M. SOKOLOVSKY, M. WILCHEK, AND A . PATCHORNIK
1202
Vol. 86
Experimental
,
I
Fig. 12.-Speculative
structure of segment of gymnosperm lignin molecule.
attached to side chain carbon atoms per phenylpropane unit; and (d) about eight aliphatic hydroxyl groups, of which about three are benzylic, are present per ten phenylpropane units. The presence of the large scale cross-linking or ring systems indicated in Fig. 12 as being present in lignin molecules is not proved, but the existence of a somewhat rigid structure is suggested by the broadness of the signals observed in the n.m.r. spectra. Evidence for hydrogen bondings of the types shown in BC and GH has been found with model compounds in chloroform solution, and such bonding may also occur in lignins in other solvents and in the dry state.
[COXTRIBUTIOX FROM THE
DEPARTMEXT OF
The n.m.r. spectra were taken using the instrumentation and procedures previously described.' Solutions of 7 to 15cc by weight of the acetylated lignins in CDCI, containing hexamethyldisiloxane internal reference were used in standard degassed and sealed 5-mm. tubes. The integrations of n.m.r. spectra were done using an integrator described by Varian Associates.'* Prior t o integration of the n . m . r . spectrum of each lignin preparation a sample of ethanol was integrated to determine t h a t the integrator was adjusted correctly. T h e integrator was considered t o be properly adjusted only when the ratio of methylene protons t o methyl protons in the ethanol was found t o be 2.00 to 3.00 within confidence limits of 0.01. The stability of the integrator was further checked by allowing the sweep t o cover several parts per million a t the beginning and end of each spectrum t o be sure there was no appreciable drift. Lignin solutions of lorc or greater concentration were found to give more reliable integration results than weaker solutions. X correction for the small amount of signal from chloroform protons in the deuteriochloroform solvent was made in each case as follows. By integrating an n . m . r . spectrum of our solvent deuteriochloroform containing a known weight of hexamethyldisiloxane it was found to contain about 0 . 0 0 4 2 5 protons by weight. Using this figure the percentage of the total integrated signal coming from chloroform protons, A , was found for each sample by applying the followirig equation in which x = the weight per cent of lignin preparation in the sample, 100 - x = the weight per cent of deuteriochloroform in the sample, and H = fraction of hydrogen in lignin. ( T h e small amount of hexamethyldisiloxane was neglected.)
Acknowledgment.-The authors are grateful to Professor Eric Adler and Dr. F. E. Brauns for their generous gifts of Bjorkman and Brauns lignins used in these studies and to Professor K. V. Sarkanen for his many helpful suggestions in the preparation of this manuscript. (12) T h e N M R - E P R Staff of Varian Associates, " N M I I and E P K Spectroscopy," Pergamon Press, New York, i X Y , 1960, Chapter 15
BIOPHYSICS, THE
WEIZMAXN IXSTITCTE
O F SCIENCE,
REHOVOTH, ISRAEL]
On the Synthesis of Cysteine Peptides' BY
L ~ O R D E C H A SOKOLOVSKY, I XIEIR
ifTILC€IEK,AND
a B R A H A M PATCHORNIK
RECEIVED AUGVST14, 1963 The preparation of various peptides of cysteine is described The basis of the method used is the selective removal of the carbobenzoxy group from either the amino or thiol group without causing racemization The action of hydrogen bromide in acetic acid a t room temperature removes the carbobenzoxy moiety from the amino group almost quantitatively, without affecting the S-carbobenzoxy group On the other hand, the action of excess sodium methoxide in anhydrous solvent causes rapid alcoholysis of the S-carbobenzoxy group with a quantitative liberation of the free thiol group a s determined iodometrically The method described was successfully applied t o the total synthesis of glutathione in 25yo over-all yield
In the course of recent studies on the development of a method for the nonenzymatic cleavage of peptide a t cysteinyl residues2 it was found necessary to prepare cysteine-containing peptides as model compounds. Methods for the preparation of peptides containing S-protected cysteinyl residues have been known for a long time.3 However, the methods used so far for the selective removal of the S-protecting group suffer from a number of disadvantages: e.g., low yields, racemization, and side reactions such as splitting of peptide b0nds.l The problems arising during the synthesis of such peptides have been reviewed recently by Young.5 (1) A recent report by L Zervas, I Photaki, a n d P;. Ghelis ( J . A m ('hem. Sor , 86. 1337 (1963)). which reached us after this work had been completed. contains results similar t o some of those described in this paper. ( 2 ) A Patchornik a n d M . Sokolovsky, "Vth European Peptide S y m posium, Oxford. 1962," Pergamon Press, 1963, p 253, a n d in the following papers ( 3 ) J P. Greenstein a n d M U'initz, "Chemistry of t h e Amino Acids," John Wiley and Sons, Inc , New York, S Y . , 1961 (4) S Sarid and A . Patchornik, Zs?ael J C h e m . , 1, 63 (1063).
