A convergent approach to cyclopeptide alkaloids: Total synthesis of sanjoinine G1

Article abstract of DOI:10.1021/ja0170807

A general strategy for the synthesis of cyclopeptide alkaloids containing an endocyclic aryl-alkyl ether bond has been developed featuring a key intramolecular S_NAr reaction. The importance of the N-terminal protective group in the realization

Full text of DOI:10.1021/ja0170807

Published on Web 01/03/2002
A Convergent Approach to Cyclopeptide Alkaloids: Total
Synthesis of Sanjoinine G1
Taoues Temal-La¨ıb, Jacqueline Chastanet, and Jieping Zhu*
Contribution from the Institut de Chimie des Substances Naturelles, CNRS,
91198 Gif-sur-YVette, France
Received September 14, 2001. Revised Manuscript Received November 5, 2001
Abstract: A general strategy for the synthesis of cyclopeptide alkaloids containing an endocyclic aryl-
alkyl ether bond has been developed featuring a key intramolecular SNAr reaction. The importance of the
N-terminal protective group in the realization of such a strategy is documented. From the appropriate amino
acid constituents, the natural sanjoinine G1, a 14-membered para cyclophane, has been synthesized in
seven steps with 21% overall yield.
Introduction
Cyclopeptide alkaloids are para or meta cyclophanes with a
polypeptidic tether. The widespread occurrence of these 13-,
14-, and 15-membered macrocyclic molecules in different plants
such as rhamnaceae, pendaceae, and rubiaceae has made them
an important class of natural products.1,2 Since the structure
elucidation of pandamine (1, Figure 1) by Goutarel and Pa¨ıs in
1964,3 this family of natural product has grown rapidly and
encompasses nowadays a group of over 200 compounds.
Although they displayed various interesting biological activities,
their limited availability from natural sources has hampered
extensive pharmacological investigations.4 However, Han’s
group5 has convincingly demonstrated that sanjoinine-A (fran-
gufoline) (2) is the major bioactive component of “Sanjoin”, a
plant seed of clinical importance in oriental medicine.
Total synthesis of cyclopeptide alkaloids has been investigated
by a number of groups, notably those of Pa¨ıs,6 Rapoport,7
Schmidt,8 Joullie´,9 Lipshutz,10 and Han.11 The moderately
complex chemical structure of these natural products is in fact
a challenging synthetic target. Various strategies based upon
macrolactamization, intramolecular Michael addition, intramo-
lecular aldol condensation, and intramolecular amide alkylation
have been examined. A seminal contribution results from
Figure 1.
Schmidt’s group who discovered that activation of a carboxyl
group as a pentafluorophenyl ester is particularly efficient for
the desired macrolactamization. On the basis of this methodol-
ogy, the same group has accomplished the first total synthesis
of zizyphine-A (4),8a mucronin-B8e in 1981, and finally fran-
gulanine,8h a 14-membered para cyclophane in 1991. Subse-
quently, other natural products have been synthesized in
Joullie´’s9 and Han’s groups.11 Alternatively, an intramolecular
N-alkylation using amino oxazole as a masked dipeptide has
been elegantly developed by Lipshutz10 for the construction of
natural product analogues. However, intramolecular Michael
addition (and its variants), a possible biomimetic approach, failed
to produce any cyclic compound probably due to the instability
of starting materials and the inherent ring strain associated with
the target molecules.12 It has been thus concluded that cycliza-
(1) For reviews, see: (a) Schmidt, U.; Lieberknecht, A.; Haslinger, E. In The
Alkaloids; Brossi, A., Ed.; Academic Press: New York, 1985; Vol. 26, pp
299-326.
(2) Joullie´, M. M.; Nutt, R. F. In Alkaloids: Chemical and Biological
PerspectiVes; Pelletier, S. W., Ed.; Wiley: New York, 1985; Vol. 3, pp
113-168.
(3) Pa¨ıs, M.; Jarreau, F.-X.; Lusinchi, X.; Goutarel, R. Ann. Chim. (Paris) 1966,
83-105.
(4) (a) Gournelis, D. C.; Laskaris, G. G.; Verpoorte, R. Nat. Prod. Rep. 1997,
75-82. (b) Itokawa, H.; Takeya, K.; Hitotsuyanagi, Y.; Morita, H. In The
Alkaloids; Cordell, G. A., Ed.; Academic Press: New York, 1997; Vol.
49, pp 301-387.
(5) (a) Han, B. H.; Park, M. H.; Han, Y. N. Phytochemistry 1990, 29, 3315-
3319. (b) Han, B. H.; Park, M. H.; Park, J. H. Pure Appl. Chem. 1989, 61,
443-448.
(6) (a) Frappier, F.; Rocchiccioli, F.; Jarreau, F. X.; Pa¨ıs, M. Tetrahedron 1978,
34, 2911-2916. (b) Rocchiccioli, F.; Jarreau, F. X.; Pa¨ıs, M. Tetrahedron
1978, 34, 2917-2926.
(7) (a) Lagarias, J. C.; Houghten, R. A.; Rapoport, H. J. Am. Chem. Soc. 1978,
100, 8202-8209. (b) Goff, D. G.; Lagarias, J. C.; Shih, W. C.; Klein, M.
P.; Rapoport, H. J. Org. Chem. 1980, 45, 4813-4817.
