F. IRREVERRE, K. MORITA,A. V. ROBERTSON, AND B. WITKOP
2824
added t o a boiling solution of 22 g. of sodium sulfate in 15 ml. of water and 15 ml. of concentrated sulfuric acid. This solution was steam distilled and the distillate made alkaline with sodium hydroxide and extracted with ether. The aqueous fraction was then acidified and extracted with ether. Evaporation of the ether and vacuum distillation of t h e residue gave 1.2 g. of phenol, b.p. 76-80" (2.5 m m . ) . The product solidified on cooling and was recrystallized from benzene-ligroin; m.p. 63.5-64.5'. Anal. Calcd. for C7HjFaOS: C , 43.30; H , 2.60. Found: C, 43.1;; H , 2.55. The phenol was converted to the phenoxyacetic acid as usual. After recrystallization from water and vacuum sublimation, it melted a t 91.5-92.5'. Anal. Calcd. for C S H ~ F ~ O J C S :, 42.85; H , 2.80. Found: C , 42.56; H , 2.74. 3-Trifluoromethylsulfonylphenoxyacetic Acid.-Our attempts t o oxidize 3-trifluoromethylthionitrobenzene t o the sulfone according t o Yagupolsky and mare net^^^ gave negligible yields. If, however, enough glacial acetic acid was used t o increase the solubility of the sulfide in the chromic acid solution, good yields were obtained. In a typical run, 21 g. of 3-trifluoromethylthionitrobenzene was placed in a solution of 16 g. of concentrated sulfuric acid, 25 ml. of water, 10 ml. of glacial acetic acid, and 16 g. of CrO3. T h e mixture was refluxed for 50 hr., then poured into ice water and extracted with ether. Evaporation of the ether and vacuum distillation of the residue gave 22.5 g., b.p. 108-112" (1 mm.). After two recrystallizations from ethanol, 16 g. of sulfone of melting point 55-56' was obtained. The 3-trifluoromethylsulfonylnitrobenzene was reduced to the amine and this in turn converted t o the phenol. In a solution of 6 . 3 g. of concentrated sulfuric acid and 22 ml. of water, 6.3 g. of 3-trifluoromethylsulfonylaniline was dissolved. Diazotization was completed with 2.3 g. of sodium nitrite in T ml. of water. T h e diazonium solution was then added dropwise t o a boiling solution of 6 g. of sodium sulfate in 45 ml. of water and 45 ml. of concentrated sulfuric acid. The mixture was steam distilled and the distillate extracted with ether. T h e ether was extracted with sodium hydroxide solution which was then treated with Korit. T h e alkaline solution was acidified and extracted with ether. Evaporation of the ether and distillation of the residue gave 5.5 g., b . p . 118' (1 m m . ) . After crystallization from benzene-hexane, the product melted a t 72-
73".
[ COXTRIBUTION FROM
THE
Vol. 85
Anal. Calcd. for C ~ H S F ~ O ~C,S :37.18; H , 2.23. Found: C, 36.90; H , 2.61. The phenol was converted to the phenoxyacetic acid, which, after recrystallization from water, melted a t 101-102". The yield was very low because of hydrolysis of the -S02CF, function. Anal. Calcd. for CsHiFaOjS: C , 38.03; H, 2.54. Found: C , 38.33; H , 2.73. 3-Trifluoromethylphenoxyaceticacid was made from a commercial sample of the phenol. After recrystallization from water, it melted a t 94.5-95.5'. Anal. Calcd. for CsH~F303: C , 49.10; H , 3.21. Found: C , 48.89; H , 3.46. 3-Trifluoromethoxyphenoxyacetic Acid.-A simple of 3-trifluoromethoxyphenol was kindly supplied by Dr. W. .4. Sheppard. This was converted as usual to the phenoxyacetic acid. ,liter recrystallization from water, this substance melted a t 89-90 . Anal. Calcd. for CsHiOaFs: C , 45.77; H , 2.99. Found: C, 46.07; H , 3.27. 4-Trifluoromethoxyacetic acid was prepared by Dr. P. E. Aldrich of d u Pont. Its melting point was 87-89'. 3-n-Propylphenoxyacetic Acid.-3-n-Propylphenol was converted as usual t o the phenoxyacetic acid. After recrystallization from water and sublimation, this substance melted at 70.5- 7 2 " . Anal. Calcd. for C11Hla03: C, 68.02; H , 7.27. Found: C, 67.85; H , 7.23. 3-n-Butylphenoxyacetic Acid.-3-n-ButylphenoI u'as converted as usual to the phenoxyacetic acid. After recrystallization from hexane and subsequent sublimation, the substance melted a t 76.5-77 .5 O . Anal. Calcd. for ClaHlaOs: C , 69.21; H , 7.74. Found: C , 68.90; H , 7.78.
Acknowledgment.-This work was supported in part under Research Grant Ghl 07492-04 from the National Institutes of Health. \Ye are grateful t o Professor R. Nelson Smith for many helpful discussions during the course of this work. We also wish to thank hlargaret Martin for assistance in the data processing.
SATIONAL IXSTITUTE OF ARTHRITIS A N D METABOLIC DISEASES,NATIONAL INSTITUTES OF HEALTH, BETHESDA14, MD.]
Isolation, Configuration, and Synthesis of Natural cis- and trans-3-Hydroxyprolines BY F. IRREVERRE, K.
MORIT.4,'
A. lr.ROBERTS ON,^
AND
B. WITKOP
RECEIVED M A Y2, 1963 3-Hydroxy-~-prolinehas been isolated from Mediterranean sponge a n d identified in hydrolysates of collagen of various sources. In addition, a diastereoisomeric 3-hydroxy-~-prolinehas been found in the antibiotic telomycin. Hydroboration and oxidation of the boron complex of r\;-carbobenzyloxy-3,4-dehydro-~~-proline methyl ester gave, after saponification and hydrogenolysis, about 70% uniform 3 - h y d r o x y - ~ ~ - p r o l i nbesides e 107% tians-4-hS.droxy-DL-proline. From the absence of 4-allohydroxy-DL-proline and t h e homobeneitg of the synthetic 3 - h y d r o x y - ~ ~ - p r o l i nite , is concluded t h a t the hydroboration is stereospecific and trans with respect to the carboxyl function. The X-tosylmethyl esters of the synthetic cis- a n d tians-3-hydroxy-~~-proline and of the natural amino acids from sponge and telomycin, respectively, had identical infrared spectra. cis-3-Hydroxy-DL-proline was prepared via the 3-ketoproline derivative IY by reduction with sodium borohydride and was found t o be identical with 3-hydroxy-~-proline from teloinycin with regard t o column, paper, and gas chromatographic analysis. This procedure byas also used for the preparation of selectively tritiated c i s - 3 - h y d r o x y - ~ ~ proline-3-H3. Enzymatic studies with o-amino acid oxidase and n.1n.r. d a t a confirmed these assignments.