In the present paper, the preparation of various Sand Xi-carbobenzoxy derivatives of cysteine and their application to the synthesis of cysteine peptides is described. The basis of this method is the selective removal of the carbobenzoxy group from either the amino or thiol group in high yields without causing racemization. The action of hydrogen bromide in acetic acid for 15 min. at room temperature removes the carbobenzoxy moiety from the amino group almost quantitatively without affecting the S-carbobenzoxy group. On the other hand, the action of excess sodium methoxide ( 5 equiv.) for 5-10 min. a t room temperature, under nitrogen, causes rapid alcoholysis of the Scarbobenzoxy group with almost quantitative liberation of the free thiol group, as determined iodometrically. Such selective removal of the protecting group may be conveniently used in the synthesis of long peptide ( 5 ) G T Young, Collectton Czech Chem C o m m u n , 24, 11-1 (I9.5Y)
SYNTHESIS OF CYSTEINEPEPTIDES
March 20, 1964
1203
tetrapeptide corresponding to the sequence C-terminal of oxytocin : N ,S-dicarbobenzoxy-L-cysteinyl-L-prolylL-leucylglycine amide (VIII) in 86% yield. SCHEME 1 Upon addition of 5 equiv. of sodium methoxide per cbz-NHCHCOOH SHpCHRCOOR' + mole of peptide for 10 min. (step d) all these derivatives resulted in the formation of the corresponding peptides (a) CH2 containing a free thiol group. Thus, compounds I I , ' I I I , S-cbz VII, and poly-S-carbobenzoxy-L-cysteine6yielded Ncbz-I\"CHCOSHCHRCOOR' carbobenzoxy-L-cysteinylglycine ethyl ester (IX), NI carbobenzoxy-L-phenylalanine-L-cysteinylglycine ethyl CHz ester ( X ) , N-carbobenzoxy-L-cysteinyl-L-phenylalanine HzO S-cbz methyl ester ( X I ) , and poly-L-cysteine, respectively, in high yields (80-90yo). The action of alkali on cysteine HBr-CHaCOOH - + ( b ) ---+ BrHJTHCONHCHRCOOR' derivatives is known to cause a @-eliminationreaction . g 15 min I However, under the conditions used in the procedure CH2 C~H~BT described it could be shown, by determination of the I S-cbz pyruvic acid known to be formed on acid hydrolysis of dehydroalanine containing peptides, l o that the pcbz-NHCHR"C0OH -~~~ 4 elimination reaction, in which S-protected cysteine (C) cbz-h-HCHR"CONHCHCOXHCHRCO0R' peptides are converted to the corresponding dehydroalanyl derivative^,^,^^ occurred to the extent of less CHz than 3-4y0. I S-cbz It is essential to carry out the reaction under anhySaOCHa drous so as to avoid hydrolysis of the ester -~~ --+ cbz-NHCHR"C0SHCHCOh-HCHRCOOR' bond. conditions In order to avoid transesterification the base anhydrous solvent used must be derived from the same alcohol as the ester (d) CHz I used for esterification of the carboxyl groups. Thus, SH when compound VI1 was treated with sodium methoxide 1 NaC)H - ---in methanol, N-carbobenzoxy - L - cysteinyl-L - phenyl_-__ + BrH3SCHR"CONHCHCONHCHRCOOH alanine methyl ester ( X I ) and not the benzyl ester