The Preparation of Hydroxyphenylserines from Benzyloxybenzaldehydes and Glycine1
William A. Bolhofer
JACS, 76, 1322, (1954)
Abstract:
Threo-beta-meta-Hydroxyphenylserine (X), threo-beta-p-hydroxyphenylserine (XI) and both diastereoisomers of beta-3,4-dihydroxyphenylserine (XII) have been prepared. Glycine, when allowed to react with m-benzyloxybenzaldehyde (I) in alcoholic potassium hydroxide, yielded threo-beta-m-hydroxyphenylserine (X) after acidification and hydrogenolysis of the intermediate N-benzylidene potassium salt. Likewise, from glycine and p-benzyloxybenzaldehyde (II), threo-beta-p-hydroxyphenylserine (XI) was prepared. The condensation of 3,4-dibenzyloxybenzaldehyde (III) with glycine yielded beta-3,4-dibenzyloxyphenylserine (IX) as a mixture of diasteroisomers after acidification of the intermediate N-benzylidene potassium salt. These isomers were separated by fractional crystallization and hydrogenolyzed to yield both pure diasteroisomers of beta-3,4-dihydroxyphenylserine (XII).
The involved, indirect procedure required for the preparation of erythro-beta-p-hydroxylphenylserine from ethyl-p-benzyloxybenzylacetate2 and the fact that the threo isomer was not obtained at all indicated that any attempt to extend this method of synthesis to the preparation of even one of the diastereoisomers of beta-3,4-dihydroxyphenylserine would meet with failure. It could be expected that the intermediates in the 3,4-dihydroxy series would be even less stable than those in the p-hydroxy series. For these reasons, a modification of the original Erlenmeyer3 synthesis was examined for its applicability to the preparation of various beta-benzyloxyphenylserines which, on hydrogenolysis, would yield the desired beta-hydroxyphenylserines. In this modification, the aqueous alkali used by Erlenmeyer for the condensation of the aldehyde with glycine, is replaced by alcoholic alkali. Such substituted serines as beta-(2-thienyl)-serine,4 beta-(2-furyl)-serine,5 and beta-3,4-methylenedioxyphenylserine6 have been synthesized by this modified procedure.
In this laboratory, benzyloxybenaldehydes were found to react rapidly with glycine in alcoholic potassium hydroxide to give N-benzylidene-beta-benzyloxyphenylserines. Acidification and hydrogenolysis yielded the desired beta-hydroxyphenylserines. Initial experiments were carried out in the m-hydroxyl series because of the ease of identification of the diastereoisomeric beta-m-hydroxyphenylserines by MP alone.2
The condensation of two moles of m-benzyloxybenzaldehyde (I) with one mole of glycine was brought about in an alcoholic potassium hydroxide solution. A solid, crystalline potassium salt (IV) could be isolated, but a better over-all yield was obtained when the entire mixture was acidified. Acidification of the benzylidene compound (IV) caused the regeneration of approximately half of the starting m-benzyloxybenzaldehyde (I) with the simultaneous formation of beta-m-benzyloxyphenylserine (VII). This compound appeared to be a single diastereoisomer and its melting point did not change after a number of recrystallizations. Although no evidence of the presence of other diastereoisomer was obtained, this does not constitute absolute proof that the reaction proceeds to yield only one isomer. Debenzylation of the beta-m-benzyloxyphetnylserine (VII) was achieved by catalytic hydrogenolysis and the product was identified by its melting point as threo-beta-m-hydroxyphenylserine (X).
Threo-beta-p-hydroxyphenylserine (XI) was prepared from p-benzyloxybenzaldehyde (II) in a similar manner. Two moles of (II) were condensed with one mole of glycine and the resulting potassium benzylidene salt (V) was isolated as a crystalline solid in good yield. Acidification of the salt (V) gave beta-p-benzyloxyphenylserine (VIII) which appeared to be a single isomer. Catalytic hydrogenolysis removed the benzyl group and threo-beta-p-hydroxyphenylserine (XI) was obtained. The threo configuration was assigned to the product by analogy with the m-hydroxy series. The decomposition point of the beta-p-hydoxyphenylserine prepared by this method is 15*C lower than that reported by this laboratory for the erythro isomer.2 Holland, Jenkins and Nayler7 have recently reported the preparation of threo-beta-p-hydroxyphenylserine (MP: 188*C) from threo-beta-p-nitrophenylserine.
