Enantioselective Rhodium Enolate Protonations. A New Methodology for the Synthesis of β²-Amino Acids

Abstract

Rhodium catalyzed conjugate addition of an aryl boronic acid to α-methylamino acrylates followed by enantioselective protonation of the oxa-π-allylrhodium intermediate provides access to aryl substituted β²-amino acids. The impact of the different variables of the reaction on the levels of enantioselectivity has been assessed.

Rhodium catalyzed conjugate addition of organoboron,¹ organosilicon,² and organotin³ reagents to α,β-unsaturated systems has seen tremendous advances in the past decade. Hayashi, Miyaura and others have developed highly efficient enantioselective protocols for these conjugate additions that allows for the establishment of a new chiral center at the β-carbon.⁴ In contrast, use of this strategy to establish a stereocenter at the α-carbon has met with limited success.⁵ Recently several examples of enantioselective rhodium enolate protonations leading to enantioenriched α-amino acids and succinates have been reported.⁶

Development of new methods for the synthesis of β-amino acids is important.⁷ There are a number of enantioselective methods for the synthesis of β-substituted-β-amino acids (β³-amino acids).⁸ In contrast, there are few methods for the synthesis of α-substituted-β-amino acids (β²-amino acids) enantioselectively.⁷ This substitution pattern is of interest since it is present in naturally occurring amino acids as well as compounds with potential therapeutic value.⁹

We have recently developed a novel method for the synthesis of β²-amino acids using free radical chemistry.¹⁰ The stereochemistry in these reactions was established by an enantioselective H-atom transfer after conjugate radical addition.¹¹ One deficiency of the H-atom transfer methodology was the inability to incorporate aromatic groups into the targets. We surmised that a rhodium catalyzed conjugate addition of an aryl boronic acid to 1 followed by enantioselective protonation of the oxa-π-allylrhodium intermediate 2¹ could provide access to aryl substituted β²-amino acids (Scheme 1).¹² Recently Frost and co-workers have reported a racemic version of the transformation shown in Scheme 1.¹³ In this work we have evaluated several variables for the conversion of 1 to 3 including the nature of the proton source, chiral ligand, catalyst, nitrogen protecting group, and the ester substituent and report a reasonably efficient method for the synthesis of enantioenriched β²-amino acids.

Scheme 1.

Our work began with the identification of an optimal rhodium catalyst for the addition of phenylboronic acid to compound 5a using BINAP as the chiral ligand and water as the proton source. Our initial choice of catalyst, ligand, and proton source was based on the work of Hayashi,¹⁴ Genet,^6a Reetz,^6b and Frost.^6c Results from these experiments are presented in Table 1. The catalyst rhodium (acac)bisethylene complex gave good yield of the addition product with modest enantioselectivity (entry 1). The reactions were effective at 50 °C. Increasing the reaction temperature to 100 °C did not improve the selectivity.¹⁵ Of the four other variants tested, the catalyst derived from rhodium hydroxide (entry 2) and rhodium chloride (entries 3 and 4) gave good yields but only modest selectivity.

Table 1.

Evaluation of Different Rhodium Complexes^a

entry	catalyst	yield, (SM), %a	ee, %b
1	Rh(acac)(ethylene)2	76 (0)	41
2	Rh(OH)(COD)2	65 (12)	30
3	[RhCl(COD)]2/NaHCO3	80 (0)	13
4	[RhCl(norbornadiene)]2/NaHCO3	76 (2)	25
5	[RhCl(COD)]2/AgPF6	2 (91)	nd

^aIsolated yields. Yields in parenthesis are for recovered starting materials.

^bChiral HPLC analysis; nd = not determined.

With these results at hand, we set out to determine the optimal chiral ligand and proton source for the formation of 6a. Results from these experiments are presented in Table 2. Several different proton sources have been evaluated by Genet and co-workers in their work on the synthesis of α-amino acids.^6a In our experiments, three different proton sources and several commercially available phosphine ligands were evaluated.¹⁶ Changing the proton source from water to 2-methoxyphenol led to a decrease in yield of 6a. However, there was a large improvement in enantioselectivity (entry 1).¹⁷ This observation is similar to that made by Genet.^6a The use of 2-acetylphenol as a proton source was very beneficial providing 6a in 84% yield and 77% ee (entry 2). Phthalimide, with a reasonably acidic N-H, was also functional as a proton source providing the highest ee for the product with moderate yield (entry 3).¹⁸ The chemical efficiency of the reaction was modest using tol-BINAP as a ligand but the selectivity was high (entries 4 and 5). Of the several other ligand/proton source combinations tested (entries 6–12), Synphos¹⁹ gave good levels of enantioselectivity (entries 11 and 12). More promising results were obtained using a bisphosphine, Difluorphos, recently introduced by Genet,²⁰ as a ligand (entries 13–15). A combination of this ligand and phthalimide as the proton source gave the product in excellent chemical yield and enantioselectivity (entry 15).

Table 2.