Previous preliminary from this Laboratory have described the isolation, characterization, and synthesis of the two diastereoisomers of &hydroxyproline. The tvans isomer was isolated from dried Mediterranean sponge and from the antibiotic telomycin. The cis isomer was obtained only from telomycin. Independent report^^,^ from other laboratories have also described the isolation and synthesis of 3-hydroxyprolines. K O assignments of configuration have been made. (1) Associate in the Visiting Program of the USPHS, 1961-1963.
(2) Associate i n thP Visiting Program of the USPHS, 1959-1961 ( 3 ) F Irreberre, K hlorita, A V Robertson, and B U'itkop, Biochem. F i i o p h y A Res. C ' o i i r i i i i i n , 8, 4.53 (1962). ( I ) F Irreverre, K . hIorita, S. Ishii, and B. XT-itkop, ibid., 9 , 69 (1962). J C Sheehan and J . G. Whitner, J . A t n . C h e w Soc , 84, 3980 (1962) (6) J D Ogle, R B Arlinghaus. and 51 A . Logan, J Bioi Chem , 237, 3667 (1962)
This paper gives detailed account of the work previously reported and presents additional information on new synthetic approaches. The key step in the isolation from acid hydrolysates of sponge was the separation of the cyclic amino acids from the primary amino acids by either nitrosation and hydrolysis7,8or by treatment of the hydrolysates with 2,4,6-trinitrobenzenesulfonicacid (TNBS) and ion-exchange chromatography'" on IR-120 resin of the fraction containing the cyclic amino acids. The transand cis-3-hydroxyprolines were separated quantitatively b y this method and were easily further purified by recrystallization from aqueous ethanol. (7) P B. H a r i l t o n and P J Ortiz, ibid , 184, 607 ( l a i n ) ( 8 ) B. Wi tkop and C M . Foltz, J . A m Cheln. Soc , 79, 19%(19.57) (9) T O k u y a m a and K Satake, J Biochem. (Tokyo), 41, 454 (1960) (10) D H Spackman, W . H . Stein, and S Moore, A ~ t a l C . h e n i . , 30, 1190 (1958).
Sept. 20, 1963
2825
NATURAL Cis- A N D trUnS-3-HYDROXYPROLINES
Isolation of truns-3-Hydroxy-~-proline from Mediterranean Sponge. Materials.-The sponge was a dried Mediterranean sponge bought from Yadgi and Co., Inc., Washington, D. C. The resin used for chromatography was hmberlite IR-120, particle size 47-65 p . The column dimensions were 2 in. X 36 in.; the amount of resin was 1725 ml. Sodium citrate buffer, 0.2 N , p H 3.25,1° was used for elution and was pumped into t h e column with a peristaltic pump (Sigmamotor, Middleport, N . Y.). The automatic amino acid analyzer used was from Phoenix Precision, Instrument Co., Philadelphia, P a . Hydrolysis.--.% piece of sponge after it was washed with distilled water and dried weighed 46.1 g.; it was refluxed with 750 HC1 for 24 hr. Variations in the time of hydrolysis, ml. of 6 such as 7 , 8, 10, and 24 h r , had practically no effect on t h e yield of 3-hydroxyproline from sponge. The hydrolysate was decolorized with charcoal. The filtrate was evaporated t o a sirup and redissolved in 250 ml. of water. This material was used for the preparation of N-nitroso derivatives of the cyclic amino acids as well as for reaction with trinitrobenzenesulfonic acid ( T X B S ) . N-Nitrosation of the Secondary Amino A c i d ~ . ~ ~ ~ - - F iml. f t y of the above hydrolysate was brought t o a volume of 100 ml. by the addition of hydrochloric acid so t h a t the final acidity was 1.0 1V. T o the solution, immersed in a boiling water bath, was added 14 g. of sodium nitrite and the reaction allowed t o proceed for 5 min. The flask was then cooled and the solution brought t o p H 1.0 by the addition of alkali. The nitrosamino acids were extracted with a total of 125 ml. of ethyl acetate in 5 portions of 25 ml. The combined ethyl acetate extracts containing the Iinitroso derivatives of 3- and 4-hydroxyprolines and proline were dried with anhydrous Na2SOI. Hydrolysis of the N-Nitrosamino Acids.-The dry ethyl acetate extracts were evaporated t o dryness in vacuo, redissolved in 6 N HCI, and, after addition of 2.4 g. of ammonium sulfamate, heated for 90 min. a t 10511Oo. Excess acid was removed in vacuo. The residue was dissolved in 50 ml. of H 2 0 adjusted t o p H 2.5 t o 3.0 and desalted on a column (1.25 in. X 8 in.) of Dowex-50W, X-4, 50-100 mesh, in the acid form. The amino acids were eluted from t h e column with 7.0 N ammonia. The eluate was evaporated t o dryness in zacuo at 25". The residue was dissolved in T ml. of water and t h e solution adjusted t o p H 2.5 with 0.5 N HCI. After filtration t h e clear solution (about 10 ml.) was used for column chromatography. Trinitrobenzenesulfonic Acid.-The details of t h e preparation of trinitrobenzenesulfonic acid ( T N B S ) were obtained from Prof. Kazuo Satakeg through D r . S. Ishii: T o a stirred solution of 40 g. (0.16 mole) of picryl chloride in 120 ml. of methanol, maintained a t a temperature of 5 5 6 0 " , was added 18 g. of sodium bisulfite (NaHS03). After 30 min. a second batch of 18 g. of bisulfite was added and stirring continued for a n additional 90 min. The insoluble salts were collected, the filtrate adjusted to pH 5.0 with 0.5 N HCl, decolorized with charcoal, and evaporated t o dryness in vacuo. At this stage the T K B S could be isolated as t h e acid or as t h e sodium salt. For t h e isolation of trinitrobenzenesulfonic acid; the residue from t h e above filtrate was dissolved in 400 ml. of 3.0 N sulfuric acid and extracted with ether in a liquid-liquid extractor. T h e ether layer was taken down t o dryness and the residue crystallized from hot 1.0 N HC1 (charcoalj, giving 10 g. (200, yield) of pale yellow crystals. Recrystallization from 1.0 iV HCI yielded t h e colorless monohydrate, m.p. 193-194" (loss of water a t 100"). Sodium Trinitrobenzenesu1fonate.-The crude trinitrobenzenesulfonic acid from t h e bisulfite reaction above was dissolved in 300 ml. of water and filtered. T o t h e filtrate was added a n equal volume of saturated sodium chloride solution. The sodium salt which crystallized in t h e cold overnight was collected, washed with saturated salt solution, and dried in the desiccator over PzOS. There was obtained 27.4 g. (54% yield). Trinitrophenylation of the Primary Amino Acids.