J. Org. Chem. 1983,48, 2557-2563 1 H), 5.97 (2 s, 2 H), 4.16 (e, 3 HI,4.00 ( s , 3 HI, 3.72 ( s , 3 HI, 3.59 (s, 3 H); UV (EtOH) 264,295,309,350nm; fluorescence 453 nm; HRMS, m/z 424.1151 (M+), calcd for C23H2008424.1155.
Acknowledgment. We thank Drs. N. G. Pate1 and E. L. Jenner for providing us natural justicidin P and its
2557
spectra. The NMR studies by Dr. G. S. Reddy are greatly appreciated. We also thank Prof. B. M. Trost for helpful discussions. Registry No. 1, 86012-93-3;2, 25001-57-4;4, 86012-94-4;5, 86012-95-5.
Chemical Synthesis of Some Mono- and Digalactosyl 0 -Glycopeptides+ J. M. Lacombe and A. A. Pavia* Laboratoire de Chimie Bioorganique, Facultt des Sciences d'Avignon, 84000 Auignon, France Received June 29, 1982 Chemical syntheses of several 0-glycopeptides containing 0-galactosylthreonine are reported. Two different approaches have been investigated to determine the strategy best adapted to the synthesis of any desired 0-glycopeptide. Glycosylation of an adequately protected threonine-containing peptide was shown to be less successful than a stepwise strategy using an appropriate 0-glycosyl amino acid as starting material. This route was shown to have more potential since it allows construction of complex glycopeptides containing both glycosyl amino acids and unglycosylatedhydroxy amino acids. Various a-and 0-0-glycopeptides are described in which the threonine molecule linked to galactose is either C-terminal or N-terminal or inserted in the peptide chain.
We have recently reported the one-step synthesis of 0-glycosyl amino acids relevant to g1ycoproteins.l The procedure involved the reaction of a fully benzylated reducing sugar with the appropriate derivative of a serine, threonine, or hydroxyproline molecule, in the presence of trifluoromethanesulfonic anhydride, followed by removal of all protecting groups by hydrogenolysis in the presence of 10% palladium-on-charcoal as catalyst. We now wish to report the synthesis of several threonine-containing 0-glycopeptides of known anomeric configuration in which the threonine molecule linked to galactose is either C- or N-terminal or inserted in the peptide core. In the past few years, natural abundance carbon-13 nuclear magnetic resonance has been used to gain dynamic and structural information about carbohydrate residues of large glycoprotein^.^-^ Application of this technique to the structural study of glycoproteins seems promising, but there are still some temporary limitations. One of these is the lack of 13C NMR data of relevant model compounds needed to make specific assignments in the spectra of glycoproteins. Knowledge of 13C NMR chemical shifts of appropriate model compounds may facilitate the use of this technique to study carbohydrate-peptide linkages in intact glycoproteins. In addition, the glycopeptides reported in this communication were required in the course of our research to provide a better understanding of the structural and conformational properties of the glycosidic bond in glycopeptides as well as the role of the carbohydrate moiety in the binding of metal cation with membrane glycoproteins.
Results and Discussion One approach to the synthesis of glycopeptides is the synthesis of the core peptide containing the desired hydroxy amino acid followed by glycosylation. To be suitable, this approach requires the glycosylation method to be highly stereoselective. Moreover, since the glycosylation 'This work was supported by a research grant from the Dblbgation Gbnhle I la Recherche Scientifique et Technique (D.G.R.S.T.),No. 81 F 0388.
