192
BERNHARD WITKOP AND CALVINAI. FCJLIZ
Betaine of Hydroxy-L-proline (XXI, Betonkine).-Hydroxy-L-proline (2.00 g.) was added to a suspension of 4 g. of silver oxide in 5 cc. of water. hfter 3 hours a t room temperature the silver salt had formed and half of the water was removed in ' J Q C U O and 40 cc. of methanol and 2 cc. of methyl iodide were added. The solution warmed up immediately and silver iodide formed. The mixture remained a t room temperature overnight. After that 1.5 cc. of methyl iodide was added. The mixture was refluxed for 3 hours, filtered and taken to dryness in vacuo. Trituration with acetone and ethanol yielded a crystalline residue weighing 1.11 g. (465Z!), [ O ~ ] ~ O D -37.6" (c 1.0, in water). A recrystallized sample of betonicine from ethanol had m.p. 252-253" and [ O C ] ~ ~-34.2' D ( c 1.0, in water); lit. m.p. 243' and [ c Y ] * ~ D -37" ( c 4.8, in water).28 The initial rotation of a lye solution in N NaOH was -36.0" (c 1.0, in water). After next reading (after 18 hours) it was 0.0"and after 24 hours i t was still 0.0". A d . Calcd. for CiHt3S03: C, 52.81; H, 8.23; iL', 6.80. Found: C, 52.83; H , 8.35; S,8.58. 0-Acetylbetonicine Hydrochloride (XXII).-0-Acetylhytlroxy-L-proline (0.66 g.), prepared according t o Sakami and Toennies40 was dissolved in 1.25 cc. of water and treated with 1 g . of silver oxide. After 3 hours, 10 cc. of methanol and 0.5 cc. of methyl iodide were added a t 0". This mixture w-as shaken a t room temperature for 1 hour whereupon 0.4 cc. of methyl iodide was added and shaking was continued for one additional hour. Filtration, evaporation of solvent and trituration of the residual oil with acetone, ethanol and ether afforded 0.12 g. of crystalline material which v a s largely betonicine. The mother liquor was taken t o dryness and the oil so obtained was converted to its hydrochloride with hydrogen chloride gas in ethyl acetate. After several (40) Sakami and Toennies. J . Bid. Chem., 144, 203 (1942).
\-@I. ;!I
crystallizations a t room temperature from water, ethanol and ethyl acetate, aytylbetonicine hydrochloride was obtained, m.p. 200-201 . A n a l . Calcd. for CSHl&04.HC1: C, 45.48; H , 6.79; N, 5.89. Found: C,45.36; H , 6.94; N , 5.66. Betaine of Allohydroxy-n-proline (XXIII, Turicine).Silver oxide (1.00g. ) was added t o allohydroxy-D-proline (2.00 g.) dissolved in 2.5 cc. of water. After 3 hours 15 cc. of methanol and 2 cc. of methyl iodide were added to the suspension of the silver salt. The mixture was agitated a t room temperature for 4 hours, 1.5 cc. of methyl iodide was added and shaking was continued for an additional 3 hours. Filtration, evaporation of the solvent in zmcuo and trituration with acetone and ethanol afforded 2.35 g. (975%) of turicine of 1n.p. 252Oand [ 0 1 ] 2 0 ~ 4-35.1' ( c 1.0, in water). X recrystallized sample from water, ethanol and ethyl acetate had m.p. 259-260" and [ C X ] ~ t~3D7 . 8 " ( c 1.0, inwater); lit. m.p. 249' and [ 0 1 ] 2 1 ~ $36" ( c 0.5, in u.ater).28 A mutarotation study of a 1% solution of turicine in IT XaOH at 20" gave the following values; !a]D initial, 4-51.1"; after 3 hours, $29.2'; after 5.5 hours, $20.6"; and after 22 hours, 0.0". A n d . Calccl. for Ci-ti13Xio3: C . X.81; 13, 8.22; S , 8.80. Found: C, 52.86; H,8.30; S,8.60. Betaine of 5-Hydroxy-~-pipecolic Acid.-The methylation of 5-hydroxy-~-pipccolic acid (0.300 g.) a t room tctnperature with silver oxide and methyl iodide according to the procedure for 0-acetylbetonicine hydrochloride gave on trituration with acetone arid ethanol, the betaine (0.275 g., 7 7 % ) , m.p. 265" dec. h recrystallized sample from water, ethanol and ethyl acetate showed m.p. 267-268" dec., [ C Y ] ~ @-13.9' D (c 1.0, in water). A n a l . Calcd. for C8Ht5N03: C, 35.47; H , 8.73; 9, 6.09. Found: C, 53.73; H, 8.43; N, 7.83. BETHESDA, MARYLASD
