Progressive-convergent elucidation of stereochemistry in complex polyols. The absolute configuration of (-)-sagittamide A

Article abstract of DOI:10.1021/ja063735y

The absolute stereostructure of sagittamide A (1), a O-hexacetyl long-chain hexahydroxy-α,ω-dicarboxylic acid, was assigned using a progressive-convergent approach that integrates three powerful regimens for stereochemical analysis of acyclic natural prod

Full text of DOI:10.1021/ja063735y

Published on Web 08/18/2006
Progressive-Convergent Elucidation of Stereochemistry in Complex Polyols.
The Absolute Configuration of (-)-Sagittamide A
Sarah C. Lievens† and Tadeusz F. Molinski*,†,‡
Department of Chemistry and Biochemistry, UniVersity of California, San Diego, 9500 Gilman DriVe,
La Jolla, California 92093, and Department of Chemistry, UniVersity of California, DaVis, One Shields AVenue,
DaVis, California 95616
Received June 13, 2006; E-mail: tmolinski@ucsd.edu
Scheme 1a
The structure of (-)-sagittamide A (1)1, an unprecedented
polyacetoxy, long-chain R,ω-dicarboxylic acid isolated from a
tropical didemnid tunicate, was solved by application of conven-
tional 2D NMR spectroscopic methods; however, only partial
stereochemistry could be assigned. Although configurations of the
terminal amino acids (L-ornithine and L-valine) were determined
readily by conventional methods, the contiguous 5,6,7,8,9,10-hexol
hexaacetate in 1 represented a significantly more complex NMR
problem, in part, because of isolated stereohexad C5-C10 flanked
by CH2 groups2 and equivocal interpretations of J coupling.
We now report the complete stereostructure of 1 using a
progressiVe-conVergent approach that integrates three powerful
regimens for stereochemical analysis of natural products: use of
Murata’s J-based analysis (3JHH and 2,3JHC),3 application of Kishi’s
universal database4 (pairwise-comparison of 13C NMR chemical
shifts with stereo-defined models), and highly sensitive exciton
coupling circular dichroism (ECCD).5 The integrated approach
rapidly converges upon a unique stereochemical assignment for 1
a (a) In°, allyl bromide, H2O; (b) TrCl, pyridine, reflux 53% (two steps);
(c) CSA, acetone, CH3C(OCH3)2CH3 58%, 10/11 dr 2:1); (d) SiO2-HPLC
1:19 EtOAc/hexanes; (e) H2, 1 atm, Pd/C, CF3CH2OH, 35-69%; (f) (i)
(COCl)2, DMSO, CH2Cl2, -78 °C (ii) Et3N; (g) (i) 15, DME, NaHMDS,
-78 °C, (ii) aldehyde, 25%, dr 3:1 (two steps); (h) (i) 14, THF, NaHMDS,
-78 °C, (ii) aldehyde, dr > 19:1, 16% two steps; (i) OsO4, NMO, acetone,
H2O; dr 1.7:1, 93%; (j) K3Fe(CN)6, K2OsO4, K2CO3 (DHQ)2PHAL,
t-BuOH, H2O, CH3SO2NH2, dr 3.8:1, 86%; (k) 2% TMS-Cl, MeOH, (ii)
CH2N2, ether/MeOH, (iii) Ac2O, pyridine 6 h: 22% 3 (three steps), 48% 2
(three steps), 44% 6 (four steps), 26% 7 (four steps).
with internal validation.
A basis set of JHH and
3
2,3
J
values were obtained by 2D
CH
heteronuclear 2D NMR experiments of 1 (COSY and HSQMBC,
respectively, see Supporting Information (SI)) and used to predict
an all anti relative configuration for C6-C9 for 1. Consequently,
the number of remaining possible diastereomers of 1 was reduced
from N ) 32 to 4. A synthetic route to six model compounds,
representing permutations of the six stereocenters C5-C10 congru-
ent with those proposed for 1, was conceived and executed starting
with D-xylose (see SI).6 To address an equivocal C8 3JCH value in
1, a parallel set of models 2-9 was also prepared from D-ribose as
described below (Scheme 1).
Indium-promoted Barbier reaction of D-ribose with allyl bromide
gave a 2:1 mixture of epimeric homoallylic alcohols7 10 and 11
after protection. Each acetonide was deprotected and hydrogenated
(Pd/C, CF3CH2OH, 1 atm H2)8 followed by Swern oxidation to the
corresponding C9 aldehydes and homologation using two stereo-
complementary methods (Z-selective Wittig olefination using
phosphonium salt 14 and E-selective Julia-Kocienski olefination
with tetrazole 159) to give 12 and 13.
Stereoselective OsO4 dihydroxylation10 of 12 gave diols 16 and
17. In this manner, E- and Z-olefins were converted to diol
diastereomers and purified by HPLC, prior to deprotection to the
hexaols. Peracetylation of each hexaol furnished the eight C7-C9
ribo-model compounds 2-9 and six xylo-models (SI). The correct
relative configuration of 1 emerged from 13C NMR comparisons
with the model compounds (Figure 1).4
1
The H and 13C NMR spectra of each peracetate model were
carefully assigned from COSY and HMBC spectra. Pairwise
comparisons of the differences of the 13C chemical shifts (∆δ) for
C4-C11 in model compounds and 1 clearly showed an excellent
match for the C8 epimer 6 obtained from D-ribose, but a mismatch
for the corresponding xylo-C8 epimer (e.g., C8: ∆δ ) +0.05 and
-3.93 ppm, respectively, see SI). A valuable object lesson is
revealed here that promotes a progressive-convergent approach
to stereochemistry. Although anomalous 3JCH values in 1 predicted
† University of California, San Diego.
‡ University of California, Davis.
9
11764
J. AM. CHEM. SOC. 2006, 128, 11764-11765
10.1021/ja063735y CCC: $33.50 © 2006 American Chemical Society