reaction depends upon the reactivity of the hydroxyl group, this reactivity should not be altered by the incorporation of serine, threonine, or hydroxyproline in the peptide chain. In fact, as seen.below, this latter requirement was not achieved in all cases. An alternative approach requires the synthesis of adequately protected 0-glycosyl amino acids, followed by elongation at either or both the C- or N-terminals. As done previously,58 we followed the latter strategy. This route has more potential since it allows construction of complex glycopeptides containing both glycosyl amino acids and unglycosylated hydroxy amino acids. It should be noted that Guillemin and co-workersg attempted to adapt this route to the solid-phase synthesis of N-glycopeptides related to somatostatin. In order to determine the strategy best adapted to the synthesis of any desired glycopeptide, both of these routes were investigated. Peptide blocking strategies are wellknown. However, in the case of glycopeptides, there are additional requirements: (i) Deprotection of peptide residues to allow coupling at the C- and N-terminal must be both selective and compatible with carbohydrate hydroxyl-protecting groups. (ii) The preparation of glycopeptides with the a-anomeric configuration precludes the presence of participating neighboring groups on the carbohydrate moiety. Therefore, benzyl derivatives were preferred to acetyl or benzoyl derivatives. (iii) Amino acid protecting groups, on the other hand, must be stable in acid medium because of the acid-catalyzed nature of the glycosylation reaction.l0 Ideally, the removal of glycosyl (1) J. M. Lacombe, A. A. Pavia, and J. M. Rocheville, Can. J. Chem., 59, 482-489 (1981). (2) K. Dill and A. Allerhand, J.Biol. Chem., 254,4524-4531 (1979). (3) E. Berman, A. Allerhand, and A. L. DeVries, J. Biol. Chem., 255, 4407-44 10 (1980). (4) E. Berman, D. E. Walters, and A. Allerhand, J.Biol. Chem., 256, 3853-3857 (1981). (5) J. Martinez, A. A. Pavia. and F. Winternitz, Carbohydr. Res., 50, 15-22, 148-151 (1976). (6) H. G. Gard and R. W. Jeanloz, Carbohydr. Res., 52,245-250 (1976). (7) H. G. Garg, R. W. Jeanloz, Pept. Proc. Am. Pept. Symp. 5th, 477-479 (1977); Chem. Abstr., 89, 454489 (1978). (8)M. G. Vafina and W. A. Derevitzkaya,Izu. Akad. Nauk SSSR, Ser. Khrm. 1475 (1965). (9) S . Lavielle, N. Ling, R. Saltman, and T. Guillemin, Carbohydr. Res., 89, 229-236 (1981).
0022-326318311948-2557$01.50/0 0 1983 American Chemical Society
2558 J. Org. Chem., Vol. 48, No. 15, 1983
Lacombe and Pavia
Table I. Carbon-13 NMR Chemical Shifts for Glycopeptides Carrying the Galactosyl Residue at the N-Terminal Threonine" co co co c1 Ca Ca Ca CH,' CH, comDounds ester amide urethane galactose Glv Thr Thr Thr Thr B z l , - a - ~-Gal+ Z-L-Thr-ONp( 3 ) Bzl,-p -D -Gal+ Z-L-Thr-ONp (4) Bz1,-a-D-Gal+ Z-~-Thr-Gly-OBzl (7) Bzl,-p-D-Gal-t Z-L-Thr-Gly-OBzl( 8 ) Bzl,-a-D-Gal+ Z-L-Thr-Thr-OBzl ( 9 ) Bzl,-p -D -Gal+ Z-L-Thr-Thr-OBzl( 1 0 )
168.5
156.7
99.2
60.14
19.30
168.3
156.6
101.1
59.5
16.5
40.7
56.3
15.3
41.4
57.5
16.8
168.55
169.25
155.6
97.37
169.28
169.78
156.6
103.9
169.2
170.4
155.7
98.6
56.7
57.9
15.2
19.8
169.3
170.3
156.5
104.8
57.2
58.3
15.6
20.0
a Chemical shifts in ppm relative to Me,Si as internal reference in CDCl,. Abbreviations: Bzl, = 2,3,4,6-tetra-O-benzyl; Z = N-(benzyloxycarbonyl); OBzl = benzyl ester; ONp = o-nitrophenyl ester; Bzl,-Lu-D-Gal+Z-L-Thr-Gly-OBzl= benzyl N-(benzyloxycarbonyl)-0-( 2,3,4,6-tetra-O-benzyl-a-~ -galactopyranosyl)-L-threonylglycinate. Indicates glycosylated threonine.