-To 25 ml. of sponge hydrolysate was added 13.3 g. of T S B S (sodium salt) equivalent t o 1.5 times the total nitrogen contents of t h e sample. The p H was adjusted t o 7.5 t o 7.8 over a period of 30-45 min. with 20% aqueous sodium carbonate. The reaction mixture was allowed t o stay a t room temperature overnight. The next day the trinitrophenylamino acids were removed by filtration. T h e filtrate was desalted on a column ( 6 in. X 2 in.) of Dowex-1 acetate. The eluate and water washings were collected and passed through a column of Dowex-50\&' in the acid form. The secondary amino acids were eluted from the column with 7.0 N ammonia and the solution evaporated t o dryness in vacuo at room temperature. The residue x a s used for column chromatography. Column Chromatography.-The material t o be chromatographed was dissolved in citrate buffer pH 2.28 and poured on a column of IK-120 a s described previously. The pump was adjusted t o deliver 180 ml./hr. of citrate buffer p H 3.25 and the fraction collector adjusted t o collect 14 rnl./tube. Under these conditions 3-hydroxyproline was found in tubes 82-95, 4-hydroxyproline in 1 0 5 1 3 8 , and proline in 206-245. The secondary amino acids were detected in t h e tubes by putting 10-pl. aliquots/ tube t o a piece of filter paper which mas first dipped into a 0.25y0
ninhydrin solution in acetone and then steamed for 2 min. The spots for the secondary amino acids when viewed under an ultraviolet light exhibit a characteristic reddish fluorescence. When t h e chromatographic separation was performed with the total hydrolysate from sponge untreated by nitrous acid or trinitrobenzenesulfonic acid, the large amount of aspartic acid interfered with t h e separation and identification of 3-hydroxyproline. The tubes containing the 3-hydroxyproline were pooled and desalted on a column of Dowex-BOW in the acid form. The amino acid was eluted from the column with 7.0 X ammonia and the elution was vaporated in vacuo. The combined material from 11 column runs, representing 64 g. of sponge, was dissolved in water, decolorized with charcoal, reduced t o a volume of 0 . 5 1.0 ml., brought t o a volume of 10 nil. by the addition of absolute alcohol, and kept in t h e refrigerator a t -10' for several days. The crystals (62 mg.) were collected, washed wirh cold absolute alcohol, and dried in vacuo over PzOS. T h e analytical sample was recrystallized a second time, m.p. > 200" dec.; specific rotation: [ C U ] ~ ' D-17.4 f 1.0' ( 6 1.0 in Hzo), [a]*'D +13.3 f 2.0' ( 6 0.5 in 1 A' HCI). Anal. Calcd. for C6H9NO3: C , 45.80; H , 6.92; S , 10.68. Found: C , 45.82; H,6.74; N , 10.46. With ninhydrin reagent on paper the amino acid gives a yellow color which turns t o brown when the paper is exposed to steam. U'hen viewed under an ultraviolet light t h e spot exhibils a brickred fluorescence which is characteristic of the cyclic secondary amino acids. 3-Hydroxyproline gives a bluish color with isatin reagent. I t reacts with 1,2-naphthoquinone-4-sodiumsulfonate" t o give a reddish orange color characteristic of the cyclic secondary amino acids. On t h e automatic amino acid analyzer" the peak due t o natural (trans) 3-hydroxyproline appears after 95 ml. of effluent (Fig. 4) a t 30-50". The ninhydrin color has a maximum absorption a t 440 mp, similar t o 4-hydroxyproline and proline. O n a 2-dimensional paper chromatogram12 it occupies a position directly on top of natural 4-hydroxyproline (ref. 12, Fig. 5). Preparation of Trichloroacetic Acid (TCA )-Soluble Collagen from Sponge.-Dried Mediterranean sponge (1.8 g . ) was extracted with 0.3 M trichloroacetic acid in a water bath a t 9U-100" for 30 min.13 The TCX extract was put inside a cellophane bag and dialyzed exhaustively in the cold against distilled water for 2 days with several changes of distilled water. The dialyzed material was lyophilized; total S = 15.0"/l. The sedimentation pattern of the TCA extract of sponge indicated t h a t t h e material is monodispersed in t h e ultracentrifuge in 0.15 M citrate buffer a t p H 3.65. The analysis of the hydrolysate gave values differing from those of the total sponge hydrolysate t h a t had been pretreated with trinitrobenzenesulfonic acid (Table
I). TABLE I C Y C L I C AMIS0
ACID ANALYSES O F MATERIALS SOURCES (RESIDCES/1O5 G . )
Cyclic amino acids
Bovine collagen
FROM \'ARIOCS
TSBS treated Sponge hyCell colladrolys- wall Rat gen ate of a n skin (TCA- from thrax colla- s o h total bacilgen ble) sponge Iuso
trans-3-Hydroxy-~-proline 2 2 11 6" 11 4-Hydroxy-~-proline 63 68 78 47 L-Proline 97 119 49 39 2 a In addition t o six residues of tvans-3-hydroxy-~-proline about one residue of a different hydroxyproline (presumably 4-allohydroxy-u-proline resulting from acid-catalyzed epimerization of 4-hydroxy-~-proline) was found. * The cell wall of anthrax bacillus was obtained from D r . L. Mester, Institut de Chime des Substances Naturelles, Gif-Sur-Yvette, France. Isolation of cis-3-Hydroxy-~-prolinefrom the Antibiotic Telomgcin.-One gram of t e l ~ m y c i n was ' ~ hydrolyzed with 100 ml. of 6.0 N HC1 a t 1 0 5 1 1 0 ° for 24 hr. The hydrolysate was decolorized with charcoal and concentrated t o a sirup repeatedly in aacuo. I t was dissolved in 50 ml. of water, neutralized by t h e addition of 5.0 N NaOH, and treated with a solution of 2.3 g. of T N B S (sodium salt) in 10 ml. of water. The reaction mixture was (11) D M u t i n g , Saiuvze~iss, 99, 303 (1952); K V Giri a n d A N a g a bhushanam, ibid., 39, 548 (1952). (12) F . Irreverre a n d W. Martin, A n a l . C h e m . , 26, 267 (1954); K A Pier, F Irreverre, and H L . Wolff, J . Bioi Chem., 229, 687 (1906) (13) S . M . Fitch, M L . R . Harkness, and R . D . Harkness, X a t u r e , 176, 163 (1955). (14) M . Misiek, 0. B . Fardig, A Gourevitch, D L Johnson, I . R . Hooper, and 1. Lein, Aviibiot. A n n . , 852 (1957-1958).