When 3,4-dibenzyloxybenzaldehyde (III) was allowed to react with glycine in alcoholic potassium hydroxide, the potassium salt of the N-benzylidene-beta-phenylserine (VI) separated from the alcohol solution as a heavy oil. After acidification and removal of 3,4-dibenzyloxybenzaldehyde, the beta-3,4-dibenzyloxyphenylserine (IX) was obtained as a powder. This product proved to be a mixture of two diasteroisomers which could be separated by fractional crystallization to give a high-melting and a low-melting isomer. The high-melting and a low-melting isomer. The high-melting isomer was debenzylated in 50% methanol by catalytic hydrogenolysis. The product, the low-melting more-soluble isomer of beta-3,4-dihydroxyphenylserine (XII), was obtained as a hydrate when the solution was concentrated. 8 The low-melting isomer of beta-3,4-dibenzyloxyphenylserine (IX) was obtained as a hydrate. It was debenzylated in dilute alkali and the high-melting isomer of beta-3,4-dihydroxyphenylserine (XII) was obtained. This isomer appears to be identical with the beta-3,4-dihydroxyphenylserine reported by Rosenmund and Dornsaft9 and Dalgliesh and Mann.10
The infrared absorption spectra of DL-threonine,11 DL-allothreonine11 and the racemic diastereoisomers of beta-phenylserine,12 beta-m-hydroxyphenylserine,2 beta-p-hydroxyphenylserine2 and beta-3,4-dihydroxyphenylserine were determined in Nujol. The absorption curves are not absolutely conclusive in themselves, but the data can be used in conjunction with chemical evidence as an indication of stereochemical structure. Comparison of the spectra of substances of proven structure (threonine and allothreonine and threo- and erythro-beta-phenylserine) shows that a band occurs regularly at 11.90-11.95 mu for those substances having the erythro- configuration. This band was not observed in the spectra of compounds possessing the threo configuration. A band at 11.90 mu appeared also in the absorption spectrum of the beta-m-hydroxyphenylserine assigned the erythro structure on the basis of chemical evidence alone. Likewise, examination of the spectra of the beta-p-hydroxyphenylserines indicates that the configuration assigned each diastereoisomer is correct. The presence of a band at 11.95 mu in the spectrum of the beta-3,4-dihydroxyphenylserine (MP: 199-200*C) indicates that is has the erythro structure. This band is missing from the spectrum of its isomer melting at 220-225*C, which can therefore be assigned the threo configuration.
Experimental13:
Example 1:
threo-beta-m-benzyloxyphenylserine (VII)
To a solution of 5.61g (0.1 mole) of potassium hydroxide and 3.75g (0.05mol) of glycine in 75ml of absolute alcohol, there was added a solution of 21.2g (0.2mole) of m-benzyloxybenzaldehyde14 in 25ml of absolute alcohol. The mixture was warmed until it was clear and the solution was allowed to stand at RT. A brown oil separated which crystallized slowly, MP: 125-130*C. However, it was unnecessary to obtain a crystalline product at this point. After standing overnight, the alcohol was decanted and the residual oil was dissolved in a mixture of 200ml of 2N HCl and 50ml benzene. The benzene was separated and the aqueous solution was again extracted with 50 ml of benzene. (from these benzene extracts approximately half of the starting material was recovered) The aqueous solution was concentrated ammonia and, after standing overnight at 0*C, crystalline threo-beta-m-benzyloxyphenyserine (9.0g, 62.7%) was obtained. After two recrystallizations from methanol, the product melted with decomposition at 185*C.
Example 2:
threo-beta-m-hydroxyphenylserine (X)
A solution of 14.37g (0.05mol) of the product from Example 1in 50ml of 2N ammonium hydroxide and 25ml of methanol was hydrogenated at 1atm using 1.0g of 5% Pd/C. After 24 hours, the theoretical amount of hydrogen had been absorbed and the catalyst was removed by filtration. Water (100ml) was added and the solution was concentrated to a small volume. The mixture was neutral and 7.9g (79.7%) of threo-beta-m-hydroxyphenylserine had crystallized, MP: 215*C with decomposition. After two precipitations from alkali (with acid), the product melted with decomposition at 225*C. This compared well with the literature. The threo isomer was obtained almost exclusively.
Example 3:
threo-beta-p-benzyloxyphenylserine (VIII)
To a solution of 5.61g (0.1mol) of potassium hydroxide and 3.75g (0.05mol) of glycine in 75ml of absolute alcohol, there was added a solution 21.2g (0.1mol) of p-benzyloxybenzaldehyde15 in 50ml of alcohol. The mixture was warmed until it was absolutely clear. (ca. 60) and the solution was allowed to stand at RT. A crystalline solid precipitated which was filtered after 5 hours and washed thoroughly with alcohol and ether. This product was the potassium salt of N-p-benzyloxybenzylidene-beta-p-benzyloxyphenylserine and it was obtained in 95% yield, MP: 161-163*C.