9
10.1021/ja0170807 CCC: $22.00 © 2002 American Chemical Society
J. AM. CHEM. SOC. VOL. 124, NO. 4, 2002 583

A R T I C L E S
Temal-La¨ıb et al.
Scheme 1
tion via formation of an aryl-alkyl ether bond is not feasible
in the presence of various functional groups present in the
peptidic precursor.
In connection with our interests in the total synthesis of
vancomycin and related antibiotics,13 we have been working
on the intramolecular SNAr reaction for the construction of
macrocycles via formation of an endo aryl-aryl ether bond.14-17
This cycloetherification has since been demonstrated to be a
powerful methodology for the synthesis of complex molecules.18
We have attributed the success of this remarkable cycloetheri-
fication to an intramolecular recognition phenomenon that favors
a folded conformer conducive to cyclization.19,20 Several
structural elements found in our previously studied substrates
(8) (a) Schmidt, U.; Lieberknecht, A.; Bo¨kens, H.; Griesser, H. Angew. Chem.,
Int. Ed. Engl. 1981, 20, 1026-1027. (b) Schmidt, U.; Griesser, H.;
Lieberknecht, A.; Talbiersky, J. Angew. Chem., Int. Ed. Engl. 1981, 20,
280-281. (c) Schmidt, U.; Lieberknecht, A.; Grisser, H.; Ha¨usler, J. Angew.
Chem., Int. Ed. Engl. 1981, 20, 281-282. (d) Schmidt, U.; Lieberknecht,
A.; Grisser, H.; Talbiersky, J. J. Org. Chem. 1982, 47, 3261-3264. (e)
Schmidt, U.; Schanbacher, U. Angew. Chem., Int. Ed. Engl. 1983, 22, 152-
153. (f) Schmidt, U.; Lieberknecht, A.; Bo¨kens, H.; Griesser, H. J. Org.
Chem. 1983, 48, 2680-2685. (g) Schmidt, U.; Schanbacher, U. Liebigs
Ann. Chem. 1984, 1205-1215. (h) Schmidt, U.; Za¨h, M.; Lieberknecht,
A. J. Chem. Soc., Chem. Commun. 1991, 1002-1004.
(9) (a) Nutt, R. F.; Chen, K. M.; Joullie´, M. M. J. Org. Chem. 1984, 49, 1013-
1021. (b) Heffner, J.; Joullie´, M. M. Tetrahedron Lett. 1989, 30, 7021-
7024. (c) Heffner, R. J.; Jiang, J.; Joullie´, M. M. J. Am. Chem. Soc. 1992,
114, 10181-10189. (d) Jiang, J.; Li, W. R.; Przeslawski, R. M.; Joullie´,
M. M. Tetrahedron Lett. 1993, 34, 6705-6708. (e) Williams, L.; Zhang,
Z.; Shao, F.; Carroll, P. J.; Joullie´, M. M. Tetrahedron 1996, 52, 11673-
11694. (f) East, S. P.; Joullie´, M. M. Tetrahedron Lett. 1998, 39, 7211-
7214. (g) Xiao, D.; East, S. P.; Joullie´, M. M. Tetrahedron Lett. 1998, 39,
9631-9632. (h) East, S. P.; Shao, F.; Williams, L.; Joullie´, M. M.
Tetrahedron 1998, 54, 13371-13390.
could indeed help their preorganization21 via possible intramo-
lecular attractive forces such as π-π stacking,22 formation of
a charge-transfer complex,23 backbone H-bonding,24 and elec-
trostatic interactions.25 To evaluate the influence of a π-π
interaction on the outcome of cyclization and to further expand
the generality of this methodology, we were interested in
investigating the cyclization involving an alcohol (alkoxide) as
an internal nucleophile instead of a phenoxide. The cyclopeptide
alkaloids were selected as synthetic targets for this purpose. A
unified synthetic scheme featuring a key cycloetherification of
the linear peptide 6 is shown in Scheme 1. Besides a significant
methodological issue, the advantages of employing such a
strategy are as follows: first, two difficult synthetic steps, aryl-
alkyl ether bond formation and macrocyclization, are reduced
into a single operation; second, it uses an intact peptidic linear
precursor making the strategy more convergent. The 14-
membered para cyclophane was chosen since it is more strained
and less accessible than the 13-membered meta cyclophane. The
successful implementation of this strategy and its application
in a total synthesis of sanjoinine G1 are the subjects of the
present paper.26
(10) (a) Lipshutz, B. H.; Hungate, R. W.; McCarthy, K. E. Tetrahedron Lett.
1983, 24, 5155-5158. (b) Lipshutz, B. H.; Hungate, R. W.; McCarthy, K.
E. J. Am. Chem. Soc. 1983, 105, 7703-7713. (c) Lipshutz, B. H.; Huff, B.
E.; McCarthy, K. E.; Miller, T. A.; Mukarram, S. M. J.; Siahaan, T. J.;
Vaccaro, W. D.; Webb, H.; Falick, A. M. J. Am. Chem. Soc. 1990, 112,
7032-7041.
(11) (a) Han, B. H.; Kim, Y. C.; Park, M. K.; Park, J. H.; Go, H. J.; Yang, H.
O.; Suh, D. Y.; Kang, Y. H. Heterocycles 1995, 41, 1909-1914. (b) Suh,
D. Y.; Kim, Y. C.; Han, Y. M.; Han, B. H. Heterocycles 1996, 43, 2347-
2351.
(12) Lawton et al.; see ref 2.
