P.
Satyanarayana
a,
H.
Maheswaran
*a,
M.
Lakshmi Kantam
a and
H. P. S.
Chawla
b
aInorganic and Physical Chemistry Division, Indian Institute of Chemical Technology, Hyderabad, India. E-mail: maheswaran_dr@yahoo.com; Fax: +91 40 27193510; Tel: +91 40 27191532
bUnimark Remedies Limited, Mumbai, India
First published on 23rd August 2012
Pregosin's complex Ru(H)(p-cymene)((R)-DTBM-Segphos)(SbF6) has been shown to be an efficient catalyst for stereoselective asymmetric hydrogenation of 2-benzamido-methyl-3-oxobutanoate to syn-(2S,3R)-methyl-2-(benzamido-methyl)-3-hydroxybutanoate in high diastereoselectivity and enantioselectivity in ethanol.
![]() | ||
Scheme 1 |
The above result has led Takaya and co-workers to investigate various factors controlling the catalytic activity and stereoselectivity of this asymmetric hydrogenation by using a variety of cationic dihalo-BINAP–RuII(arene) complexes as catalysts.4b,5 During these studies it was discovered that the diastereoselectivity of the hydrogenation depends strongly on the solvent and on the nature of halide counter anions (I > Br > Cl) in the BINAP–RuII as well as the presence of m,m′-dialkyl substituents on the four phenyl rings of the BINAP ligands (meta-dialkyl effect), though the optical purities of the products are less sensitive to these factors. Notably, the results from these studies enabled successful development of an asymmetric hydrogenation process for synthesis of chiral syn-(2S,3R)-2-benzamidomethyl-3-hydroxybutanoates (2) for industrial scale carbapenem synthesis via DKR.6
Since, the choice of counter anions plays a pivotal role in determining diastereoselectivity, we have conceived the use of non-halogen containing bulky counter anions in cationic bis(phosphine)–RuII(p-cymene) complexes in an attempt to develop a new asymmetric hydrogenation process for syn-(2S,3R)-2. First, we have carefully studied the active patent literature6 governing asymmetric hydrogenation of 1, and then have narrowed down to use SbF6 counter anions for bis(phosphine)–RuII(p-cymene) complexes for our studies. This is because use of this counter anion is not adequately protected in the patent literature. A further literature search revealed that Pregosin et al. have already synthesized and characterized many bis(phosphine)–RuII(arene) complexes with SbF6 counter anions but have not evaluated these complexes as catalysts for asymmetric hydrogenations.7
Pregosin et al. have shown that dichloro-bis(phosphine)–RuII(p-cymene) complexes readily react with two equivalents of AgSbF6 salts and form their corresponding bis-(SbF6) complexes with a biaryl ligand double bond chelating RuII in a η2-fashion (see Fig. 1).7 This novel η2-chelating feature has been shown to be characteristic of such complexes involving a variety of chiral biaryl ligands containing SbF6 counter anions. Pregosin et al. have also shown that these complexes readily react with methanol or alcoholic solvents which displace η2-fashion bonding mode of the olefinic bond of the biaryl backbone and form stable yellow colored Pregosin's hydrido complexes of type [(bis-phosphine)-Ru(H)(arene)]-(SbF6)] in good yields7 (see Fig. 1).7c
![]() | ||
Fig. 1 Synthesis of Pregosin's hydrido-Ru(bis-phosphorus)donor (p-cymene) complexes. |
In this work, we have synthesized several Pregosin's hydrido complexes containing a variety of chiral biaryl phosphine ligands, and evaluated their utility as catalysts for the synthesis of syn-(2S,3R)-2 involving DKR assisted asymmetric hydrogenation of racemic 2-benzamidomethyl-3-oxo-butanoate (1). We have carried out these studies in order to develop a non-patent infringing process, if possible, for the exclusive synthesis of syn-(2S,3R)-2 in good yields under DKR conditions. The results from our studies involving various Pregosin's hydrido-(bis-phosphine)Ru(arene)-(SbF6) catalysts are the focus of this paper.
First, Pregosin catalyst 6c containing chiral (S)-DTBM Segphos ligand is chosen as the model catalyst system for asymmetric hydrogenation of racemic 2-benzamido-methyl-3-oxobutanoate (1) for optimization of reaction conditions. The model hydrido catalyst 6c was synthesized in situ by reacting 25 mg of (S)-5c with two equiv. of AgSbF6 in 3 mL methanol at 65 °C for 30 minutes under a nitrogen atmosphere. Upon the reaction, a dark brown solution of the catalyst (S)-5c turned pale yellow which is followed by the precipitation of AgCl salts. The mass balance of the precipitated AgCl salts indicated that the reaction occurred quantitatively, and all the chloride anions in 5c were completely removed by reaction with AgSbF6 salts under our experimental conditions. The mother liquor (3 mL) containing catalyst 6c was then pipetted out using a syringe and transferred to a 50 mL Parr autoclave containing 1.0 mmol of racemic 2-benzamidomethyl-3-hydroxybutanoate dissolved in 12 mL methanol under an inert atmosphere. Then the reduction reaction was conducted at 808 psi of hydrogen gas in a Parr autoclave at 70 °C for six hours with a substrate to catalyst ratio (S/C) of 60.
