methyl ether, 0.1 L * 10 % Hydrochloric acid, 1.5 L Synthetic steps: 1. In a stainless steel reaction vessel (40 L) equipped with a mechanical stirrer, a thermometer and an immersion cooler with a temperature controller set to 4 °C, dissolve hydroquinine 1,4- phthalazinediyl diether, (DHQ) 2PHAL, (9.738 g, 12.5 mmol), in t- butanol (7 L) with stirring (150 rpm).

2. Add water (10 L) and start cooling.

3. In an Erlenmeyer flask (2 L), dissolve lithium hydroxide monohydrate (56.12 g, 1.337 mmol) in water (1.25 L) with stirring. Add KZ [Os02 (OH) 9] (6.909 g, 18.75 mmol) and continue stirring until you obtain a pink solution.

4. Pour the osmate solution into the reaction vessel.

5. Add isopropyl cinnamate (273.8 g, c. 1.25 mol). Rinse the residual olefin with t-butanol (0.5 L). Add N-bromoacetamide (195.6 g, c. 1.375 mol) quickly in a single portion at 4 °C.

Continue stirring of the green reaction mixture at this temperature for 3.5 hours.

6. At this point, a thin layer chromatogram shows the absence of starting material (Rf = 0.85) and the mixture is red. Add sodium sulfite (150 g), remove the immersion cooler and stir for 12 hours.

7. Add 1.5 kg NaCl and stir for 15 min. Stop stirring and allow phase separation. Siphon the upper organic layer through a PVC tubing (1 m, 10 mm diameter) into two Erlenmeyer flasks (4 L)

which contain anhydrous sodium sulfate (300 g each).

8. Filter the dried solution into a single-necked, round- bottomed flask (2 L) and concentrate the mixture on a rotary evaporator set to 60 °C.

9. Extract the remaining aqueous layer with ethyl acetate (1 x 6 L and 3 x 4 L). Decant the organic extract into the Erlenmeyer flasks.

10. Concentrate the organic extracts until you obtain a dark, red oil. Add a mixture of warm ethyl acetate-chloroform (1.6 L, 3: 1 v/v). Filter through a sintered glass filter funnel containing a 5-cm layer of SiO2, covered with a 1-cm layer of Na2SO4. Wash with ethyl acetate (3 L). Evaporate the filtrate in a round-bottomed single-necked flask (2 L) to a volume of c. 0.8 L, add hexane (250 mL) and a magnetic stirbar. Cool with stirring to 4 °C with an ice-water bath.

11. After 1 h, filter the suspension through a sintered glass filter funnel (1 L) and wash with ethyl acetate/hexane (250 mL, 3: 2 v/v). Dry the filter cake under high vacuum for 4 h to yield 233 g of the AA product.

12. Concentrate the mother liquor and triturate with diethyl ether/t-butyl methyl ether (200 mL,. 1 : 1 v/v). Filtration after 2 h and drying yields another 27 g of product. The combined crystalline, white solid, 260 g is of 99 % ee as determined by HPLC on Chiralcel OD-H, Daicel, i-PrOH/hexane 40: 60 v/v, 0.5 mL/min, 254 nm; retention times: 8.2 min (2S, 3R), 12.7 min (2R,3S).

13. Heat the combined material in a 2 L round-bottomed flask immersed in an oil bath and equipped with a reflux condenser and a magnetic stirbar in 10 % HCl (1.5 L, 4 h, 100 °C).

14. Concentrate the mixture to c. 20 % of the initial volume on a rotary evaporator set to 60 °C.

15. Filter the product through a sintered glass funnel (1 L), wash with cold diethyl ether (0.5 L), and dry the white crystals of hydrochloride (2R, 3S) for 12 h under high vacuum at 40 °C ;

m. p. 224-6 °C, [a] o (ZS, D =-14. 8 (c = 0.55,6 M HC1), 209.5 g, 77 % (based on 96 % GC-pure cinnamate).

