‘March, 1970) BULLETIN OF THE CHEMICAL SOCIETY OF JAPAN vou. 43 ‘Syntheses of Optically Active «Amino Acids 921 ‘921—925 (1970) Syntheses of Optically Active a-Amino Acids from Esters of @-Keto Acids by Hydrogenolytic Asymmetric Transamination, A Solvent Effect’ Kaoru Harapa and Takao Yosuipa Institate of Molecular Evolution and Department of Chemistry, University of Miami, Coral Gables, Florida 33134, U.S. A, (Received September 29, 1969) Optically active a-amino acids were synthesized from esters of a-keto acid azomethines by catalytic hydrogenation. are discussed. Several studies on the asymmetric syntheses of a-amino acids from their corresponding a-keto acids have been reported.!-! Hiskey and Northrop? demonstrated the synthesis of optically active amino acids by catalytic hydrogenation and subsequent hydrogenolysis of the Schiff bases of a-keto acids with optically active a-methylbenzylamine (hydro- genolytic asymmetric transamination). ‘The steric course of the hydrogenolytic asymmetric transami- nation reactions has been studied in this labora- tory." It was proposed that the possible con- formation of the substrate in the catalytic hydro- genation could be a five-membered cyclic complex with the catalyst, as shown in structure 1” The fee 1) Sterically Controlled Synthesis of Optically Active Organic Compounds VIII. For pare VIT. K. Matsumoto and K. Harada, J. Org. Chem, 33, 4526 (1968). Contri- bution No. 124 of the Institute of Molecular Evolution, University of Miami. 2) F. Knoop and G. Martius, Z. Plys. Chem., 258, 238 (1939). 3) J. B. Herbst and E, A. Swart, J. Org. Ghem., 11, 366 (1946) 4) RG. Hiskey and R. C, Northrop, J. Amer. Chem, Soc., 83, 4798 (1961). 5) A. Kanai and S. Mitsui, Nippon Kagaku Zasshi, 89, 183 (1966). 6) K. Matsumoto and K. Harada, J. Org. Glem., 31, 1956 (1966). 7) K. Harada, Nature, 212, 1571 (1966); J. Org. Ghen., 32, 1790 (1967). 8) K. Harada and K. Matsumoto, J. Org. Chem, 32, 1794 (1967). 9) K. Harada and K. Matsumoto, ibid, 33, 4466 (1968), 10) K. Matsumoto and K. Harada, ibid., 38, 4526 (1968), Solvent effect of the reaction was studied and the possible steric courses proposed intermediate I was supported by the study of a solvent effect of the asymmetric synthesis. In this investigation, the Schiff bases of a-ketor acid esters with optically active amines were sub- Jected to hydrogenation in order to examine the existence of structure I type chelated intermediates in these reactions. a-Keto aoids used were ethyl pyruvate and ethyl a-ketobutyrate. Optically active amines used were (S)(—) and (R)(+)-2 methylbenzylamine, _(R)(-|-)-2-ethylbenzylamine, and (R)(-+)-I-(J-naphthyl)ethylamine. Solvents used in these reactions were benzene, ethyl acetate, isopropyl alcohol, ethanol, and methanol, Results are summarized in Table 1. In each six sets of reactions, a solvent effect of the asymmetric synthesis was examined by the use of methanol, ethanol, propanol-2, ethyl acetate, and. benzene (dielectric constant ranging from 32.6 to 23; Table 1). It was found that when a polar solvent was used, the optical purity of the resulting amino acid was lower and when less polar solvent was used the optical purity increased steadily. ‘These solvent effects in the reactions suggest that the conformations of the substrate molecule changed depending on the solvent used. ‘These results are similar to those obtained in the study. of solvent effect in the hydrogenolytic asymmetric transami- nation between z-keto acids and optically active