[ C O X T R I B U T I O N FROM THE NATIONAL INSTITUTE O F ARTHRITIS AND hIETAROLIC L)ISEASl%, PUBLIC HEALTHSERVICE]
S.41lo h A L lIv5TIl CTES O F $ i E A I . l € I
The Configuration of 5-Hydroxypipecolic Acid from Dates BY BBKNHAKDWITKOPA X D CALVINh l . IJoLrz RECEIVED JULY 9, 1956 5-Hydroxypipecolic acid ( I ) was isolated on a preparative scale from the fruits of Phoenix dactylijeera. The mixture of the free amino acids was treated with nitrous oxides which deaminated the primary amino acids and converted the secondary amino acids to the ether-soluble N-nitroso acids. The latter were hydrolyzed to the secondary amino acids and separated on a column of Dowex-50 ion exchange resin, The oxidation of S-carbobenzyloxy-5-hydroxy-~-pipecolic acid (11) with chromium trioxide in sulfuric acid yielded the &keto compound I11 which was reduced with sodium borohydride to give, 0x1 treatment with acetic anhydride, the lactone of ~-carbobenzyloxy-5-allohydroxy-~-pipecolic acid (VI). The hydrobromide of the free lactone VI1 opened up in water with mutarotation to give the salt of 5-allohydroxy-~-pipecolicacid (v). The formation of the lactone in the a110 series established the cis-relationship of the functional groups and the applicatiori of the rule of Lutz and Jirgensons made possible the complete steric assignments to natural and allo-5hydroxypipecolic acids. Hudson's lactone rule was applied to the mutarotation of the lactone VI1 and found to be valid and applicable. The formation of the two racemic amino acids ( I and V) was observed in the internal opening of diethyl 2-(3',4'-epoxybutyl)-2-formamidomalonate (IX), after removal of the formyl group by base, together with the five-membered isomers X I and XII.
The configuration of 6-hydroxy-L-lysine, an important building stone of collagen,2is not known. Its cyclization t o the two diastereoisomeric 5hydroxypipecolic acids has been achieved3 in the same manner as the conversion of the y-hydroxyornithines to the normal and allo-hydroxy proline^.^ 'This cyclization, done on a preparative scale, would allow of the exact steric correlation of the two asymmetric centers carrying the hydroxyl of 6hydroxylysine and 5-hydroxypipecolic acid. As a step in this direction, this paper describes the (1) Cf.W. S. Fones, THIS J O U R N A L , 75, 4865 (1953). Cf,K . H. Gustavson, "The Chemistry and Reactivity of Collagen," Academic Press, Inc., Kern York, PIT.Y . . 1956. (3) I,. A. Cohen, F. Irreverre, K. A. Piez, B. Witkop and If. I N H for >CH2 is generally considered to cause only minor conformational changes.'* The total configuration of 5-hydroxy-~-pipecolic acid was established by making C(2) the absolute reference point (Lutz-Jirgensons rule) for C(5) whose relative position was deduced from the lactone VII. With these two centers fixed a model was available to test the applicability of Hudson's rule. Just as in the case of the lactones of allohydroxy-D and +proline the rule was obeyed. This extension of Hudson's rule to y-and &hydroxyamino acids is a useful new method and has been applied to 6-hydroxy-L-lysine. Synthesis.