C O M M U N I C A T I O N S
combination of methodologies lies in judicious interpretation of
3
homonuclear J and heteronuclear 2,3J to provide partial stereo-
chemical information which is then used to inform correct choices
for synthesis of model compounds to be used in the next stage:
13C NMR comparative analysis.
A significant advantage is gained by a requirement for only a
limited subset of stereomodel compounds without the necessity for
synthesis of all 64 possible permutations. The progressive-
convergent approach succeeds where other singular methods based
on NMR may become irreducibly complex13 or rendered equivocal
by second-order effects that militate against reliable stereochemical
assignments.
Acknowledgment. We thank J. DeRopp (UC Davis) for
assistance with NMR measurements and R. New (UC Riverside)
and Y.X. Su (UCSD) for MS data. This work was supported by
the NIH (Grant CA85602).
Supporting Information Available: Preparation of ribo- and xylo-
model model compounds, and their stereochemical assignments, ∆δ’s
1
of xylo-models, H, 13C NMR, and MS spectral data. This material is
available free of charge via the Internet at http://pubs.acs.org.
References
Figure 1. 13C NMR (125 MHz, d6-DMSO, T ) 298 K) ∆δ values
(δC 1-δC model) of ribo-model compounds 2-9.
(1) Lievens, S. C.; Molinski, T. F. Org. Lett. 2005, 7 (11), 2281-2284.
(2) All attempts to convert 1 to a C1-O5 δ-lactone (potentially useful for
stereochemical assignment) using acid catalysis were unsuccessful.
(3) Murata, M.; Nakamura, H.; Tachibana, K. J. Org. Chem. 1999, 64, 866-
876.
(4) (a) Kobayashi, Y.; Hayashi, N.; Kishi, Y. Org Lett. 2002, 4, 411-414.
(b) Kobayashi, Y.; Tan, C.-H.; Kishi, Y. J. Am. Chem. Soc. 2001, 123,
2076-2078. (c) Kobayashi, Y.; Lee, J.; Tezuka, K.; Kishi, Y. Org. Lett.
1999, 1, 2177-2180.
(5) (a) Vazquez, J. T.; Wiesler, W. T.; Nakanishi, K. J. Am. Chem. Soc. 1987,
109, 5586-5592. (b) Zhou, P.; Berova, N.; Nakanishi, K.; Knani, M.;
Rohmer, M. J. Am. Chem. Soc. 1991, 113, 4040-4042. (c) Harada, N.;
Nakanishi, K. Circular Dichroic Spectroscopy: Exciton Coupling in
Organic Stereochemistry; University Science Books: Mill Valley, CA,
1983.
(6) The carbons numbered C7, C8, and C9 in 1 map to C4, C3, and C2 of
ribose or xylose, respectively. Thus, the stereochemical descriptors ‘xylo-’
and ‘ribo-’ in the context of this work refer to C7-C9 of 1.
(7) The configuration of the major isomer was assigned by analogy with the
well-known 1,2-syn-stereopreference for In°-promoted allylation of aldo-
hexoses [(a) Kim, E.; Gordon, D. M.; Schmid, W.; Whitesides, G. M. J.
Chem. Org. 1993, 58, 5500-5507. (b) Kobayashi, S.; Nagayama, S. J.
Org. Chem. 1996, 61, 2256-2257] and subsequent conversion to the
acetonides 10 and 11.