(13) (a) Williams, D. H.; Bardsley, B. Angew. Chem., Int. Ed. 1999, 38, 1173-
1193. (b) Zhu, J. Expert Opin. Ther. Pat. 1999, 9, 1005-1019.
(14) Selected examples: (a) Beugelmans, R.; Singh, G. P.; Bois-Choussy, M.;
Chastanet, J.; Zhu, J. J. Org. Chem. 1994, 59, 5535-5542. (b) Bois-
Choussy, M.; Neuville, L.; Beugelmans, R.; Zhu, J. J. Org. Chem. 1996,
61, 9309-9322. (c) Beugelmans, R.; Bois-Choussy, M.; Vergne, C.;
Bouillon, J. P.; Zhu, J. J. Chem. Soc., Chem. Commun. 1996, 1029-1030.
(d) Bois-Choussy, M.; Vergne, C.; Neuville, L.; Beugelmans, R.; Zhu, J.
Tetrahedron Lett. 1997, 38, 5795-5798. (e) Islas-Gonzalez, G.; Zhu, J. J.
Org. Chem. 1999, 64, 914-924. (f) Bigot, A.; Tran Huu, Dau, M. E.; Zhu,
J. J. Org. Chem. 1999, 64, 6283-6296. (g) Neuville, L.; Bois-Choussy,
M.; Zhu, J. Tetrahedron Lett. 2000, 41, 1747-1751. (h) Bigot, A.; Bois-
Choussy, M.; Zhu, J. Tetrahedron Lett. 2000, 41, 4573-4577.
(15) Other groups’ contribution, selected examples: (a) Boger, D. L.; Zhou, J.
J. Org. Chem. 1996, 61, 3938-3939. (b) Boger, D. L.; Zhou, J.; Borzilleri,
R. M.; Nukui, S.; Castle, S. L. J. Org. Chem. 1997, 62, 2054-2069. (c)
Rama Rao, A. V.; Gurjar, M. K.; Lakshmipathi, P.; Reddy, M. M.;
Nagarajan, M.; Pal, S.; Sarma, B. V. N. B. S.; Tripathy, N. K. Tetrahedron
Lett. 1997, 38, 7433-7436. (d) Carrington, S.; Fairlamb, A. H.; Blagbrough,
I. S. J. Chem. Soc., Chem. Commun. 1998, 2335-2336.
Results and Discussion
Model Studies IsImportance of the N-Terminal Protect-
ing Group. To validate the proposed strategy, a model linear
dipeptide 13 was synthesized as shown in Scheme 2. Coupling
(20) (a) Illuminati, G.; Mandolini, L. Acc. Chem. Res. 1981, 14, 65-102. (b)
Menger, F. M. Acc. Chem. Res. 1985, 18, 128-134. (c) Mandolini, L. Bull.
Soc. Chim. Fr. 1988, 173-176.
(21) Cram, D. J. Angew. Chem., Int. Ed. Engl. 1986, 25, 1039-1057.
(22) (a) Hunter, C. A.; Sanders, J. K. M. J. Am. Chem. Soc. 1990, 112, 5525-
5534. (b) Hunter, C. A. Chem. Soc. ReV. 1994, 101-109. (c) Jones, G. B.;
Chapman, B. J. Synthesis 1995, 475-497.
(23) (a) Mutai, K.; Tukada, H.; Nakagaki, R. J. Org. Chem. 1991, 56, 4896-
4903. (b) Abramovitch, R. A.; Shi, Q. Heterocycles 1994, 38, 2147-2151.
(c) Abramovitch, R. A.; Ye, X.; Pennington, W. T.; Schimek, G.; Bogdal,
D. J. Org. Chem. 2000, 65, 343-351.
(16) Metal-arene complex: (a) Janetka, J. W.; Rich, D. H. J. Am. Chem. Soc.
1997, 119, 6488-6495. (b) Pearson, A. J.; Zhang, P.; Bignan, G. J. Org.
Chem. 1997, 62, 4536-4538.
(17) Reviews: (a) Rama Rao, A. V.; Gurjar, M. K.; Reddy, L.; Rao, A. S. Chem.
ReV. 1995, 95, 2135-2167. (b) Burgess, K.; Lim, D.; Martinez, C. I. Angew.
Chem., Int. Ed. Engl. 1996, 35, 1077-1078. (c) Zhu, J. Synlett 1997, 133-
144. (d) Nicolaou, K. C.; Boddy, C. N. C.; Bra¨se, S.; Winssinger, N. Angew.
Chem., Int. Ed. 1999, 38, 2096-2152. (e) Sawyer, J. S. Tetrahedron 2000,
56, 5045-5065.
(24) Intramolecular hydrogen-bond-directed macrocyclization, see for ex-
amples: (a) Carver, F. J.; Hunter, C. A.; Shannon, R. J. J. Chem. Soc.,
Chem. Commun. 1994, 1277-1280. (b) Neumann, W. L.; Franklin, G. W.;
Sample, K. R.; Aston, K. W.; Weiss, R. H.; Riley, D. P. Tetrahedron Lett.
1997, 38, 779-782. (c) Adria´n, F.; Burguete, M. I.; Luis, S. V.; Miravet,
J. F.; Querol, M.; Garc´ıa-Espan˜a, E. Tetrahedron Lett. 1999, 40, 1039-
1040.