was 2, HBr-CHaCOOH I obtained. (e) CHB In order to estimate whether racemization occurred SH by this procedure the following determinations were abbreviations: cbz = carbobenzoxy group, R, R" = amino acid carried out : (a) N,S-dicarbobenzoxyglutathione( X I I ) residue, R ' = carboxyl protecting group was treated with 5 equiv. of sodium methoxide in absolute methanol for 10 min. ; the p H was lowered to 8, In step a , N,S-dicarbobenzoxy-L-cysteine is coupled and recarbobenzoxylation was effected by addition of 5 with an amino acid ester by known methods of coupling.3 equiv. of benzyl chloroformate. The isolated comIn a similar fashion, S-carbobenzoxy-L-cysteine ester pound had a specific rotation identical with t h a t of the can be coupled with carbobenzoxy amino acids yielding original N,S-dicarbobenzoxylglutathione. (Identical peptides of the type cbz-NHCHRCOXHCHCOOR'. results were obtained when the base used was aqueous I sodium hydroxide.) (b) Similarly, compound VI11 was CH2 transformed to the corresponding ?;-carbobenzoxy-Sbenzyl derivative in 90% yield, and this compound had S-cbz By the use of the carbodiimide method,6 for example, [cyIz5D -61" ( c 2, dimethylformamide); the reported values are [CX]~'D-60°,12 -620.13 These tests clearly high yields (S5-90Y0) of pure protected peptides can be indicates t h a t step d does not cause racemization. obtained. Treatment of these peptides with 3 equiv. of hydrogen bromide in glacial acetic acid yields the When a peptide contains N-terminal cysteine with a corresponding hydrobromides (step b) . This reaction free amino group and a S-carbobenzoxy group, it must goes to completion within 15 min. a t room temperature. be borne in mind t h a t under alkaline conditions (step For example, N,S-dicarbobenzoxy-L-cysteinylglycine d ) , the carbobenzoxy group may migrate from the sulfur benzyl ester (I) gives the corresponding hydrobromide to the nitrogen. l 4 Indeed, when S-carbobenzoxycysin 7 5 7 , yield. Under these conditions, the S-carboteine was treated with sodium methoxide as described benzoxy moiety was not removed as could be shown tiabove, and then oxidized with iodine under acidic trimetrically.' It should be mentioned t h a t even under conditions, dicarbobenzoxycysteine was isolated in 61% more drastic conditions (100' for 30 min.) only 30YG yield. Analogously, S-carbobenzoxy - L - cysteinyl- Lremoval of the S-carbobenzoxy group occurred. phenylalanine benzyl ester hydrobromide ( X I I I ) gave Step c is carried out similarly to step a. By this N-carbobenzoxy-L-cysteinyl- L - phenylalanine in 60% procedure, various protected peptides were synthesized yield. This rearrangement may occur as the result of in SO-90yc yields : e.g., N,S-dicarbobenzoxy-L-cysteinylglycine benzyl ester ( I ) ,N ,S-dicarbobenzoxy-L-cysteinylglycine ethyl ester ( I I ) , carbobenzoxy-L-phenylalanyl(8) A . Berger, J . Soguchi, and E. Katchalski, J . A m Chem. SOL, 