2nd example of homogeneous procedure: (example using isopropyl cinnamate) In 3 mL of an aqueous solution of LiOH. H20 (42.8 mg, 1.02 mmol), K2Os02 (OH) 4 (14.7 mg, 0.04 mmol, 4 mol %) was dissolved with stirring. After addition of tBuOH (0.17 Molar; 6 mL), (DHQ) 2-PHAL (39 mg, 0.05 mmol, 5 mol %) was added and the mixture was stirred for ten minutes to give a clear solution. Water (0.17 Molar; 6 mL) was added subsequently, and the mixture was immersed in a cooling bath set to 4 °C. After addition of isopropyl cinnamate (190 mg, 1 mmol), N-bromoacetamide (151.8 mg, 1.1 mmol; N-bromoacetamide is commercially available (e. g. from Lancaster), but. it should be recrystallized (CHCl3/hexane 1: 1) before use. We recommend preparation via a published procedure: E. P. Oliveto, C. Gerold, Org. Synth., Coll. Vol. IV 104-105. The purity of this oxidant was checked via acid-base titration: Bachand et al. J. Org. Chem. 1974,39,3136-3138. (b) S. C. Virgil (N-Bromoacetamide) in Encyclopedia of Reagents for Organic Synthesis, Vol. 1 (Ed.: L. A. Paquette), John Wiley & Sons, 1995, p. 691) was added in one portion (which resulted in an immediate color change to green) and the mixture was vigorously stirred at the same temperature. The reaction was monitored by tlc and pH control (full conversion is indicated when the mixture reaches pH 7). After 20 h, the reaction mixture was treated with Na2SO3 (0.5 g) and, after stirring at room temperature for 30 min, ethyl acetate (5 mL) was added. The organic layer was separated, and the water layer was extracted with ethyl acetate (3 x 10 mL). The combined organic extracts were washed with brine (5 mL), and dried over MgSO4. After evaporation of the solvent, the crude product was purified by chromatography on silica gel (hexanes/ethyl acetate 1: 1) to give 215 mg (81 % yield, 99 % ee) of isopropyl (2R, 3S)-3-

(acetylamino)-2-hydroxy-3-phenylpropanoate (1).

Homogeneous procedure with modifications of the above procedure: see Figure 3. For styryl olefins (Figure 6), 1.0 eq. KOH and a 1: 1 solvent/water ratio (15 mL) was employed. For reactions with anthraquinone based ligands, (DHQD) 2-AQN (42.8 mg, 0.05 mmol, 5 mol %) instead of the PHAL ligand were used.

Homogeneous procedure for large scale synthesis using the example of (2R, 3S)-3-phenylisoserine hydrochloride (9) : The above procedure was followed with 120.0 g isopropyl cinnamate (0.631 mol, not corrected for purity, 96 %, Lancaster), tBuOH (3.79 L), water (5.68 L), 99.71 g (0.694 mol, 96 % pure upon titration) N-bromoacetamide, 28.32 g (1.07 eq., 0.675 mol) LiOH. H2O, [6] 4.914 g (DHQ) 2-PHAL (6.31 mmol, 1 mol %) and 3.486 g (9.08 mmol, 1.5 mol %) K2OsO2 (OH) 4. Reaction time: 4 h. To separate the ligand after completion of the reaction, the crude product was taken up in 500 mL ethyl acetate and passed through a 4.5-inch sintered glass funnel covered with a one inch layer of silica gel. Product purification was accomplished by recrystallization from ethyl acetate/hexane 1: 2 (4 mL/g). A second crop of material was obtained by trituration of the previously evaporated mother liquor with 100 mL Et2O and subsequent filtration. Yield 119.5 g (71 % yield, 99 % ee).

Hydrolysis of this material (10 % HC1, reflux, 4 h, followed by concentration and filtration) provided 92.5 g (68 % over two steps) enantiomerically pure (2R, 3S) 9 (m. p. 224-226 °C, [a] \o (25, D) =-14. 9 (c = 0.55 in 6N HC1, lit.-14.8); correct elemental analysis.

Heterogeneous Example: Synthesis of methyl (2S, 3R)-2- (benzamido)-3-hydroxy-3- (4-fluoro-3-nitrophenyl) propanoate (16) wherein ligand used is (DHQD) 2-PHAL) as shown in Figure 8.