(25) Cation-π interaction, see: (a) Dougherty, D. Science 1996, 271, 163-
167. (b) Nishio, M.; Umezawa, Y.; Hirota, M.; Takeuchi, Y. Tetrahedron
1995, 51, 8665-8701. Electron-rich (donor) atom-electron-deficient
aromatic systems, see: (c) Alkorta, I.; Rozas, I.; Elguero, J. J. Org. Chem.
1997, 62, 4687-4691. (d) Breinlinger, E. C.; Keenan, C. J.; Rotello, V.
M. J. Am. Chem. Soc. 1998, 120, 8606-8609. (e) Gallivan, J. P.; Dougherty,
D. A. Org. Lett. 1999, 1, 103-105.
(26) Preliminary communication: (a) Zhu, J.; La¨ıb, T.; Chastanet, J.; Beugel-
mans, R. Angew. Chem., Int. Ed. Engl. 1996, 35, 2517-2519. (b) La¨ıb,
T.; Zhu, J. Tetrahedron Lett. 1999, 40, 83-86.
(18) Total synthesis of vancomycin aglycon: (a) Evans, D. A.; Wood, M. R.;
Trotter, B. W.; Richardson, T. I.; Barrow, J. C.; Katz, J. L. Angew. Chem.,
Int. Ed. 1998, 37, 2700-2704, 2704-2708. (b) Nicolaou, K. C.; Natarajan,
S.; Li, H.; Jain, N. F.; Hughes, R.; Solomon, M. E.; Ramanjulu, J. M.;
Boddy, C. N. C.; Takayanagi, M. Angew. Chem., Int. Ed. 1998, 37, 2708-
2714, 2714-2716, 2717-2719. (c) Boger, D. L.; Miyazaki, S.; Kim, S.
H.; Wu, J. H.; Loiseleur, O.; Castle, S. L. J. Am. Chem. Soc. 1999, 121,
3226-3227.
(19) Beugelmans, R.; Bourdet, S.; Zhu, J. Tetrahedron Lett. 1995, 36, 1279-
1282. (b) Zhu, J.; Beugelmans, R.; Bourdet, S.; Chastanet, J.; Roussi, G.
J. Org. Chem. 1995, 60, 6389-6396.
9
584 J. AM. CHEM. SOC. VOL. 124, NO. 4, 2002

Total Synthesis of Sanjoinine G1
A R T I C L E S
Scheme 2a
Scheme 3
a Reagents: (a) N-Boc-Phe, Et3N, EDC, HOBt, 85%; (b) (i) HCl-MeCN,
(ii) Et3N, N-Boc-Ser (TBS)OH, EDC, HOBt, 68%; (c) CsF, DMF, 86%.
of 2-(3′-nitro-4′-fluoro) phenylethylamine (10)14a with N-Boc-
Phe afforded dipeptide 11. Removal of N-tert-butyloxy car-
bamate under mild acidic conditions (HCl-MeCN) followed
by its coupling with N-Boc Ser (OTBS) gave the dipeptide 12
in 68% overall yield. Selective deprotection of silyl ether (CsF,
DMF) furnished compound 13, ready for cyclization studies.
Unfortunately, attempted ring closure of 13 under sets of
conditions varying solvents (THF, DMF, DMSO), bases (NaH,
KHDMS, K2CO3, and CsF), temperature, and concentration was
uniformly unsuccessful. In most cases, only degradation of the
starting material was observed leading to intractable tars.
nitrogen is less basic.29 For this reason, cyclization of a dipeptide
with a terminal primary amine and a terminal N,N-dialkylated
amino function was next investigated. Indeed, it has been
demonstrated that the bulky phenylfluorenyl and trityl protecting
groups protect the R-center of amino acids from deprotonation,
thus preventing the racemization30 and â-elimination.31
A simultaneous deprotection of N-carbamate and O-silyl ether
of 12 with HCl-MeCN gave amino alcohol 18 (Scheme 4). In
principle, cyclization of compound 18 could lead to the
formation of two chemically distinct products resulting from
O- and/or N-arylation. Because of the hindered rotation around
the newly formed Csp2-heteroatom bond and the presence of a
nitro group ortho to this linkage, a planar chirality will be created
after cyclization. Thus theoretically, four cyclophanes could be
produced from 18. The results of cyclization under various
conditions are summarized in Table 1. As it is seen, neither
NaH nor K2CO3 was able to promote the cyclization in different
reaction media. However, when TBAF in THF was used as a
base, a mixture of two separable products 19a and 19b (one-
to-one ratio) was isolated in 65% yield. In terms of the yield,
THF was found to be a better solvent than was DMF in this
particular case (entries 6 vs 7).32 Although addition of molecular
sieves was beneficial to the cyclization, other additives such as
sodium sulfate and cesium carbonate blocked completely the
reaction.