F. IRREVERRE, K. MORITA,A. V. ROBERTSON, A N D B. WITKOP
2826
maintained a t p H 7.5-7.8 by the addition of 20% sodium carbonate solution. When (after 1-2 hr.) t h e p H no longer changed, t h e reaction mixture was allowed t o stand in the dark a t room temperature overnight. The yellow precipitate of the trinitrophenylamino acids was removed by filtration and washed with water. The combined filtrate and washings were adjusted t o a p H of 3.0 with 1.0 N HCI a n d desalted on a column (2 in. X 6 i n . ) of Dowen-1, X-4 resin in the acetate form. The effluent and washings, about 370-450 ml., were adjusted t o p H 3.0 and desalted on a column (1.25 i n . X 9 in.) of Dowex-50W, X-4 resin in the hydrogen form. T h e cyclic secondary amino acids were eluted with i . 0 iV ammonia a n d the solution evaporated in z m u o a t room temperature. The TNBS-treated hydrolysate from 2 g. of telomycin in a volume of 5.0 ml. of water was chromatographed on a column of IR-120 following t h e same technique as t h a t used for the isolation from sponge. The fractions containing the trans-3-hydroxyproline were located in tubes 78-111, while t h e cis isomer was contained in tubes 169-190. The respective fractions of t h e two isomers were pooled, passed through a column of Dowex-BOW [ H + ] , eluted with ammonia, and evaporated t o dryness. Recrystallization of the solids from aqueous ethanol (containing 357, water) yielded, respectively, 62 mg. of trans-3-hydroxy-~-proline and 70 mg. of cis-3-hydroxy-~-proline. Further recrystallization of the two isomers from aqueous ethanol yielded analytically pure samples. trans-3-Hydroxy-~-proline: [ ~ ] ' O D - 15.3 f 1.0" ( c 1.00 in H'O), [,]?OD +17.4 f 2.0" ( c 0 . 5 0 i n 1.0 s H C 1 ) . Anal. Calcd. for C S H ~ S O SC, : 45.80; H , 6.92; AT, 10.68. Found: C, 45.88; H,6.99; Iv, 10.60. The physical and chemical properties and the color reactions were identical with those given by trczns-3-hydroxy-~-proline from sponge. cis-3-Hydroxy-~-proline: [ ~ ] * O D -91.5 f 1.5' ( c 0.61 in H20). When t h e same solution was analyzed by a Rudolph recording spectropohrimeter it gave [a] 2 0 ~-90.2" (in water), [ a ]2 0 -54.3 ~ i 2.0" ( c 0.50 in 1 A' HCI). Anal. Calcd. for CsHQN03: C , 45.80; H , 6.92; S , 10.68. Found: C,45.68; H , 6 . 6 8 ; X , 10.64. The cis isomer gave t h e same color reactions with ninhydrin, isatin, and 1,2-naphthoquinone-4 sodium sulfonate as t h e tvans isomer. On the automatic amino acid analyzer the peak due t o the c i s isomer overlapped t h a t of threonine. T h e yellow color with ninhydrin absorbs maximally a t 440 m w .
Ia, R=R'=H b, R = Cbz, R'=CH3
-I
-.
68%
-5
n
Cbz
R
R
IIa, R = Cbz, R'= Me
IV NaBH4 80-90%
I11
b, R = R'=H
Brown and Zweifel.I7 The resulting alkylborane was oxidized with alkaline hydrogen peroxide, and the product was hydrogenolized to a mixture of isomeric 3- and 4hydroxyprolines and separated by column chromatography. On the automatic amino acid analyzer1* the reaction mixture was resolved into 68YG3-hydroxyproline, 10% 4-hydroxyproline, and a trace of 4-allohydroxyproline. T h e other diastereoisomer of 3hydroxyproline was not detected on the amino acid analyzer which separates the diastereoisomeric 3- and 4-hydroxyprolines accurately. alternate procedure was discovered when conditions for the oxidation of the alkylborane were used which did not lead to saponification of the ester. In that case, the mixture of the N-carbobenzyloxy-trans3- and -4-hydroxy-m-proline methyl esters was resolved on a silicic acid column. This procedure offers preparative advantages. The production of only the trans diastereoisomers in this synthetic procedure extends Brown's observations on the stereospecificity of the hydroboration reaction to the heterocyclic series. The carbomethoxy group in I not only directs the approach of the boron to the less hindered trans side but also preponderantly to the 3-position. Apparently the carbomethoxy group exercises both a steric and inductive control leading to a preferred polarization of the 3,4-double bond. I t has been pointed out previously t h a t this 3,4-double bond in dehydroproline and derivatives is stable to reagents known to isomerize double bonds.lg There is again no indication of noticeable double bond isomerization under the conditions of hydroboration. The product expected from hydroboration (or hydrolysis) of such a rearranged olefin, v i z . , 2,3-dehydroproline, would be &amino-a-ketovaleric acid. The trans assignment to the major product from the hydroboration is further strengthened by the sequence which of reactions leading to cis-3-hydroxy-~~-proline proceeded in analogy to the preparation and reduction of 4-keto-DL-proline. The methyl ester of trans-Ncarbobenzyloxy-3-hydroxy-~~-proline (IIIa) in acetone on oxidation with 8.0 ;V chromium trioxide solution in 50% sulfuric acid gave the 3-oxo compound IV, which was obtained as a chromatographically pure oil after fractionation on a silica gel column. By the criteria of gas, as well as thin-layer chromatography, compound I'c' proved to be homogeneous. The infrared spectrum of IV contained a band a t 5.63 I.( characteristic of five-membered ring ketones. Reduction of the ketone IV with a large excess of sodium borohydride in dimethoxyethane in the absence of water yielded a mixture (about 4 : 1) of cis (1') and trans (IIa) isomers as shown by gas chromatography. Separation was possible a t the stage of the X-carbobenzyloxy esters I I a and 1' by fractionation on a silicic acid column and elution with benzene-ethyl acetate (4 : 1). In 50% aqueous dimethoxyethane the amount of trans isomer decreased to less than 10% as assayed by both gas chromatography and separation on silicic acid. Inspection of space-filling models suggests that only two relatively unhindered conformations, VI1 and VIII, exist for the ester function which retain trans geometry about the ester C-0 bond.*l In one of these
aooCH3 2, 1,W-C, H o e HI *
N
O H o o H N H
I
Cbz
VI
V
The synthesis of one diastereoisomer of 3-hydroxym-proline was accomplished by stereospecific hydroboration of the methyl ester of X-carbobenzyloxy3 , 4 - d e h y d r o - ~ ~ - p r o l i (I) n e Ififollowing the procedure of \
(15) H C Broun V 1562 p 110
Hydroboration
\V A BenJamin Tnc
\ e u York,
Vol. 85
(16) N . Izumiya. J . E . Francis, A . V. Robertson, and B LX'Vitkop, J . A m Cheni Soc., 84, 1702 (1962). (17) H . C . Brown and G Zweifel, i b i d . , 83, 2814 (1961) [lS) D. H Spackman, \V H Stein, and S Moore, Anal C h e m , 30, 1190 (1958) (19) A V Robertson, J E Francis, and B W i t k o p , J A m Chem S a c , 84, 1709 (1962). (20) A . A . P a t c h e t t and B Witkop, i b i d . . 7 9 , 179 (1957). ( 2 1 ) C',f R. Huisgen and H O t t , Terrohedvoii, 6 , 2 5 3 (1959). All r-