amine.” Therefore, similar conformation of the substrate in the hydrogenolytic asymmetric trans- amination between optically active amine and aeketo acid ester could be assumed as shown in structure II. Because electrostatic attraction be- tween substrate and catalyst in a less polar solvent is stronger than in a polar solvent and also the solvation of the substrate in the less polar solvent is weak, so that the substrate could react easily with the catalyst to form the intermediate complex II. On the other hand, when polar solvent was used, electrostatic attraction between substrate and cata~ lyst was weaker and the stronger solvation of the922 Kaoru Harapa and Takao Yostpa [Vol. 43, No. 3 ‘Tani 1, AMINO ACIDS OBTAINED BY HYDROGENOLYTIC ASYMMETRIC TRANSAMINATION BETWEEN Free amino DNP-Amino ; ‘acd acid Amino acid KE0SS!4 Amine? Solvent Catalyst YEN Config’n (55 Optical (ajF Optical Sx purity Tx purity HCI (%)® 201 ‘Ala 1 EvPy (R)(+)-Me MeOH Pd(OH),/C 74 oR —3.1 21 2 EvPy (R)(+)-Me EtOH Pd(OH),/C 77 R 4.7 32 3) EtPy (R)(+)-Me i-PrOH Pd(OH),/C 73 R 32 4 BLPy (R)(+)-Me AcOEt PAH) JC 78 R “4 5 BePy (R) (4)-Me Benzene PAOH)/C 70 R 82 6 EePy ($)(—)-Me McOH PAOH)IG 75S 31 ; 7 Er-Py (8)(—)-Me EtOH = Pd(OH),/C 74 Ss 40 +59. 8 EePy (S)(=)-Me POH PAH) IC 79 Sc e364 9 EePy (R)(—)-Me AcOEt PA(OH)IG 73S 45.6 89 +91 10 EePy (S)(—}-Me Benzene PA(OH)/C 75S 46.7 46482. 11 Et-Py (R) (+)-Et ‘McOH = Pd(OH),/C 83 Ro-1.5 100 —21 2 EePy (R)(4-E HOH PaOH)JC 8% R -20 14 —33 13 EePy (R)(4)-Et PrOH Pd(OH)/C 80 R —3.3 23 —46.! 14 BePy (R)(+}Et AcOEt PA(OH)G 77 RK =7.0 48 80. 15 BePy (R)(+)-Et Benzene PA(OH)IG 73° R 5.9 41 72. 16 Et-Py (R) (+)-Naph McOH Pd(OH),/C 83 R -7.0 49 —82. 17 EtPy (R)(+)-Naph EtOH Pd(OH),/C 82 oR -8.2 56 —93 18 EePy (R) (4)-Naph iPrOH PA(OH)JG ¢ RK =7.3 30 95.3 19 BLPy (R)(+)-Naph AcOEt PA(OH)J/C 68 RR -7.9 5 118. 20 EePy (R)(+)-Naph Benzene PAOH) IC 70 R 8.760 —100. 2 BePy (8)(—}-Me MeOH BAG APE gy 51 95 469. 2 EePy (8)(—)-Me BOK Puan. ozs 32 4738.2 5 PalC and 38 478.6 55 23 EuPy (8) (—)-Me bPOH RYIGAM 65S Mt Bery @()Me Acom MGA mS nee aT ExPy (8) (—}Me Benzene pa G Ante 57 40 498.2 65 Et-Bu (S)(—)-Me = McOH Pilonic sl 7 427.9 8 EvBu (S)(—)-Me EOH BGS o 45 27 +306 31 2 BeBe (8)(—)-Me EPO PYGAPd 4g 2453.5 9h 29 ELBu (8) (—)-Me —AcOEL PUG AN 42 2% 437.6 38 20 EABa (6) (Me Bomene FHGAM 99s a) EtPy: ethyl pyruvate; Bt-Bu: ethyl acketobutyrate. b) (8) (—)-Me, (8) (—)-smethylbenzylamine; R(-+)-Me, (R) (-+)-xmethylbenzylamines (S) (—)- Bx, (8) (—)-zeethylbenzylamine; R(+)-Ee, (R) (+ )-a-ethylbenzylamine; R(-+)-Naph, (R) (+)= 1-(I-naphthyl)ethylamine, €) The samples are lost. 4d) Defined as {4}. obsd[2}» lit 100, _(S)-Alanine, [2] -+14.6° (5x HCl); (S)=x-Aminovr-butyric acid, [23 +206" (58 HCl). J. P. Greenstein and Winite, “Chemistry of the Amino Acids,” Vol. 8, John Wiley & Sons, Inc, New York, N. Y. (1961) (alanine, p. 1819; 2-amino-r-butyric acid, p. 2401). €) Defined as [2]» obsd)fJo lit 100, DNP-(S)-ala, [2p +143.9° (IN NaOH); DNP-(S)-2-NE~ But, [2]o +-98.8° (Ix NaOH), K, R. Rao and H. A, Sober, J. Amer, Chem Soc. 76, 1328 (1954).