Knowing the propensity of N-Boc Serine derivatives to
undergo â-elimination under basic conditions,27 we surmised
that dipeptide 13 may not be the right substrate for the
cyclization studies. The reasoning is that the cyclic compound
14, if produced at all, would be unstable and would decompose
to the acyclic compound 15. This â-elimination process is likely
to be facile due to (1) release of ring strain, and (2) the fact
that ortho nitro substituted phenol is a good leaving group
(Scheme 3). While we were unable to isolate compound 15 at
this stage of work, we did verify later that such a degradation
process is indeed operating. Thus, treatment of dipeptide 16
with NaH in THF afforded a rather complex reaction mixture
from which a dehydroalanine derivative 17 was isolated in 30%
yield.28
It is reasonable to assume that the suspected â-elimination is
initiated by deprotonation of the terminal serine R-CH. Con-
sequently, reducing the thermodynamic and/or kinetic acidity
of this proton should help avoid such a process. Since the
carboxylic group of serine is engaged within the cyclophane
unit, the amino group is the only function that can be
manipulated for this purpose. The acidity of the R-CH and hence
the rate of the racemization of a given amino acid derivative
under basic conditions are highly dependent on the nature of
the nitrogen protecting group and are higher when the protected
The cyclic structure of these two products is readily deduced
from both 1H NMR and mass spectrometry. However, the
available spectroscopic data did not allow us to distinguish
between the N- and O-cyclized compounds (19 and 20,
(29) Hobbs, D. W.; Still, W. C. Tetrahedron Lett. 1987, 28, 2805-2808.
(30) Lubell, W.; Rapoport, H. J. Org. Chem. 1989, 54, 3824-3831.
(31) (a) Cherney, R. J.; Wang, L. J. Org. Chem. 1996, 61, 2544-2546. (b)
Sowinski, J. A.; Toogood, P. L. J. Org. Chem. 1996, 61, 7671-7676.
(32) In accord with the ionic mechanism of SNAr, cycloetherification proceeded
generally faster in DMF and DMSO than in THF. However, in some cases,
THF was found to be a solvent of choice although longer reaction time
was required. For examples in the biaryl ether series, see refs 14g and
19b.
(27) Examples, see: (a) Photaki, I. J. Am. Chem. Soc. 1963, 85, 1123-1126.
(b) Olsen, R. K. J. Org. Chem. 1970, 35, 1912-1915. (c) Cheung, S. T.;
Benoiton, N. L. Can. J. Chem. 1977, 55, 906-910. (d) Wojciechowska,
H.; Pawlowicz, R.; Andruszkiewicz, R.; Grzybowska, J. Tetrahedron Lett.
1978, 4063-4064. (e) Bajgrowicz, J. A.; El Hallaoui, A.; Jacquier, R.;
Pigiere, C.; Viallefont, P. Tetrahedron 1985, 41, 1833-2843.
(28) La¨ıb, T.; Zhu, J. Synlett 2000, 1363-1365.
9
J. AM. CHEM. SOC. VOL. 124, NO. 4, 2002 585

A R T I C L E S
Temal-La¨ıb et al.
Scheme 4
Scheme 5
(Ac2O, DMAP, Et3N, CH2Cl2) of individual cyclic compounds
19a and 19b gave mono acetylated derivatives whose structures
were determined to be the 13-membered cyclophanes 21a and
21b, respectively, on the basis of the following facts. First, in
the 1H NMR spectra of cyclophanes 21a and 21b, both protons
Ha moved to low field while Hb was almost unchanged as
compared to their starting materials. If it was 22 resulting from
the acylation of the O-cyclized cyclophane 20, we would expect
the low-field shift of proton Hb after N-acylation. Second, the
appearance of a strong absorption peak at 1737 and 1749 cm-1
in the IR spectra of 21a and 21b indicated the presence of an
ester rather then an amide bond. Furthermore, treatment of the
cyclic compound 19a with 1,1′-carbonyldiimidazole produced
an oxazolidinone 23 which definitively discarded the structure
20 as a primary cyclic compound. From these studies, we
concluded that cyclization of 18 produced the two atropisomers
of the 13-membered para cyclophanes 19a and 19b at the
expense of the 14-membered para cyclophane, presumably
because the nitrogen nucleophile is more accessible. Attemps
to convert the compound 19 into 20 via Smiles rearrangement
failed under various basic conditions.
Depending on the reaction conditions, an intermolecular SN-
Ar reaction between an amino alcohol and a fluoro nitro
aromatic compound can produce chemoselectively the N- or
O-arylated product. Thus a strong base (NaH) promotes the
O-arylation, while a weaker base (e.g., K2CO3) favors the
N-arylation.33 In view of the peculiar property of ionic fluoride
and its known ability to promote the O-alkylation and O-
arylation,34 the exclusive N-cyclization of 18 was intriguing.
To learn whether this chemoselective cyclization is a substrate-
controlled or a reagent-controlled process, an intermolecular
condensation between 2-fluoro nitrobenzene 24 and (R)-phenyl
glycinol 25 was performed using TBAF as a base in THF. As
shown in Scheme 5, TBAF is capable of promoting both the
N- and the O-arylation of amino alcohol. The result of this
control experiment clearly indicated that the high chemoselec-
tivity observed in the cyclization of 18 is inherent to this
particular substrate and may not be a general phenomenon. The
observed shorter distance of N-CF than that of O-CF in an
energy-minimized conformation may explain in part the ob-
served selectivity.
Table 1. Cyclization of 18 to 19: Survey of Cyclization
Conditionsa
base
additive
solvent
T°C
time (h)
yield (%)
NaH
NaH
NaH
DMF
DMF
THF
DMF
DMSO
THF
DMF
THF
THF
20
60
20
60
20
20
20
90
20
20
20
24
48
24
24
24
48
48
48
48
48
12
0
0
0
0
0
65
50
0
0
0
K2CO3
K2CO3
TBAF
TBAF
TBAF
TBAF
TBAF
TBAF
Na2SO4
Cs2CO3
3 Å
THF
THF
68
a Five equivalents of base was employed.