Sept. 20, 1963
NATURAL Cis-
AND
I
VI1
2827
trans-3-HYDROXYPROLINES I
I
IO0
150
VI11
conformations (VII), the borohydride ion is repelled electrostatically by the negative charge of the ester carbonyl; in the other (VIII), the bulk of the methyl group shields the ketonic carbonyl from attack cis to the ester. In aqueous media2' the size of the boron hydride ion is probably increased by conversion to hence the cis product predominates more [HOBH3]e,z3 than in dimethoxyethane where the reducing species is [BH4Ie. In addition, the presence of a basic or nonbasic nitrogen affects the steric course of reduction of ketoprolines and ketopipecolic acid derivativesz4 in a manner not fully rationalized. The difference in the reduction of 4-ketoprolineZ0 and 3-ketoproline raises also the question of enolization. In our first experiments the mixture of V and I I a was converted to the free amino acids by saponification and hydrogenolysis and analyzed by ion-exchange technique. Ratios approaching 1 : 1 were then ~ b s e r v e d . ~However, in this case the complete oxidation of the starting material was in doubt, In addition the effect of base in the process of saponification is not known. The cis isomer is saponified in about one-third the time of the trans isomer and may not be stable to the prolonged action of base. LVhen the reduction of the 3-ketoproline derivative IV was carried out with sodium borohydride-H3 followed by alkaline hydrolysis, hydrogenolytic removal of the carbobenzyloxy residue, and-column separation, were obtained cis- and trans-3-hydroxy-~~-proline-3-H~ which were utilized for incorporation studies into the peptide moiety of actinomycin analogs. 25 I t was ascertained by gas and column chromatography that no interconversion took place between the cis and trans isomers under conditions of hydrolysis, viz.,on prolonged treatment (24 hr.) with mineral acid or alkali. Destruction by acid under these conditions does not exceed The synthetic and natural cis-3-hydroxyprolines proved to be inseparable b y paper chromatography in five solvent systems (Table I I ) , by ion-exchange chromatography, and by high voltage paper electrophoresis. The infrared solution spectra of N-tosylcis- and -trans-3-hydroxy-~~(synthetic) and -L-prolines (from sponge and telomycin) were identical. Enzymatic Oxidation of 3-Hydroxyprolines.-The absolute configuration of the amino acids was determined by measuring the oxygen consumption of the natural and synthetic trans- and cis-3-hydroxyproline in enzymatic dehydrogenation with D-amino acid oxidase. Figure 1 shows that the synthetic substrates and D-alanine consumed almost the theoretical amount carbobenzyloxymethyl ester derivatives of cas- and trans-hydroxyproline, 3-ketoproline and 3,4-dehydroproline show a characteristic doubling of t h e T h e two components are of about 0 - m e t h y l group in t h e n m r . spectra equal intensity and cannot be due t o splitting because the separation is solvent dependent (8-9 c.,/sec. in deuteriochloroform, 4-5 c./sec. in acetonitrile a t 60 .\IC ,set, 40' probe temperature). Temperature dependence studies are in progress t o prove t h e existence of a conformational equilibrium due t o restricted rotation either a t the ester carbonyl a n d l o r t h e amide carbonyl. ( 2 2 ) Cf. H 0 House, H. Babad, R B . Toothill, and A W h701tes, J . Oi,g Chein , 27, 1141 (1962). ( 2 3 ) H C . Brown and K . Ichikawa, Tetrahedron, 1, 214, 221 (1957) (24) A . V. Kobertson, E . K a t z , and B Witkop, J . Org. Chem., 27, 2676 (1962) ( 2 5 ) C j E K a t z and H Weissbach, J . Bioi. Chem., 238, 666 (1963).
50
TIME (min.).
Fig. 1.-Oxidation of various secondary cyclic amino acids by D-amino acid oxidase in relation t o malanine: A, D-alanine; B, 4-al1o-hydroxy-~-proline;C , cis-3-hydroxp-DL-proline; D, trans3 - h y d r o x y - ~ ~ - p r o l i n E, e; 3,4-dehydro-~~-proline.
of oxygen (based on the D-form), whereas the natural cis- as well as trans-3-hydroxyprolines and L-alanine consumed no oxygen a t all. TABLE I1 A-ATURAL AND SYSTHETIC cas- ASD trens-3HYDROXYPROLISES (Paper chromatography, descending technique)
R f VALUESOF
THE
Rf
Solvent system
--trans-----. NatuSynral" thetic (I.) (DL)
i--c~s----
Satural
Synthetic
(L)
(DL)
Methanol-pyridine-HzO 0.39 0.39 0.35 0.35 (20 : 1 : 5 ) l-Butanol-pyridine-H20 (1 : 1:1) .36 .36 .35 35 Ethanol-formic acid-H*O ( 1 2 :3 : 5 ) .57 .57 .62 61 t-Amyl alcohol-2,4-lutidine-H~0 6fjh .65' 43' 43b (178:178:110) &Butyl alcohol-formic acid-HS0 .35 .36 .29 .27 (70:15:15) 5 Both compounds from sponge and telomycin. Instead of the solvent front the Ri of valine is taken as a reference for the mobility of the 3-hydroxyprolines in this solvent system of high resolving power.
I t is clear, therefere, that the natural amino acids belong to the L-series, a fact which is further supported by-the rotational data (Table 111) which obey the rule of Lutz and Jirgensons.26 Oxidation of trans-3-hydroxy-~~-proline b y D- imino acid oxidase produces a dehydro compound, pi-esumably I X , whose catalytic hydrogenation yields mainly cis-3-hydroxy-~-proline,accompanied by proline (hydrogenolysis of the "allylic" hydroxyl). Oxidation of trans-3-hydroxy-~~-proline may also be effected b y the action of hydrogen peroxide and C u + + ions, .producing DL-IX, which, upon hydrogenation, forms cis-3(26) 0. Lutz and B. Jirgensons, Ber., 63, 448 (1930); cf. M. Winitz a n d J. P. Greenstein, J . A m . Lhem Soc , 77, 716 (1955).
F. IRREVERRE, K. MORITA,A. V. ROBERTSON, A N D R.I ~ I T K O P
2828
Vol. 85
,OH
COOH
H
OH Fig 2 -Preponderant disappearance of S-carbobenzyloxv-cis3-hydroxyproline methyl ester relative t o the trans compound on oxidation with chromic acid
hydroxy-DL-proline Sodium borohydride carries out this reduction more selectively and without the formawas tion of proline When czs-3-hydroxy-~~-proline oxidized with D-amino acid oxidase the resulting mixAPPLICATION OF
THE
TABLE I11 LUTZ-JIRGEZSOTS RULETO czs-
H The rate of oxidation of the two 4hydroxyprolines X and X I 2 9is less affected by the relative position of the
[ ~ I ~ O inD 1 0 Y [ ~ I ~ O in D
Hs0
s-3- hydroxy-$, 2-pyrroline-2-carboxylic acid
H Cl
~is-3-hydroxy--proline
trens-
AND
3-HYDROXY-L-PROLINE
Formula
B &3-hydroxy-~-proline
A
OH
/-(,COH -91 5 5t 1 5 " -54 3
+2O
+37 2"
H
HO
cis( from telomycin)
trails,
trans( from sponge)
ture was easily separable on a column of IR-120 ionexchange resin into cis-3-hydroxy-~-proline and s3-hydr0xy-A~~~-pyrroline-2-carboxylic acid. Reduction with sodium borohydride gave cis-3-hydroxy-~-prohe.