In 2 ml of an aqueous solution of phase transfer reagent tetra- n-butylammonium hydroxide (170.3 mg of a 40 % aqueous solution, 0.2625 mmol; commercially available-see above for other available phase transfer catalysts), K2OsO2OH4 (3.68 mg, 0.01 mmol, 4 mol-%) was dissolved with stirring. After addition of chlorobenzene (2 ml), (DHQD) 2-AQN (10.72 mg, O. OX mmol, 5 mol-%) was added with stirring, and the mixture was immersed in a cooling bath set-to 4 °C. After addition of methyl 4-fluoro-3- nitrocinnamate (56.3 mg, 0.25 mmol), N-bromobenzamide (55.0 mg, 0.275 mmol) was added in one portion, (color change to green) and the mixture was vigorously stirred at the same temperature.

After 16 h, the reaction mixture was treated with 0.2 g Na2SO3 and after stirring at room temperature for 30 min, ethyl acetate (5 ml) was added. The organic layer was separated, and the water layer was extracted with ethyl acetate (3 x 10 rnL). The combined organic extracts were washed with brine (5 mL), and dried over MgS04. After evaporation of the solvent, the crude product was purified by chromatography on silica gel (hexanes/ethyl acetate 1: 2) to give 46.4 mg (53 %) methyl (2S,3R)-N- (benzamido)-2amino-3-hydro-3- (4-fluoro-3-nitro- phenyl) propanoate (16).

Experimental Data for selected compounds: sopropyl (2R, 3S)-3- (acetylamino)-2-hydroxy-3-phenylpropanoate (1) M. p. 112-113 °C ; [a] o (25, D) = +20. 0 (c = 1.16 in 95 % EtOH) ; IH NMR

(400 MHz, CDCl3) 6 = 1.21 (d, J = 6.3 Hz, 3H), 1.29 (d, J = 6.3 Hz, 3H), 1.97 (s, 3H), 4.46 (d, J = 2.3 Hz, 1H), 5.09 (septet, J = 6. 3 Hz, 1 H), 5.54 (dd, J = 9.3,2.3 Hz, 1H), 6.46 (d, 9.3 Hz, 1H), 7.26-7.40 (m, 5 H) ; 13C NMR (100 MHz, CDCl3) 6 = 21.4,21.6, 23.0,54.4 (2C), 70.6,73.3,126.8,127.7,128.5,138.8,169.5, 172.3; FT-IR (neat): o (-, #) = 3344, 1732,1656,1103 cm-1 ; HR-MS (NBA/NaI): m/z : exp. 288.1216 [M+Na] +, calc. 288.1212 for C1gH,9NO, Na ; HPLC: Chiralcel OD-H, 40 % iPrOH/hexane, 0.5 mL min-', 254 nm, 8.2 min (2S, 3R), 12.7 min (2R, 3S). The absolute configuration was established, after hydrolysis, by comparison to the known (2R, 3S) aminoalcohol hydrochloride,"" [a] 0 (25, 14.9 (c = 0.55 in 6N HC1).

Isopropyl (2S, 3R)-3- (acetylamino)-2-hydroxy-3- (2- methoxyphenyl) propanoate (ent-2) (ligand = (DHQD) 2-PHAL) M. p. 161-162 °C ; [a] O (2SED) =-16. 9 (c = 1.41 in 95 % EtOH) ; 1H NMR (400 MHz, CDCl3) 6 = 1.12 (d, J = 6. 3 Hz, 3H), 1.24 (d, J = 6. 3 Hz, 3H), 1.98 (s, 3H), 3.43 (d, J = 5.4 Hz, 1H), 4.53 (dd, J = 5.4,4.1 Hz, 1H), 5.00 (septet ; 6. 3 Hz, 1H), 5.70 (dd, J = 8.8, 4.1 Hz, 1H), 6.87 (br d, J = 8.8 Hz, 1H), 6.84-6.95 (m, 2H), 7.17-7.28 (m, 2H) ;"C NMR (100 MHz, CDC13) 6 = 21.3,21.6,23.2, 52.0,52.1,55.4,69.9,72.5,110.6,120.6,126.0,128.1,129.0, 156.6,169.6,172.6; FT-IR (neat): o (-, #) = 3326, 1727,1643, 1242 cm-l ; HR-MS (NBA/NaI): m/z : exp. 318.1323 [M+Na] *, calc.