(33) Beugelmans, R.; Bigot, A.; Zhu, J. Tetrahedron Lett. 1994, 35, 5649-
5652. (b) Neuville, L.; Chastanet, J.; Zhu, J. Tetrahedron Lett. 1999, 40,
7087-7090.
respectively, Scheme 4). A chemical transformation was thus
carried out to gain more structural information. Acetylation
(34) Clark, J. H. Chem. ReV. 1980, 80, 429-452.
9
586 J. AM. CHEM. SOC. VOL. 124, NO. 4, 2002

Total Synthesis of Sanjoinine G1
A R T I C L E S
Scheme 6a
Figure 2.
Figure 3.
showed that compounds 32a and 32b did not undergo equilib-
rium under the cyclization conditions, indicating therefore that
they are kinetic products. The removal of the nitro group was
carried out by a two step sequence. Reduction of a mixture of
35
31a and 31b with sulfurated NaBH4
in THF gave the
corresponding amino compounds 33, and the two atropisomers
became separable at this stage by flash chromatography.
Diazotization of 33a and 33b under Doyle’s conditions36
followed by FeSO4-mediated reduction of the crude diazonium
salt37 gave the desired cyclophane 35 (R ) allyl) in 65% overall
yield. The same sequence applied to 32a and 32b gave the amino
compounds 34a and 34b and then the reduced cyclophane 36
1
(R ) Bn) in 67% yield. As is seen in the H NMR spectra of
a Reagents and Conditions: (a) THF/DMSO (4/1), allyl bromide, or
BnBr, NaHCO3, 80 °C, 70-85%; (b) TBAF, DMF, molecular sieves, room
temperature, 4 h, 60-75%; (c) NaBH4, elemental sulfur, THF, reflux, 81%;
(d) (i) t-BuONO, BF3‚Et2O, CH2Cl2, -15 °C, (ii) FeSO4, DMF, room
temperature, 65-75%.
compounds 35 and 36, the four protons in the para disubstituted
benzene ring are chemically inequivalent, each being an apparent
doublet of doublet. Such a splitting pattern is indicative of the
severe ring strain, which hindered the free rotation around aryl-
alkyl ether bond.
Although cylization of 18 provided the wrong regioisomers,
occurrence of such a reaction under defined conditions was
encouraging, and we set out to study the cyclization of substrates
with a terminal N,N-dialkyl amino function. Chemoselective
N,N-bisallylation or N,N-bisbenzylation of 18 in the presence
of primary hydroxyl function was realized by using NaHCO3
as a base in mixed solvent (DMSO/THF ) 1/4) to afford 29
and 30 in 78% yield (Scheme 6).
When a DMF solution of 29 (0.01 M) was treated with TBAF
in the presence of molecular sieves at room temperature for 4
h, a mixture of two cyclized monomers 31a and 31b was isolated
in 75% overall yield. THF could also be used as solvent for the
cyclization; however, a significantly prolonged reaction time
(4 days) was required to complete the cyclization. Similar
treatment of compound 30 afforded two separable products 32a
(32%) and 32b (48%), whose structure was assigned as two
atropisomers from spectroscopic studies and was confirmed by
subsequent chemical transformation. A control experiment
The key SNAr-based cyclization was routinely carried out in
0.01 M DMF. A higher dilution was unnecessary; however,
variable amounts of cyclic dimer 37 (Figure 2) were observed
when the cycloetherification of 30 was performed at a concen-
tration of 0.05 M under otherwise identical conditions. To gain
information regarding the solution conformations of the cy-
clization precursor, we carried out a computational simulation
(macromodel, Batchmin version 3.5 a, OPLS force field,38 water
set). The calculated lowest energy conformer of 30 (Figure 3)
has a bent orientation placing the two reactive sites (HO and
CF) within a distance of 4.67 Å. Such a “near attack conforma-
tion”39 should thus reduce the entropy loss of the cyclization
(35) Lalancette, J. M.; Freˆche, A.; Brindle, J. R.; Laliberte´, M. Synthesis 1972,
526-532.
(36) Doyle, M. P.; Bryker, W. J. J. Org. Chem. 1979, 44, 1572-1574.
(37) Wassmundt, F. W.; Kiesman, W. F. J. Org. Chem. 1995, 60, 1713-1719.
(38) Jorgensen, W. L.; Tirado-Rivers, J. J. Am. Chem. Soc. 1988, 110, 1657-
1666.
(39) Bruice, T. C.; Lightstone, F. C. Acc. Chem. Res. 1999, 32, 127-136.
9
J. AM. CHEM. SOC. VOL. 124, NO. 4, 2002 587

A R T I C L E S
Temal-La¨ıb et al.
Scheme 7a
Scheme 8
Scheme 9
a Reagents and Conditions: (a) L-N-Boc Leu (39), EDC, HOBt, 92%;
(b) (i) HCl-MeCN (15% v/v), (ii) EDC, HOBt, L-N,N-dibenzyl Ser (41),
80%; (c) (i) TBAF, DMF, molecular sieves 3 Å, (ii) Ac2O, Et3N, DMAP,
70%; (d) NaBH4, elemental sulfur; (e) (i) t-BuONO, BF3‚OEt2, (ii) Fe,
FeSO4, DMF, 65%.
process, resulting in a low activation energy for cyclization.
Besides protecting the R-CH of serine from deprotonation, it is
also conceivable that the presence of the bulky bisalkylated
amino substituent may impart an element akin to a gem-dimethyl
effect40 which preorganizes the substrate into a productive
conformation.