The potassium salt (41.6g, 0.08mol) was stirred vigorously with 400ml of 1N HCl acid and the mix was filtered immediately. The solid p-benzyloxybenzaldehyde was washed with 200ml of 0.5N HCl acid and then with water. It weighed 19.7g (theory 17g) and probably contained some beta-p-benzyloxyphenylserine. The acid filtrates were combined and, on neutralization, beta-p-benzyloxyphenylserine crystallized, 17.7g (77%). A sample, recrystallized from a mixture of alcohol (6 parts), water (2 parts) and DMF (2 parts) melted at 190-192*C.
Example 4:
threo-beta-p-hydroxyphenylserine (XI)
To 100ml of 0.5N sodium hydroxide there was added 4.0g of threo-beta-p-benzyloxyphenylserine and 0.25g of 5% Pd/C catalyst. The benzyloxy compound did not dissolve completely. The mixture was hydrogenated at 1atm and, when absorption of hydrogen ceased (305ml, theory 330ml), there was no undissolved material present except the catalyst. After removal of the catalyst by filtration, the filtrate was neutralized with HCl and then concentrated under reduced pressure to 30ml. On cooling, a crystalline product weighing 1.8g (65.5%) was obtained. This product was recrystallized from water and washed until chloride free, MP: 195-205*C/w/decomp.
Example 5:
beta-3,4-dibenzyloxyphenylserine (IX)
A warm (70*C) solution of 89.0g (0.28mol) of 3,4-dibenzyloxybenzaldehyde16 was added to a solution of 15.7g (0.28mol) of potassium hydoxide and 10.5g (0.14mol) of glycine in 140ml of alcohol. A crystalline solid formed rapidly but this soon dissolved and an oil separated from solution. After standing overnight at 0*C, the alcohol was decanted from the viscous oil.
The oil was dissolved in 350ml of carbon tetrachloride and the clear solution was acidified with 15 ml of glacial acetic acid. Potassium acetate was removed by filtration and the filtrate was stirred vigorously with 500ml of water for 2hours. The water was decanted and the carbon tetrachloride was washed 4x500ml of water. The carbon tetrachloride slurry was allowed to evaporate in a flat tray and a dry yellow powder remained. After extraction of the powder with three portions of 500ml of ether, 30.5g (55.4%) of insoluble beta-3,4-dibenzyloxyphenylserine remained. Evaporation of the ether solution yielded 51g of 3,4-dibenzyloxybenzaldehyde.
The beta-3,4-dibenzyloxyphenylserine obtained by this method was a mixture of stereoisomers which were separated by fractional crystallization from 50% t-butyl alcohol. The high-melting (180-185*C) isomer was the more insoluble substance and it could be obtained in almost pure state by eluting the more-soluble, low-melting (146-147*C) isomer from the mixture with minimal quantities of boiling 50% t-butyl alcohol. Fairly pure isomer melting at 146-147*C could be recovered by cooling the hot t-butyl alcohol extracts to RT. The original crude mixture consists of about 30% of isomer melting at 146-147*C. The high-melting isomer was obtained pure by recrystallizing it from 50% t-butyl alcohol (1 L. for 8g), MP: 180-185g.
Example 6:
beta-3,4-dihydroxyphenylserine (XII)
from High-melting beta-3,4-dibenzyloxyphenylserine
A suspension of 3.94g (0.01mol) of beta-3,4-dibenzyloxyphenylserine (MP: 180*C) in 50ml of 50% methanol was hydrogenated at 1 atm using 0.2g of 5% Pd/C. In the course of the reduction, it was necessary to add 0.2g of fresh catalyst. After eight hours, 465ml of hydrogen had been absorbed (theory 480ml) and the reduction was stopped. Product had crystallized but it redissolved on warming and the catalyst was removed by filtration. The alcohol was removed by vacuum concentration and the aqueous solution was treated with acid-washed Darco. The almost colorless filtrate was concentrated to 5ml and 15ml of ethyl alcohol was added slowly. After standing at –20*C for 18hours, the crystalline product was collected and washed with 50% alcohol. It weighed 1.83g (85.9%) and melted with decomposition at 199-200*C. Analyses showed that the compound was hydrated.
Example 7:
from Low-melting beta-3,4-dibenzyloxyphenylserine
A solution 12.22g (0.03mol) of beta-3,4-dibenzyloxyphenylserine monohydrate (MP: 146*C) in a mixture of 30ml of water, 30ml of ethanol and 15ml of 2N LiOH was hydrogenated at 1 atm using 2.0g of 5% Pd/C. After 1490ml of hydrogen had been absorbed (1470ml theory). the reduction ceased and 6ml of conc. HCl acid was added to the reaction mix. After filtration through acid-washed Darco, the light yellow filtrate was neutralized with 2N LiOH (22ml). Crystallization was rapid and after cooling the mix at 0*C for 24hours, 5.94g (93%) of product was obtained. Due to the insolubility of the compound, it was purified by dissolving it in alkali and then reprecipitating by the addition of acid. The product had a gray tinge and melted with decomposition at 220-225*C.