A caveat in the application of the exciton chirality method to N,N- dialkyl amides. Synthesis and structural revision of AT2433-B1 [12]

Full text of DOI:10.1021/ja9842924

J. Am. Chem. Soc. 1999, 121, 3801-3802
3801
Scheme 1
A Caveat in the Application of the Exciton Chirality
Method to N,N-Dialkyl Amides. Synthesis and
Structural Revision of AT2433-B1
John D. Chisholm,† Jerzy Golik,‡ Bala Krishnan,‡
James A. Matson,‡ and David L. Van Vranken*,†
Department of Chemistry, UniVersity of California
IrVine, California 92697
Bristol-Myers Squibb Pharmaceutical Research Institute
5 Research Parkway, Wallingford, Connecticut 06492-7660
Scheme 2
Scheme 3
ReceiVed December 14, 1998
Circular dichroism (CD) is a powerful technique for absolute
stereochemical assignment.1,2 Dibenzoyl derivatives of vicinal 1,2-
amino alcohols exhibit strong anisotropic absorption of circularly
polarized light due to the interaction between gauche benzoyl
chromophores. This differential absorption is expressed as an
exciton couplet in the CD spectrum. The diagnostic relationship
between exciton couplet and absolute stereochemistry is the basis
for the exciton chirality method. The benzamide configuration is
critical for stereochemical assignment.3 For amino alcohol deriva-
tives, E and Z amides possess an enantiotopic relationship between
chromophores that produces opposite CD curves. This deceptive
effect of N,N-dialkyl amide conformation was revealed during
the first total synthesis of the antitumor antibiotic AT2433-B14,5
from Actinomadura melliaura.
fundamental patterns of reactivity, and expanding the range of
available derivatives. Derivatives have been particularly important
for understanding and improving the efficacy of topoisomerase
inhibitors such as camptothecin, rebeccamycin, and NB-506.11,12
Mannich cyclization of N-methylarcyriarubin A 1 is ac-
companied by an undesired opening of the indoline ring to give
carbazole 2.13,14 The corresponding bisindolylsuccinimide, how-
ever, cyclizes efficiently in neat trifluoroacetic acid (TFA) to
afford indoline (()-3 as the cis-syn-cis diastereomer (Scheme 1).
The two carbohydrate subunits of AT2433 were ultimately
derived from D-glucose15 and D-serine. The stereochemistry of
the amino sugar was established by Roush allylboration of
D-Garner’s aldehyde.16
Coupling of the primary alcohol 5 was performed under
Mukaiyama-Nicolaou conditions with little control of anomeric
stereochemistry (Scheme 3).17 For the synthesis of AT2433-B1,
carbamate 7 was N-methylated to give 8. Disaccharide 8 was
debenzylated and coupled with 3 equiv of indoline 3 in refluxing
methanol. Oxidation with DDQ afforded the fully aromatic
indolocarbazole glycoside. Deprotection of the Boc group com-
pleted the synthesis but was accompanied by a significant amount
of hydrolysis, the 2-deoxyglycoside being labile.
The Mannich dimerization of indoles6 provides an efficient
method for the construction of indolocarbazole glycosides7
because the indoline products can efficiently capture unprotected,
unactivated carbohydrates.8-10 This mild glycosylation is stereo-
selective and produces the equatorial N-glycoside. The amino-
disaccharide moiety of the AT2433 natural products challenges
traditional methods for indole glycosylation and provides an
opportunity to validate the efficiency of carbohydrate capture by
indolines. Total synthesis serves an important role in new drug
development by confirming structural assignment, revealing
† University of California.
‡ Bristol-Myers Squibb Pharmaceutical Research Institute.
(1) Harada, N.; Nakanishi, K. Circular Dichroic Spectroscopy: Exciton
Coupling in Organic Stereochemistry; University Science Books: New York,
1983.
Both glycoside 10 and AT2433-B1 exist in two conformations
in d6-DMSO. At room temperature (slow exchange), or at 150
(2) Nakanishi, K.; Berova, N. In Circular Dichroism: Principles and
Applications; Nakanishi, K., Berova, N., Woody, R. W., Eds.; VCH: New
York, 1994; p 361.