‘March, 1970} if cH. HB ade She AOE 0 Pr’ So a 0 m substrate interfered with the attraction between substrate and catalyst to form the intermediate complex IT. Therefore, it could be assumed that the conformation of substrate in polar solvent could be more in the form of structures III and IV. Structure IV is rather sterically hindered so that the structure ITT could be increased in proportion Optical purity, w FT) FT) Dielectric constant Fig. 1. Relationship between optical activities of synthesized amino acids and dielectric constants, of the solvents used. . ala, reactions 1—5 (see Table 1) + ala, reactions 6—10 . ala, reactions 11—15, . ala, reactions 16—20 ala, reactions 21—25 ; eNH,-Bu, reactions 26—30 mmo Om> ‘Syntheses of Optically Active x Amino Acids 923, by the use of polar solvent. Structures If and IIT would be adsorbed at the less bulky side of the molecule and the hydrogenation takes place. When (8)-amine was used, (S)- and (R)-amino acid would. be expected from structures I and IIT respectively. ‘Therefore the use of polar solvent increases the proportion of structure III, and this would result in the formation of amino acid of lower opticat purity. From the results obtained, structure IE could be the major conformation throughout the reactions. Figure 1 shows the relationship between the dielectric constants of the solvents used and the optical activities of synthesized amino acids. The effect of the catalyst, palladium hydroxide om charcoal or palladium on charcoal, was observed. By the use of palladium hydroxide on charcoal, optical activity of amino acids dropped when benzene was used as a solvent (Fig. 1). Optical activities of amino acids increased rapidly when less polar solvents were used. On the other hand, by the use of palladium on charcoal optical acti ties of amino acids obtained increased slowly in the less polar solvents. A decrease in optical activity of amino acid by the use of benzene was not observed. High optical activity of amino acid (82%) was observed when R(+)-1-(I-naphthyl)~ cethylamine was used. The reason for this is under= standable from the steric effect in the intermediate complex I. In the previous studies, hydrogenation of oximes: and Schiff bases of a-keto acid [-menthyl esters has been studied. It was discussed that the conforma tion of the substrate might be transoidal, as shown in structure IV."*) From the results obtained in the present investigation, it could also be possible: RY to explain the steric course by a cisoidal conforma~ tion which forms chelate intermediate with the catalyst as shown in structure V. The menthyb group would be situated as in structure V because of the steric hindrance between menthyl residue and catalyst. Then, the structure V could be ad- sorbed on the catalyst at the less bulky side of the molecule, and hydrogenation would take place. From structure V, R-amino acid would be expected, which agrees with the experimental results.) In order to examine the newly postulated inter- mediate (V) in the hydrogenation of a-keto acid Lmenthyl ester oxime, a similar study of solvent ‘effect was carried out. The substrates used were L-menthyl pyruvate oxime, -menthyl a-ketobutyrate oxime, and the Schiff base of /-menthyl pyruvateot Kaoru Hanapa and Takao Youups [Vol 43, No. 3 ‘Tantus 2, AmINo AciDs OBTAINED FROM G-KETO ACID METHYL ESTERS? DNP-Amino acid Amino acid Keeg5% Solvent YSI6° Config'n jp Optical pag Optical ° purity glam purity (ay ON NAOH) [)4 ‘Ala 31 PyMOX MeOH 87 18 20.220 9 PyMOX HOH 82 7 33 33° PyM-OX ——i:PrOH R "7 BS 16 34 Py:M-OX AcOEt R 7 28.1 20 35 PyM-OX Benzene R 3-922 0 -D 2NH-Ba 36 BuM-OX MeOH. R ww -164 7 37 BuM-OX ROH R rs) 38 BuM-OX i: PrOH R 4 16 39 BuMOX AcOEt R 20007 40 BuM-OX Benzene R 2 23 Ala Ho PYM-R)-Me MOH 46 R 0 37 2 EOH 31 RK 559 43° Py-M-(R)-Me —i-PrOH 50 R 3 46 4 PyY-M(R}-Mce AOE 48 RK 7 46 45 Py-M-(R)-Me Benzene 47 R 8 8 2) All hydrogenation reactions were carried out by the use of 5°, palladium on