318.1317 for C15H21NO5Na ; HPLC: Chiralcel OG, 15 % iPrOH/hexane, 1 mL min-1, 254 nm, 12.6 min (2S, 3R), 16.7 min (2R, 3S). The absolute

configuration is assumed to be (2S, 3R) in analogy to the AA product from isopropyl cinnamate.

..

Isopropyl (2R, 3S)-3- (acetylamino)-2-hydroxy-3- (4- methoxyphenyl) propanoate (3) M. p. 159-160 °C; ; [a] o (25, D) = +38. 9 (c = 1.00 in 95 % EtOH) ; 1H NMR (400 MHz, CDCl3) = 1.27 (2 d, J = 7.0,7.0 Hz, 6H), 1.97 (s, 3H), 3.31 (d, J = 3. 9 Hz, 1H), 3.78 (s, 3H), 4.43 (dd, J = 3. 4, 2.4 Hz, 1H), 5.11 (septet, J = 6.3 Hz, 1H), 5.48 (dd, J = 9.3, 2.1 Hz, 1H), 6.22 (d, 9.3 Hz, 1H), 6.84-6.91 (m, 2H), 6.22-6.33 (m, 2H) ; 13C NMR (100 MHz, CDCl3) # = 21.5,21.6,23.2,53.8, 55.3,70.7,73e3,113.9,128.1,131.0,159.1,169.2,172.4; FT- IR (neat): o (-, v) = 3324, 1711,1650,1247 cm-1 ; HR-MS (NBA/NaI): m/z : exp. 318.1324 [M+Na] +, calc. 318.1317 for C15H2, NO, Na ; HPLC: Chiralcel OG, 20 % iPrOH/hexane, 1 mL min-1, 224 nm, 10.4 min (2R, 3S), 14.6 min (2S, 3R). The absolute configuration is assumed to be (2R, 3S) in analogy to the AA product from isopropyl cinnamate.

(1R, 2R)-2- (Acetylamino)-1, 2-diphenylethanol (ent-4) (ligand = (DHQD) 2-PHAL) M. p. 158-159 °C; [α]O (25,D) = +3. 1 (c = 0.713 in 95 % EtOH) (enantiomerically pure product, obtained by recrystallization from diethyl ether) ; 1H NMR (400 MHz, CDCl3) 6 = 1.90 (s, 3H), 3.14 (br s, 1H), 4.94 (d, J = 4.6 Hz, 1H), 5.16 (dd, J = 7.8,

4.6 Hz, 1H), 7.03-7.40 (m, 10 H) ; 13C NMR (100 MHz, CDCl3) 6 = 23.1,59.6,77.1,126.0,126.1,126.9,127.8,128.3,128.6, 139.4,140.7,170.6; HPLC: Chiralcel ODH, 15 % iPrOH/hexane, 0.5 mL min-', 254 nm, 21.8 min (1R, 2R), 31.7 min (1S, 2S). The absolute configuration was established, after hydrolysis (3N HCl, reflux, 1 h, followed by evaporation), by comparison to the known aminoalcohol hydrochloride.

Ethyl (S)-3-(acetylamino)-2-hydroxypropanoate (ent-5) (ligand = (DHQD) 2-PHAL) GC: Cyclodex B, J&W Scientific, initial time 5 min, initial temperature: 120 °C (5 min), rate: 0.5 °C/min, final temperature: 140 °C, 40.3 min (R), 41.3 min (S) ; [α]O(25,D)= +17.8 (c = 0. 5 in CHCl3 90 % ee), lit.'"'for () [a] o (", J =- 18.7 (c = 3 in CHCl3).

(R)-2-(Acetylamino)-1-phenylethanol (7a) HPLC: Chiralpak AD, 5 % iPrOH/hexane, 1.5 mL min-1, 254 nm, 12.7 min (R), 16.2 min (S).

(R)-2- (Acetylamino)-2-phenylethanol (8a) HPLC: Chiralpak AD, 5 % iPrOH/hexane, 1.5 mL min-', 254 nm, 11.6 min (R), 10.1 min (S).