TLC studies indicated complete conversion of 1 to products. Then, the solvent was removed under reduced pressure and the crude reaction mixture was subjected to chiral HPLC analysis to determine both diastereoselectivity and enantioselectivity of the products. As can be seen from Table 1 (entry 1) that a 52% de and 69% ee was obtained. Interestingly, when the commercial cationic dichloro catalyst 5c was used without any additives under identical reaction conditions, we have obtained very poor de and ee for the desired product (Table 1; entry 2). It is to be noted that the corresponding cationic diiodo catalyst (5c analogue) has been shown to outperform and give high de and ee for the desired product in a 9:
1 DCM–MeOH mixture on a commercial scale.4b,6 We could not get any products in such a mixed solvent system for our case for the reduction of 2-benzamidomethyl-3-hydroxybutanoate using Pregosin's 6c catalyst. Therefore, the best solvents for the reaction with our catalysts systems are highly polar alcoholic solvent systems like methanol or ethanol. When the reaction was carried out with the catalyst prepared in situ using 2.0 equiv. of AgPF6 in methanol, 44% of de and 81% of ee was obtained for the syn-(2R,3S)-2 product (Table 1; entry 3). When the same reaction was carried out in EtOH as solvent about 63% de and 87% ee were obtained for the syn-(2S,3R)-2 product (Table 1; entry 4).
Entry | Catalystc | S/Ce | Solvent | de (%) | ee (%) | Configd of syn-2 |
---|---|---|---|---|---|---|
a Hydrogenations were carried out in 1.0 mmol of the substrate in 15 mL of solvent at 808 psi of hydrogen gas at 70 °C in 15 mL of appropriate solvent present in a 50 mL Parr reactor.
b Conversions of starting material for all the reactions are complete or quantitative.
c All catalysts were prepared in situ by using appropriate dichlorocatalyst (1a–1c, and 2a–2c) by adding 2.0 equiv. of Ag salts in appropriate solvent at 65 °C for 30 minutes.
d HPLC analysis was performed on a Chiralpak AD (4.6 × 250 mm) using 9![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
||||||
1 | (S)-6c | 60 | MeOH | 52 | 69 | (2R,3S) |
2 | (S)-5c | 60 | MeOH | 25 | 16 | (2R,3S) |
3 | (S)-5c + 2.0 equiv. AgPF6 | 60 | MeOH | 44 | 81 | (2R,3S) |
4 | (R)-5c + 2.0 equiv. AgPF6 | 60 | EtOH | 63 | 87 | (2S,3R) |
5 | (S)-5c + AgPF6 + AgSbF6 | 60 | MeOH | 61 | 84 | (2R,3S) |
6 | (S)-5c + AgPF6 + AgSbF6 | 60 | EtOH | 90 | 94 | (2R,3S) |
7 | (R)-6c | 60 | EtOH | 85 | 98 | (2S,3R) |
8 | (R)-4c | 60 | EtOH | 02 | 73 | (2S,3R) |
9 | (R)-6b | 60 | EtOH | 03 | 87 | (2S,3R) |
10 | (R)-4b | 60 | EtOH | 02 | 74 | (2S,3R) |
11 | (R)-4a | 60 | EtOH | 06 | 78 | (2S,3R) |
12 | (R)-6a | 60 | EtOH | 18 | 74 | (2S,3R) |
When mixed counter anions were employed for the reaction during catalyst preparation; i.e., 1.0 equiv. AgPF6 and 1.0 equiv. AgSbF6 instead of 2.0 equiv. of AgSbF6, 61% de and 84% ee were obtained for syn-(2R,3S)-2 in methanol (Table 1; entry 5). When the same reaction is carried out in ethanol, we have obtained 90% de and 94% ee for the syn-(2R,3S)-2 product (Table 1; entry 6). It can be seen that an increase in de is obtained along with moderately high ee for the desired product in this mixed counter-anion system. However, when the catalyst (6c) containing (R)-DTBM Segphos ligand was prepared in ethanol with 2.0 equiv. of AgSbF6 and the reaction solvent changed from methanol to ethanol (total volume 15 mL), high 85% de and >98% ee values were obtained for the desired syn-(2S,3R)-2 product (Table 1; entry 7). This result matches well with the one reported by Noyori using (R)-BINAP containing [Ru2Cl4{(R)-BINAP}2(NEt3)] catalyst.4 For carbapenem synthesis ee of the product is more important than de of the product. Therefore, we consider Table 1; entry 7 as the best result in comparison to those obtained with the mixed counter anion system (Table 1; entry 6).