charcoal. b) Py-M-OX: /-menthyl pyruvate oxime; Bu-M-OX: L-menthy! s-ketobutyrate oxime; Py-M-(R)- ‘Me: Schiff base of /-menthyl pyruvate with R(-+)-z-methylbenzylamine. ©) Reactions 31—40: the yields are calculated from oximes of arketo acid menthy] esters. Reactions 4) 41-45: the yields are calculated from /-menthyl pyruvate. Defined as [2]» obsd/f2], lit 100. (S)-Alanine, [2] +14.6° (SN HCI); (S)-2-amino-n-butyrie acid, [2] +20.6° (5x HC). J. P. Greenstein and Winitz, “Chemistry of the Amino Acid,” Vol. 3, John Wiley & Sons, Inc, New York, N. acid, p. 2401). Defined as [2]» obsd {ay lit 100. Bu, [a}» +98.8° (In 2 ©) with (R)(+)-e-methylbenzylamine. Results are summarized in Table 2. However, in these reactions, the solvent effect ‘observed was rather small compared with that of the previous study” or with the results summarized in Table 1. Also observed was the discontinuation ‘of optical purities of amino acids in these series of reactions (Fig. 2). The optical purities increased by the use of ethanol compared to that use of meth- anol. However, optical purities dropped by the use of isopropyl alcohol and increased steadily by the use of less polar solvent. The factors which affect the optical purity of amino acids in these reactions seem to be complex and it is difficult to explain these results at the present time. However, from the results obtained from the hydrogenolytic transamination summarized in Tables 1 and 2 the ‘existence of intermediate complex structure V seems Tkely. On the other hand, catalytic reduction of a-keto acid esters with optically active alcohols has been discussed and the reduction was found to follow the Prelog rule” The results obtained in the present study suggest that the formation of a-hydroxy 11) V. Prelog, Bull. Soc. Chim. Fr., 1956, 987. DNP-(S)- OH). K.R. Rao and H. A. Sober, J. Amer. Chem, $c, 76, 1328 (1954) Y. (1961) (alanine, p. 1819; a-amino-r-butyric ala, (2Jy +143.9° (1x NaOH); DNP-(S)-2-NH, acid from acketo acid ester with optically active alcohol by catalytic hydrogenation could also be explained by the chelated intermediate hypo- thesis.” ‘The study on the synthesis of optically active a-hydroxy acid from a-keto acid amide with optically active amine also supports the chelation hypothesis." Optically active free amino acids were obtained by the use of a Dowex 50 column, However, these amino acids still contain some impurities. To avoid fractionation of optically active isomers during the purification process, all amino acids were treated with 2,4-dinitrofluorobenzene."*) ‘The resulting DNP-amino acids were purified by the use of celite column"? treated with pH. 7.0 citrate buffer without fractionation of partially optically active amino acids. ‘Therefore, the values of optical purities of DNP-amino acids are more reliable and consistent (Table 1). 12) _K,Harada and K. Matsumoto, Presented at the Ist American Peptide Symposium at Yale University (1968). 13) _F. Sanger, Biochem. J., 39, 507 (1945); K. Rao and H, A. Sober, J. Amer. Chem. Soc. 76, 1328 (1954). 14) J. G. Perrone, Nature, 167, 513 (1951); A. Court, Biochem, J, 58, 70 (1954).