(8) Deprotection of 10 and 11 to the corresponding primary alcohols was
rapidly effected when CF3CH2OH was used as solvent for hydrogenolysis.
No reaction was observed in ethanol, even after several days at 3 atm H2.
(9) Blakemore, P. R.; Cole, W. J.; Kocienski, P. J.; Morley, A. Synlett 1998,
(1), 26-28. Both 14 and 15 were prepared from δ-valerolactone in three
and four steps, respectively (see SI).
Figure 2. CD spectra of sagittamide A derivative 18 (s), together with
models 19 (‚‚‚) and 20 (- - -), (CH3CN, c ) 10 µM).
(10) Diastereomeric assignments of 5,6-diols were based on the expectation
of anti-selectivity of OsO4 addition to allylic alcohols and confirmed by
the outcomes from double-diastereoselection using the Sharpless asym-
metric dihydroxylation (Kolb, H. C.; VanNieuwenhze, M. S.; Sharpless,
K. B. Chem. ReV. 1994, 94, 2483-547) and observed pseudo-C2 symmetry
in the 1H and 13C NMR spectra of 2 and 8 (see SI).
an erroneous xylo-configuration during J-based analysis,11 this was
readily rectified in the progressive 13C ∆δ analysis allowing
reassignment of C8 configuration to that of 6.
(11) This observation suggests caution in using J-based methodology and over-
reliance on the underlying assumption of all-staggered conformations and
the accuracy of J’s measured in strongly coupled contiguous polyols that
may not be amenable to first-order spin analysis.
(12) The lactam-mono methyl ester that formed spontaneously upon treatment
of 1 (CH2N2, MeOH-ether, ref 1) and the hexaols corresponding to 6 and
7 were each converted (excess BzCl, pyridine, 40 °C) to hexabenzoates
17, 18, and 19, respectively, after HPLC purification. Benzoylation at
higher temperatures (60-90 °C) led to significant formation of tetraben-
zoyloxy-tetrahydrofuran.
(13) The similarity of CD spectra of diastereomeric 19 and 20 reflect the
dominance of the C7-C10 configuration on the Cotton effects.
(14) Rychnovsky, S. D.; Rogers, B.; Yang, G. J. Org. Chem. 1993, 58,
3511-3515.
The absolute stereochemistry of 1 was secured by transformation
of the natural product, and hexaol diastereomers corresponding to
6 and 7, to the per-benzoate ester derivatives, 18, 19, and 20,
respectively,12 and comparison of their corresponding CD spectra
(Figure 2). Since the fingerprint Cotton effects observed in the CD
spectra of 18 and 19 were equal in magnitude but opposite in sign,
the absolute configuration of 1 corresponds to ent-19 and is related
to L-ribose.12 Thus, the complete configuration of sagittamide A
(1) is depicted as (5S,6S,7S,8R,9R,10S).
In summary, we have deployed an integrated approach to solve
the configuration of sagittamide A (1). The power of this triple-
JA063735Y
9
J. AM. CHEM. SOC. VOL. 128, NO. 36, 2006 11765