(11) Anizon, F.; Belin, L.; Moreau, P.; Sancelme, M.; Voldoire, A.;
Prudhomme, M.; Ollier, M.; Severe, D.; Riou, J. F.; Bailly, C.; Fabbro, D.;
Meyer, T. J. Med. Chem. 1997, 40, 3456.
(12) Ohkubo, M.; Kawamoto, H.; Ohno, T.; Nakano, M.; Morishima, H.
Tetrahedron 1997, 53, 585.
(3) Polonski, T. J. Org. Chem. 1993, 58, 258.
(4) Golik, J.; Doyle, T. W.; Krishnan, B.; Dubay, G.; Matson, J. A. J.
Antibiot. 1989, 42, 1784.
(5) Matson, J. A.; Claridge, C.; Bush, J. A.; Titus, J.; Bradner, W. T.; Doyle,
T. W.; Horan, A. C.; Patel, M. J. Antibiot. 1989, 42, 1547.
(6) Bergman, J.; Koch, E.; Pelcman, B. Tetrahedron Lett. 1995, 36, 3945.
(7) Gribble, G. W.; Berthel, S. J. In Studies in Natural Products Chemistry;
Atta-ur-Rahman, Ed.; Elsevier Science Publishers: Dodrecht, 1993; Vol. 12.
(8) Chisholm, J. D.; Van Vranken, D. L. J. Org. Chem. 1995, 60, 6672.
(9) Gilbert, E. J.; Van Vranken, D. L. J. Am. Chem. Soc. 1996, 118, 5500.
(10) Gilbert, E. J.; Ziller, J. W.; Van Vranken, D. L. Tetrahedron 1997,
53, 16553.
(13) Steglich, W. Pure Appl. Chem. 1989, 61, 281.
(14) Joyce, R. P.; Gainor, J. A.; Weinreb, S. M. J. Org. Chem. 1987, 52,
1177.
(15) Samuelsson, B.; Johansson, R. J. Chem. Soc., Perkin Trans. 1 1984,
2371.
(16) Roush, W. R.; Hunt, J. A. J. Org. Chem. 1995, 60, 798.
(17) Groneberg, R. D.; Miyazaki, T.; Stylianides, N. A.; Schulze, T. J.;
Stahl, W.; Schreiner, E. P.; Suzuki, T.; Iwabuchi, Y.; Smith, A. L.; Nicolaou,
K. C. J. Am. Chem. Soc. 1993, 115, 7593.
10.1021/ja9842924 CCC: $18.00 © 1999 American Chemical Society
Published on Web 04/07/1999

3802 J. Am. Chem. Soc., Vol. 121, No. 15, 1999
Communications to the Editor
Scheme 6
Figure 1. (a) CD spectrum of 11. (b) CD spectra of 14a and 14b. All
spectra were recorded in acetonitrile at a 40 mM concentration.
Scheme 7
Scheme 4
Scheme 5
was confirmed by Mosher’s ester analysis. The discrepancy
between 10 and AT2433-B1 was now clear; the puzzling results
of the CD were not.
To explain the discrepancy between the CD of 14a and 11,
we directed our attention to the tertiary amide moiety. Molecular
modeling indicates only a small energy difference between E and
Z amide conformations of 14a, and this is supported by the
existence of a 1:1 ratio of signals in the 1H NMR (in CD3CN).21
As shown in Scheme 6 the aryl chromophores in the E amide
adopt a clockwise orientation, whereas the aryl chromophores
in the Z amide adopt a counterclockwise orientation. Thus,
E and Z amides should give opposite Cotton effects. The
resulting CD spectrum is dictated by two factors that are difficult
to predict: (i) the relative population of conformers and (ii) the
degree to which the two chromophores interact in each conformer.
These combined effects can generate either positive or negative
chirality, depending on which conformation dominates the CD
spectrum.