From the results of these studies it is apparent that for the reaction both the choice of non-halogen counter anions (SbF6 is superior to PF6) and the choice of alcoholic solvent (ethanol is superior to methanol) play a vital role in efficient synthesis of syn-(2S,3R)-2. Therefore, the optimized reaction conditions for development of an effective process for the syn-(2S,3R)-2 product involving DKR assisted asymmetric hydrogenation of racemic 1 should involve utilization of SbF6 counter anions in the RuII complex along with a suitable (R)-(bisphosphorus-donor) ligand at 808 psi H2 at 70 °C in ethanol. To verify this proposition, we have subjected the (R)-DM-BINAP ligand containing Pregosin complex (4c) as the catalyst for the hydrogenation of 1 in ethanol. In this catalyst, all four phenyl rings of the ligand contain methyl groups at m,m′-positions. Given well known meta-dialkyl effects augmenting de and ee in many chemical reactions,8 to our surprise, we have obtained much lower 02% de and 73% ee for the syn-(2S,3R)-2 product (Table 1; entry 8). It is to be noted that with halogen (iodo) groups containing cationic commercial catalysts, the m,m′-methyl disubstitution in all the phenyl rings of the ligand framework or the meta dialkyl effect in the Ru complex has been shown to be very effective for the DKR mechanism to occur and give syn-(2S,3R)-2 exclusively.3c,d,4b
To verify whether improved results could be obtained with superior Segphos type ligands, we have studied use of 6b (m,m′-dimethyl substituted ligand in all four phenyl rings = (R)-DM-Segphos containing complex) as the catalyst for the DKR assisted asymmetric reduction of 1. As can be seen from Table 1 (entry 9), we have obtained lower 03% de and 87% ee for the desired syn-(2S,3R)-2 product. We have also performed the reaction with a Tol-BINAP ligand containing 4b Pregosin catalyst as well, and again, we have obtained much lower 02% de and 74% ee for the syn-(2S,3R)-2 product (Table 1; entry 10). Similarly, poor results with low de and moderate ee were obtained when simple (R)-BINAP and simple (R)-Segphos containing catalysts 4a and 6a were evaluated as catalysts under our experimental conditions (Table 1; entries 11 and 12).
The above results show that unlike cationic dihalo-(bis-phosphorus)donor–RuII(arene) complexes studied by Takaya et al. the presence of m,m′-di-alkyl substituents on the four phenyl rings of the bis-phosphorus-donor ligand does not play a significant role in facilitating high de and ee for the product under our experimental conditions. Interestingly, we have obtained excellent results with a (R)-DTBM ligand using Pregosin's 6c catalyst that contains m,m′-disubstitution of bulky tBu groups, which is very surprising. However, the (R)-DTBM ligand in catalyst 6c also contains p,p′-substitution of OMe groups in all the phenyl rings of the ligand. It is likely that the presence of such strongly electron donating groups in the ligand motif may be very important for an efficient DKR mechanism to occur with adequate rate constant numbers upon asymmetric reduction of 1 to give syn-(2S,3R)-2 with high de and high ee.
In conclusion, we have synthesized a variety of Pregosin's complexes of formula [(bis-phosphine)–Ru(H)(arene)]-(SbF6)] and subjected them to DKR assisted asymmetric reduction of racemic 1 to the syn-(2S,3R)-2 product. The conclusion drawn from the study includes that unlike cationic dihalo complexes the m,m′-substitution pattern of dialkyl groups in all the four phenyl rings of the bis(phosphorus)-donor ligand may have little or no effect in inducing high de and ee for the desired product. Our studies show that a sterically hindered (R)-DTBM Segphos ligand with structural features like strongly electron-donating OMe groups at the p,p′-position in all phenyl rings in Pregosin's catalyst 6c is essential for asymmetric reduction of 1 to give syn-(2S,3R)-2 with high de and high ee. Finally, given the fact that the (R)-DTBM ligand is still under patent protection, we were unable to develop a non-infringing process for the desired syn-(2S,3R)-2 product using Pregosin's hydrido complexes.
Footnote |
† Electronic supplementary information (ESI) available: Experimental procedure and HPLC data. See DOI: 10.1039/c2cy20335a |
This journal is © The Royal Society of Chemistry 2012 |