Identification of Optimal Chiral Ligand and Proton Source for the conversion of 5a to 6a.

entry	ligand	proton sourcea	yield, (SM) %b	ee, %c
1	(S)-BINAP	2-Methoxyphenol	31 (49)	81
2	(S)-BINAP	2-Acetylphenol	84 (2)	77
3	(S)-BINAP	Phthalimide	56 (34)	82
4	(S)-Tol-BINAP	2-Methoxyphenol	21 (39)	73
5	(S)-Tol-BINAP	2-Acetylphenol	42 (40)	77
6	(S,S)-DIOP	2-Methoxyphenol	43 (31)	36
7	(R,R)-CHIRAPHOS	2-Methoxyphenol	8 (80)	nd
8	(R,S)-JOSIPHOS	2-Methoxyphenol	0 (95)	-
9	(S)-MethylBOPhoz	2-Methoxyphenol	8 (87)	nd
10	(S)-SYNPHOS	2-Methoxyphenol	1 (73)	nd
11	(S)-SYNPHOS	2-Acetylphenol	30 (67)	71
12	(S)-SYNPHOS	Phthalimide	25 (50)	70
13	(S)-DIFLUORPHOS	2-Methoxyphenol	8 (83)	nd
14	(S)-DIFLUORPHOS	2-Acetylphenol	71 (15)	88
15	(S)-DIFLUORPHOS	Phthalimide	91 (0)	88

^aOne equivalent of the proton source was used. The reactions were carried out at 50 °C using dioxane as a solvent and 2 mol% of the chiral rhodium catalyst.

^bIsolated yields. Yields in parenthesis are for recovered starting materials.

^cChiral HPLC analysis; nd = not determined.

Having identified an optimal ligand/proton source combination, we evaluated the effect of the nitrogen protecting group and the ester substituent on efficiency and selectivity (Table 3). These two variables had a significant impact on the course of the reaction. Changing the ester substituent from tert-butyl to others with a phthalimido nitrogen protoecting group led to either inefficient reactions or low selectivity (entries 1–4). Thus a bulky ester substituent is essential for obtaining high selectivity. Succinimide and tosyl protecting groups seem promising (entries 5 and 6).

Table 3.

Effect of Nitrogen Protecting Group and Ester Substituent on Selectivity^a

ent	nitrogen PG	ester R	R1	yield, (SM) %a	ee, %b
1	Phthalimide	t-Butyl 5a	Ph 6a	91	88
2	Phthalimide	Methyl 5b	Ph 6b	35 (35)	10
3	Phthalimide	Cyclohexyl 5c	Ph 6c	0	-
4	Phthalimide	Benzyl 5d	Ph 6d	85	20
5	Succinimide	t-Butyl 5e	Ph 6e	61	71
6	Tosyl	t-Butyl 5f	Ph 6f	86	81
7	Tosyl	t-Butyl 5f	4-BrPh 6g	23 (67)	84

^aIsolated yields. Yields in parenthesis are for recovered starting materials.

^bChiral HPLC analysis.

We have carried out preliminarily work on the scope of the aryl boronic acid component in the enantioselective protonation experiments and these results are shown in Table 4. The reaction conditions which were found to be best for phenylboronoic acid addition was employed for these studies. In general, the enantioselectivity in these experiments were high (entries 1, 2, 4, 5, and 6). However, the chemical yield for the reaction was variable. For example, while reaction with 4-methoxyphenylboronic acid was highly efficient (entry 4), reaction with 4-methylphenylboronic acid gave the product in only 16 % yield (entry 3). Of the different substrates evaluated, reaction with 2-naphthylboronic acid gave the highest chemical yield and enantioselectivity (entry 6).

Table 4.

Preparation of Different β²-Amino Acids

entry	R	yield, (SM)%a	ee, %b
1	Phenyl 6a	91	88
2	4-Chlorophenyl 6h	71 (10)	84
3	4-Methylphenyl 6i	16 (62)	63
4	4-Methoxyphenyl 6j	84	86
5	3,5-Bistrifluoromethylphenyl 6k	70 (23)	90
6	2-Naphthyl 6l	95	91

^aIsolated yields. Yields in parenthesis are for recovered starting materials.

^bChiral HPLC analysis.

The absolute stereochemistry of the enolate protonation product 6a derived from reaction with phenylboronic acid and phthalimide as a proton donor was determined to be (S) by converting it into a known compound.²¹ A catalytic cycle as postulated by Hayashi and Miyaura^1b appears to be operative in our experiments also. Genet observed a strong correlation between the level of enantioselectivity and the nature of the proton donor.^6a A functionality capable of coordinating the rhodium which is located ortho to the proton donor was optimal in their work. In our experiments we suggest that phthalimide containing a carbonyl donor coordinates the rhodium and transfers a proton to the oxa-π-allyl complex. The present work and the prior results in the literature suggest that enantioselective rhodium enolate protonations require a proper matching of all the variables and development of a general protocol is yet to be achieved.

In conclusion, we have developed a reasonably practical method for the preparation of β²-amino acids. Furthermore, we have also demonstrated that enantioselective rhodium enolate protonations can be carried out with good selectivity. The extension of the present protocol to more complex susbtrates is underway in our laboratory.

Supplementary Material

1si20050323_04. Supporting Information Available.

Characterization data for compounds and experimental procedures. See any current masthead page for ordering information and Web access instructions.