Figure 1b highlights the deceptive effects of N,N-dialkyl amide
conformation. The negative chirality of N-alkyl amide 14b
corrrectly correlates with the amino alcohol configuration; the
positive chirality of N,N-dialkyl amide 14a does not. The
diminished amplitude of the CD curve of 14a corroborates the
opposing effects of E and Z conformers.22
Having established the correct configuration of the amino sugar
as 3S, 4S (rather than 3R, 4R), we repeated our original synthesis
with L-Garner’s aldehyde and prepared the enantiomer of pyra-
nosyl fluoride 6. Glycosylation with glucoside 5, installation of
the methyl group, and debenzylation provided disaccharide 15.
N-Glycosylation, oxidation, and deprotection of the Boc group
provided material identical with AT2433-B1 (Scheme 7).
This work has capitalized upon a Mannich cyclization/
glycosylation route for the synthesis of indolocarbazole glycosides.
It has led to a structural revision of the AT2433 series of antitumor
antibiotics. Finally this work exposes the need for caution when
N,N-dialkyl amides are involved in determination of absolute
stereochemistry of amino alcohols using exciton chirality.
°C (rapid exchange), the NMR spectra were similar, but did not
match. In addition, glycoside 9 (prepared by the same route) did
not match authentic AT2433-B2. The tetraacetate of 9 was
prepared and the stereochemistry of the â-glucoside moiety and
the connectivity were confirmed using J analysis and HMBC,
respectively. At this point we were confident that 9 corresponded
with AT2433-B2 in the following ways: (i) the connectivity was
the same, (ii) the structure of the 4-O-methyl-â-D-glucoside was
secure, and (iii) the relative stereochemistry of the amino sugars
was the same.
We were led to reconsider the absolute stereochemistry of the
amino sugar moiety. Other natural products isolated from Acti-
nomadura (calicheamicin and esperamicin) possess the enantiomer
of the AT2433 amino sugar.18,19 The absolute stereochemistry of
the AT2433 amino sugar was originally assigned by application
of the exciton chirality method to a dibenzoyl derivative 11
derived from methanolysis of AT2433-A1.4 The absolute stereo-
chemistry was assigned as 3R, 4R based on the negative chirality
of the exciton couplet (Figure 1a).
To verify the absolute configuration of the amino sugar moiety,
independent synthesis of pyranoside 11 was required. Lactol 12
was converted to 14a by the route shown in Scheme 5. Suprisingly
the exciton couplet for synthetic 14a exhibited positive chirality
(Figure 1b), whereas amino alcohol derivatives with the same
absolute configuration as 14a should exhibit negative chirality.20
To remove any doubts about the configuration of amino alcohol
derivative 14a, the absolute stereochemistry of intermediate 13b
(18) Lee, M. D.; Dunne, T. S.; Chang, C. C.; Ellestad, G. A.; Siegel, M.
M.; Morton, G. O.; McGahren, W. J.; Borders, D. B. J. Am. Chem. Soc. 1987,
109, 3466.
(19) Golik, J.; Clardy, J.; Dubay, G.; Groenewold, G.; Kawaguchi, H.;
Konishi, M.; Krishnan, B.; Ohkuma, H.; Saitoh, K.; Doyle, T. W. J. Am. Chem.
Soc. 1987, 109, 3461.
Acknowledgment. This work was supported by the American Cancer
Society, PRF, and the UC Cancer Research Coordinating Committee.
D.V.V. also gratefully acknowledges the support of GlaxoWellcome.
(20) Kawai, M.; Nagai, U.; Katsumi, M. Tetrahedron Lett. 1975, 16, 3165.
(21) One of the reviewers has suggested an alternative E amide conforma-
tion for 14a in which the aroyl group is “axial”. This conformation should
also lead to a positive couplet. MM2 and AM1 calculations (gas, H2O) predict
this conformation to be unfavorable relative to E and Z amide conformations
shown in Scheme 6.
Supporting Information Available: Procedures and characterization
data for compounds 3, 5-10, 12-15, and results of mechanics and
semiempirical calculations for conformations of 14a (PDF). This material
is available free of charge via the Internet at http://pubs.acs.org.
(22) Lo, L. C.; Berova, N.; Nakanishi, K.; Morales, E. Q.; Vazquez, J. T.
Tetrahedron Asymm. 1993, 4, 321.
JA9842924