.OH
,OH acid oxidase
H lrans
N IX
_HN_ cis
T h i s method provides a n e w way f o r the preparation of c i s - 3 - h y d r o x y - ~ -a n d -L-proline in addition to the coneientional enzymatic resolution b y leucine a m i n o peptidase (hog kidney acylase). Conformational Considerations.-Relative rates of saponification of the trans-3-hydroxyproline ester, according to preliminary results obtained by gas chromatograpic analysis (see above), were higher than for the cis-ester." Similarly, when a mixture of the cis and trans isomers V and I I a was exposed to brief chromic acid oxidation, the analysis of the resulting mixture by gas chromatography (Fig. 2) as a function of time showed the preponderant disappearance of the c i s isomer V, in which nonclassical (Pitzer) stra,in between OH and COOR leads to steric acceleration, s (271 C i T C Bruice and T. H . Fife, J . A m C h e i i i . Soc , 84, 1973 (1962) 128) J Schreiher and A Eschenmoser. H e l v C h i m A d a , 28, 1529 (1953)
cis,
XI
carboxyl group. The results on oxidation of a mixture of the two hydroxyprolines reflect this situation (Fig. 3 ) . Although cis- or allohydroxyproline disappears somewhat faster, there is not the same striking difference between oxidation of the cis and trans isomer as with the two 3-hydroxyprolines. Incomplete permanganate oxidation gave glycine and (3-alanine in amounts not significantly different for cisand trans-3-hydroxy-~-proline (Table IV). I t was riot OXIDATION OF cis3-Hydroxy.~proline
COOH-
-D
__
.OH
H
X
TABLE 11 ~YU~S-HYDROXY-L-PROLINES TVITH AQUEOUS PERMAXGAXATE
ASD
Glycine,
yo
p- Hydrr,xk--L~ glutamic acidh
21-35"
Traces of threo (and f r y t h r o ) trans1 1-3 2 10-18 Trace of evythro The results reflect variations in procedure. The lower figures result from oxidations a t room temperature, the higher one a t 4 0 4 0 ' for 1-4 hr. ; in each experiment 8.6 pmoles of coinpound in 0.5 ml. of 1.0 S HzSOd was mixed with 0.1 1111. of 0.1 iV K M n 0 4 in 1.0 N H2S04. After the discharge of color the reaction mixture was desalted on Dowes-jO\V ( H +) and examined on an automatic amino acid analyzer. Cf.T. Kaneko a n d I< 170shida,Japanese Patent 2174 (1961); Chem. Abstr., 56, 8837 (1962). LVe are greatly indebted t o Prof. Kaneko for his kind donation of reference samples.
cis-
3.9-5.2
@-Alanine,Oc
possible to establish clearly the expected relationship between tram-3-hydroxy-L-proline and erythro-8-hydroxy-L-glutamic acid ( X I I I ) , on the one hand, and cis-3-hydroxy-~-prolineand threo-/3-hydroxy-L-glutamic acid ( X I I I ) , on the other hand, although only XI11 was observed (in traces) in the oxidation of cis-%hydroxyr,-proline. X more thorough investigation of this kind, e . g . , the exact relative rates of chromic acid oxidation of :3(20) R J Abraham and K A liIc1,auchlan. .Moi. P h y s
, 6 , ,513
I
101i2)
NATURAL CZS-
Sept. 20, 1963 COOH
COOH
COOH XI1
COOH XI11
2829
A N D t7'UnS-3-HYDROXYPROLINES
\H 2 N H
and 4-hydroxyproline derivatives by the methods of Westheimer30 and Kwart, 3 1 may permit certain conclusions with regard to the relative strain between functional groups, i.r., steric acceleration and, indirectly, their conformations in the five-membered ring system. The n.m.r. spectra of cis- and trans-3-hydroxyproline in D 2 0 show a clear difference with regard to the splitting of the peak due to the C-2 proton. Decoupling a t appropriate frequencies confirmed the assignment of peaks to individual protons. The C-2 proton peak in the cis compound was a doublet indicating Jz,a = 4 c.P.s., while that of the trans compound was an unresolved but broader peak, indicating J2,3 < 1 c.p.s. From thc general relationship between coupling constant and dihedral angles these findings offer support, in a qualitative manner, for the assignments of cis and trans structures independent of the stereochemistry of hydroboration. In a quantitative sense the numerical values for the coupling constants are much lower than expected on the assumption of a planar five-membered ring and applicability of the Karplus equation. There are two possible conclusions: either the Karplus equation is not readily applicable to this 5-membered ring system, or the pyrrolidine ring in 3- and 4-hydroxyprolines is considerably puckered, 2 9 a conclusion which has also been reached for cyclobutane derivative^.^^ The infrared spectra iO.005 -11in CC14) of trans-Ncarbobenzyloxy-3-hydroxyproline methyl ester (strong band a t 3610 cm.-' low half-intensity band width 22 cm.-l, slight shoulder a t 3580 crn.-l) and of cis-Ncarbobenzyloxy-3-hydroxyproline methyl ester (similar but much broader band a t 3603 cm.-', A y l , * = 58 cm.-'. shoulder a t h . 3 5 0 0 ) indicate a t best weak bonding of the hydroxyls to a n-electron center but do not provide further evidence for cis and trans assignments.
Experimental Benzyloxycarbonyl-3,4-dehydro-~~-proline Methyl Ester ( I b ) . Benzyloxycarbonyl-3,4-dehydro-~~-proline~~ (25.0 g., 0.9 mole) was dissolved in ether (50 ml.) and treated gradually with a n excess of diazomethane in ether during 30 min. Esterification proceeded rapidly and the excess of reagent was destroyed by dropwise addition of acetic acid. The reaction mixture was evaporated, leaving the methyl ester as a colorless oil (26.0 g., 997c) which has not yet been obtained crystalline. r? sample was pumped a t 80" ( 1 m m . ) overnight and analyzed without further purification. \.apor phase chromatography showed t h a t the oil was a t least 997; pure. The nuclear magnetic resonance spectrum contained only the expected peaks whose area agreed with the ester formulation Ia.33 A n a l . Calcd. for CIIH,jSO,: C, 64.36; H, 5.79; S , 5.36. Found: C, 63.49; H , 5 . 9 i ; S , 5.88. ( 3 0 ) F H Westheimer and N . Nicolaides, J A m C h e m . Soc., 71, 2 5 (1 9-19). 131) H . K n a r t and P S. Francis. ibid , 81, 2116 (19.59) (1(2) J .\I Conia, J -1- Ripoll. I.. A Tushaus, C L Neumann, and S I. A l l i n ~ e r ,: b i d 84, 4982 (1962) ( 3 3 ) T h e interesting splitting present i n the n . m I. spectrum will be discussed separately (I,. F. Johnson. A . V. Robertson. J N. Shoolery, and B Witkop, in preparation).
l 2
/
l 4
/
/ 6
,
l l I 8 2 MINUTES,
,
I 4
,
I 6
,
I
,
I
8
Fig. 3.--Almost identical rate of disappearance of N-carbobenzyloxy-cis-(allo)-4-hydroxy-and S-carbobenzylox~--trens-4hydroxyproline methyl esters on oxidation t o ?;-carbobenzylox>-4-ketoproline methyl ester with chromic acid.