(R)-2- (Acetylamino)-l- (3-nitrophenyl) ethanol (7b): M. p. 124-125 °C ; [a] o (25, D)-+6. 1 (c = 0.655 in 95 % EtOH) ; 1H NMR (400 MHz, DMSO-d6) 6 = 1.76 (s, 3H), 3.16-3.23 (m, 1H), 3.26-3.33 (m, 1H), 4.73-4.78 (m, 1H), 5.81 (d, J = 4.5 Hz, 1H), 7.62 (t, J = 7. 9,1 H), 7.76 (d, J = 7.7 Hz, 1H), 7.97 (t, J = 5.6 Hz, 1H), 8.11 (dd, J = 8.1,1.5 Hz, 1H), 8.16-8.18 (m, 1H) ;"C NMR (100 MHz, DMSO-d6) 6 = 22.5,46.4,70.5,120.6,122.0,129.6,132.9, 146.1,147.7,169.6; FT-IR (neat): o (-, #) = 3345, 3274,1602, 1524,1346 cm-1 ; HR-MS (NBA/NaI): m/z : exp. 225.0871 [M+H] +, calc.

225.0875 for C10H13N2O4 ; HPLC: N, O-diacetate derivative: Chiralcel ODH, 7.5 % iPrOH/hexane, 1 mL min-1, 254 nm, 26.1 min (R), 28.8 min (S).

(R)-2- (Acetylamino)-2- (3-nitrophenyl) ethanol (8b): M. p. 152-153 °C; ; [a] o (25, D) =-82. 6 (c = 1.05 in 95 % EtOH) ; 1H NMR (400 MHz, acetone-d6) 6 = 1.96 (s, 3H), 3.82 (t, J = 5.6 Hz, 2H), 4.19 (t, J = 5.6 Hz, 1H), 5.08-5.16 (m, 1H), 7.60 (t, J = 7.9,1H), 7.71 (bs, 1H), 7.79-7.83 (m, 1H), 8.08-8.12 (m, 1H), 8.24 (t, J = 1.9 Hz, 1H) ; 13C NMR (100 MHz, acetone-d6) 6 = 22.9,55.6,65.6,122.6 (2C), 130.2,133.3,134.7,144.8,170.1; FT-IR (neat): o (~, v)-3374, 3295,1652,1524,1346 cm-1 ; HR-MS (NBA/NaI): m/z : exp. 247.0690 [M+Na] +, calc. 247.0695 for C10H12N2O4Na ; HPLC: N, O-diacetate derivative: Chiralcel ODH, 7.5 % iPrOH/hexane, 1 mL min-1, 254 nm, 36.4 min (R), 32.5 min (S).

(R)-2- (Acetylamino)-l- (4-methoxyphenyl) ethanol (7c) HPLC: N, O-diacetate derivative: Chiralcel ODH, 7.5 % iPrOH/hexane, 1 mL min-1, 254 nm, 19.1 min (R), 23.6 min (S).

(R)-2- (Acetylamino)-2- (4-methoxyphenyl) ethanol (8c) M. p. 99-100 °C ; [a] o (ZS, D) =-126. 8 (c = 1.1 in 95 % EtOH) ; 1H NMR (400 MHz, acetone-d6) 1.92 (s, 3H), 3.69 (d, J = 6.2 Hz, 2H), 3.74 (s, 3H), 4.38 (bs, 1H), 4.93-5.00 (m, 1H), 6.82- 6.87 (m, 2H), 7.24-7.28 (m, 2H), 8.08-8.12 (m, 1H), 8.24 (d, J = 7.4 Hz, 1H) ; 13C NMR (100 MHz, acetone-d6) 6 = 23.0,55.4, 55.8,66.3,114.3,128.9,133.7,159.6,170.4; FT-IR (neat): o (-, v) = 3374, 3288,1645,1546,1246,1026 cm-' ; HR-MS (NBA/NaI): m/z : exp. 232.0954 [M+Na] +, calc. 232.0950 for C11H15NO3Na ; HPLC: N, O-diacetate derivative: Chiralcel ODH, 7.5 % iPrOH/hexane, 1 mL min-1, 254 nm, 26.2 min (R), 21.6 min (S).

Transformation of R-COOH to R-COOMe Procedure as adapted from Chan et al. Synthesis 1983,201.