Scheme I and amino acid protecting groups should require one step. (iv) Finally, since 0-glycopeptides readily undergo @-elmN(Et1) C ~ Z N H C H C O O C ~ H ~ N O Z -tO NH2CHCOzBzl ination, none of the reaction conditions should be very I I THF I I basic. CHCH3 RZ I The choice of threonine instead of serine was determined I 5, R2 = H by the presence of the methyl group on the former. This 0 R' 6, R2= CHOHCH, methyl signal represents an extremely sensitive probe, 3, R' = Bzl,a:-D-Gal which allowed monitoring both the stereochemistry of the 4, R' = Bzl,-p-D-Gal anomeric linkage" and the racemization a t either or both H 2 , Pd/C the asymmetric amino acid centers by 13C NMR specC bzNHC HCONHCHCOzBzl EtW, Hz0, AcOH troscopy.12 Scheme I shows the synthetic route used to I 12 CHCH3 R prepare glycopeptides with galactose linked to the NI terminal threonine. The synthesis began with a triI OR' fluoromethanesulfonic anhydride condensation' of N (benzyloxycarbonyl)-L-threonineo-nitrophenyl ester (l), 7, R' = Bzl,-a-D-Gal; Rz= H 8, R' = Bzl,$-D-Gal; R2 = H prepared according to Bodans~ky,'~ with 2,3,4,6-tetra-O9, R' = Bzl,-a-D-Gal; R2 = CHOHCH, benzyl-D-galactopyranose (2). Pure cy- and @-galactosyl 10, R' = Bzl,-p-D-Gal; R2 = CHOHCH, derivatives (3 and 4) were obtained after silica gel column chromatography and condensed respectively with glycine HzNCHCONHCHCOzH and threonine benzyl esters (5 and 6) as in a previous FI H C H 3 R12 reports and in the presence of a small amount of 1I hydroxybenzotriazole (HOBT). The reaction mixture was I OR1 worked up as usual, and the crude residue was purified by column chromatography. Pure 7, 8, 9, and 10 were ob11, R' = a-D-Gal; R2= H 12, R' = p-D-Gal; R2= H tained in good yield. The corresponding 13C NMR data, ; = CHOHCH, 13, R' = a - ~ - G dR2 reported in Table I, are in excellent agreement with ex14, R' = p-D-Gal; R' = CHOHCH, pectations. Removal of protecting groups by hydrogenation in a water-ethanol-acetic acid mixture using palladium-on-charcoal as catalyst afforded pure O-(CX-D- less satisfactory since the hydrolysis of methyl N-(benzyloxycarbonyl)-0-(2,3,4,6-tetra-0-benzyl-a-~-galactogalactopyranosy1)-L-threonylglycine(1 l),O-(@-D-galactopyranosy1)-L-threoninateand the corresponding @ isomer pyranosy1)-L-threonylglycine (12), O-(a-D-galactoby treatment with a basic resin to produce 15 and 16 was pyranosy1)-L-threonyl-L-threonine (13), and 0(B-Daccompanied by degradation of the starting material (the galactopyranosy1)-L-threonyl-L-threonine(14). @-elimination mentioned before). Moreover, the peptide A second approach to compounds 7 and 8 involved a coupling reaction did not work as well as above. For inBOP ((benzotriazolyl-1-oxy)tris(dimethylamino)stance, compound 16 was obtained in only 40% yield (see phosphonium hexafluorophosphate) synthesis14employing N-(benzyloxycarbony1)-0- (2,3,4,6-tetra-O-benzyl-a-~- Experimental Section, synthesis of 8,Procedure B). The synthesis of glycopeptides carrying the galactosyl galactopyranosy1)-L-threonine(15)and the corresponding residue a t the C-terminal threonine is shown in Scheme 0 isomer (16)with glycine benzyl ester (5). This route was 11. In this particular case, ((9-fluorenylmethy1)oxy)carbonyl (Fmoc) was preferred to (tert-buty1oxy)carbonyl (IO) A. A. Pavia and S. N. Ung-Chhun, Can. J. Chem., 69, 482-489 (Boc) for temporary protection of the amino group. Sev(1981). eral attempts to glycosylate Boc-threonine benzyl ester in (11) A. A. Pavia, S. N. Ung-Chhun, and J. M. Lacombe, Nouo. J. order to prepare Boc analogues of compounds 20 and 21 Chim., 6, 101-108 (1981). (12) A. A. Pavia and J. M. Lacombe, J. Ora. Chem.. followinn Dauer were unsuccessful. This behavior was ascribed to the -. in this issue. presence of trace amounts of free amino acid resulting from (13) M. Bodanszky, M. Kondo, M. L. Link, and G . F. Sigler, J. Org. partial hydrolysis of the Boc substituent. Previous work16 Chem., 39.444-447 (1974): M. Bodanszkv. K. W. Funk. and M. L. Link.