To further validate the potential of this strategy in the
synthesis of a cyclopeptide alkaloid, linear peptide 42 incor-
porating an optically pure (R)-2-amino-1-(4′-fluoro-3′-nitro)
ethanol41 was synthesized (Scheme 7). Treatment of a DMF
solution of 42 (0.01 M) with TBAF in the presence of molecular
sieves (3 Å) at room temperature gave two separable atropiso-
mers 43a (35%) and 43b (35%) after acylation. Reductive
removal of a nitro function provided then the cyclophane 45 in
65% overall yield. Thus, the presence of a hydroxyphenethyl-
amine unit did not affect the efficiency of the cycloetherification
reaction.
To verify whether the stereochemistry of the chiral secondary
benzyl alcohol can exert any influence on the macrocyclization
reaction, compound 46, an epimer of 42, was synthesized
starting from (S)-2-amino-1-(4′-fluoro-3′-nitro) ethanol (Scheme
8). Cyclization of 46 under the standard conditions (TBAF,
DMF, MS 3 Å) provided two separable cyclophanes 47a and
47b in 65% yield. The same efficiency observed between the
cyclization of 42 and 46 indicated that the SNAr-based cyclo-
etherification was insensitive to the chirality of the tether chain,
expanding further its general applicability in organic synthesis.
Overall, by simply tuning the N-protective group from N-tert-
butyloxycarbamate to N,N-dialkyl amino function, we were able
to bypass the otherwise dead-end of a synthetic strategy.42
Model Studies IIsActivation Level of Aromatic Ring for
SNAr Reaction. The presence of electron-withdrawing groups
facilitates the nucleophilic aromatic substitution.43 We thus
reasoned that introduction of a carbonyl function at the benzylic
position of 30 should further improve the reaction outcome. If
(40) Jung, M. E.; Gervay, J. Tetrahedron Lett. 1988, 29, 2429-2432.
(41) La¨ıb, T.; Ouazzani, J.; Zhu, J. Tetrahedron: Asymmetry 1998, 9, 169-
(42) For an interesting review, see: Sierra, M. A.; Torre, M. C. Angew. Chem.,
Int. Ed. 2000, 39, 1538-1559.
178.
9
588 J. AM. CHEM. SOC. VOL. 124, NO. 4, 2002

Total Synthesis of Sanjoinine G1
A R T I C L E S
Scheme 10a
a Reagents and Conditions: (a) EDC, C6F5OH, CH2Cl2; (b) DMF, 50, 60 °C, 75%; (c) TBAF, DMSO, 85 °C, then Ac2O, Et3N, DMAP, CH2Cl2, 45%;
(d) SnCl2, DMF, 60 °C; (e) NaNO2, H3PO2, Cu2O, THF-H2O, 73%; (f) Pd(OH)2, THF-t-BuOH, quantitative; (g) L-N,N-dimethyl phenylalanine, EDC,
HOBt; (h) K2CO3, MeOH-H2O, 85%.
cyclization occurred, the reagent-controlled reduction of the
carbonyl function44 would provide directly the core structure
of natural product with a stereochemically defined hydroxy-
phenethylamine unit.9g
Vulgaris) and was shown to possess an interesting sedative
effect.5 In fact, sanjoin has been frequently used in oriental
medicine as an important and reliable hypnotic or sedative agent
for the treatment of insomnia. Two total syntheses have been
accomplished by Han11a and Joullie´9f,h employing a key mac-
rolactamization reaction.
A PCC-mediated oxidation of 48 in buffered dichloromethane
solution provided linear peptide 49 (Scheme 9). Surprisingly,
treatment of 49 under conditions identical to those established
earlier provided the dipeptide amide 50 as the only isolable
compound in 46% yield. The structure of 50 was fully
determined by comparison with an authentic sample obtained
by amidation of the corresponding dipeptide acid. The fact that
the C-terminal of compound 50 was an amide instead of an
acid eliminated the simple hydrolysis mechanism and implied
a rather complex reaction manifold. Failure to isolate the other
part of the degraded molecule prohibited any meaningful
mechanistic speculation. Nevertheless, oxidation of the carbon
R to the ketone function by adventitious oxygen followed by
fragmentation of the so-produced hemiaminal could be one of
the possible reaction pathways leading to 50. Other conditions
varying the base and the solvent were briefly surveyed for the
cyclization of 49. Unfortunately, none of them gave the expected
cyclophane. In view of the successful cyclization of compound
42, we turned our attention on its application in the total
synthesis of natural products.
Our synthesis of sanjoinine G1 was shown in Scheme 10.
The (2S,3S)-N,N-dibenzyl hydroxyleucine 51 was prepared from
(2S,3S)-hydroxyleucine,45 by a sequence of esterification, N,N-
bisbenzylation, and saponification under classic conditions in
excellent overall yield. The coupling of acid 51 with amine 52
proved to be more difficult than originally anticipated. At-
tempted one pot activation/coupling using various coupling
reagents under diverse conditions provided the desired peptide
in low yield. The â-lactone 53 resulting from the intramolecular
attack of the â-hydroxyl group onto the activated carboxylic
function was found to be the major product in most of the cases.
Eventually, the best condition found was activation of the
carboxylic acid as its pentafluorophenol ester 54.46 Heating a
mixture of 52 and 54 in DMF at 60 °C then reproducibly
provided the dipeptide 55 in higher than 75% yield. Activation
of 51 as its acyl fluoride47 followed by coupling with 52 gave
a lower yield of 55.