A . Hydroboration with Saponification of the Ester I b . Benzyloxycarbony1-3,4-dehydro-~~-proline methyl ester ( I b , 10.5 g., 40 mmoles) was dissolved in dry tetrahydrofuran (50 m l . ) and treated with diborane (15 mmoles) a t 0'. The diborane was generated by the dropwise addition ( 1 h r . ) of a solution of sodium borohydride (0.84 9.) in diglyme (36 m l . ) to a magnctically stirred solution of boron trifluoride ethcrate (5.6 ml.) in diglyme ( 7 ml.).17 A4slow stream of dry nitrogen was uscd t o carry the diborane into the olefin solution. When the borohydride addition was complete, the diborane generator was heated a t 70-80' for 1 hr. and then disconnected. The olefin reaction mixture was left a t room temperature for another hour. Excess of diborane was decomposed cautiously with water and the alkylborane was oxidized by the addition of 3.0 AV sodium hydroxide (6.0 ml.) followed by 30% hydrogen peroxide ( 5 ml.) drop by drop (10 m i n . ) . During the peroxide addition, the temperature of the magnetically stirred reaction mixture was kept below 40" by using a bath of water a t room temperature. Then 14 ml. of 3.0 N sodium hydroxide was added to complete the hydrolysis of the methyl ester. Initially two layers formed, but a homogeneous solution was obtained by stirring rapidly for a n hour and the alkaline solution was left for another hour. The reaction mixture was diluted with water (200 nil.) and extracted with ether ( 3 X 100 ml.) t o remove neutral material. The aqueous phase was acidified a t 0' with concentrated hydrochloric acid (40 ml.) and the oil which precipitated was extracted into ether ( 3 X 50 ml.). The combined ether extracts were dried over magnesium sulfate and evaporated, leaving the crudc mixture of benzyloxycarbonylhydroxyprolinesas an oil (6.8 g . ) . The oil was dissolved in a mixture of ethanol (200 ml.), water (60 ml.), and acetic acid (40 ml.) and the protecting group was hydrogrnoll-zed with hydrogen a t room temperature and atmospheric pressure in the presence of 105; palladium-on-charcoal ( 2 g . ) . Gas uptake was initially very rapid, but the reaction rate then slowed and the hydrogenation was left overnight for completion. The catalyst was filtered off and the filtrate was evaporated. The residue of crude hydroxyprolines, 3.2 g., could not be readily crystallized and was purified by filtration on 100 ml. of D o w x 50W (H+). The impurities were removed from the column by washing with water until the eluate was neutral. The amino acids were then eluted with 2 N ammonium hydroxide (250 m l . ] . Evaporation of this fraction in uncuo gave an off-white solid residue ( 2 . 3 g . ) . The purified amino acid mixture was recrystallized by dissolving it in the minimum volume of boiling ethanol and adding acetone ( c n . 50 ml.) a t the boiling point; yield of white crystals, 1.9 g. The composition of this mixture was determined on an automatic amino acid analyzer t o be: tmns-3-hydr 4 - h y d r o x y - ~ ~ - p r o h n e10yc , ; allo-4-hydro ninhydrin-negative material, 21 5 ; . trnns-3-Hydroxy-~~-proline by Direct Recrystallization .--Shw recrystallization of the above product from ethanol-water gave 49yc of a colorless crystalline product which contained 94.4( trans-3-hydrox~.-~~-proline and 5 l - h y d r o x y - ~ ~ - p r o l i n e ..A second recrystallization gave 47 of l00c/;; pure ticin.r-3-hyd r o x y - ~ ~ - p r o l i nidentical e, with the material prepared by column chromatography (see below). Anal. Calcd. for CjHSnTO3: C, 45.80; H , 6.92; N, 10.68. Found: C , 46.01; H , 6.98; S ,10.62. The same analytically pure homogeneous t r a n s - 8 - h y d r o x y - ~ ~ proline was also obtained by preparative chromatography on a 2 in. X 36 in. column of Xmberlite IK-120, 47-65 p , arid the use of
2s30
F.IRREVERRE, K. MORITA,A. V. ROBERTSON, A N D B.I ~ I T K O P
i701 85
citrate buffer p H 3.25 for elution as described under the isolation gave a crystalline product which on further recrystallization from procedure from sponge. ethanol-water 180:7 0 ) yielded 510 mg. (80';,;) of colorless crysN-Tosyl-tvuns-3-hydroxy-~~(and L)-proline Methyl Ester.analytically and cliromatotalline trans-3-hydros}--~~-proline, T o a stirred solution of trans-3-hydroxy-DL-proline (15 mg.) in 4 graphically pure. ml. of 0.2 ' I' KOH ivas added pulverized p-toluenesulfonyl chloN-Carbobenzyloxy-3-keto-~~-proline Methyl Ester (IV).--X ride (30 ing.) and the mixture \vas stirred a t 0" for 2 hr. The solution of 600 nig. (1.5 mmoles) of N-carbobenz\loxv-tl.ans-3mixture \\-asextracted with ether and the ether extract was dishydroxy-DL-proline methyl ester ( I I a ) in 180 nil. d acetone \vas carded The aqueous solution was then acidified, extracted with oxidized with 1.5 nil. of 8.0 S CrOs in 5 0 ' ~ iH2SO4and allowed t o ether, a n d the ethereal solution was washed with water, dried, stand a t room teiriperature for 65 min. A small quantity of and evaporated to give S-tosyl-tiuns-3-hydros)--DL-proline(27 ethanol \vas added t o decompose excess oxidant. Evaporation of nig.). After methylation with diazomethane and recrystallizathe solution under reduced pressure gave an oily substancc v-hich tion from ether, N-tosyl-tunns-3-hydro~y-DL-proline methyl ester was extracted with 200 nil. of ether. The ether solution \\-ai was obtained as colorless cubes, 1n.p. 99-100",yield 21 rng. washed with salt water, dried over anhydrous Sa?S04, arid Annl. Calcd. for CisHiiNC,S: C, 52.17; H , 5.73; S , 4.68. evaporated to give the ketone (590nig., 1-ield 93' j. 