-
ibid., 38, 3565-3570.(1973).
-,
(14) B. Castro, J. R. Dormoy, G . Evin, and C. Selve, Tetrahedron Lett., 14,1219-1222 (1974); B. Castro, G. Evin, C. Selve, and R. Seyer, Synthesis, 413 (1977).
(15) M. Bodanszky and V. Du Vigneaud, J. Am. Chem. SOC.,81, 5688-5691 (1959).
Synthesis of Some 0-Glycopeptides
J. Org. Chem., Vol. 48,No. 15, 1983 2559
Table 11. Carbon-13 NMR Chemical Shifts for Glycopeptides Carrying the Galactosyl Residue at the N-Terminal Threoninea
co
c1 Gal
170.6
156.9
98.6
54.9
19.2
47.4
170.3
156.6
102.2
59.18
17.6
47.3
ester
Bzl 230 OC; [ctI2OD +97' (c 0.9, H20);13CNMR 6 177.3, 172.1 and 169.2 (CO acid and amides), 100.4 (Cl-Gal), 76.9 (CP-Thr), 72.5 (C5-Gal),70.8,70.7 (C3- and C4-Gal),69.9 (C2-Gal),62.5 (C6-Gal), 60.5 (Ca-Thr), 43.8 and 41.9 (Ca-Gly), 19.6 (CH3-Thr). Glycyl-0-(a-D-galactopyranosy1)-I,-threonylglycine (44): mp 230 "c dw; [aIa0~ +49" (c 1.0, H@); l3C NMR 6 176.8, 172.4, 169.2 (CO acid and amides), 100.85 (C1-Gal), 76.0 (CP-Thr), 72.6 (CSGal), 70.6 (C3- and C4-Gal),67.9 (C2-Gal),62.5 (C6-Gal), 59.1 (Ca-Thr), 44.8 and 42.0 (Ca-Gly), 19.1 (CH3-Thr). Registry No. 1, 62087-89-2; 2, 53081-25-7; 3, 86161-21-9; 4, 86117-93-3; 5 tosylate, 1738-76-7;6,33640-67-4; 7,86088-43-9; 8, 86117-94-4; 9, 86088-44-0; io, 86117-95-5; 11, a4m-99-6; 12, 84710-39-4; 13, 84654-00-2; 14, 84710-40-7; 15, 86088-45-1; 16, 86117-96-6; 17, 73724-48-8; 18, 86088-46-2; 19, 86117-97-7; 20, 86088-47-3; 21, 86117-98-8; 22, 1138-80-3; 23, 6154-41-2; 24, 86088-48-4; 25, 86117-99-9; 26, 86088-49-5; 27, 86118-00-5; 28, 84654-01-3; 29, 84710-41-8; 30, 84710-42-9; 31, 84654-02-4; 32, 86088-50-8; 33, a4m-03-5; 34, 16305-79-6; 35, 86088-51-9; 36, 86088-52-0; 37, a5774-46-5; 38,75530-95-9; 39,86088-53-1; doa, soaa-54-2; 408, siia-01-6; 41, a6oaa-55-3;d i p , a6iia-02-7; aa, 86088-56-4; 43, 86088-57-5; 44, 85774-48-7; methyl N-(benzyloxycarbonyl)-0- (2,3,4,6-tetra-O-benzyl-a-~-galactopyranosyl)-~threoninate, 77942-97-3; methyl N-(benzyloxycarbony1)-0(2,3,4,6-tetra-0-benzyl-~-~-galactopyranosyl)-~-threoninate, 77942-98-4; ((9-fluorenylmethy1)oxy)carbonyl chloroformate, 86088-58-6; benzyl L-threoninate hemioxalate, 86088-59-7; dibenzofulvene, 4425-82-5; N-(benzyloxycarbonyl)-L-threonine, 19728-63-3; o-nitrophenyl N-(benzyloxycarbonyl)glycinate, 6154-41-2; Cbz-Gly-Gly-ONO, 86088-60-0; N-(((9-fluorenylmethy1)oxy)carbonyl)-0-tert-butylthreonine,71989-35-0.