‘March, 1970] 4 NO SOS Optical purity 10] ° 0 20 0 Dielectric constant Fig. 2. Relationship between optical activity of synthesized amino acids and dielectric constants of the solvents used. G. ala, reactions 31—35 (see Table 2) H. «-NH,-Bu, reactions 36—40 T. ala, reactions 41—45 Experimental) Alanine from Ethyl Pyruvate and § (—)-a-Meth- ylbenzylamine. Ethyl pyruvate (1.16 g, 0.01 mol) and. §(—)-xmethylbenzylamine (1.21 g, 0.01 mol) were dis solved in 20 ml of benzene at room temperature. After about ten minutes the solution became cloudy by sepa- ration of water. Anhydrous sodium sulfate was added. and the mixture was stirred for one hour at room tem= 15) All hydrogenations and hydrogenolyses were carried out by the use of the Parr 3910 shaker type hydrogenation apparatus. All optical rotations were measured by the use of Durrum-JASCO ORD/UV-5 automatic spectropolarimeter. Some of the optical rotations of free amino acids were measured by the use of the Rudolph model 80 polarimeter with a PEC-101 photometer. Syntheses of Optically Active «Amino Acids 925 perature. Sodium sulfate was removed by filtration and the sodium sulfate was washed with benzene. ‘The com- bined benzene solution was evaporated to dryness under reduced pressure on a water bath (45°C). A light yellow ilwas obtained. ‘This was dissolved in 30 mf of methanol and the solution was subjected to hydrogenation and hydrogenolysis by the use of 1.0 g of palladium hydroxide cn charcoal” for 12 br. The first mole of hydrogen was absorbed in a few minutes. However, the second mole of hydrogen for hydrogenolysis required a much longer time. After the reaction was over, the catalyst was removed by filtration. The catalyst was washed with hhot methanol several times. The combined methanol solution was evaporated under reduced pressure at 40°C. The residue was hydrolyzed under reflux with 60 mi of 83 hydrochloric acid for 4h. The hydrolyzate was extracted once with ether and the hydrochloric acid solution was evaporated to dryness under reduced pres- sure. Water was added to the residue and it was evaporated again to minimize the remaining free hydro- dhlorie acid. "The residual amino acid hydrochloride ‘was dissolved in 10 mi of water and the solution was applied to a Dower 50%2 column (hydrogen form, 100-200 mesh, 25> 1.8m). The acidic component was washed with water, then alanine was eluted with Ix aqueous ammonia. Fractions containing alanine were combined and were evaporated to dryness. Free alanine, 0.67 g, was obtained. This is pure on paper chromatograph. +4.5° (¢ 40, 58 HC), optical purity 31%. After one recrystallization with water and Cthanol, clemental analysis was carried out. Found: C, 40.65; H, 8.01; N, 15.58%. GyH,ONy: C, 40.44; H, 7.92; Ny 15.72%. ‘A part of the unrecrystallized alanine (100 mg) was converted to DNP-alanine in a conventional manner. DNP-Alanine obtained was purified by the use of celite column treated with pH 7 citrate buffer (0.22 ‘These procedures are similar to those described in carlier reports." Caled for DNP-Alanine, mp 172—174°C dee, [al= +46." (€ 1.1, 1x NaOH), optical purty 32%. Found: ©, 42.42; H, 3.66; N, 16.64%, Caled for GHN,O,: C, 42.36; H, 8.55; Ny 16.47%, ‘Other experiments were carried out in the same way as described above, Results are summarized in Table 1. ‘Amino Acids from a-Keto Acid [-Menthyl Ester. ‘The experimental procedures for the synthesis of optically active amino acids from aketo acid Lmenthyl esters were similar to these already described in the carlicr papers except forthe use of different solvents. Alanine and zaminobutyrie acid obtained by saponification of renthyl esters were converted to their DNP-derivatives. ‘These DNP-amino acids were similarly purified by the use of elite column. Results are summarized in Table 2. ‘This work was supported by grant number NGR 10-007-052 from the National Aeronautics and Space Administration.