'The ketone I\., after chromatographic purification oii hlalliiickrodt silicic Found: C , 52.11; H , 5.54; K , 4.70. acid, failed t o crystallize. However, thin-layer chromatography Likewise, frcins-3-hydroxy-~-proline,obtained from telornycin, and gas chromatography indicated the sample t o be honiogenewas converted t o S-tos~l-trcins-3-hydroxr-L-proline methyl ous and the infrared spectrum of the compound revealed a strong ester, which crystallized as colorless platelets, m.p. 101-102°. peak a t 5 . M p , characteristic of a ketone carbonyl in a fivc-nicrriThe infrared spectra of synthetic and natural S-tosyl-truns-3bered ring. hydroxyproline methyl ester in chloroform solution were superReduction of N-Carbobenzyloxy-3-keto-~~-proline Methyl posdble. B. Hydroboration of N-Carbobenzyloxy-3,4-dehydro-~~- Ester ( I V ) with NaBH4.-To a solution of 2.0 g. (7.2 nimoles) of S-carbobenzyloxy-3-keto-~~-prolinemethyl ester (11.) in 100 proline Methyl Ester ( I b ) without Saponification.-To a stirred ml. of ditnethoxyethane was added 100 mg. of sodium borohydride solution of S-carbobenzyloxy-3,4-dehydro-~~-proline methyl in 1 .0 inl. of water, and the reduction was allowed t o proceed at ester [ I b , 1.0 g . ) in 50 ml. of dimethoxyethane ( D M E ) a t 0' was room temperature for 15 miii. Xcetic acid \vas added t o deadded a n excess of diborane (generated from 300 mg. of 825, compose the excess reagent and the solution evaporated in m i t ~ ~ . LiBH, and 3 nil. of 20';i BF3 in D M E ) over a period of 5 min. The residue \\-as extracted with ether and the ether solution was After 20 min., an aliquot of the reaction mixture was examined by mashed with water, dried, and evaporated to give a11 oily residue gas chromatography (SE-30, 1(, on GCI'; 190°)34which indicated (2.0 g . ) . Gas chromatographic analysis of this oil, after acetylathe complete disappearance of the starting material. Careful tion with acetic anhydride in pyridine, revealed the presence of addition of water decomposed the excess diborane. The borane 80' 1 of S-carbobenzyloxy-cis-a-hydroxy-DL-proline methyl ester complex was oxidized by addition of 20 g . of K2COsand 5 inl. of (\'I and less than 20;' of the trans isomer I l a . 30' 1 H.02 and allowed t o stand a t room temperature for 2 days Separation of the Esters IIa and V on Silicic Acid.-The oily with occasional agitation. Inorganic salt was removed by mixture ( 2 . 0 g.) was dissolved in 20 nil. of benzene-ethyl acetate filtration and the cake was washed with dimethoxyethane. The ( 4 : 1)aiid chroniatographed in four equal portions on a column (18.5 combined filtrate and washings were evaporated in t'urz40, reX 6 c m . ) of hlallinckrodt silicic acid (mesli: 100-200) with the dissolved in ether, and the ethereal solution was washed, dried, and same solvent mixture. Fractions of 22 rril./tube were collected evaporated, giving 0.9 g. of an oily residue. Gas chrornatoon a fraction collector. The cis-ester 1. (1.46 g.) appeared in graphic examination of the oil, after acetylation with acetic a n fractions 62-85, the trans-ester ( 150 nig.) in fractions 94-120. hydride in pyridine, showed the presence of S-carhobrnzyloxl-The total yield of the two isomeric esters was 81 trnns-3-hydrouy-~~-pro~ine methyl ester and of S-carhobenzylversion of the pure esters I I a and 1. t o trcins- and o x y - l - h y c i r o x \ - - ~ ~ - p r o l i nmethyl e ester in a ratio of 3 : 1. A . Isolation of N-Carbobenzyloxy-trans-3-hydroxy-DL-proline DL-prolines is described in the following. Preparation of cis- and tvnns-3-Hydroxy-~~-proline.-The Methyl Ester (IIa).-Elution of the mixture from a column of mixture of cis- and truns-3-hydroxy-S-carbobenzyloxy-n~-proline silicic acid (mesh 100-200) wit11 benzene-ethyl acetate ( 4 : 1) gave methyl ester from another S a B H r reduction of 0.4 g. of unpuria s the more rapidly moving component 440 mg. of S-carbobenzylfied keto ester 11. was dissolved in 50 nil. of dimethoxyethane, ouy-lrcins-3-1iydrouy-nL-proiine methyl ester, which crystallized treated with 2.5 n i l . of 1.0 S KOH, and allowed to stand a t room in colorless prisms, m . p . 81-82", undepressed on admixture with temperature for 2 hr. m'ater was carefully added and the Inisa synthetic sarnplc prepared under B. ture was cxtracted with ether and the ether extract discarded. Arzcii. Calcd. for CL&II7SOj:C, 60.20; H , 6.14; S ,5.02. After acidification, the aqueous s \\-as extracted with Found: C, 60.49;H , 6.20; S, 5.01. ether and the ether mlution was w t h water, dried, and N-Carbobenzyloxy-4-hydroxy-~~-proline Methyl Ester (IIIa). evaporated t o give a mixture of trans-3-hydroxy-S-The s!ower-moving component (93 mg.) was then eluted from carbobenzylox!--DL-proliiies (350 m g . ) , which was dissolved in the coluiiin with the same solvent system. T h e identity with an methanol-water ( 2 1) and hydrogenated over palladiutn to give authentic sample was confirmed by direct comparison on thina mixture of 147 mg. of cis- and trans-3-hydroxy-DL-proline. layer clirornatography (Brinkmann Silica Gel -l DL-proline, as determined by titration of 5-mg. samples of the ester as :in oil (182 1ng.l which became crystalline on standing. amino acids in 2.0 nil. of water with both 1.0 HCI and 1.0 S Iiecrystallizatioii from isopropyl ether gave cubic crystals, n1.p. S a O H in a Radiometer titrator with a n Auto-Burette, Ivcre in80-81". 'This product is identical with t h a t obtained by the 0.05, P K Z H= 9.73 + 0.05. hydrohoration of S-carbobenzyloyy-3,4-dehydro-DL-proline distinguishable: p K c ( 1 0 ~= 1.92 For comparison tlie pK values for proline and 3,4-dehydro-o~methyl ester. were determined independently through the courtesy of Conversion of N-Carbobenzyloxy-tmns-3-hydroxy-~~-proline proline Dr. Simon, E T H , Zurich. Methyl Ester to tran.~-3-Hydroxy-~~-proline.--Xsolution of 1.42 g . ( 5 minoles) of N-c;rrbobenzylosy-tmns-3-liydroxy-n~-proli1ie --pKx(.s----pKlt*(l--tnetll? I e s t c i - ( I i a j i n 1 0 1111. of dimethoxyethane was treated with HCI (CHslrS?? HCI (CH~)IN@ Substance 10 nil. of 1 .O iV K O H arid allowed t o stand a t room temperature