Cyclization of 55 was best realized in DMSO in the presence
of TBAF and 3 Å molecular sieves at 85 °C. Under these
conditions, the desired 14-membered ansa cyclophane 56 was
obtained in 45% yield after O-acylation. The reacting secondary
alcohol in compound 55 is highly hindered since it is flanked
Total Synthesis of Sanjoinine G1. Sanjoinine G1 (3) (Figure
1) has recently been isolated from sanjoin (seed of Zizyphus
(43) (a) Terrier, F. Nucleophilic Aromatic Displacement: The Role of the Nitro
Group; VCH: New York, 1991; Chapter 1. (b) Paradisi, C. Arene
Substitution via Nucleophilic Addition to Electron Deficient Arenes.
ComprehensiVe Organic Synthesis; Pergamon Press: Oxford, 1991; Vol.
4, pp 423-450. (c) Vlasov, V. M. J. Fluorine Chem. 1993, 61, 193-216.
(44) Cyclopeptide alkaloid with a ketone function has been isolated, see: Pais,
M.; Marchand, J.; Ratle, G.; Jarreau, F.-X. Bull. Soc. Chim. Fr. 1968, 2979-
2984.
(45) La¨ıb, T.; Chastanet, J.; Zhu, J. J. Org. Chem. 1998, 63, 1709-1713.
(46) Kisfaludy, L.; Scho¨n, I. Synthesis 1983, 325-327.
(47) Carpino, L.; Sadet-Aalaee, D.; Chao, H. G.; DeSelms, R. H. J. Am. Chem.
Soc. 1990, 112, 9651-9652.
9
J. AM. CHEM. SOC. VOL. 124, NO. 4, 2002 589

A R T I C L E S
Temal-La¨ıb et al.
by a bulky isopropyl and a N,N-dibenzylamino group at the
adjacent positions. This unfavorable steric effect may account
for the moderate cyclization yield. It is interesting to note that
only one atropisomer was isolated from the reaction mixture.
From detailed NMR studies, especially the observation of the
NOE cross-peak of protons Ha and Hb, a M configuration was
assigned for the newly created planar chirality.48
for the formation of an aryl-alkyl ether bond under mild
conditions. Thus treatment of linear peptide (generic structure
6, Scheme 1) with TBAF in polar aprotic solvents (THF, DMF,
or DMSO) provided the 14-membered para cyclophane in good
yield. The importance of the N-terminal protective group in the
realization of such a strategy has been demonstrated. Indeed,
cyclization of peptide 13 containing a terminal N-Boc serine
residue afforded intractable tars, while that incorporating a N,N-
dialkyl serine terminal (29, 30, 42, 46, and 55) cyclized to
provide the desired macrocycles. Reducing the kinetic/thermo-
dynamic acidity of the R-CH proton of serine, thus protecting
the strained para cyclophane from the destructive â-elimination,
has been the rationale behind such a productive N-protective
group tuning. The overall synthetic scheme is highly convergent
since only two conventional couplings are required to reach the
cyclization precursor. The nitro group, after serving as an
activating group for the SNAr-based cycloetherification, can be
efficiently removed according to literature procedures. From the
appropriate amino acid constituents, the natural sanjoinine G1
has been synthesized in seven steps with 21% overall yield
which is superior to previous syntheses.
Reduction of nitro to amino group (SnCl2, DMF) followed
by one-pot reductive-deamination provided compound 57.
Lalancette’s conditions (NaBH4, S8) were found to be inefficient
for the reduction of nitro in contrast to our model studies.
Hydrogenolysis of the N-benzyl group was realized using
Pearlman’s catalyst in THF-t-BuOH to provide 58 in quantitative
yield. When hydrogenolysis was carried out in methanol or
ethanol, partial transesterification occurred to give a deacetylated
compound, complicating thus the synthetic operations. The
benzylic acetoxy group was perfectly stable under both the tin
chloride (acidic)-mediated reduction of nitro group and the
hydrogenolysis conditions. Examination of a Dreiding model
as well as molecular modeling studies showed that the σ bond
Cc-Cd in macrocycles lies out of the plan defined by the
aromatic ring. Such an unusual bond organization, inherent to
the strained 14-membered para cyclophane, led to the loss of
benzylic carbon character of Cd, accounting for the stability of
the Cd-oxygen bond under the reducing conditions. Finally,
coupling of the crude hydrogenolysis product with L-N,N-
dimethyl phenylalanine followed by saponification gave san-
joinine G1 (3) in 85% overall yield whose spectroscopic data
is in all respect identical to those reported in the literature.
In conclusion, we have devised and executed an efficient and
unified synthetic strategy to the family of cyclopeptide alkaloids.
This novel synthesis features a key intramolecular SNAr reaction
Experimental Section
Full experimental details and compound characterizations are
provided in the Supporting Information.
Acknowledgment. We thank CNRS for support of this
research. A doctoral fellowship from “Ligue Nationale Contre
le Cancer” to T.T.L. is gratefully acknowledged.
Supporting Information Available: Experimental procedures
and spectroscopic and analytical data for all compounds (PDF).
This material is available free of charge via the Internet at
http://pubs.acs.org.
(48) A P configuration was erroneously assigned in our preliminary communica-
tion (ref 26b).
JA0170807
9
590 J. AM. CHEM. SOC. VOL. 124, NO. 4, 2002