7 8 , S-carbobenzoxy-L-cysteinylglycineethyl ester ( I I I ) , 4483 (1956) (9) H . T. Clarke and J . M . lnouye, J . Biol. C h e m . , 94, 541 (1931); J . M N,S-dicarbobenzoxyglutathione dibenzyl ester (IV) Swan, Nature, 18,643 (1957); M Sokolovsky, M Wilchek, and A PatchorN-carbobenzoxy - L - alanyl - S - carbobenzoxy - L - cysteine nik, Bull. Res. Council Israel, H A , 79 (1962) benzyl ester (V) , N,S-dicarbobenzoxy-L-cysteinyl-L- (10) A. Patchornik and M. Sokolovsky, J . A m Chem. S O L , 86, 1206 leucine methyl ester (VI), and N,S-dicarbobenzoxy-L(1964). (11) M . Sokolovsky, T. Sadeh, and A. Patchornik, ibid.,86, 1212 ( 1 9 6 4 ) cysteinyl-L-phenylalaninebenzyl ester (VII). Further(12) M . Bodansky and V . du Vigneaud, i b i d . , 81, 2504 (3959) more, the procedure was applied t o the synthesis of a (13) H . C. Beyerman, J . S. Bontekoe, and A . C . Koch, Rec. lvaa chrm . . 1 8 , chains containing cysteine residues as exemplified by Scheme 1.
+
+
+
->
~
(6) J C . Sheehan and G . P . Hess, J . A m . Chem. Soc., 7 7 , 1067 (1955) (7) A Patchornik and S. Ehrlich-Rogozinski, A n a l . Chem., 8 8 , 803 (1961).
935 (1959). (14) H . P . Burchfield, .Vature, 181,49 (1958); L. A . Cohen and B Witkop, A n g e w . C h e m . , 78, 253 (1961).
1204
-
M. SOKOLOVSKY, M. WILCHEK, A N D A. PATCHORNIK
Vol. 86
Anal. Calcd. for C Z ~ H ~ ~ S C, ~ O62.68; ~ S : H , 5.26; N,3.22; S, 5.96. Found: C,62.90; H , 5.32; S , 5.35; S,5.98. S-Carbobenzoxy-L-cysteine Benzyl Ester Hydrochloride.-Ihy phosgene was passed a t room temperature through a suspension of 25.1 g . IJf S-carbobenzoxy-L-cysteine (0.1 moles) in anhydrous dioxane (300 1111.)until a clear solution was obtained (about 1 h r . ) , Phosgene was removed by a stream of dry nitrc~genand the solvent was distilled off in w c u o a t 45'. Benzyl alcohol (60 1111.) and dry ether (300 nil.), previously saturated with gaseous HCI (10 9 , ) a t O", were added and the solution was left overnigl~ta t room temperature. The ester which separated was filtered off and washed with ether. The ester was recrystallized from hot water or methyl alcohol-ether; yield 30 g. (T8[';j , m . p . l~16--l(17°, A n d . Calcd. for C18Hy,,SO&Cl: C, 36.09; H , 5.24; S , 3.67; S,8.39. Found: C, 56.T5; H , 5.37; S , 3.50; S, 8.51. L-Phenylalanine benzyl ester hydrochloride was prepared in the same manner as described for S-carbi)benzosy-L-c!.steine benzyl ester hydrochioride, and i t was recrystallized from water2 yield 80%, n1.p. XM", [ a I z 6-23" ~ i r 2, 0.25 .\- hydrochloric acid); r e p ~ i r t e d m ' ~ . p . 203', [ a I z b D -22.8' ( 1 - 1 , (J.25 hydrochloric acid). A n a l . Calcd. for C 1 ~ H , , S O ~ C 1N: , 4.81; '21, 12.03 Found: n-, 4.84;CI, 12.18. SCHEME 2 The procedure described for the synthesis of compound I was cbz-NHCHCOOH NHzCH~COOC~H~ used for the preparation of the following compounds: N,S-Dicarbobenzoxy-L-cysteinylglycineethyl ester (11): CH2 + yield ( 7