-The starting material for the synthesis of 5-hydroxypipecolic acid was diethyl 2(3',4'-epoxybutyl)-2-formamidomalonate(IX)l9 obF tained by perbenzoic acid oxidation of diethyl 2-(3'-butenyl)-2-forniamidomalonate. &Hydroxylysine was obtained by opening of the epoxide with ammonia.Ig The internal opening of the epoxide by the amino group of X was visualized as a straightforward reaction (arrows B) leading to I and V. An excess of N alkali a t 37" for 3 days converted IX into X which was decarboxylated by refluxing in acid solution. The reaction mixture was assayed on a 150-cin. column of Dowex-50 using O.l5y0 ninhydrin solution in glacial acetic 350 ~JJL, according to the procedure acid and A,, NUMBER OF T U B E S OR ML. developed for the analysis of 5-hydroxypipecolic Fig. 1.-Analysis on Dowex 50 of cyclization mixture acid by Piez.?o Figure 1 shows the results. The from alkali treatment of diethyl 2-(3 ',4'-epoxybutyl)-2- two smaller peaks are those of nornial and alloformamidomalonate. Peaks A and B are probably the hydroxy-D,L-pipecolic acid whose identities were indiastereoisomeric 2-hydroxymethylprolines (measured a t 375 dependently established by paper chromatography. mp) and peaks C and D are normal and allohydroxy-D,L- The other two peaks belong to the two secondary pipecolic acid (measured a t 350 mg). The dotted parts of amino acids (presumably the diastereoisomers X I and XII) which gave an orange color with ninhythe curves are calculated. drin and which were not deaminated by nitrous For the debenzylation of the N-carbobenzyloxy- acid. Pathway A is apparently favored over pathlactone VI hydrogen bromide in glacial acetic acid way B. An analogous case is the alkali-catalyzed rather than the catalytic method was used. Pure cyclization of unsaturated tertiary amines in the preparations of VI gave directly the beautifully conessine series to yield quaternary pyrrolidines crystalline hydrobromide of the lactone VII. rather than piperidines.21 It may be recalled in Paper chromatograms of preparations from the de- this connection that a hydroxylnethylproline has benzylation of less pure VI showed the presence of recently been isolated from the wood of apple glutamic acid, a well-known oxidation product of trees.22 A third compound present in the cycliza5-hydroxypipecolic acidI6possibly a small amount tion mixture gave a blue color with ninhydrin, was of baikiainl' resulting from dehydration of IV (or deaminated by nitrous acid and behaved like serine 11) and of two further spots indicative of norvaline on paper chromatography. It is probably the and of norleucine. glycol, 6,edihydroxy-a-aminocaproic acid. The Rapid lactone opening, as detected by change of approximate yields were 10% of I V, 30-40% of rotation, is observed with the lactone VI1 hydro- XI XI1 and 50-6070 of a hydroxyamino acid, bromide in aqueous solution. This lactone in- presumably the glycol. stability, even under acidic conditions, is greater The open-chain analog of 5-hydroxypipecolic than normally found in non-nitrogenous y- or 6- acid (I) is erythro-6-hydroxy-L-lysine (XIII)) and lactones,17even bicyclic ones. An important factor affecting the equilibrium VIIaHBr e V . H B r is that of the allo isomer V the threo diastereoisomer XIV. the latter as an amino acid hydrobromide is in equi- There seems t o be a preponderance for DG-configlibrium with its zwitterion. This view is supported (18) D. H. R. Barton and R . C.Cookson, Qzrartrriy Reviews, 10, 72 by the much greater stability of the N-carboben- (1956). (19) J. K. Weisiger, J. B i d . Chem., 186, 591 (1950). zyloxy lactone VI which, for example, is recrystal(20) IC. Piez, F. Irreverre and H. L. Wolff, ibid.. December
+
+
(17) Cf.L. N. Owen and P. A. Robins, J . Chcm. Soc.. 326 (1949); N. R. Campbell and J. H. Hunt, i b i d . , 1379 (1950); K. Fisher, W. H. Perkin, ibid.. 93, 1884 (1908).
(1956) (21) R. D. Haworth, ref, 10, pp. 6-7. (22) F. Urbach, h'afurc, 115, 171 (1955).
+
195
5-HYDROXYPIPECOLIC ACID
Jan. 5, 1957
i
